Simulation of a modified cyclone separator with a novel exhaust

合集下载

se

Simulation Environment Help™Product Version 5.1.41June 20041999-2004 Cadence Design Systems, Inc. All rights reserved.Printed in the United States of America.Cadence Design Systems, Inc., 555 River Oaks Parkway, San Jose, CA 95134, USATrademarks:Trademarks and service marks of Cadence Design Systems, Inc. (Cadence) contained in this document are attributed to Cadence with the appropriate symbol. For queries regarding Cadence’s trademarks, contact the corporate legal department at the address shown above or call 1-800-862-4522.All other trademarks are the property of their respective holders.Restricted Print Permission:This publication is protected by copyright and any unauthorized use of this publication may violate copyright, trademark, and other laws. Except as speciﬁed in this permission statement, this publication may not be copied, reproduced, modiﬁed, published, uploaded, posted, transmitted,or distributed in any way,without prior written permission from Cadence.This statement grants you permission to print one (1) hard copy of this publication subject to the following conditions:1.The publication may be used solely for personal, informational, and noncommercial purposes;2.The publication may not be modiﬁed in any way;3.Any copy of the publication or portion thereof must include all original copyright,trademark,and otherproprietary notices and this permission statement; and4.Cadence reserves the right to revoke this authorization at any time, and any such use shall bediscontinued immediately upon written notice from Cadence.Disclaimer: Information in this publication is subject to change without notice and does not represent a commitment on the part of Cadence. The information contained herein is the proprietary and conﬁdential information of Cadence or its licensors, and is supplied subject to, and may be used only by Cadence’s customer in accordance with, a written agreement between Cadence and its customer. Except as may be explicitly set forth in such agreement, Cadence does not make, and expressly disclaims, any representations or warranties as to the completeness,accuracy or usefulness of the information contained in this document. Cadence does not warrant that use of such information will not infringe any third party rights,nor does Cadence assume any liability for damages or costs of any kind that may result from use of such information.Restricted Rights:Use,duplication,or disclosure by the Government is subject to restrictions as set forth in FAR52.227-14 and DFAR252.227-7013 et seq. or its successor.ContentsPreface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 Related Documents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71Setting Up SE Help. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 About SE Help. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 Finding the Information Y ou Want . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 Before Y ou Can Run a Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11 Displaying the Simulation Menu. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11 Initializing a New Run Directory in the Graphical Environment . . . . . . . . . . . . . . . . . . . .11 Reinitializing an Existing Run Directory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 Setting Up Simulation in the UNIX Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 Setting Up Remote Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 Customizing Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 Using the.simrc File To Customize SE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 Customizing Netlisting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16 Customizing Scale Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .172Creating the Input Stimulus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 Creating the Input Stimulus in the Control File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 Using Substitution Functions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .203Customizing Netlisting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 SE Netlisting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..23 Specifying a Hierarchy of Netlisting Views. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 Switching Views . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 Selecting a Netlisting View from a Hierarchy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 Overriding Default View and Stop Lists. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24Controlling Renetlisting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .254Running a Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 Choosing Simulation Run Options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 Simulation in the Graphical Environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28 Simulation in the UNIX Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29 Full Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 Simulation in Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 simInitRunDir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30 netlist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..31 simin. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 runsim. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 exit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..31 Simulation in Batch Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32 Interactive Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32 Netlist and Simulate Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34 Simulation Environment Options Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .355Displaying Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 Displaying Waveform Results in the Graphical Environment. . . . . . . . . . . . . . . . . . . . . .37 Displaying Waveform Results in Register Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 Displaying T ext Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39 Displaying Netlisting Errors for Speciﬁc Nets or Instances . . . . . . . . . . . . . . . . . . . . . . .39 Displaying Other Netlisting Errors (SILOS II Only) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39 Displaying a Speciﬁed T ext File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .396Controlling Job Status. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41 Accessing the Job Monitor Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41 Checking Current Simulation Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42 Adjusting Job Priority. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43 T erminating a Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43Interrupting or Restarting a Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43 Editing the Job Monitor Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .447SE Functions Reference. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45 Initialize . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45 Initialize Forms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46 Using Initialize for a New Run Directory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47 Using Initialize for an Existing Run Directory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48 T op-Level SKILL Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48 Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49 Options Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49 Using Options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50 Stimulus – Edit File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50 Edit File Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50 Netlist/Simulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51 Netlist/Simulate Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52 Using Netlist/Simulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 T op-Level SKILL Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 Interactive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 Using Interactive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54 T op-Level SKILL Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54 Show Outputs – Show Run Log. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .54 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 Using Show Run Log . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 Show Outputs – Show Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 Using Show Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 Show Outputs – Show Global Error. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56 Using Show Global Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56Show Outputs – Highlight Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .56 Using Highlight Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57 Show Outputs – Show Run File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57 Show Run File Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57 Using Show Run File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57 Show Waveforms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 Show Waveforms Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 Using Show Waveforms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 T op-Level SKILL Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59 Show Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59 Show Registers Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .59 Using Show Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 Job Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..60 Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 Job Monitor Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 Using Job Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .61 T op-Level SKILL Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .628Sample Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63 Sample control File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63 Sample si.inp File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .64 Sample si.inp File Generated for Cadence SILOS II. . . . . . . . . . . . . . . . . . . . . . . . . . . .65Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67PrefaceSE gives you two basic ways to run simulations: using the graphical interface or using a command line. In the graphical environment, you use SE menus and forms. In the nongraphical environment,you use SE commands in a UNIX®xterm window running si or SKILL commands in the Command Interpreter Window (CIW).Related DocumentsThe Simulation Environment is often used with other Cadence products during the design process.The following manuals give you more information about the tools used to enter and verify your design.s Before you begin using the Open Simulation System, you should be familiar with the Design Framework II™ environment. See Cadence Design Framework II UserGuide.s If you want to enter or modify your design,see Virtuoso Schematic Composer User Guide.s If you want to integrate a simulator into the Cadence System,see the Open Simulation System Reference.Preface1 Setting Up SE HelpIn this chapter, you can ﬁnd information abouts About SE Help on page9s Finding the Information Y ou Want on page10s Before Y ou Can Run a Simulation on page11s Displaying the Simulation Menu on page11s Initializing a New Run Directory in the Graphical Environment on page11s Reinitializing an Existing Run Directory on page13s Setting Up Simulation in the UNIX Environment on page13s Setting Up Remote Simulation on page15s Customizing Simulation on page15s Customizing Scale Factors on page17s Customizing Netlisting on page16About SE HelpSE allows you to run simulations from the graphical interface or from a command line.In the graphical environment,you use SE menus and forms.In the nongraphical environment,you use SE commands in a UNIX®xterm window running si or you use SKILL commands in the Command Interpreter Window (CIW).SE supports user-deﬁned simulators and the following standard simulators:s System HILOs SPICE/HSPICEs Verilog-XL SimulatorSetting Up SE HelpFinding the Information You WantIf you have used Simulation Environment before, but need detailed information about a speciﬁc Simulation Environment feature,click on any menu choice in the following diagram to go to reference information about that choice.Misc->Probe->SimulationInitialize...Options...Application Options...StimulusNetlist/Simulate...Interactive...Show Outputs->Show Stimulus WaveformsShow Waveforms...Edit File...Show Registers...Show Run Log->Show Foreground Run LogJob Monitor...Show Output Show Background Run LogShow Global ErrorsHighlight ErrorsShow Run File...Application OptionsThis menu choice lets you view forms that are speciﬁc to your simulator. Y ou must write the necessary SKILL code to activate this menu choice.Show Foreground Run LogThis command lets you view the run log of the currently running foreground job.Show Background Run LogThis command lets you view the run log of the currently running background job. Before You Can Run a SimulationAfter you complete your design, you extract it, correct errors, and save the design for simulation input. Y ou must correct all errors you ﬁnd during extraction before you simulate your design. For information on extraction, refer to Virtuoso Schematic Composer User Guide.Displaying the Simulation MenuTo bring up the Simulation menu when you are displaying a schematic➤In the Schematic window, select T ools – Simulation – Other.The system changes the menu banner to include the Simulation menu. This menu has the commands you need to simulate your design.Initializing a New Run Directory in the Graphical EnvironmentThe ﬁrst step in simulation is setting up the simulation environment. When you initialize the environment, you specify the following:s Designs Simulators Simulation run directoryInitially,the system turns off all commands on the Simulation menu except Initialize.After you run the Initialize command, you can use the remaining simulation commands.To initialize a new simulation run directory1.In the Schematic window, select Simulation –Initialize.The following form appears:2.T ype the name of the simulation run directory.Y ou can type either a full or a relative path. If you type a relative path, the system puts the run directory under the directory in which you started the software. The default run directory is spice.run1.The system stores all simulation input and output ﬁles in the simulation run directory.As the system initializes the environment, it lists the ﬁles it has loaded and anyoverridden variables in the si.foregnd.log ﬁle3.Click OK.The following form replaces the ﬁrst Initialize form:The values on this form are the current window and default SE values. Y ou can edit these ﬁelds by typing in the form or by using the Browser.1.From the Simulator Name cyclic ﬁeld, select the simulator.If you use a simulator that is not listed,select other and type the name of the simulator in the adjoining text entry ﬁeld. Y ou must select other before you can type in the text entry ﬁeld.2.Enter the name of the library containing the top level of your design.3.Enter the cell name of the top level of your design.4.Enter the view name of your design (for example, schematic).5.Click OK.Reinitializing an Existing Run DirectoryYou must reinitialize an existing run directory for each new simulation session with the following procedure:1.In the Schematic window, select Simulation –Initialize.The following form appears:2.T ype the name of the simulation run directory.3.Click OK.Setting Up Simulation in the UNIX EnvironmentCadence recommends that you run simulation using the menus and forms in the graphical environment. However, you can also run a simulation using SE commands in the CIW or in UNIX using the si binary.With these commands,you can run simulation in either interactive or batch mode.Before you start a simulation in the UNIX environment,you must create the followingﬁles in your simulation run directory:s si.envs controlIf you run simulation in UNIX, you must create the si.env ﬁle. (This ﬁle is createdautomatically when you use the graphical environment.)The si.env ﬁle tells SE which design to simulate and what simulator to use.The following table lists the variables you must deﬁne in the si.env ﬁle.Each interface might store additional speciﬁc variables in the si.env ﬁle Required Properties in si.env.Optional Property in si.envThe following is a sample si.env ﬁle:simLibName = "testLib"simCellName = "74169"simViewName = "schematic"simSimulator = "silos"simHost = "cds642"T o set up SE to run simulations in the UNIX environment, follow the steps below:1.Change to the directory that will contain the simulation run directory.2.Create the run directory using the UNIX command mkdir directoryname , where directoryname is the name of the simulation run directory. For example, if your simulation run directory is spice.run1, type the following:mkdir spice.run13.Change to the newly created directory.cd spice.run14.Create the simulation environment ﬁle si.env .VariableDescription simSimulatorSimulator to be run simLibNameName of the library containing the top-level cellview simCellNameName of the top-level cell to be simulated simViewName View name of the top-level cell to be simulatedVariableDescription simHost Host name of remote simulatorSetting Up Remote SimulationYou can set up the system to run remote simulation using the Verilog-XL, HSPICE, and SPICE simulators. The local machine and the remote host must both run the X Window System TM.T o set up remote simulation, perform the following steps to deﬁne the necessary variables (you can set these variables using the Options command on the Simulation menu by typing them in the CIW or in the.simrc ﬁle):1.Set the SE variable simHost to the name of the remote workstation.For example:simHost = "cds17"2.Set the SE variable simHostDiffers to true(t)if the host computer has a different binarystorage format than the local computer.For example:simHostDiffers = tAfter you set these variables,you can run a remote simulation and view the results as if you were running the simulation locally.Customizing SimulationUsing the.simrc File To Customize SEWhen you initialize SE, it ﬁrst loads the si.env ﬁle. This ﬁle tells SE what design to simulate and what simulator to use.SE then loads the.simrc ﬁle if it exists. The.simrc ﬁle lets you override any netlisting or simulation environment variables. SE searches the following directories and loads the ﬁrst .simrcﬁle it ﬁnds:$SIMRC/.simrc$ossSimUserSiDir/.simrcinstall_dir/tools/dfII/local/.simrc./.simrc~/.simrcIf you set a variable in.simrc that also sets options in the graphical environment (with the Options command), SE uses the.simrc ﬁle settings and ignores the Options settings.The .simrc ﬁle must be in SKILL syntax.The following is a sample .simrc ﬁle.The ﬁrst line overrides the default view list used for view switching with SILOS.The second line overrides the default stopping list that stops hierarchy expansion for SILOS.hspiceSimViewList = ("hspice" "cmos_sch" "schematic")hspiceSimStopList = ("hspice" "cmos_sch")The following table shows you some SE variables you can set in your .simrc ﬁle tocustomize simulation.Y ou can also see the Open Simulation System Reference Manual for a complete list of SE variablesFor a description of SKILL syntax and further information about SKILL functions, see the SKILL Language User Guide and the SKILL Language Reference Manual.Customizing NetlistingThe way the design hierarchy is traversed to produce the netlist and the syntax of the netlist depends on your simulator.For example,you might want the netlist for a Verilog simulation to VariableDescription simSimulatorSimulator to run simControlFilePath of default control ﬁle simDefaultControlName of default control ﬁle if stored in install_dir/etc/s simTimeUnit Scaling factor for delay times. This valueshould match the ﬁrst argument of thedeftiming command.simCapUnit Scaling factor for capacitancesimNlpGlobalLibName Name of library containing globalformatting instructions for ﬂat netlistersimNlpGlobalCellName Name of cell containing global formattinginstructions for ﬂat netlistersimNlpGlobalViewName Name of view of the cell containing globalformatting instructions for ﬂat netlistersimNotIncremental Speciﬁes incremental netlisting when setto nil .simReNetlistAll Speciﬁes non-incremental netlistingsimNetlistHierSpeciﬁes hierarchical netlistingbe at the logic gate level because Verilog can simulate primitives such as AND gates and AOIs. Y ou might want the netlist for a SPICE simulation of the same design to be at the transistor level because SPICE cannot simulate logic gates.Click on the topics below to go to information about customizing netlisting.s SE Netlistings Specifying a Hierarchy of Netlisting Viewss Selecting a Netlisting View from a Hierarchys Overriding Default View and Stop Listss Controlling RenetlistingCustomizing Scale FactorsThe netlister can scale time and capacitance values. The scale factors for time and capacitance are deﬁned by two SE variables: simTimeUnit and simCapUnit. With both variables, the value to be scaled is divided by the scale factor. The default value of variable simTimeUnit is1e-9(nanoseconds),and the default value of variable simCapUnit is1e-15 (femtofarads). Y ou can customize the scale factors by typing new simTimeUnit and simCapUnit values in your.simrcﬁle.Y ou can change the SILOS time unit,for example,to a picosecond by setting simTimeUnit to 1e-12.。

multi2sim

The Multi2SimSimulation FrameworkA CPU-GPU Model for Heterogeneous Computing(Version3.1)Authors:Rafael UbalJulio SahuquilloSalvador PetitPedro L´o pezZhongliang ChenDavid R.KaeliContents1Introduction51.1The Simulation Paradigm (5)1.1.1The Functional Simulator (5)1.1.2The Detailed Simulation (7)1.2Getting Started (8)1.2.1Installation (8)1.2.2Simple execution (8)1.3Context Conﬁguration File (9)1.3.1Format (9)1.3.2Example (10)1.4The Processor Pipeline (10)1.5Compiling and Simulating Your Own Source Code (11)1.5.1Static and Dynamic Linking (11)1.5.2Observing the Diﬀerences (12)1.5.3Execution Variability (14)1.5.4Error Messages when Simulating your Program Binaries (15)2The Processor Front-End162.1Branch Predition (16)2.1.1Perfect branch predictor (16)2.1.2Taken branch predictor (16)2.1.3Bimodal branch predictor (17)2.1.4Two-level adaptive predictor (17)2.1.5Combined predictor (17)2.2Multiple Branch Prediction (18)2.3Macroinstruction Decoding (19)2.4Trace Cache (20)2.4.1Creation of traces (21)2.4.2Trace cache lookups (21)2.4.3Trace cache statistics (21)2.5The Fetch Stage (22)2.6The Decode Stage (23)3The Processor Back-End253.1Integer Register Renaming (25)3.1.1Logical Registers (25)3.1.2Physical Register File (25)23.1.3Renaming Process (26)3.2Floating-Point Register Renaming (26)3.2.1The x86Floating-Point Stack (26)3.2.2Two-Stage Renaming Process (27)3.3The Dispatch Stage (27)3.4The Issue Stage (28)3.5The Writeback Stage (28)3.6The Commit Stage (28)4Support for Parallel Architectures304.1Deﬁnitions (30)4.2Multithreading (31)4.2.1Conﬁguration of storage resources (31)4.2.2Conﬁguration of bandwidth resources (31)4.3Multicore Architectures (32)4.4The Context Scheduler (32)4.4.1The Static Scheduler (32)4.4.2The Dynamic Scheduler (33)4.5Statistics Report (34)4.5.1Global statistics (34)4.5.2Statistics related to all pipeline stages (34)4.5.3Statistics related to the dispatch stage (35)4.5.4Statistics related to the execution stage (35)4.5.5Statistics related to the commit stage (35)4.5.6Statistics related to hardware structures (36)5The Memory Hierarchy375.1Memory Hierarchy Conﬁguration (37)5.1.1Conﬁgurationﬁle format (37)5.1.2Sections,variables,and values for the conﬁgurationﬁle (39)5.2Multi-ported caches (39)5.3Cache coherence (40)5.3.1Blockﬁelds and directory entries (40)5.3.2Deadlocks (40)5.4Statistics Report (41)5.4.1General statistics (42)5.4.2Statistics related with access retries (42)6The Pipeline Debugger446.1Obtaining Timing Diagrams (44)6.2Pipeline Debugger Elements (44)6.3Timing Diagram Navigation (45)6.4Comparing Several Executions (46)7Graphics Processing Units487.1Basic Concepts about GPU Computing (48)7.1.1The OpenCL Programming Model (48)7.1.2The AMD Evergreen GPU Architecture (48)7.1.3Data Structures Modeling (50)7.2Emulation of an OpenCL Program at the ISA Level (51)7.2.1Evergreen Assembly (51)7.2.2Control Flow and Thread Divergence (52)7.2.3Program Loading and Emulation Loop (53)7.2.4The Power of Heterogeneous Simulation at the ISA Level (54)7.3Trying it out (54)7.3.1First Executions (55)7.3.2The OpenCL Trace (55)7.3.3The Evergreen ISA Trace (57)7.4The Execution Model for OpenCL Programs (58)7.4.1Native vs.Emulated Execution of OpenCL (58)7.4.2Execution of an OpenCL Program on an AMD-based Native Environment..597.4.3Execution of an OpenCL Program on Multi2Sim (59)7.4.4Implementation of the Multi2Sim OpenCL Library (60)7.5Building and Simulating Your Own OpenCL Program (61)7.5.1Building the Multi2Sim OpenCL Library (61)7.5.2Compiling Source Files (61)7.5.3Linking Object Files for Native Execution (62)7.5.4Linking Object Files for Execution on Multi2Sim (62)7.5.5Simulating an OpenCL Program Linked for Native Execution (63)7.5.6The Multi2Sim OpenCL Kernel Compiler (63)8The GPU Architectural Simulation658.1The GPU Device Architecture (65)8.2The Compute Unit Architecture (66)8.2.1The Control-Flow(CF)Engine (67)8.2.2The Arithmetic-Logic(ALU)Engine (69)8.2.3The Texture(TEX)Engine (71)8.3The GPU Memory Hierarchy (72)8.3.1Private Memory (72)8.3.2Local Memory (72)8.3.3Global Memory (73)8.3.4Memory Access Queuing and Coalescing (75)8.4Properties of the Global Memory Hierarchy (77)8.4.1Interconnection Networks (77)8.4.2Cache Properties (78)9Tools809.1The INIﬁle format (80)9.1.1The iniﬁle.py tool (80)9.1.2Reading INIﬁles (81)9.1.3Writing on an INIﬁle (81)9.1.4Using scripts to edit INIﬁles (81)9.2McPAT:Power,Area,and Timing Model (82)9.2.1McPAT inputﬁle (82)9.2.2Interaction with Multi2Sim (82)9.2.3McPAT output (82)Chapter1IntroductionMulti2Sim is a simulation framework for heterogeneous computing,including models for super-scalar,multithreaded,multicore,and graphics processors.Multi2Sim is an application-only simu-lator,which allows one or more applications to be run on top of it without booting a guest operating systemﬁrst.In this chapter,an introduction to Multi2Sim is presented,and it is shown how to perform basic simulations and extract performance results.1.1The Simulation ParadigmThe simulation paradigm can be divided into two main modules:the functional simulation and the timing or detailed simulation.The functional simulation is just an emulation of the input program. Given an executable ELF(Executable and Linkable Format)ﬁle,the functional simulator provides the same behavior as if the program was executed natively on an x86machine.The detailed simulator provides a model of the hardware structures of an x86-based machine;it provides timing and usage statistics for each hardware component,depending on the instructionﬂow supplied by the functional simulator.Throughout this document,the term guest will be used to refer to any property of the simulated program,as opposed to the term host,used to refer to the simulator properties.For example,the guest code is formed by instructions of the program whose execution is being simulated,whereas the host code is the set of instructions executed by Multi2Sim natively in the user’s machine.The next sections describe the main actions performed by the functional and detailed simulators,and their interactions.1.1.1The Functional SimulatorThe executable generated when building Multi2Sim is m2s,which is a uniﬁed tool for both functional and detailed simulation.Adding the option--cpu-sim functional will enable the functional simu-lator,which is also the default.This conﬁguration provides an implementation for the functional simulation.It takes as an input one or more ELFﬁles,and emulates their execution,providing a basic set of statistics based on the guest code run.The functional simulation is mainly imple-mented in the Multi2Sim kernel library(src/libcpukernel).The main actions performed by the functional simulator can be classiﬁed into program loading,x86instructions emulation,and system calls emulation,described next:•Program Loading.The state of a guest program’s execution,referred to as context,is basically represented by a virtual memory image and a set of logical register values(Figure5Figure1.1:Initialization and central loop of the functional simulation of an x86program. 1.1a).The former refers to the values stored in each memory location in the context’s virtual memory,while the latter refers to the contents of the x86registers,such as eax,ebx,etc. The Linux Application Binary Interface(ABI)speciﬁes an initial value for both the virtual memory image and register values,before control is transferred to the new context.The initial state of the context is inferred mainly from the program ELF binary,and the command-line run by the user to launch it,during the process called program loading.In a real system, program loading is performed by the operating system after an execv system call or any of its variants.In the simulation environment,Multi2Sim is in charge of performing program loading for each guest context run on top of it.Program loading is implemented inﬁle src/libcpukernel/loader.c,with the following steps:–First,the x86binary is analyzed with an ELF parser.An ELFﬁle contains sections of code(x86instructions)and initialized data,jointly with the virtual address where they should be initially loaded.For each ELF section,the program loader obtains its virtual oﬀset and copies it into the corresponding location in the virtual memory image.–The context stack is initialized.The stack is a region of the guest virtual memory image pointed to by register esp.Initially,it contains a set of program headers copied from the ELFﬁle,followed by an array of environment variables,and the sequence of command-line arguments provided by the user.–The x86registers are initialized.The esp register is set to point to the top of the stack, and the eip register is set to point to that memory location containing the code to run when control isﬁrst transfered to the new context.•Emulation of x86instructions.Once the initial image of the new context is ready,its emulation can start.Iteratively,the functional simulator reads a sequence of bytes at the guest memory address pointed to by guest register eip.Then,the represented x86instruction is obtained by calling the Multi2Sim x86decoder and disassembler(library src/libdisasm). The instruction is emulated by updating accordingly the guest virtual memory image andregisters1.Finally,guest register eip is updated to point to the next x86instruction to be executed.Instruction emulation is implemented inﬁles src/libcpukernel/machine*.c.•Emulation of system calls.A special case of machine instruction is the software interrupt x86instruction int.Speciﬁcally,instruction“int0x80”is used to perform a system call.When Multi2Sim encounters a system call in the emulated application,it updates the context status accordingly depending on the system call code and its arguments,providing the guest program with the view of actually having performed the system call natively.In most cases,Multi2Sim will need to perform the same host system call as the guest program is requesting,by making a pre-and post-processing of the arguments and result,respectively.For example,when a guest program runs an open system call,it provides in speciﬁc x86 registers a pointer to the string containing the path to be opened,and it expects aﬁle descriptor as a return value,also in a speciﬁc x86register.Roughly,Multi2Sim deals with this by locating the path string in guest memory,performing its own host open system call,and placing the resultingﬁle descriptor into guest register eax,where the guest context expects it to be once execution resumes at the instruction following the system call.The emulation of system calls is implemented inﬁle src/libcpukernel/syscall.c.The high-level actions performed by the functional simulator loop are represented in Figure 1.1b,including the emulation of both x86instructions and system calls.1.1.2The Detailed SimulationAdding the option--cpu-sim detailed to m2s will enable the detailed simulator.This program provides a model of CPU structures,such as caches,functional units,interconnection networks, etc.,most of them described in the central sections of this guide.Most importantly,it models a CPU pipeline with support for branch prediction and speculative execution.Theﬁrst stage of this pipeline(fetch stage)interacts with the functional simulator module or Multi2Sim kernel library,by using a simple interface.Iteratively,m2s asks the functional simulator to emulate the next guest x86instruction and return some information about it.Based on this information,m2s canﬁgure out which hardware structures are activated and performs a timing simulation.The functional simulator knows at each time which is exactly the next instruction to execute. In contrast,real hardware obtains the address of the next instruction to fetch from the output of a branch predictor.This address can be correct or might be the start of a sequence of mispredicted instructions,followed by a pipeline squash and recovery.This process is modeled in m2s as follows.As long as the predicted address for the next instruction matches the functional simulator state-ment,both functional and detailed simulation are synchronized.However,a branch misprediction will make m2s start fetching instructions through the wrong execution path.At this time,m2s forces a new value for the eip register in the functional simulator,which automatically checkpoints the context state.m2s keeps fetching instructions and forcing the functional simulator to execute them, until the mispredicted branch is resolved in an advanced stage of the pipeline.The branch resolu-tion leads m2s to squash the modeled pipeline contents,and makes the functional simulator return to the last valid checkpointed state,from which correct execution is normally resumed.1For example,an add instruction would read the source operands from guest memory or registers,perform an addition,and store its result back into guest memory or registers,depending on the location of the destination operand.1.2Getting Started1.2.1InstallationTo install Multi2Sim,download the package with the source code for the latest version,unpack it, and compile it.You can do this with the following sequence of commands,assuming version2.2: wget /files/multi2sim-2.2.tar.gztar-xzvf multi2sim-2.2.tar.gzcd multi2sim-2.2./configuremakeAfter compiled,the executableﬁles associated with the simulator are located within the src directory.Among all generated executables,the most complete simulator is the m2sﬁle,which provides the entire model for the processor pipeline.The syntax of the m2s command is:m2s[<options>][<x86_binary>[<arg_list>]]where options is a list of parameters that deﬁne the modeled processor,x86binary is an executable ﬁle to run on top of the simulator,and arg list is the list of the command-line arguments passed to the simulated program.If you try to run m2s without arguments,you will obtain the list of all possible options and their default values,followed by an error message telling that there was no executableﬁle to simulate.1.2.2Simple executionIn what follows,a brief example is shown as a guide to launch simple simulations.You canﬁnd it in the samples directory,where some minibenchmarks,that is,small programs statically precompiled for the x86architecture,are included.The minibenchmarks are thoseﬁles with the i386extension. First,run the test-args.i386program,followed by any number and kind of arguments you wish. As you can see,this is a very simple program that dumps the command line arguments it receives. Now,let us run it on top of Multi2Sim with the following command:m2s--cpu-sim detailed test-args.i386how are youThis command provides two diﬀerent outputs.The standard output(stdout)is the output of the simulated program,that is,the same output provided by test-args.i386if it was run without the simulator.On the other hand,the standard error output(stderr)is the output of Multi2Sim, which consists in a list of parameters and their actual values,followed by a list of statistics of the performed simulation.The most important statistics are Cycles and Instructions,which provide the total number of simulation cycles,and the total number of committed instructions,respectively.Likewise,the statistic InstructionsPerCycle is an important performance metric,which represents the number of instructions committed per cycle(IPC).You can type the following commands to obtain the IPC for the execution of test-args:tempfile=$(mktemp)m2s--cpu-sim detailed test-args.i3862>$tempfileinifile.py$tempfile read CPU.DetailedSimulation InstructionsPerCyclerm-f$tempfileThe output of a simulated program is a textﬁle with the INIﬁle format,which includes one or more sections speciﬁed with a name enclosed in brackets.In each section,a set ofﬁelds formed of pairs<key>=<value>represent diﬀerent metrics and their values.Please see Section9.1for more details.1.3Context Conﬁguration FileBoth multithreaded and multicore processors provides for the operating system the view of having multiple logical processors where several tasks can be run at the same time.In Multi2Sim,the software tasks are referred to as contexts,and their number is limited to the number of threads multiplied by the number of cores in the modeled processor.When a superscalar processor is modeled(i.e.,a processor with one thread and one core),the single context run on top of it can be speciﬁed as an executable program in the m2s command-line arguments,as shown above.However,a context conﬁgurationﬁle should be used when several executableﬁles are initially mapped each to one thread or core.1.3.1FormatThe context conﬁgurationﬁle is a plain textﬁle using the INIﬁle format(see Section9.1),which contains a list of sections[Context<num>],each with the following list of possibleﬁelds:[Context<num>]Exe=<path>Args=<arg_list>Env=<env_list>Cwd=<path>StdIn=<file>StdOut=<file>This is the meaning for eachﬁeld:•Context<num>:context unique identiﬁer,starting at0.The Multi2Sim context scheduler maps software contexts to hardware threads,so a3-core,2-threaded processor can run up to 2×3=6contexts without software context switches.More contexts can be run if the dynamic context scheduler is activated.See Section4.4for further details about context scheduling.•Exe=<path>:path of the executableﬁle.Thisﬁle will be mapped into the initial process memory image before starting the simulation.Thisﬁeld is the only mandatory one.Any otherﬁeld can be omitted.•Args=<arg list>:command line arguments for the simulated program.They are placed into the process stack before starting the simulation.•Env=<env list>:additional environment variables for the simulated program.These variables are also placed into the program stack at startup,in addition to the environment variables of the system on top of which the simulator is run.They are enumerated using single or double quotes.•Cwd=<path>:current working directory for the simulated program.Whenever the simulated program uses relative paths,this will be the directory used to build absolute paths.If not speciﬁed,the current working directory for the simulator will be also used for the simulated program.Figure1.2:Processor pipeline.•StdIn=<file>:standard input for the program.If the simulated program reads data from a ﬁle as standard input,specify a value here.If none speciﬁed,the simulator standard input is selected.•StdOut=<file>:standard output and standard error output.If thisﬁeld is speciﬁed,both the standard and the error output of the simulated program will be dumped into thisﬁle.If omitted,the standard output of the simulator will be used.1.3.2ExampleThe following text corresponds to a context conﬁgurationﬁle creating two contexts.In theﬁrst one, the program test-args.i386is used with three arguments,and the output is sent toﬁle context-0.out. In the second one,the program test-sort.i386is used,and its output is dumped in context-1.out.——————————————————————————————[Context0]Exe=test-args.i386Args=how are youStdout=context-0.out[Context1]Exe=test-sort.i386Stdout=context-1.out——————————————————————————————This context conﬁgurationﬁle could be named,for example,ctxconfig.args-sort.Now,let us try to run these programs on top of a processor model that allows two contexts,like,for example, a1-core2-threaded processor,using the CPU conﬁgurationﬁle cpuconfig.args-sort.This is the required command line:m2s--ctx-config ctxconfig.args-sort--cpu-config cpuconfig.args-sortAfter a couple of seconds,the simulation ends and the output of each context can be found in the speciﬁedﬁles with extension.out.Likewise,the simulation results are dumped as a list of statistics in the standard error output.1.4The Processor PipelineFigure1.2shows a block diagram of the processor pipeline modeled in Multi2Sim.The gray-painted boxes represent hardware structures,whereas the round shapes represent pipeline stages.Six stages are modeled in Multi2Sim,named fetch,decode,dispatch,issue,writeback,and commit.In the fetch stage,instructions are read from the instruction or the trace cache.Depending on their origin,they are placed either in the fetch queue or the trace queue.The former contains rawmacroinstruction bytes,while the latter stores pre-decoded microinstructions(uops).In the decode stage,instructions are read from these queues,and decoded if necessary.Then,uops are placed in program order into the uop queue.The fetch and decode stages form the front-end of the pipeline, and are detailed in Chapter2.The dispatch stage takes uops from the uop queue,renames their source and destination regis-ters,and places them into the reorder buﬀer(ROB)and the instruction queue(IQ)or load-store queue(LSQ).The issue stage is in charge of searching both the IQ and LSQ for instructions with ready source operands,which are schedule to the corresponding functional unit or data cache. When and uop completes,the writeback stage stores its destination operand back into the register ﬁle.Finally,completed uops at the head of the ROB are taken by the commit stage and their changes are conﬁrmed.A detailed report of the simulation statistics related with the processor pipeline can be obtained with option--report-cpu-pipeline.Please refer to Section4.5for a detailed description of its format and meaning.1.5Compiling and Simulating Your Own Source CodeDay after day,Multi2Sim provides a more and more robust and complete support for the x86 instruction set and Unix system calls.This means that every new release of the simulator makes it more likely for your own compiled sourceﬁles to be supported,regardless of your gcc,glibc,or Linux kernel versions.The main test scenarios for Multi2Sim have been executions of the pre-compiled benchmark suites provided in the website,but also support has been added for missing features, based on reports sent by users in the past years.The next sections show some considerations when simulating your own program sources.1.5.1Static and Dynamic LinkingWhen compiling a program,there are two main approaches to link the objectﬁles into theﬁnal executable,called dynamic and static linking.It is important to understand the characteristics of each approach and their impact on the program execution,either native or simulated.•Static linking.The gcc linker can be conﬁgured to generate a statically linked executable by adding the-static option into the command line.In this case,the code of any shared library used by the program(such as the mathematic library,the POSIX thread library,glibc library, etc.)is linked together with the program.This code includes,for example,the implementation of the printf function,along with many other program initialization procedures.Even for the simplest hello world program,a huge executableﬁle is generated.The advantage thereof is that thisﬁle can be used on any Linux machine with a compatible version of the kernel, regardless of the versions of the remaining installed development packages and libraries.•Dynamic linking.This is the default behavior for gcc.When a program is linked dynami-cally,the library code is not attached for theﬁnal executable.Instead,every reference to an external symbol,such as the printf function,is left unresolved initially.The compiler adds into the executable some code to load the dynamic loader,which is also a dynamic library present in your system,usually under the/etc directory.When the program is executed,the guest code itself copies the dynamic loader code into its own context image,and then jumps into it to transfer control.Then,the dynamic loader tries toﬁnd all shared libraries required by your program(usually*.soﬁles under the/lib or/usr/lib directories),and loads theircode into the process image as well.Finally,control is transferred back to the program code, which continues with other initialization actions.Based on the previous description,the following diﬀerence can be noted between the static and dynamic linking approaches,regarding the creation of the process executable code image.In the case of static linking,this initialization is exclusively performed by the program loader,implemented by the OS in a real machine,and by a simulator library in the Multi2Sim environment.In contrast, dynamically linked programs follow two steps in the code initialization:ﬁrst,the OS(or Multi2Sim) creates an initial program image;second,once the program starts running,it continues to update its image by loading the shared libraries.Thus,it is important to note that the OS(or the simulator)is not involved in the dynamic linking process.Since the update of the program image relies on the dynamic loader and dynamic libraries provided by your distribution,it is much more likely for a given pre-compiled,dynamically linked executable program to generate incompatibility issues.Thus,all benchmark packages available for download include statically linked programs.1.5.2Observing the DiﬀerencesThis section shows a practical example to observe the implications of static versus dynamic linking in the program execution,using the Multi2Sim functional simulator m2s--cpu-sim functional(which is also the default).Let us base this example on the execution of a hello world program,stored in aﬁle called hello.c,and containing the following code:#include<stdio.h>int main(){printf("hello world\n");return0;}First,let us generate a statically linked version of the program,and run it on top of Multi2Sim.A good clue of the program behavior is going to be given by the system calls.Similarly to the output provided by the strace tool,a trace of the performed system calls,their arguments,and return values can be obtained with Multi2Sim by using command-line option--debug-syscall<file>, being<file>the name of theﬁle where to dump the trace.A special value of stdout dumps the trace into the standard output:$gcc hello.c-o hello-static$m2s--debug-syscall stdout hellosyscall’newuname’(code122,inst418,pid1000)syscall’brk’(code45,inst738,pid1000)syscall’set_thread_area’(code243,inst850,pid1000)[...]syscall’open’(code5,inst911,pid1000)filename=’/dev/urandom’flags=0x0,mode=0x0return=0x3syscall’read’(code3,inst932,pid1000)guest_fd=3,pbuf=0xfffdffbd,count=0x3return=0x3syscall’close’(code6,inst948,pid1000)guest_fd=3return=0x0[...]syscall’fstat64’(code197,inst7973,pid1000)fd=1,pstatbuf=0xfffdfe58return=0x0syscall’mmap2’(code192,inst8028,pid1000)addr=0x0,len=4096,prot=0x3,flags=0x22,guest_fd=-1,offset=0x0prot={PROT_READ|PROT_WRITE},flags={MAP_PRIVATE|MAP_ANONYMOUS}return=0xb7fb0000syscall’write’(code4,inst8881,pid1000)guest_fd=1,pbuf=0xb7fb0000,count=0xcbuf="hello world\n"return=0xcsyscall’exit_group’(code252,inst9475,pid1000)The system call trace should look similar to the one shown above(it may vary across systems or even executions).We can observe that the program is retrieving some kernel information(newuname), updating the heap size and allocating memory(brk,mmap2),getting some random numbers for initialization purposes(open,read,close),getting information about the standard output(fstat64), displaying the hello world string(write),and exiting the program(exit group).Now let us try the same with the dynamically linked version of the program:。

非线性有限元建模仿真

The George Washington University
1
General FEA Process

Background and History of LS-DYNA

Model Development - Pre-processing

1976

LS-INGRID, FEM-B I-DEAS, , True-Grid, , EasiCrash PATRAN, HyperMesh LS-DYNA, PamCrash, RADIOSS NASTRAN, ANSYS, Algor LS-TAURUS, LS-POST HyperMesh
Department of Civil and Environmental Engineering
Department of Civil and Environmental Engineering
The George Washington University
Applications

Nonlinear Finite Element Modeling and Simulation
CE 264 Dhafer Marzougui Pradeep Mohan
Week 1

General FEA Process LS-DYNA Background

LS-DYNA Input Format
1993

1988

Keyword Format Automatic Single Surface Contact 1st International LS-DYNA User Conference Training Lab Established at West Coast - LSTC Training Class Started at East Coast - NCAC/GWU Release of Version LS970, Many New Features

混凝土热开裂数值描述方法1(英文)

Thermal induced stress and associated crackingin cement-based composite at elevated temperatures––Part I:Thermal cracking around single inclusionY.F.Fu a ,Y.L.Wonga,*,C.A.Tang b ,C.S.PoonaaDepartment of Civil and Structural Engineering,The Hong Kong Polytechnic University,Hong Kong,ChinabLab of Numerical Test of Material Failure,Northeastern University,Shenyang 110006,ChinaAccepted 25April 2003AbstractThis paper presents the development and veriﬁcation of 2-D mesoscopic thermoelastic damage model used to numerically quantify the thermal stresses and crack development of a cement-based composite subjected to elevated temperatures.The program is then used to study the thermal fracture behavior of a cement-based matrix with a single inclusion.The results show that the mechanisms of thermal damage and fracture of the composite depend on (i)the diﬀerence between the coeﬃcients of thermal ex-pansion (CTE)of the inclusion and the cement-based matrix,(ii)the strengths of materials,and (iii)the heterogeneity of materials at meso-scale.The thermal cracking is an evolution process from diﬀused damage,nucleation,and ﬁnally linkage of cracks.If the CTE of the inclusion is greater than that of the matrix,radial cracks will form in the matrix.On the other hand,inclusion cracks and tangential cracks at the interface between inclusion and matrix will form if the CTE of the inclusion is smaller than that of the matrix.Ó2003Elsevier Ltd.All rights reserved.Keywords:Thermal stress;Thermal induced cracking;Heterogeneity;Numerical simulation1.IntroductionThermal cracking induced by thermal mismatch has been one of the problems in a cement-based composite material under elevated temperatures.For a multi-phase material,the eigenstrains deriving from the heteroge-neous deformations among phase components inevita-bly cause cracking in the composite,even though it is under a uniform temperature ﬁeld.Experimental results [1]have shown that this type of cracking signiﬁcantly reduces the strength and elastic modulus of a cement-based composite.However,the entire thermal cracking process (initiation,propagation and linkage of cracks)and the associated stress distributions under elevated temperatures are diﬃcult to quantify experimentally,mainly because of the limitation of equipment and the complex structure of a composite material.In order to understand the failure mechanism of a composite material due to thermal eﬀects,many math-ematical models have been proposed [1–3].In these models,the driving stresses for the crack initiation and propagation are the heterogeneous eigenstresses,which develop in and around the restraining inclusion.These eigenstresses might be caused by thermal expansion,shrinkage [4,5],initial strains and misﬁt strains.Timo-shenko and Goodier in 1970[6]proposed a closed-form solution for the axisymmetric problem of a circular in-clusion concentrically embedded in the circular disc of another phase material with diﬀerent thermal and mechanical properties.Hsueh et al.[7]transformed a composite with a microstructure of square-array,hexagon-array,brick-array grains,as well as the actual microstructure of random-array grains into a simple composite-circle analytical model.The residual thermal stresses were predicted reasonably well using the pro-posed linear elastic solutions except for the model mi-crostructure of brick-array grains.A modiﬁed version of Timoshenko and Goodier Õs solution incorporating the longitudinal strain proposed by Gentry and Husain*Corresponding author.Tel.:+852-2766-6009;fax:+852-2334-6389.E-mail address:ceylwong@.hk (Y.L.Wong).0958-9465/$-see front matter Ó2003Elsevier Ltd.All rights reserved.doi:10.1016/S0958-9465(03)00086-6Cement &Concrete Composites 26(2004)99–111[2]was also used to study the diﬀerential pressure de-veloped in the interface between concrete and a com-posite rod.As for a40°C temperature increase,the concrete was modeled with a linear-elastic and nonlin-ear tension-softening material model using aﬁnite ele-ment approach.The calculated results showed that the large spacing of the rods and the thick concrete cover were helpful to reduce the tensile stress in concrete as well as the potential for thermally induced cracking. Based on a fracture mechanics model,Timoshenko and GoodierÕs solution was adopted by Dela and Stang[3] to calculate the crack growth with time in a high-shrinkage cement paste with a single aggregate disc.The experimentally measured stresses in the selected circular aggregate were employed to predict the stresses dis-tributed in cement paste and the crack growth at a crack tip close to the aggregate in terms of a given stress in-tensity factor.Although the above-mentioned models deepen the understanding on thermal stress and cracking,essen-tially,none of them can simulate the entire thermal cracking process from crack initiation to propagation. HsuehÕs and RussellÕs models can determine the stress distribution around a single inclusion in the composite before crack initiates.DelaÕs model was suitable to cal-culate the critical stress value when an existing crack starts to grow.The stress distribution represented by this model would be invalid as soon as the crack is ex-tended.A fracture mechanics model is able to study the growth of existing single crack,but it is not suitable to explain the initiation and coalescence of cracks.More importantly,the phase materials of a cement-based composite are often heterogeneous so that the eﬀect of change in microstructure(mesostructure)on the mac-roscopic behavior is diﬃcult to be studied by using an analytical model.Consequently,a numerical method appears to be an eﬀective tool to model cracking processes.Substantial progress[8,9]has been achieved in numerical simulation of failure occurring in a cement-based composite at ambient temperatures.However,a satisfactory model to simulate the cracking processes caused by the thermal induced stresses in a heated cement-based composite is still not available.The aim of this paper is to propose and verify a mesoscopic thermoelastic damage(MTED)model,that can numerically simulate the formation,extension and coalescence of cracks in a cement-based composite ma-terial(cement-based matrix+aggregate inclusion), caused by the thermal mismatch of the matrix and the inclusion under uniform temperature variations and free boundary conditions.Numerical studies of the eﬀects of the thermal mismatch between the matrix and a single circular inclusion on the stress distribution and crack development are also presented.2.Numerical modelIn the MTED model,phase materials of a composite are considered to be heterogeneous following the Wei-bull distribution.Tensile and shear cracking at meso-scale occur if the stress in the composite subjected to high temperatures satisﬁed with the failure criteria of Coulomb–Mohr with tensioncutoﬀ. 100Y.F.Fu et al./Cement&Concrete Composites26(2004)99–1112.1.Material modelFor a cement-based composite material,the phase materials are cement mortar matrix and aggregate in-clusions.Although the composite material is regarded as an isotropic elastic-brittle solid at a macroscopic scale, while the individual grains in the matrix and inclusions are distinguished at microscopic or mesoscopic scales [8].The eﬀect of heterogeneity on the stress distribution has been studied[10],and much of the behavior ob-served at a macro-level can be explained in terms of the material structure at a meso-level.As a result,the matrix and the inclusions are considered as disorder solids in a meso-scale in this study.To account for the heterogeneity of the matrix and inclusions,their statistical distributions of properties (elastic modulus,compressive strength and Poisson ratio)are assumed to follow the Weibull distribution:uðh;bÞ¼hb0Ább0hÀ1ÁeÀðb=b0Þhð1aÞwhere uðh;bÞis the distribution density of parameter b which is a material property(such as strength,elasticity and Poisson ratio)of a representative volume element (RVE)in the mesh divisions,and b0is the mean value of the material property under consideration.h is the ho-mogeneity index of the RVE which represents the degree of homogeneity.The statistical distribution function Uðh;bÞis expressed by Eq.(1b)after integrating Eq.(1a):Uðh;bÞ¼1ÀeÀðb=b0Þhð1bÞThe randomness of the mechanical properties of RVE can be simulated using the distribution function with given parameters h and b0,i.e.Uðh;b0Þ.The relationship of distribution density of RVE strength and homoge-neity index is shown in Fig.1.With increasing h,the material is more homogeneous or vice versa.For in-stance,we consider a material with a mean strength of 200MPa.If the material has a homogeneity indexðhÞof 30,the distribution probabilities Uð30;200Þwill be close to zero and unity for the strengths of RVE less than160 and210MPa,respectively as shown in Fig.1b.In an-other case,if it has a homogeneity indexðhÞof1.1,the corresponding distribution probabilities Uð1:1;200Þwill become0.54and0.64for the strengths of RVE less than 160and210MPa,respectively.From these strength distributions,it is evident that increasing heterogeneity of a material will increase both the diﬀerence in me-chanical properties among the RVEs,and the popula-tion of the RVE with lower strengths.The strength,elastic modulus and Poisson ratio are randomly allocated to each RVE so to account for the inherent variability in phase materials,using the Monte-Carlo method.A more detailed introduction and ex-planation to the material model were reported in our previous publications[11–13].The thermal properties (CTE)of the phase materials are assumed to be uniform and location-invariant,and only depend on the indi-vidual phase.2.2.Mesoscopic thermoelastic damage(MTED)modelIt has been known that the thermal damages of a heated concrete is a complex problem.There are a number of aﬀecting factors,such as thermal mismatch, temperature gradient,degradation of mechanical prop-erties of cementitious materials due to chemical de-composition,and pore water pressure,that cause such damages.However,the focus of this paper is on the damage caused by diﬀerential thermal strains as aresultof diﬀerent CTEs of the phase materials(matrix and inclusion).Studies[8,14]show that the macroscopic fracture of materials is always related to the initiation and propagation of cracks at a meso-scale.Hence,it is assumed that the damage of a cement-based composite is due to the cracking caused by thermal induced stresses at a meso-scale.The bond between the matrix and the inclusion is considered to be perfect.In fact,the pro-posed model can be further modiﬁed to incorporate the eﬀects of temperature gradients and temperature-de-pendent mechanical properties,pending on the avail-ability of experimental data to quantify the associated simulation functions,details of which are under inves-tigation by the authors of this paper.In the numerical modeling,each phase material is discretized into many RVEs with a suitable charac-teristic length.In general,the precision of computa-tional results will increase with decreasing RVE size,at the expense of longer computational time.The RVE has the same size as the meshedﬁnite element in this paper.It is also assumed that the stress–strain rela-tionship of a RVE is linearly elastic till its peak-strength is reached,and thereafter follows an abrupt drop to its residual strength.Cracking is treated as a smeared phenomenon.That is,a crack is not consid-ered as a discrete displacement jump,but rather changing the properties of the RVE according to a continuum law,such as damage mechanics.Although this modeling approach might appear to be crude, however,the complex failure phenomenon(such as compressive and tensile failure)and the nonlinear be-havior in a macro-scale have been proved to be suc-cessfully simulated using the material heterogeneity [11].The behavior laws of the RVE are implemented by introducing a MTED variable D into a constitutive relationship.Based on the above-mentioned ideas and the damage mechanics[21],the general form of an eﬀective stress for a given state of damage for a RVE can be expressed as follows:r¼ð1ÀDÞÁE0Áe rð2Þwhere r is the eﬀective stress,D is the damage variable, E0is the elastic modulus at a reference/undamaged condition(such as at reference temperature),and e r is the strain.Under a uniform temperatureﬁeld,the damage is induced both by diﬀerential thermal strains and by the temperature increment D T,the general ex-pression of damage variable is D¼Dðe r;D TÞ.Let D m and D T denote the damages by the thermal strain and temperature increment,respectively.They can be ex-pressed in terms of the stiﬀness degradation as follows: D m¼1ÀEðe rÞE0ð3ÞD T¼1ÀEðD TÞE0ð4Þwhere Eðe rÞand EðD TÞare the elastic modulus corre-sponding to a given thermal strain e r and the elastic modulus at temperature increment of D T,respectively. If they are independent,the damage variable Dðe r;D TÞcan be expressed as follow:Dðe r;D TÞ¼1Àð1ÀD mÞÁð1ÀD TÞDðe r;D TÞ¼1ÀEðe rÞE0ÁEðD TÞE0ð5ÞSince the temperature-dependent properties are not considered,the damage D T is equal to zero.According to the description of the damage process of a material by Mazars[14]and Yu[15],the thermal induced damage before and after the peak-strength can be determined by the thermal strain and the temperature increment D T102Y.F.Fu et al./Cement&Concrete Composites26(2004)99–111through a separation function,respectively.Fig.2shows a general constitutive relationship of a RVE under thermal loading.At a temperature increment of D T ,the initial thermal strain e thermal is equal to a ÁD T ,and the damage at any given thermal strain can be calculated from Eq.(6)D ðe ;D T Þ¼0;e thermal 6e 6e r 01Àn ðe r 0Àa ÁD T Þðe Àa ÁD T Þ;e P e r 0under compression1;e P e r 0under tension8<:ð6Þwhere D ðe r ;D T Þrepresents the thermal damage with respect to the thermal strains.e r 0is the strain at peak-strength;n ð¼S r =S Þis the coeﬃcient of residual strength for a RVE,S and S r are the peak-strength and residual strength,respectively.Under compression,n is less than 1but greater than 0.Under tension,n is equal to 0.When the strain e becomes smaller than or equal to e r 0,the RVE is undamaged and intact,and D ¼0.When the strain e is larger than e r 0,and under a compressive state,the RVE is damaged,i.e.D >0,and damage variable shall be calculated by the residual strength.Under a tensile state,the RVE is fully damaged and does not sustain any load,and D ¼1.The behavior for a given state of thermal induced damage can be represented by r ¼½1ÀD ðe r ;D T Þ ÁE 0Áðe r Àe thermal Þð7ÞHence,substituting Eq.(6)into Eq.(7),a mesoscopic nonlocal damage model,which can describe the com-plete thermal induced damage process,is expressed as:r ¼E 0Áðe Àa ÁD T Þ;e thermal 6e 6e r 0n ÁE 0Áðe r 0Àa ÁD T Þ;e P e r 0under compression0;e P e r 0under tension8<:ð8ÞIn order to simulate the thermal damage induced by thermal tensile or compressive stresses,a failure crite-rion,which can consider the eﬀects of both tension and compression,is necessary.In this study,the Mohr–Coulomb criterion with tension cutoﬀ[16]is chosen as the criterion of cracking:r 1À1þSin h r 2P S c if r 1P S c 1À1þSin h Á1ÀÁor r 26ÀS t if r 16S c 1À1þSin h 1ÀSin h Á1k ÀÁ8<:ð9Þwhere S c and S t are the uniaxial compressive strengthand tensile strength respectively,S t ¼Àk ÁS c ,and k is the ratio of tensile strength to compressive strength.h is the friction angle of the material.All these parameters can be obtained experimentally.r 1and r 2are the maximum and minimum principal stresses respectively.A compressive stress is positive,and a tensile stress is negative.Finally,a ﬁnite element program T-MFPA,incor-porating the above-mentioned MTED model and failurecriteria,was developed based on the Material Failure Process Analysis (MFPA)program [11,12],using a four-node isoperimetric element.2.3.Numerical specimensNumerical tests of ﬁve specimens (one circular spec-imen and four square specimens)using the T-MFPA program are reported in following sections.Let a i de-note the CTE of the inclusion and a m be the CTE of the matrix.The specimens were analyzed under a plain stress condition without external loading.Specimen no.1is a circular specimen comprising two diﬀerent homogeneous phase materials (matrix and in-clusion,see Fig.3a).It is numerically heated under a uniform temperature ﬁeld of 50°C,and free boundary conditions.The numerical thermal stresses determined from the proposed program are compared with those derived from the classical theory of thermoelasticity,from which the validity of the MTED model in an elastic and undamaged state can be justiﬁed.The me-chanical and geometrical properties of the phase mate-rials are listed in Table 1.In this case,the CTE of the inclusion is greater than that of the matrix.A homoge-neity index h ¼300is chosen so that the phase materials are basically homogeneous in nature.The numerical results are shown in the following section.Y.F.Fu et al./Cement &Concrete Composites 26(2004)99–111103In order to determine the eﬀects of material hetero-geneity,material strength,and CTE on the stress de-velopment and the process of thermal cracking around a single inclusion,four square specimens (Specimens no.2to no.5,see Fig.4)with diﬀerent thermal and me-chanical properties (see Table 2)are numerically stud-ied.Basically,the specimens can be classiﬁed into two groups.In Group 1(Specimens no.2and no.3),the CTE of the matrix is smaller than that of the inclusion.In Group 2(Specimens no.4and no.5),the CTE of matrix is larger than that of the inclusion.Within a group,the only variable is the mean strength of the in-clusion.The four specimens have the same homogeneity index h equal to 3,representing a high degree of heter-ogeneity.They are subjected to a uniform temperature increment from 20to 620°C at an incremental step of 10°C.3.Model validationFig.3b shows the comparison of the thermal stresses around the single inclusion of Specimen no.1calculated from the T-MFPA program,and from the analytical solutions (Eqs.(10)and (11))derived from the classical theory of thermo-elasticity [6,17].It is evident that under an elastic and undamaged state,an excellent agreement between the stresses ob-tained from the two diﬀerent approaches has been ob-tained.4.Thermal cracking history of square specimens Fig.5shows the eﬀect of thermal mismatch on the thermal induced damages and fracture processes of Specimen no.2(Group 1)and Specimen no.4(Group 2)due to increasing temperatures.Fig.6illustrates the inﬂuence of the mean strength of the inclusions on the crack development in each group.Detailed descriptions of crack formation of the specimens are shown below.4.1.Thermal cracking in composite of a i >a mIn the case of Specimen no.2,since the a i (CTE)of the inclusion is greater than that of the matrix ða m Þ,the incompatibility of thermal deformation at the interface between the matrix and the inclusion leads to the stress concentration around the inclusion (see Fig.5a(a)).The inclusion is under a statistically hydrostatic compres-sion,and the matrix is under a combination of com-pression and tension.When the temperature reaches 200°C,a few broken elements randomly occur (due to heterogeneity)in the high stress zone around the inclu-sion.With increasing temperatures,the number of the diﬀused damaged elements increases.The damaged ele-ments exist in both the high stress zone and in the low stress zone,but most of them are located near the for-Table 1Material properties of circular Specimen no.1ParameterValue Matrix Inclusion Heterogeneity index (h )300300Mean elastic modulus (MPa)60,000100,000Mean compressive strength (MPa)3060Poisson ratio0.250.20Coeﬃcient of thermal expansion (/°C) 1.0E )5 1.1E )5Temperature increment (°C)1010Tension cutoﬀ0.10.1Frictional angle (°)3030Diameter (mm)10020Number of elements31,4001256Fig.4.Numerical square specimen with single inclusion.Table 2Material properties of square Specimens no.2to no.5ParameterValue Matrix Inclusion Heterogeneity index (h )33Mean elastic modulus (MPa)60,000100,000Mean compressive strength (MPa)Specimen no.2200300Specimen no.3150Specimen no.4300Specimen no.5150Poisson ratio0.250.20Coeﬃcient of thermal expansion (/°C)Specimen no.2 1.0E )51.1E )5Specimen no.3 1.1E )5Specimen no.40.9E )5Specimen no.50.9E )5Temperature increment (°C)1010Tension cutoﬀ0.10.1Frictional angle (°)3030Dimension (mm)100Â100U 30Number of elements200Â2001412104Y.F.Fu et al./Cement &Concrete Composites 26(2004)99–111merly broken elements in the high stress zone (see Fig.5a(c)).When the temperature increases to 430°C,a macro-crack is formed ﬁrstly at the top-left area around the inclusion.At the same time,only a few of cracks nucleate far away from the high stress zone around the inclusion (see Fig.5a(d)).As the temperature further increases,the broken elements around the inclusion nucleate into several discontinuous macro-cracks (see Fig.5a(e)and (f)),and simultaneously corresponding tensile stress zones are formed at the tips of these cracks.Bridges are formed between the cracks due to the fact that many small cracks simultaneously grow at diﬀerent locations caused by the heterogeneity.This phenomenon is also described by Van Mier [19].As the temperature rises to 570°C,all these macro-cracks further propagate under the tensile stresses at their tips,followed by the occurrence of dispersed damaged elements in the frac-ture process zone.During the heating process,the macro-cracks are formed in the way that the discontin-uous cracks continue to grow and bridges are formed.The shapes of these cracks are irregular,rough and bi-furcate (see Fig.5a(e)–(h)).The macro-cracks formed along the radial direction around the inclusion can be called ‘‘radial cracks’’,which were also evident intheFig.5.(a)Thermal cracking of cement-based composite of Specimen no.2(inclusion diameter ¼30mm and a i ¼1:1Â10À6/°C).(b)Thermal cracking of cement-based composite of Specimen no.4(inclusion diameter ¼30mm and a i ¼0:9Â10À6/°C).Y.F.Fu et al./Cement &Concrete Composites 26(2004)99–111105experiments reported by Zhou et al.[18]and Golter-mann [5].It is also noted that when the main macro-cracks begin to propagate,the pace of minor crack develop-ment is slow down (see Fig.5a(g)and (h)).Fig.6a shows the thermal fracture process of the companion Specimen no.3with a lower mean strength ðr i3Þof the inclusion than that ðr i2Þin Specimen no.2.It is evident that the variation of the mean inclusion strength does not aﬀect the patterns of thermal damage initiation and propagation.4.2.Thermal cracking in composite of a i <a mIn the case of Specimen no.4,since the a i (CTE)of the inclusion is smaller than that of the matrix ða m Þ,azone of stress concentration also occurs around the in-clusion.The inclusion is stressed under tension and the matrix is under a combination of tension and com-pression (see Fig.5b(a)).When the temperature reaches 160°C,a few of the damaged elements distribute dis-orderly inside the inclusion.With increasing tempera-tures,the number of broken elements grows,and a few of them occur in the stress concentration zone outside the inclusion (see Fig.5b(c)).When the temperature increases to 400°C,the broken elements at the interface between the matrix and the inclusion nucleate and form several small discontinuous cracks.As the temperature becomes further higher,the discontinuous cracks at the interface propagate gradually and coalesce with the stress transferring from the inclusion and the matrix nearby the inclusion to the tips of the cracks (seeFig.Fig.6.(a)Thermal cracking processes of specimens in Group 1.(b)Thermal cracking processes of specimens in Group 2.106Y.F.Fu et al./Cement &Concrete Composites 26(2004)99–1115b(e)–(g)).Eventually,after the temperature has reached 620°C,most of all the elements around the interface between the matrix and the inclusion are broken and a nearly close circular macro-crack is formed at the in-terface.The high stress distributing inside the inclusion is transferred into the crack tips.This kind of crack is called‘‘tangential crack’’,which is also observed in the experiments by Zhou et al.[18]and Goltermann[5].Fig.6b demonstrates the thermal fracture process of the companion Specimen no.5with a lower mean strengthðr i5Þinclusion than thatðr i4Þin Specimen no.4. Although the thermal damage initiation and crack propagation of the two specimens are similar,the number of the broken elements and the kinds of cracks at each temperature level are diﬀerent.At a lower tem-perature,more elements in the inclusion of Specimen no. 5are damaged than those in Specimen no.4(see Fig. 6b(a)and(a0)).When the temperature reaches360°C, the macro-cracks pass through partly or wholly the in-clusion,and high stresses previously distributed around the inclusion are transferred into the tips of these cracks (as shown in Fig.6b(b0)).With increasing temperatures, these discontinuous cracks nucleate and coalesce with the redistributing stressﬁeld(as shown in Fig.6b(c0)and (d0)).This kind of crack occurred inside the inclusion is called‘‘inclusion crack’’.5.Thermal stressﬁelds of square specimens5.1.Eﬀect of thermal mismatchAlthough the four specimens are subjected to uniform temperature changes,local stress concentration occurs around the inclusion due to the thermal mismatch be-tween the matrix and the inclusion.When the CTE of the inclusion is greater than that of the matrix,the inclusion in Specimen no.2is stressed under a state of statistically hydrostatic compression due to the restriction from the matrix,and the matrix is under a general bi-axial state of stresses(tensile/com-pressive and shear stresses)due to the outward expan-sion from the inclusion.The distribution of maximum and minimum principal stresses and the maximum shear stress along the mid-section of Specimen no.2can be shown in Fig.7a(a).Although the maximum and mini-mum principal stresses in the inclusion are high,the maximum shear stress is much smaller so that few ele-ments with lower strength in this area reach their failure strength.The absence of tensile stresses in the inclusion delays the attainment of the Mohr–Coulomb with ten-sion cutoﬀfailure criterion.Unless the inclusion is ab-normally weak in compression,the strength of inclusion has no eﬀect on damage initiation(see Fig.6a).As a result,most of the diﬀused damages distribute in the high stress zone of the matrix around the inclusion for Group1specimens.With increasing temperatures,these broken elements nucleate and form several discontinu-ous cracks due to the stress redistribution at the crack tips.Since the minimum principal stress is nearly per-pendicular to the radial direction of the inclusion and is in tension,these cracks are developed in the manner of radial cracks in the matrix.When the CTE of the inclusion is smaller than that of the matrix,the inclusion in Specimen no.4is stressed under bi-axial tension,and the matrix remains in a state of compressive/tensile and shear stresses(see Fig.7b). Since a bi-axial tension leads to early attainment of the failure criterion,it is not surprised that the initiation of damage takes place only in the inclusion of Group2 specimens.In such a case,the strength of the inclusion has considerable eﬀects on the crack formation.That is, a weaker inclusion will have damage initiated at a lower temperature and grow more rapidly at high tempera-tures(see Fig.6b).The minimum principal stress in the matrix is parallel to the radius direction of the inclusion and is in tension,so that the main cracks propagate in the manner of tangential cracks at the matrix–inclusion interfacial region.5.2.Eﬀect of heterogeneity at meso-scaleThe thermal stressﬁelds are shown in Figs.4–6.The bright color indicates the higher stress,and vice versa.It is found that the points with diﬀerent scale colors exist in a same zone.It means that there are existing points subjected to diﬀerent stresses due to the heterogeneity at meso-scale in such zone,where the stressﬁeld is statis-tically uniform at macro-scale.The ratio of the local stress to the local strength is a very important parameter which can be used to decide whether or not an element fails.The eﬀect of the heterogeneity at meso-scale can be reﬂected by the stressﬂuctuation shown in Fig.7.In comparison with the results from Fig.3,the curves of stress distribution along the mid-section E–E in Speci-mens no.2to no.5before crack initiating are charac-terized by an irregular variation of stress values(see Fig. 7a and b).Such a strong thermal stressﬂuctuation in a heterogeneous composite,which can be quantitatively identiﬁed in our numerical study,is diﬃcult to be de-termined by experiments.Taking into account of the material heterogeneity, the failure of a material is dependent both on the in-duced stress level and on the strength itself.An element subjected to high stress may not break due to the fact that this element has higher strength;whereas an ele-ment subjected to low stress may break because of its low strength.These kinds of failure are deﬁnitely dif-ferent,since their released energies are diﬀerent.Con-sequently,some RVE can still remain un-fractured in a zone of high stress,if these elements have higherY.F.Fu et al./Cement&Concrete Composites26(2004)99–111107。

直流导叶式旋风分离器内气相流动的数值模拟_王建军

文章编号:1000-7466(2011)05-0013-04直流导叶式旋风分离器内气相流动的数值模拟王建军1,陆文龙2,高文山1,金有海1(11中国石油大学机电工程学院,山东东营257061;2.中海油气电集团有限公司,北京100027)摘要:利用Fluent软件对直流导叶式旋风分离器内气相流场进行了数值模拟研究。

结果表明,在分离空间内,气流流动比较稳定,切向速度呈兰金组合涡分布;在排尘环隙内出现气流分层,一部分气流携带颗粒进入灰斗,而另一部分气流在此处进行二次流动;在排气管内,在变径的开始区域存在二次涡流,而在后边段流动比较稳定。

关键词:直流导叶式旋风分离器;气相流场;数值模拟;三维速度中图分类号:TQ051.804文献标志码:ANumerical Simulation of Flow Filed in Uniflow Guide Vanes TubeW ANG Jian-jun1,LU Wen-long2,GAO Wen-shan1,JIN You-hai1(1.Colleg e of M echanical and Electronic Eng ineer ing,China U niversity of Petr oleum,Do ng ying257061,China; OOC Gas&Pow er Gro up,Beijing100027,China)Abstract:With the help o f Fluent softw ar e,the flow filed in uniflow g uide vanes tube w as in-v estig ated w ith numerical simulatio n.The numerical result indicated that:in the separate zone, the air flow stability in the separ ate zone,the tangential velocity is like Rankine vo rtex,the air separate in the ring-g ap in two parts,one part flow in the ash hoo per w ith the dust,and the sec-o nd par t flow in the seco ndar y backset,in the v ent-pipe,due to the chang e of diam eter,ther e are secondary backset in the first par t,and the fluid flow stability in the behind zone.Key words:ax ial flow cyclone;flow field;numerical simulatio n;thr ee-dimensional velocity在天然气的净化过程中,分离装置主要用于清除天然气中的岩屑、沙粒、液滴和其他有害杂质[1]。

橡胶密封元件轴对称结构在ABAQUS中的数值模拟---HOFF

橡胶密封元件轴对称结构在ABAQUS中的数值模拟和强王宇火进（北京霍夫技术服务有限公司北京100083）摘要：运用ABAQUS/Standard模拟了橡胶密封元件装配过程中的受力特性；实现了用数值方法解决包含超弹性材料，轴对称结构，摩擦接触以及多步骤分析的问题；为复杂工程问题的数值解法提供了参考价值。

关键词：橡胶密封超弹性轴对称多步骤分析引言橡胶是在使用温度下处于高弹态的高分子材料，变形中表现出很强的几何物理非线性，与其他材料最基本的区别是其弹性模量低，具有很高的伸缩性和储能能力，因此，被广泛的用来制成密封，减振，防护等用品。

传统的橡胶元件的设计方法是根据试验和试验所取得的规律进行，造成产品开发周期长，成本高，复杂结构或工况试验难以进行等后果。

近年来，计算机仿真技术和现代非线性理论的快速发展为设计人员提供了良好的基础。

本文借助有限元理论和分析软件ABAQUS对汽车转向轴橡胶密封圈的安装过程进行了研究。

1，油封结构简介目前中国已经标准化的产品有内包骨架油封（B型），外露、半外露骨架油封（W型），装配式油封（Z型），有副唇内包骨架油封（FB型），有副唇外包骨架油封（FW型），有副唇装配式油封（FZ型）以及流体动力型旋转油封等等。

其中外露、半外露骨架油封具有油封定位准确，同轴性好，安装性能高，摩擦升热小，导热性能好和材质消耗小等优点。

油封结构包括结构形式和结构参数两部分。

结构形式（油封结构和几何形状）是由配合及安装要求，密封介质及轴的旋转方向等使用要求来决定；而结构参数的内容包括工作情况参数，性能参数，胶料性能参数和剖面结构参数等。

2，橡胶变形特点及超弹性本构模型2．1 橡胶变形特点橡胶材料在承受拉压过程中，体积的变化量很小，可以忽略不计，但是作为超弹性体，其最明显也是最重要的物理特性是在较小的外力作用下就能产生很大的变形，其伸长率可达100％～1000％。

2．2 超弹性本构模型弹性材料的变形过程是可逆的，如无其他不可逆伴随，单纯的弹性变形过程的熵产率为零，也就是单位质量的热力学能等于单位质量的应变能，对于等温过程，单位质量的自由能便是单位质量的应变能，存在应变能的材料称为超弹性材料，因此橡胶材料属于超弹性材料。

面向虚拟制造的塑料注射成形CAE模拟

迅速，它在全面提升塑件及模具设计水准中的显著效果已逐渐为模具界所认识，促进了模具行业的进步。但从总体上看，现有的 CAD/CAE/CAM 软件仍存在着塑件设计与模具设计相互独立、 CAD 与 CAE 模型不统一、缺乏对设计过程的指导作用、集成度差等不足。塑料注射件虚拟制造系统的研究与开发是解决上述缺点、推动计算机辅助设计技术在塑件及模具设计中的应用更为广泛与深入的有效途径。 CAE 分析是以设计为中心的虚拟制造系统的重要组成部分，作为对设计进行精确评价和分析的工具，可以使设计过程不再仅仅依靠以往积累的经验知识，通过 CAE 与 CAD 等技术的有效集成，共同实现“设计—评价—再设计”任务的自动化，提高设计的精确程度和效率。但在注射成形 CAE 软件集成到 VM 系统的过程中，目前存在的主要困难有： ① 如何实现基于产品实体/表面模型的 CAE 分析；② 如何从 CAE 分析的结果出发评价和修改设计。针对上述两个问题，本文建立了基于实体/ 表面模型的智能 CAE 分析系统。
x,y —— 中性层面上的二维坐标 z —— 壁厚方向 δ—— 型腔半壁厚 u, v —— x,y 方向的速度
2
Cm-I = Cm-II (10) 式中下标Ⅰ,Ⅱ —— 同一截面处对应的两部分 Cm-I, Cm-II —— 两部分的自由移动流动前沿（如图 4 所示）
_______________________________________________________________________________ 中国科技论文在线
型腔和型芯中间的层面，其模拟过程如图 2 所示。基于这种型腔模型的 CAE 软件在应用中具有很大的局限性，主要表现在：① CAE 软件的使用人员

Optimization of the cyclone separator geometry for minimum pressure drop

a b s t r a c t
The response surface methodology has been performed based on the Muschelknautz method of modeling (MM) to optimize the cyclone geometrical ratios. Four geometrical factors have signiﬁcant effects on the cyclone performance viz., the vortex ﬁnder diameter, the inlet width and inlet height, and the cyclone total height. There are strong interactions between the effect of inlet dimensions and vortex ﬁnder diameter on the cyclone performance. CFD simulations based on Reynolds stress model are also used in the investigation. A new set of geometrical ratios (design) has been obtained (optimized) to achieve minimum pressure drop. A comparison of numerical simulation of the new design and the Stairmand design conﬁrms the superior performance of the new design compared to the Stairmand design. & 2010 Elsevier Ltd. All rights reserved.

Simulationoftemp...

Simulationoftemp...Trans. Nonferrous Met. Soc. China 24(2014) 2168?2173Simulation of temperature and stress in6061 aluminum alloy during online quenching processMeng-jun WANG 1,2, Gang YANG, Chang-qing HUANG 2, Bin CHEN 11. Key Laboratory of Nonferrous Metal Materials Science and Engineering of Ministry of Education,Central South University, Changsha 410083, China;2. State Key Laboratory of High-performance & Complicated Manufacturing,Central South University, Changsha 410083, ChinaReceived 17 October 2013; accepted 17 April 2014Abstract: The cooling curves of 6061 aluminum alloy were acquired through water quenching experiment. The heat transfer coefficient was accurately calculated based on the cooling curves and the law of cooling. The online quenching process of complex cross-section profile was dynamically simulated by the ABAQUS software. The results suggest that the heat transfer coefficient changes during online quenching process. Different parts of the profile have different cooling velocity, and it was verified by water quenching experiment. The maximum residual stress of the profile was predicted using FEM simulation based on ABAQUS software. The relations between the temperature and stress werepresented by analyzing the data of key points. Key words: 6061 aluminum alloy; quenching; complex cross-section profile; residual stress; ABAQUS1 Introduction6061 aluminum alloy is widely used in the production of large scale complex cross-sections architecture profiles and industrial profiles due to its characteristics of moderate intensity, nice plasticity, favorable solderability and corrosion resistance [1,2]. The complex cross-section profile has a wide range of application in the aerospace, train and hull structure. As for variations in thickness, shape (especially asymmetric structure) and the inability to cool inside the hollow section, there is non-uniform cooling both across the section and along the length of the section during the online quenching process [3,4]. This non-uniform cooling may lead to large temperature gradients, and cause high residual stresses and thermally induce distortions, such as warping and twisting [5?8]. This may severely affect the property and the precision of the products, and reduce the production. Therefore, it is very necessary to investigate the residual stresses of 6061 alloy parts after quenching. In order to predict the temperature and stress distributions of the complex cross-section profile in the online quenching process, ABAQUS/standard was applied to simulate the performance of the profile [9?11]. YANG et al [12] simulated the temperature field and residual stress of large complicated thin-wall workpieces by finite element method. LI et al [13] studied the temperature and stress fields of Ti-alloy thin-well barrel during quenching process. The heat transfer coefficient (HTC) is an important factor for quenching process, so it must be taken into account [14?16].In this work, the study combining FEM with experimental methods was performed to analyze the temperature and stress distributions of a circular pipe of 6061 aluminum alloy. The cooling curves of 6061 aluminum alloy were obtained in water quenching. Based on the cooling curves, the HTC curves were solved. The temperature and stress distributions of a circular pipe of 6061 aluminum alloy were simulated by ABAQUS software, and the basis for preferable quenching techniques was offered.2 Experimental2.1 Quench experiment6061 aluminum alloy was used for quenchingFoundation item: Project (zzyjkt2013-10B) supported by the Foundation of State Key Laboratory of High-performance & Complicated Manufacturing,China; Project (51275533) supported by the National Natural Science Foundation of ChinaCorresponding author: Meng-jun WANG; Tel: +86-731-88836408; E-mail: 347468230@/doc/0a13963566.html DOI:10.1016/S1003-6326(14)63328-8Meng-jun WANG , et al/Trans. Nonferrous Met. Soc. China 24(2014) 2168?2173 2169experiment to acquire heat transfer coefficient. Table 1 lists its chemical compositions. The shape and size of sample are shown in Fig. 1. Holes A and B were installed with thermocouple to get temperatures in the quenching process. Recycling water of 25 °C was used as quenchant. Figure 2 shows the corresponding cooling curves of Points A and B .Table 1 Chemical compositions of tested 6061 aluminumalloy(mass fraction, %) Si Mg Fe Cu Mn0.62?0.67 0.95?1.00 ≤0.35 0.17?0.21 0.10?0.15Cr Zn TiAl0.05?0.10 ≤0.05 0.02?0.05 Bal.Fig. 1 Schematic diagram of quenching sample (unit: mm)Fig. 2 Cooling curves of 6061 aluminum alloy in water quenching2.2 Calculation of HTCPrecise HTCs are the key boundary conditions in simulating the quenching process, which was acquired from the past research results [17].The quenching sample fits one-dimensional heat transfer model. Based on one-dimensional unsteady heat conduction differential equation, the temperature of the quenching surface can be defined as follows:)]()([)()()(2)(112210t T t t T tx C t T t T t T p ?Δ+ΔΔ+?λρ＝ (1)where T 0(t ), T 1(t ) and T 2(t ) are the temperatures of quenching surface, the Points A and B , respectively; T 1(t+?t ) refers to the temperature of Point A at momentt+?t ; λ, ρ, C p are the thermal conductively, the mass density and the specific heat capacity, respectively; Δt and Δx refer to the temperature and displacement change amount.Based on the Fourier law and Newton’s law of cooling, the relationship between the quenching time t , face temperature T 0(t ) and HTC h w (t) can be described as])([)()()(w 001w s w T t T x t T t T T T q t h ?Δ=λ＝ (2)where T s is the temperature before quenching; T w is the temperature of quenchant; q is the internal heat source density. According to the cooling curves, together withEqs. (1) and (2), h w ?T 0 curve can be got, as shown in Fig.3.Fig. 3 HTC curves of 6061 aluminum alloy in quenching surface2.3 Numerical simulationABAQUS/standard with coupled thermal displacement analysis was applied to simulate the quenching process. Since the cooling along the length axis was uniform, a 2D model was used to reduce the computational time. The scale of the model was in agreement with the real profile, as shown in Fig. 4. The physical properties of the alloy were considered to be constant in the models and listed in Table 2. The yield stresses at differenttemperatures are listed in Table 3 [18]. Element type was CPE4T and mesh size was properly selected. Reference Points A ?H were also defined to extract data for analyzing the process.3 Results and discussion3.1 Temperature distributionThe temperature distributions in the quenching process at different time are shown in Fig. 5, from which, the maximum cooling velocity occurs in external edge of four ribs around the profile, while the joints of the internal face and ribs have the slowest cooling velocity.Meng-jun WANG , et al/Trans. Nonferrous Met. Soc. China 24(2014) 2168?21732170Fig. 4 Profile (a) and its FEM model (b)Table 2 Material parameters of 6061 aluminum alloyConductivity/(W·m ?1·K ?1)Density/ (g·cm ?3) Specific heat capacity/ (J·kg ?1·K ?1) ExpansioncoefficientElastic modulus/GPa Poisson ratio180 2.789610?5 40 0.35Table 3 Yield stresses of 6061 aluminum alloy at differenttemperaturesTemperature/°C 24 100 150 200 260 316 371 Yield stress/MPa276 262 214 105 34 19 12Fig. 5 Temperature distributions of profile model during quenching process at different time: (a) 0.5 s; (b) 1 s; (c) 2 s; (d) 3 s; (e) 5 s;(f) 7 sAs the temperature difference between the quenching medium and profile increases, the heat exchanges rapidly and the temperature drops sharply. The temperature difference of the profile maintains 100 °C in the first 2 s and then decreases apparently. Finally, the temperature of the entire profile tends to be uniform. As largetemperature difference is prone to form residual stress, theheat induced stresses will concentrate in the rib, and the peak value appears within 2 s.Figure 6 shows the cooling curves of key Points A ?H of 6061 aluminum alloy profile at the water flow of 0.32 m 3/h. Among the internal points A ?D , the maximumMeng-jun WANG, et al/Trans. Nonferrous Met. Soc. China 24(2014) 2168?2173 2171cooling velocity appears in Points A and D, and the cooling velocity of Point C is smaller than that of the other three points. However, among the external Points E?G, Point G has the maximum cooling velocity, and point H has the smallest cooling velocity. It fits well with the cooling law which reflects in the temperature distributions of the profile model, as shown in Fig.5.3.2 Stress distributionThe stress distributions of the profile in different quenching time are shown in Fig. 7. At the beginning, the external walls of the pipe and ribs suffer tensile stress, however, the internal walls are under compressive stress, those stresses reverse as the temperature decreases. From the stress contours, it can be seen that stress concentration appears mainly in the joints of the pipe wall and ribs. The maximum residual stress exists around the joints of internal wall and ribs when the quenching process is over.To further illustrate the relations between the temperature and the stress of the profile during the quenching process, data of the key Points A?H areFig. 6 Cooling curves of key Points A?H of 6061 aluminum alloy profile: (a) Internal Points A?D; (b) External Points E?GFig. 7 Stress distributions of profile model during quenching process at different time: (a) 0.1 s; (b) 0.5 s; (c) 1 s; (d) 2 s; (e) 4 s;(f) 6 sMeng-jun WANG, et al/Trans. Nonferrous Met. Soc. China 24(2014) 2168?2173 2172extracted. The temperature?stress curves are shown in Fig. 8. Points A?D suffer compressive stress firstly and then change to tensile stress. Finally, residual tensile stresses exist after quenching. Points A?D get to the peak compressive stress before500 °C. When Points B and C get to the peak stress value, the corresponding temperature is even higher than that of Points A and D, while the peak stress values of Points A and D are larger than that of Points B and C. The peak compressive stresses of Points A and D in the initial stage are 1.5 and 1.3 MPa, respectively. And the residual tensile stresses after quenching are 8.2 and 7.7 MPa, respectively. On the contrary, stress states of Points E?H are tensile firstly, then change to compressive stress. In the first place, Points E and G reach the peak stress values of 1.7 and 2.8 MPa, respectively. Finally, residual compressive stresses reach 17.4 and 21.7 MPa, respectively.Fig. 8 Temperature?stress curves of key points during quenching process: (a) Internal Points A?D; (b) External Points E?G4 Conclusions1) The value of heat transfer coefficient is low at the beginning of the quenching process, however, with the temperature decreasing, it rises till the peak of 20 kW/(m2·K) and then decreases.2) The maximum cooling velocity appears in the external edge of ribs around the profile, while the joints of the internal surface and ribs have the slowest cooling velocity. The temperature difference of the profile is about 100 °C in the first2 s, then decreases apparently. And the temperature of entire profile tends to be uniform. The cooling curves of key points fit well with the cooling law which reflects in the temperature distributions of the profile model.3) The external walls of the pipe and ribs suffer tensile stress firstly, while the internal walls are under compressive stress, with the temperature decreasing, the stresses reverse. The maximal residual compressive stress and tensile stress in the quenching process are 28.7 and 21.7 MPa, respectively.References[1]JI Yan-li, GUO Fu-an, PAN Yan-feng. Microstructuralcharacteristics and paint-bake response of Al?Mg?Si?Cu alloy [J].Transactions of Nonferrous Metals Society of China, 2008, 18(1):126?131.[2]PAN Dao-zhao, WANG Zhi-xiu, Li Hai, ZHENG Zi-qiao. Effects oftwo-step ageing treatment on tensile properties and intergranularcorrosion of 6061 aluminum alloy [J]. The Chinese Journal of Nonferrous Metal, 2010, 20(3): 435?441. (in Chinese)[3]SHANG Bao-chuan, YIN Zhi-min, WANG Gang, LIU Bo.Investigation of quenching sensitivity and transformation kineticsduring isothermal treatment in 6082 aluminum alloy [J]. Materialsand Design, 2011, 32(7): 3818?3822.[4]BIKASS S, ANDERSON B, PILIPENKO A, LANGTANGEN H P.Simulation of the distortion mechanisms due to non-uniformcoolingin the aluminum extrusion process [J]. International Journal ofThermal Sciences, 2012, 52(1): 50?58.[5]LI Hong-ying, ZENG Cui-ting, HAN Mao-sheng, LIU Jiao-jiao, LUXiao-chao. Time?temperature?property curves for quench sensitivityof 6063 aluminum alloy [J]. Transactions of Nonferrous MetalsSociety of China, 2013, 23(1): 38?45.[6]RUUD C O. Residual stresses and their measurement, quenching anddistortion control [C]//Proceedings of the First International Conference on Quenching and Control of Distortion. Chicago: ASMInternational, 1992: 193?198.[7]SARMIENTO G S, CASTRO M, TOTTEN G E, HARVIS L,WEBSTER G., CABR M F. Modeling residual stresses in spring steelquenching [J]. Heat Treatment of Metals, 2003, 28(11): 47?52.[8]TANNER D A, ROBINSON J S. Residual stress prediction anddetermination in 7010 aluminum alloy forgings [J]. ExperimentalMechanics, 2000, 40(1): 75?82.[9]ZHANG Qing-feng, JIAO Si-hai, MA Zhao-hui. FEM simulation oftemperature field in plate during quenching process [J]. Material andHeat Treatment, 2010, 39(6): 157?160. (in Chinese)[10]XIAO B W, WANG Q G, PARAG J, LI K Y. An experimental studyof heat transfer in aluminum castings during water quenching [J].Journal of Materials Processing Technology, 2010, 210(1): 2023?2028.[11]LI H P, ZHAO G Q, NIU S T, HUANG C Z. FEM simulation ofquenching process and experimental verification of simulation results[J]. Material Science and Engineering A, 2007, 452?453: 705?714. [12]YANG Xia-wei, ZHU Jing-chuan, LAI Zhong-hong, LIU Yong, HEDong, NONG Zhi-sheng. Finite element analysis of quenchingMeng-jun WANG, et al/Trans. Nonferrous Met. Soc. China 24(2014) 2168?2173 2173temperature field, residual stress and distortion in A357 aluminumalloy large complicated thin-wall workpieces [J]. Transactions of Nonferrous Metals Society of China, 2013, 23(6): 1751?1760.[13]LI Yan-zeng, YAN Mu-fu, WU Kun. Numerical simulation oftemperature and stress fields of Ti-alloy thin-well barrel during quenching process [J]. Transactions of Meterials Heat Treatment, 2004, 25(5): 769?772.[14]SHI Zhi-yu. Online quenching process practice of industrialaluminum [J]. Metallurgical Collections, 2010, 185(1): 43?45.[15]CHEN Nai-lu, GAO Chang-yin, SHAN Jin, PAN Jian-sheng, YEJian-song, LIAO Bo. Research on the cooling characteristic and heattransfer coefficient of dynamic quenchant [J]. Transactions of Materials and Heat Treatment, 2001, 22(3): 41?44.[16]LI H P, ZHAO G Q, NIU S T. Technological parameters evaluationof gas quenching based on the finite element method [J].Computational Materials Science, 2007, 40(2): 282?291.[17]CHEN Bin, WEN Liu, PAN Xue-zhu, WANG Yin-xin, GAO Meng,WANG Meng-jun. The experiment research and simulation analysison the online quenching process of the 6061 aluminum alloy [J].Aluminum Fabrication, 2011, 5(1): 18?21. (in Chinese)[18]WANG Zhu-tang, TIAN Rong-zhang. Handbook of aluminum alloyand processing [M]. Changsha: Central South University of Technology Press, 1988: 1015?1017. (in Chinese)6061铝合金在线淬火温度场和应力场模拟王孟君1, 2，杨刚1，黄长清2，陈彬11. 中南大学有色金属材料科学与工程教育部重点实验室，长沙410083；2. 中南大学高性能复杂制造国家重点实验室，长沙 410083摘要：通过淬火实验获得6061铝合金的冷却曲线，实验根据冷却曲线并结合数值方法获得在线淬火换热系数，运用ABAQUS有限元软件动态模拟复杂截面型材在线淬火过程。

simulation (2)精品PPT课件

第2章系统建模的基本方法与模型处理技术
系统是我们研究的对象，模型是系统行为特性的描述仿真则是模型试验。可以说，系统建模是系统仿真的基础，系统模型化技术是系统仿真的核心。计算机仿真的前提
是建立数学模型，本章只要介绍系统数学模型的建模原理，建立方法、以及模型之间的转换和处理方法。
本章教学目的及要求
电力电子与电力传动实验室
Lab of PEED
Bring Ideas Together
5 建模的一般过程
目
目的
标协
调
先验知识
演绎分析
模型构造
可信度分析
最终模型
归
纳程
数据
序
电力电子与电力传动实验室
Lab of Biblioteka EEDBring Ideas Together
2.2确定型系统的数学模型
2.2.1连续时间系统的模型 1.微分方程
熟悉常用的数学模型及其特点
掌握微分方程,传递函数,结构图,状态空间法的基本表达
熟悉各种数学模型之间的相互转换
电力电子与电力传动实验室
Lab of PEED
Bring Ideas Together
概述
相似：相似是一种认识，我们认识的世界是真实客观世界的相似物。是普遍、绝对的。
相似性原理：就是按照某种相似方式或相似原则对各种事物进行分类，获得多个类集合；在每一个类集合中选取一个具体事物对它进行综合性的研究，获得相关规律，这些规律可以推广到集合中的其它事物中去。相似具有自反性、对称性、传递性。
电力电子与电力传动实验室
Lab of PEED
Bring Ideas Together
3 建模方法
• 演绎法通过定理、定义、公理以及已经验证了的理论推演得出数学模型。这是最早的一种建模方法，这种方法适用于内部结构和特性很明确的系统，可以利用已知的定律，如：力、能量等平衡关系来确定系统内部的运动关系，大多数工程系统属于这一类。电路系统、动力学系统等都可以采用这种演绎法来建立数学模型。

Autodesk Simulation CFD 2013 September 24, 2012 Re

September 24, 2012Release Notes - Read Me fileAutodesk, Inc.Installation PreparationDownload LocationInstallation RequirementsInstallation InstructionsIssues Addressed in this Service Pack∙Results∙CAD Connection∙Meshing∙Materials∙Boundary Conditions∙API∙Solver∙Documentation∙InstallationLegal NoticeReturn to Top Prior to installing Autodesk Simulation CFD Service Pack 1, you will need to uninstall any previous versions of Autodesk Simulation CFD 2013.To check if an earlier version has previously been installed, open the Control Panel (Start > Control Panel), select Programs and Features (Add/Remove Programs on Windows XP systems) and look for entries titled Autodesk® Simulation CFD 2013.To uninstall an earlier version of Autodesk® Simulation CFD 2013:1.Verify that Autodesk® Simulation CFD is not running.2.Open the Control Panel (Start > Control Panel).3.Select Programs and Features (Add/Remove Programs on Windows XP systems).4.Select the entry you want to remove and click Uninstall/Change (Change/Removeon Windows XP systems) to launch the uninstall procedure.5.Restart the computer to complete the uninstall of the software.Return to TopYou can download the Autodesk Simulation CFD Service Pack 1 installation by clicking here.Return to Top ∙Autodesk Simulation CFD supports Windows® 7 Home Premium, Professional, Enterprise, Ultimate (SP0 or SP1 for x86 and x64); and Windows XP Professional (x86: SP3, .net 4.0; x64: SP2, .net 4.0 SP1).∙Verify that you have administrator privileges on your local machine to install Autodesk Simulation 2013 SP1.For a complete list of installation requirements, click here.The 32-Bit and 64-Bit versions of Autodesk® Simulation CFD are delivered in separate installation packages. It is important to install the correct version for your system. The 64-bit version of Autodesk® Simulation CFD cannot be installed on a 32-bit system, nor can the 32-bit version be installed on a 64-bit system.Return to Top Autodesk® Simulation CFD uses an Installation Wizard to guide you through the installationprocess.1.Log in to the computer using an account with Administrator privileges.2.Exit any Autodesk® programs that are currently running.3.You can install Autodesk® Simulation CFD from either a downloaded installationexecutable or from an installation DVD.∙If installing from a downloaded executable file, navigate to the file, right click on it, and select Run as Administrator. (For an XP computer, just double-click on theinstallation executable to start the installation.)∙If installing from the DVD, insert it into the DVD drive. If Autorun is enabled, the DVD main menu displays automatically. Otherwise, navigate to the DVD drive folder, right-click on setup.exe, and select Run as Administrator. (For an XP computer, double-click on setup.exe to start the installation.)4.After the Installation Wizard initializes, the Welcome page appears.∙Click Install to begin the installation process.5.The License Agreement page appears.∙Select the appropriate entry in the Country or Region list.∙Review the agreement.∙Select the option to accept the agreement.Note: If you reject the license agreement, you cannot install the software.6.The Product Information page appears.∙Enter a Serial number and Product key for one of the modules that you purchased. (It does not matter which module you use.)∙Select the License Server Model. There are three options:∙Single License Server: In the single license server model, the Network License Manager is installed on a single server. License management and activity is restricted to this server. A single license file contains the total number of licenses available on the server.∙Distributed License Server: In the distributed license server model, licenses are distributed across multiple servers. A unique license file is required for each server.To create a distributed license server, you must run the Network License Manager on each server that is part of the distributed server pool.∙Redundant License Server: In the redundant license server model, you use three servers to authenticate a single license file. One server acts as the master, while the other two provide backup if the master server fails. Licenses are monitored andissued as long as at least two servers are functional. The license file on all threeservers is the same. You must install the Network License Manager on each server. Note:Network License is the only license type available for Autodesk® Simulation CFD. Check with your system administrator if you are unsure which model to select.∙Enter the name of the license server computer what will run the Network License Manager (NLM).∙If the NLM is installed on another computer, enter the name of that machine. (Do not enter the name of the local machine.)∙If the NLM is installed and configured to run on the local machine, enter 127.0.0.1 as the license server name. If the "FLEXnet License Finder" dialog opens whenAutodesk® Simulation CFD is first started, enter 127.0.0.1 as the Computer Name of the License Server System.Note: Check with your system administrator if you are unsure which computer tospecify.7.The Configure Installation page appears.∙Select Autodesk Simulation CFD.∙Optionally, select Autodesk Inventor Fusion 2013. Fusion does not require licensing, and provides unique capabilities for editing 3D geometry data, regardless of the source. It is a powerful design exploration tool.∙Accept the default Installation Path or click Browse to select a different folder.∙Review the installation settings. You can change settings by clicking Back until the relevant screen appears. When you are satisfied with the settings, click Install tobegin the installation.8.The Installation Progress page appears.∙The Wizard begins installing Autodesk® Simulation CFD.∙ A progress indicator shows how much of the installation has been completed.9.The Installation Complete page appears.∙The successfully installed products are listed, as are any products that failed to install.∙Click Finish to close the Setup Wizard.Notes:∙Before you can use Autodesk® Simulation CFD, the Network License Manager must be installed, and you must be able to access your network. If the Network LicenseManager was already installed for an earlier version of Autodesk Simulation CFD2013, then no further action is required to install it for the service pack.∙You must have activated your licenses on the license server machine.∙To change the language of the installed version, open SimCFDConfig, located in the Autodesk Simulation CFD installation folder.Return to TopResults∙Temperature results reported in the ".sol" file are reported in Celsius instead of Kelvin for steady state heat conduction simulations run in SI units.∙The area of fluid-fluid interfaces (and in a specific case, the solid-solid interface) computed by the Wall Calculator is double the actual area.∙The resultant total heat to surface values computed from residuals for interior and exterior surfaces for steady state heat conduction analysis are higher than theyshould be.∙The function that maps CFD results to FEA loads fails.∙The heat flux reported by the Wall Calculator at a solid-solid interface for a specific conduction simulation is incorrect.∙It is not possible to export certain mesh or output files. An error message is issued, but the requested file is not saved.∙The surface temperature displayed on a Heat Exchanger Device after the simulation is complete incorrectly appears in Kelvin.∙Massed particle trace bounce data is not saved when the flag to enable it, write_trace_bounce_data, is enabled.Return to Top CAD Connection∙The UGNX CAD connection now supports a direct launch in the same manner as Autodesk Simulation 2012.∙In SolidWorks 2012, CAD Connection Launchers are added in the "Office Products"tab as well as to the designated tabs. The result is duplicate launchers in theSolidWorks 2012 UI.∙On a Japanese OS, Japanese Parasolid part names are not transferred correctly into the CFD design study environment.∙Part names from Revit MEP 2013 on a French OS do not appear on the model after launching into Simulation CFD.∙Part Names are not transferred when an assembly is launched from SpaceClaim. All the parts come as "CAD Volume."∙After launching a model from Pro/E into CFD, the part or assembly can no longer be saved in Pro/Engineer. The Save object path is incorrect and it cannot be changed.∙ A design that originated in Pro and launched through Granite is cloned. If the geometry in Pro/E is updated with parts substituted in the assembly from the original design, the cloned design in the CFD design study will not update when launchedfrom Pro/E.∙Part names are not transferred into the CFD model correctly when a SolidWorks model contains a 3D body and a 2D shell is launched into CFD.∙For some Pro/E-based models launched through Granite, using Geometry tools causes part names to change and causes associativity errors between appliedconditions and geometric entities.∙ A crash occurs on some computers when loading a model unless thesoftware_rendering flag is enabled. If your system was affected by this issue, youshould disable the "use_dropshadow" flag with this service pack to ensure graphicsstability.∙When a multiple configuration launch is used to update a design study, the cloned designs do not update based on the modified design geometry.Return to TopMeshing∙Meshing fails for certain motion simulations that have contacting grouped parts and a specified initial position.∙The localized message dialog that appears when applying Automatic Mesh Sizing after having modified applied mesh sizes shows "" instead of the correct text.Return to TopMaterials∙When a CoCreate-based simulation containing a blower material and a surface part is continued, the solver exits unexpectedly and the simulation does not run.∙For certain materials, the vapor pressure displayed by the Material Editor is different from the value saved in the Material database.∙The displayed units of vapor pressure are incorrect when a material is saved in a custom database.∙ A crash occurs on Unicode versions when a custom PCB material is saved to the "My Materials" material database.Return to TopBoundary Conditions∙ A crash occurs for a specific model containing several transient boundary conditions when an attempt is made to clone either the design or scenario or run thesimulation.∙Slip/Symmetry boundary conditions that not aligned to a Cartesian axis behave as non-slip walls instead of slip conditions.∙ A crash occurs for design studies that contain multiple designs when a pressure boundary condition is marked as a summary plane and results planes are createdduring the simulation.Return to TopAPI∙ A crash occurs when running a simulation using a Python script if the "solverComputer" parameter is not specified.∙It is not possible to change the 'continueFrom' Scenario property in a Python-based API script.Return to TopSolver∙ A crash occurs if the Thermal Comfort output quantity is enabled without also enabling the radiation solver.∙The memory footprint of the SimCFD executable grows substantially when switching between active scenarios. Also, the memory allocation increases when the UI is left open over a significant period of time.∙The first job submitted on a HPC cluster runs and finishes, but subsequent jobs in the queue do not start and remain in the queue. Such jobs do not appear in the HPC Cluster Manager Job Scheduler.∙ A crash occurs before the first iteration completes when geometry tools are used to prepare the model for a Motion simulation.∙Flag settings that control Motion are not saved to the scenario when they are enabled.Return to TopDocumentation∙The "MaxRadMatrixSize" flag is missing from the flag manager, and there is no corresponding help topic that describes this flag.∙Several buttons on the Getting Started tab are linked to the legacy CFdesign Customer Portal. These have been modified to link to the Autodesk Simulation CFD WikiHelp site.∙The description of the DRSmoothing flag in the Flag Manager is incorrect. To disable smoothing, a value of -2 is required, not 2.Return to TopInstallation∙Deployment installation fails with the following error: "The system cannot find the specified path when copying files from the install kit."∙Deployment installation fails with the following error: "1: 5 2:adlmPITSetProductInformation failed: 3: 26". The installation rolls back and fails.Return to TopAutodesk Simulation CFD, Autodesk Simulation, Autodesk Inventor, Autodesk Fusion, Autodesk Vault, Autodesk Moldflow Insight, Autodesk Moldflow Advisor, AutoCAD, and Mechanical Desktop are trademarks or registered trademarks of Autodesk, Inc. in the United States and/or other countries.Windows 2000, Windows Server 2003, Windows XP, Windows Vista, Microsoft Office XP, Microsoft Office 2000, Microsoft Office 2003, Microsoft Office 2007, Microsoft Word, Microsoft PowerPoint, and Microsoft Excel are either trademarks or registered trademarks of Microsoft Corporation in the United States and/or other countries.All other trademarks are the property of their respective owners.Return to TopWe thank all our customers who identify issues and report them to us. These reports give us the opportunity to improve the product and provide you with the best solution in simulation. We also thank you for your continued business and for the feedback regarding the content of this release.Respectfully,Autodesk Simulation CFD TeamReturn to Top Copyright© 2012 Autodesk, Inc.。

plant simulation

Plant SimulationIntroductionPlant simulation is a powerful tool used in various industries to model, analyze, and optimize the performance of manufacturing systems. It helps in understanding and improving the efficiency of production processes by simulating the behavior of different components within a plant.Benefits of Plant SimulationSome of the key benefits of using plant simulation include:1.Visualization: Plant simulation provides a visual representation ofthe manufacturing system, allowing stakeholders to understand complexprocesses easily. It helps in identifying bottlenecks, optimizing layouts, andimproving overall productivity.2.Process Optimization: By simulating different scenarios, plantsimulation helps in identifying inefficiencies in the production process. Itallows users to test and evaluate alternative strategies, optimize schedules, and reduce cycle times.3.Cost Reduction: Plant simulation enables users to identifyopportunities for cost reduction by analyzing production line performance,minimizing material handling, and optimizing resource utilization. It helps in reducing inventory levels, improving throughput, and increasing overallprofitability.4.Reduced Downtime: Simulation enables users to analyze and predictthe impact of equipment failures, maintenance activities, or other disruptions on the production process. It helps in planning preventive maintenance,reducing unplanned downtime, and increasing overall equipment effectiveness.5.What-if Analysis: Plant simulation allows users to perform。

(完整版)ABAQUS中Cohesive单元建模方法讲解

复合材料模型建模与分析1. Cohesive单元建模方法1。

1 几何模型使用内聚力模型(cohesive zone）模拟裂纹的产生和扩展，需要在预计产生裂纹的区域加入cohesive 层。

建立cohesive层的方法主要有：方法一、建立完整的结构（如图1（a)所示）,然后在上面切割出一个薄层来模拟cohesive单元，用这种方法建立的cohesive单元与其他单元公用节点,并以此传递力和位移.方法二、分别建立cohesive层和其他结构部件的实体模型，通过“tie”绑定约束,使得cohesive单元两侧的单元位移和应力协调，如图1（b)所示。

（a）cohesive单元与其他单元公用节点 (b）独立的网格通过“tie"绑定图1.建模方法上述两种方法都可以用来模拟复合材料的分层失效，第一种方法划分网格比较复杂；第二种方法赋材料属性简单，划分网格也方便，但是装配及“tie"很繁琐;因此在实际建模中我们应根据实际结构选取较简单的方法。

1.2 材料属性应用cohesive单元模拟复合材料失效，包括两种模型：一种是基于traction-separation描述；另一种是基于连续体描述。

其中基于traction-separation描述的方法应用更加广泛.而在基于traction—separation描述的方法中，最常用的本构模型为图2所示的双线性本构模型。

它给出了材料达到强度极限前的线弹性段和材料达到强度极限后的刚度线性降低软化阶段. 注意图中纵坐标为应力，而横坐标为位移,因此线弹性段的斜率代表的实际是cohesive单元的刚度。

曲线下的面积即为材料断裂时的能量释放率。

因此在定义cohesive的力学性能时,实际就是要确定上述本构模型的具体形状：包括刚度、极限强度、以及临界断裂能量释放率,或者最终失效时单元的位移。

常用的定义方法是给定上述参数中的前三项，也就确定了cohesive的本构模型。

failed to read simulation model from fields

failed to read simulation model from fields"Failed to read simulation model from Fields" is an error message that indicates some kind of issue or problem when attempting to access or load a simulation model from the Fields software. In order to provide relevant reference content related to this issue, I will explain what simulation models are, how they are used in the field of engineering or science, common reasons for encountering the "failed to read simulation model from Fields" error, and potential solutions to resolve the issue.Simulation models are mathematical representations or computer programs that attempt to replicate the behavior or characteristics of a real-world system. They are widely used in various industries, including engineering, physics, chemistry, and computer science. These models help researchers, engineers, and scientists to understand complex systems, optimize designs, analyze scenarios, and make informed decisions.Now, let's explore some possible reasons for encountering the "failed to read simulation model from Fields" error:1. Incorrect file format: One common reason for this error could be that the simulation model file is in an incorrect or unsupported format. Different simulation software may have specific requirements for file formats, so it is essential to ensure that the file format is compatible with the Fields software.2. Corrupted or damaged file: Another reason could be that the simulation model file itself is corrupted or damaged. This can happen due to various reasons, such as incomplete downloads, filetransfer errors, or storage issues. In such cases, it is recommended to re-download or obtain a fresh copy of the simulation model and attempt to load it again.3. Incompatible software version: The Fields software might require a particular version or update to run certain simulation model files successfully. If the software version is outdated or incompatible with the model file, it may result in the "failed to read simulation model" error. Updating the software to the latest version or confirming compatibility is crucial in such situations.4. Insufficient computational resources: Simulation models often require significant computational resources, including memory (RAM), processing power, and disk space. If the computer running the Fields software does not meet the minimum system requirements, it can lead to failed model loading. Ensuring that the computer has sufficient resources, or optimizing the model to reduce resource usage, might resolve this issue.5. Syntax or logical errors within the model: If the simulation model itself has syntax or logical errors, such as typos or inconsistent values, it can prevent successful loading. Carefully reviewing the model or consulting the model's documentation might help identify and correct any errors.To resolve the "failed to read simulation model from Fields" error, consider the following solutions:- Verify the file format: Ensure that the simulation model file is in a compatible format with the Fields software or convert it to acompatible format if needed.- Obtain a fresh copy of the model: If the model file is suspected to be corrupted or damaged, download or obtain a new copy of the model from a reliable source.- Update the software: Check for software updates for the Fields software and install the latest version to ensure compatibility with the simulation model.- Check system requirements: Verify that the computer running the Fields software meets the minimum system requirements, including sufficient memory, processing power, and disk space.- Review the model for errors: Inspect the simulation model for any potential syntax or logical errors, and correct them based on the model's documentation or guidelines.It is important to note that the specific steps to resolve the "failed to read simulation model from Fields" error may vary based on the software, model, and system configuration. Consulting the software's official documentation, forums, or contacting the software provider's support team can provide further guidance and specific troubleshooting steps.In conclusion, the "failed to read simulation model from Fields" error can occur due to various reasons, such as incompatible file formats, software versions, system resources, or errors within the model. By verifying file formats, obtaining fresh copies of the model, updating software, checking system requirements, andreviewing the model for errors, users can potentially resolve this issue and successfully load simulation models in the Fields software.。

基于ABAQUS二次开发的巴西圆盘断裂机

第４５卷第１期２０２３年１月沈　阳　工　业　大　学　学　报ＪｏｕｒｎａｌｏｆＳｈｅｎｙａｎｇＵｎｉｖｅｒｓｉｔｙｏｆＴｅｃｈｎｏｌｏｇｙＶｏｌ４５Ｎｏ１Ｊａｎ２０２３收稿日期：２０２１－１１－１６．基金项目：国家自然科学基金面上项目（５１９７９２９２）；辽宁省教育厅项目（ＬＪＫＭＺ２０２２０４８８）；辽宁省自然科学基金项目（２０１９ＭＳ２４２）．作者简介：刘钧玉（１９７８－），男，辽宁沈阳人，副教授，博士，主要从事断裂力学数值方法和结构地基相互作用等方面的研究．ｄｏｉ：１０．７６８８／ｊ．ｉｓｓｎ．１０００－１６４６．２０２３．０１．１７基于ＡＢＡＱＵＳ二次开发的巴西圆盘断裂机理刘钧玉，张天禹，苏　艳，宁宝宽（沈阳工业大学建筑与土木工程学院，沈阳１１０８７０）摘　要：针对含中心裂纹的巴西圆盘开裂模型利用ＡＢＡＱＵＳ进行了参数化二次开发，基于扩展有限元法和最大周向应力准则对试件裂纹扩展进行数值模拟并验证，研究了围压对裂纹扩展以及裂纹尖端应力强度因子和Ｔ应力的影响．研究结果表明，试件在预制裂纹尖端发生起裂并沿最大周向应力方向扩展．随着裂纹倾角增大，Ⅰ型应力强度因子逐渐减小，Ⅱ型应力强度因子呈现先增大后减小的趋势，Ｔ应力逐渐增大．随着围压数值的升高，试件的断裂韧度增大，Ｔ应力增大，而Ⅰ型和Ⅱ型应力强度因子几乎不受影响．关　键　词：巴西圆盘；二次开发；应力强度因子；Ｔ应力；裂纹扩展；围压；扩展有限元法中图分类号：ＴＵ４５文献标志码：Ａ文章编号：１０００－１６４６（２０２３）０１－０１０６－０７ＦｒａｃｔｕｒｅｍｅｃｈａｎｉｓｍｏｆＢｒａｚｉｌｉａｎｄｉｓｃｂａｓｅｄｏｎＡＢＡＱＵＳｓｅｃｏｎｄａｒｙｄｅｖｅｌｏｐｍｅｎｔＬＩＵＪｕｎｙｕ，ＺＨＡＮＧＴｉａｎｙｕ，ＳＵＹａｎ，ＮＩＮＧＢａｏｋｕａｎ（ＳｃｈｏｏｌｏｆＡｒｃｈｉｔｅｃｔｕｒｅ＆ＣｉｖｉｌＥｎｇｉｎｅｅｒｉｎｇ，ＳｈｅｎｙａｎｇＵｎｉｖｅｒｓｉｔｙｏｆＴｅｃｈｎｏｌｏｇｙ，Ｓｈｅｎｙａｎｇ１１０８７０，Ｃｈｉｎａ）Ａｂｓｔｒａｃｔ：ＡｉｍｉｎｇａｔｔｈｅＢｒａｚｉｌｉａｎｄｉｓｋｃｒａｃｋｍｏｄｅｌｗｉｔｈｃｅｎｔｒａｌｃｒａｃｋ，ｔｈｅｐａｒａｍｅｔｒｉｃｓｅｃｏｎｄａｒｙｄｅｖｅｌｏｐｍｅｎｔｗａｓｃａｒｒｉｅｄｏｕｔｂｙｕｓｉｎｇＡＢＡＱＵＳ．Ｔｈｅｃｒａｃｋｐｒｏｐａｇａｔｉｏｎｏｆｔｈｅｓｐｅｃｉｍｅｎｗａｓｎｕｍｅｒｉｃａｌｌｙｓｉｍｕｌａｔｅｄａｎｄｖｅｒｉｆｉｅｄｉｎｔｅｒｍｓｏｆｔｈｅｅｘｔｅｎｄｅｄｆｉｎｉｔｅｅｌｅｍｅｎｔｍｅｔｈｏｄａｎｄｔｈｅｍａｘｉｍｕｍｃｉｒｃｕｍｆｅｒｅｎｔｉａｌｓｔｒｅｓｓｃｒｉｔｅｒｉｏｎ．Ｔｈｅｅｆｆｅｃｔｓｏｆｃｏｎｆｉｎｉｎｇｐｒｅｓｓｕｒｅｏｎｃｒａｃｋｐｒｏｐａｇａｔｉｏｎ，ｓｔｒｅｓｓｉｎｔｅｎｓｉｔｙｆａｃｔｏｒａｔｔｈｅｃｒａｃｋｔｉｐａｎｄＴｓｔｒｅｓｓｗｅｒｅｓｔｕｄｉｅｄ．Ｔｈｅｒｅｓｕｌｔｓｓｈｏｗｔｈａｔｔｈｅｓｐｅｃｉｍｅｎｃｒａｃｋｓａｔｔｈｅｐｒｅｆａｂｒｉｃａｔｅｄｃｒａｃｋｔｉｐａｎｄｐｒｏｐａｇａｔｅｓａｌｏｎｇｔｈｅｍａｘｉｍｕｍｃｉｒｃｕｍｆｅｒｅｎｔｉａｌｓｔｒｅｓｓｄｉｒｅｃｔｉｏｎ．Ｗｉｔｈｔｈｅｉｎｃｒｅａｓｅｏｆｃｒａｃｋｉｎｃｌｉｎａｔｉｏｎａｎｇｌｅ，ｔｈｅｍｏｄｅⅠｓｔｒｅｓｓｉｎｔｅｎｓｉｔｙｆａｃｔｏｒｄｅｃｒｅａｓｅｓｇｒａｄｕａｌｌｙ，ｔｈｅｍｏｄｅⅡｓｔｒｅｓｓｉｎｔｅｎｓｉｔｙｆａｃｔｏｒｉｎｃｒｅａｓｅｓｆｉｒｓｔａｎｄｔｈｅｎｄｅｃｒｅａｓｅｓ，ａｎｄｔｈｅＴｓｔｒｅｓｓｉｎｃｒｅａｓｅｓｇｒａｄｕａｌｌｙ．Ｗｉｔｈｔｈｅｉｎｃｒｅａｓｅｏｆｃｏｎｆｉｎｉｎｇｐｒｅｓｓｕｒｅ，ｔｈｅｆｒａｃｔｕｒｅｔｏｕｇｈｎｅｓｓａｎｄＴｓｔｒｅｓｓｏｆｔｈｅｓｐｅｃｉｍｅｎｉｎｃｒｅａｓｅ，ｗｈｉｌｅｔｈｅｍｏｄｅⅠａｎｄｔｈｅｍｏｄｅⅡｓｔｒｅｓｓｉｎｔｅｎｓｉｔｙｆａｃｔｏｒｓａｒｅｈａｒｄｌｙａｆｆｅｃｔｅｄ．Ｋｅｙｗｏｒｄｓ：Ｂｒａｚｉｌｉａｎｄｉｓｃ；ｓｅｃｏｎｄａｒｙｄｅｖｅｌｏｐｍｅｎｔ；ｓｔｒｅｓｓｉｎｔｅｎｓｉｔｙｆａｃｔｏｒ；Ｔｓｔｒｅｓｓ；ｃｒａｃｋｐｒｏｐａｇａｔｉｏｎ；ｃｏｎｆｉｎｉｎｇｐｒｅｓｓｕｒｅ；ｅｘｔｅｎｄｅｄｆｉｎｉｔｅｅｌｅｍｅｎｔｍｅｔｈｏｄ岩石普遍存在裂纹，内部裂纹的贯通会破坏岩石整体结构的稳定性［１］．巴西圆盘试验是岩石力学试验中具有代表性的试验［２－３］，最早被提出用于测定岩石材料拉伸强度．含中心裂纹的巴西圆盘试件可以通过改变加载角的方式实现Ⅰ型、Ⅱ型、Ⅰ Ⅱ复合型的断裂模式，从而根据应力强度因子等参数测算材料断裂韧度．王辉等［４］对含预制裂隙的圆盘试件进行巴西劈裂试验，对试件的破坏过程进行了研究；周北明等［５］针对含中心裂纹巴西圆盘，对其无量纲化应力强度因子的获取精度进行了研究；ＡｌＳｈａｙｅａ［６］对脆性石灰岩圆盘试件进行了试验，研究了围压和温度对裂纹起裂的影响；Ｓａｒｆａｒａｚｉ等［７］利用数字图像识别法研究了裂隙填充材料及裂纹倾角对巴西圆盘裂纹扩展的影响；栗青等［８］通过室内试验研究了围压对岩石强度和弹性模量的影响．然而，含中心裂纹的巴西圆盘试件在进行试验时，在集中荷载处可能会发生应力集中的情况，使试件在应力集中点率先开始破坏并进行裂纹扩展，这不符合试件破坏时必定在预制裂纹尖端起裂的理论假设．因此，在巴西圆盘的加载点处设置平台作为改进，可以有效缓解试件与加载处接触部分的应力集中现象，且同时可以实现不同的断裂形式［９－１０］．岩石发生断裂后，当试件的加载方向和预制裂纹存在夹角或加载位置与裂纹面呈非对称时可以测得岩石的断裂韧度［１１］．本文基于ＡＢＡＱＵＳ平台对三维平台巴西圆盘模型进行了二次开发，对其断裂参数进行了求解，并得出不同开裂情况下的断裂参数变化规律，为岩石材料断裂韧度的求解提供了指导与帮助．１　扩展有限元法及断裂准则１１　扩展有限元法基础理论图１为扩展有限元函数节点示意图．扩展有限元法的位移插值函数［１２］为Ｕ（ｘ）＝∑ｉ∈ＮＮｉ（ｘ）ｕｉ＋∑ｊ∈ＮｄｉｃＮｊＨ（ｘ）ａｊ＋　∑ｋ∈ＮａｓｙＮｋ∑４ａ＝１ａ（ｘ）ｂａｋ（１）式中：Ｎ为常规有限元节点集；Ｎｄｉｃ为被裂纹贯穿的节点集；Ｎａｓｙ为裂尖单元集；ｕｉ为常规有限元节点位移；ａｊ和ｂａｋ为被裂纹贯穿节点以及裂尖节点的位移；Ｈ（ｘ）为能反应裂纹面非连续性的跳跃函数；ａ（ｘ）为裂纹尖端渐进位移函数，能够体现裂纹尖端的应力奇异性，其表达式为ａ（ｘ）＝［槡ｒｓｉｎθ２，槡ｒｃｏｓθ２，槡ｒｓｉｎθ２ｃｏｓθ，　槡ｒｃｏｓθ２ｃｏｓθ］（２）图１　扩展有限元函数节点示意图Ｆｉｇ１　Ｓｃｈｅｍａｔｉｃｎｏｄｅｓｏｆｅｘｔｅｎｄｅｄｆｉｎｉｔｅｅｌｅｍｅｎｔｆｕｎｃｔｉｏｎ１２　裂纹扩展断裂准则在断裂力学传统理论基础上考虑了Ｔ应力情况后，裂纹尖端应力场表达式［１３］为　σｘ′＝ＫⅠ２π槡ｒｃｏｓθ２１－ｓｉｎθ２ｓｉｎ３θ()２－ＫⅡ２π槡ｒｓｉｎθ２２＋ｃｏｓθ２ｃｏｓ３θ()２＋Ｔσｙ′＝ＫⅠ２π槡ｒｃｏｓθ２１＋ｓｉｎθ２ｓｉｎ３θ()２＋ＫⅡ２π槡ｒｓｉｎθ２ｃｏｓθ２ｃｏｓ３θ２τｘ′ｙ′＝ＫⅠ２π槡ｒｃｏｓθ２ｓｉｎθ２ｃｏｓ３θ２＋ＫⅡ２π槡ｒｃｏｓθ２１－ｓｉｎθ２ｃｏｓ３θ()２（３）式中：ＫⅠ、ＫⅡ、Ｔ分别为Ⅰ型、Ⅱ型应力强度因子、Ｔ应力；σｘ′、σｙ′、τｘ′ｙ′为裂纹尖端应力．２　ＡＢＡＱＵＳ二次开发２１　ＡＢＡＱＵＳ脚本接口及ＧＵＩ插件开发ＡＢＡＱＵＳ基于Ｐｙｔｈｏｎ语言建立了二次开发环境相关的脚本结构，ＡＢＡＱＵＳ／ＣＡＥ所进行的操作均可由Ｐｙｔｈｏｎ语言编写对应命令并实现［１４］，其与脚本接口的通信关系如图２所示．图２　ＡＢＡＱＵＳ脚本接口通信关系Ｆｉｇ２　ＡＢＡＱＵＳｓｃｒｉｐｔｉｎｔｅｒｆａｃｅｃｏｍｍｕｎｉｃａｔｉｏｎｒｅｌａｔｉｏｎｓｈｉｐ本文主要实现参数化建模和ＧＵＩ插件程序的创建．在参数化建模方面，通过编写脚本对前处理建模部分进行操作，利用脚本对需求的模型尺寸、参数等进行控制，也可以对后处理分析进行数据的提取与绘制，在ＧＵＩ脚本程序方面创建了适应数值模型的图形界面．这使得数值计算过程避７０１第１期刘钧玉，等：基于ＡＢＡＱＵＳ二次开发的巴西圆盘断裂机理免了由于不断修改参数以及后处理分析提取数据等产生的繁琐操作，且创立的用户界面简洁直观，让操作过程更加便捷．ＡＢＡＱＵＳＧＵＩＴｏｏｌｋｉｔ提供了二次开发的相关工具，ＡＢＡＱＵＳＧＵＩ插件也在此基础上，通过内核执行程序Ｋｅｒｎｅｌ和ＧＵＩ的交互完成ＧＵＩ界面的创建．其中，内核程序Ｋｅｒｎｅｌ负责将用户界面输入的建模相关数据进行处理，并存储成Ｉｎｐｕｔ文件．用户在已完成创建的ＧＵＩ界面进行数据录入后，输入的结果会被脚本传输到内核执行程序Ｋｅｒｎｅｌ进行分析，此交互过程的工作原理如图３所示．图３　ＧＵＩ与Ｋｅｒｎｅｌ的交互原理Ｆｉｇ３　ＩｎｔｅｒａｃｔｉｏｎｐｒｉｎｃｉｐｌｅｂｅｔｗｅｅｎＧＵＩａｎｄＫｅｒｎｅｌ２２　平台巴西圆盘参数化建模平台巴西圆盘试件计算模型如图４所示．图４　数值模型示意图Ｆｉｇ４　Ｓｃｈｅｍａｔｉｃｄｉａｇｒａｍｏｆｎｕｍｅｒｉｃａｌｍｏｄｅｌ图４中，试件直径为５０ｍｍ，厚度为１０ｍｍ，且裂纹位置位于试件中心，试件取砂岩材料参数，弹性模量Ｅ为４７５ＧＰａ，泊松比μ为０２５．裂纹位置位于试件中心，数值模型采用竖向位移加载的方式，选择八节点单元Ｃ３Ｄ８，且将圆盘试件分区以便对多个区域进行网格划分．计算模型的上下平台选用解析刚体，在将裂纹和圆盘试件进行装配后设置接触，并对裂纹扩展相关内容进行设置．ＡＢＡＱＵＳ用户图形界面以巴西圆盘模型尺寸标注，材料参数、分析步接触设置、网格荷载控制和后处理分析等作为布局．界面整体如图５所示．图５　模型参数化分析界面Ｆｉｇ５　Ｍｏｄｅｌｐａｒａｍｅｔｒｉｃａｎａｌｙｓｉｓｉｎｔｅｒｆａｃｅ８０１沈　阳　工　业　大　学　学　报第４５卷２３　数值模型脚本构建平台巴西圆盘建模过程中需要使用圆盘、平台、裂纹这三种不同的部件，这些部件的尺寸可以直接定义，但有些部件为了方便后续装配等操作需要额外进行修改．在装配过程中需要将裂纹实例的旋转倾斜度进行定义，并将各个部件移动到相应位置，在装配完成后首先要通过指定裂纹位置以及整体关系的方式进行裂纹设定，并选择扩展有限元算法的方式，选取最大周向应力准则为判定方式应用在模拟过程中，随后定义接触属性设置平台和圆盘部件的接触［１５］．数值模拟中除了需要观察裂纹扩展形式，更重要的是提取裂纹尖端奇异参数，因此，在设置分析步和历程输出变量时，需要分别对非线性开关进行设置，并设定是否允许裂纹扩展，而在提取裂纹尖端奇异参数时，需要在裂纹设定模块对所需求的不同结果进行输出．在图形界面上方是以图片示意的巴西圆盘模型，并通过尺寸标注表示结构组成，使用户界面更加直观，下方可定义模型名称并可通过快捷键在模型创立完成后直接提交作业运算．最下方为参数化建模控制流程，包括不同部件的创建、材料参数输入、分析步和装配接触设置、网格划分、施加荷载，以及提交工作后的后处理数据提取分析．每个建模模块都有模型对应的控制需求选项，例如其中的分析步模块如图６所示，除了对增量步等参数的设置，也包含对非线性开关的设置以及场变量输出的模式及内容．图６　分析步模块设置界面Ｆｉｇ６　Ａｎａｌｙｓｉｓｓｔｅｐｍｏｄｕｌｅｓｅｔｔｉｎｇｉｎｔｅｒｆａｃｅ３　巴西圆盘断裂参数分析３１　巴西圆盘裂纹扩展数值模拟在基于ＡＢＡＱＵＳ二次开发对模型进行建模后，首先对含中心裂纹巴西圆盘进行了裂纹扩展的数值模拟，对不同裂纹倾角β的试件进行数值模拟，并将所得裂纹扩展结果和试验结果进行对比，如图７所示，其中左侧为试验结果，右侧为数值模拟结果．图７　数值模拟与试验对比结果Ｆｉｇ７　Ｃｏｍｐａｒｉｓｏｎｂｅｔｗｅｅｎｎｕｍｅｒｉｃａｌｓｉｍｕｌａｔｉｏｎａｎｄｔｅｓｔｒｅｓｕｌｔｓ对比结果表明，数值模拟结果与试验结果较为一致，预制裂纹从裂纹尖端起裂并沿最大主应力方向进行扩展直至破坏，这也符合巴西圆盘试件破坏的理论假设．在后处理中对平台压板反作用力载荷进行提取，并绘制载荷随时间变化的关系曲线，如图８所示．围压，分别对反作用力进行提取并绘制关系曲线，在数据表中找到荷载瞬时变化的时间点和荷载值，并分别提取荷载瞬时变化前的数值，绘制荷载随围压的变化关系曲线，如图９所示．图９　围压与荷载的关系曲线Ｆｉｇ９　Ｒｅｌａｔｉｏｎｃｕｒｖｅｓｏｆｃｏｎｆｉｎｉｎｇｐｒｅｓｓｕｒｅｖｅｒｓｕｓｌｏａｄ由图９可知，随着围压的增大，瞬时荷载基本呈线性增大趋势，由于瞬时荷载点对应裂纹萌生时间点，所以瞬时荷载值的大小即反应试件阻止裂纹扩展的能力．因此可以看出围压对岩石断裂韧度有很大影响，断裂韧度随围压增大而增大．３２　巴西圆盘裂纹尖端奇异参数分析在提取数值模型的裂纹尖端奇异参数时，需要对裂纹接触过程进行设置，并在历程输出变量中分别对需要的参数进行输出．通过改变裂纹长度和裂纹倾角，对不同情况下平台巴西圆盘试件计算了应力强度因子和Ｔ应力的数值．为便于描述，应力强度因子和Ｔ应力表达式为ＫⅠ＝ＦＲｔｃ槡πＺⅠα，ｃ()ＲＫⅡ＝ＦＲｔｃ槡πＺⅡα，ｃ()ＲＴ＝ＦπＲｔＴ α，ｃ()Ｒ１－ｃ()Ｒ　（４）式中：Ｆ为载荷；Ｒ为圆盘试件半径；ｔ为试件厚度；ｃ为试件中心裂纹长度；ＺⅠ为无量纲Ⅰ型应力强度因子；ＺⅡ为无量纲Ⅱ型应力强度因子；Ｔ为无量纲Ｔ应力；α为裂纹初始角度．在ｃ／Ｒ＝０２和ｃ／Ｒ＝０４情况下，裂纹尖端应力强度因子ＺⅠ和ＺⅡ的变化如图１０所示．由图１０可以看出，在裂纹倾角为０°时，ＺⅡ为０且ＺⅠ不为０，即试件为纯Ⅰ型开裂．随着裂纹倾角的增加，当倾角β为３０°时，ＺⅠ减小到０且ＺⅡ由０开始增大，此时为纯Ⅱ型开裂．当裂纹倾角继续增大直到９０°时，ＺⅠ逐步减小，而ＺⅡ先增大到极值再减小到０，此时又为纯Ⅰ型开裂．可以看出，当试件发生纯Ⅰ型开裂时的两种情况下，ＺⅠ均为极值，而当事件发生纯Ⅱ型开裂时的情况下，ＺⅡ并非极值，这表明当试件发生纯Ⅱ型破坏时并不是Ⅱ型应力强度因子为最大值的情况．随着ｃ／Ｒ增大，Ⅱ型应力强度因子随之增大，而在裂纹倾角为３０°～６０°时，不同ｃ／Ｒ值情况下Ⅰ型应力强度因子差距减小，在０°～３０°以及６０°～９０°时，Ⅰ型应力强度因子随ｃ／Ｒ增大而增大．图１０　不同裂纹倾角下的无量纲应力强度因子Ｆｉｇ１０　Ｄｉｍｅｎｓｉｏｎｌｅｓｓｓｔｒｅｓｓｉｎｔｅｎｓｉｔｙｆａｃｔｏｒｓｕｎｄｅｒｄｉｆｆｅｒｅｎｔｃｒａｃｋｄｉｐａｎｇｌｅｓ在ｃ／Ｒ＝０２和ｃ／Ｒ＝０４情况下，Ｔ应力的变化如图１１所示．图１１　不同裂纹倾角下的无量纲Ｔ应力Ｆｉｇ１１　ＤｉｍｅｎｓｉｏｎｌｅｓｓＴｓｔｒｅｓｓｕｎｄｅｒｄｉｆｆｅｒｅｎｔｃｒａｃｋｄｉｐａｎｇｌｅｓ由图１１可以看出，Ｔ应力会随着裂纹倾角的增加而增加，在裂纹倾角达到约４５°时由负值达到０值．在裂纹倾角小于４５°时，Ｔ应力会随ｃ／Ｒ值的增大而增大，而在裂纹倾角大于４５°时，Ｔ应力会随ｃ／Ｒ值的增大而减小．以裂纹倾角为４５°，ｃ／Ｒ＝０４的情况下对数值模型施加围压，为避免０１１沈　阳　工　业　大　学　学　报第４５卷围压过高导致裂纹面接触产生压剪破坏情况，围压值控制在１～１０ＭＰａ，并在每次改变围压时对应力强度因子和Ｔ应力分别进行提取．裂纹尖端应力强度因子以及Ｔ应力的无量纲数值随围压变化的关系曲线如图１２～１３所示．图１２　不同围压下的无量纲应力强度因子Ｆｉｇ１２　Ｄｉｍｅｎｓｉｏｎｌｅｓｓｓｔｒｅｓｓｉｎｔｅｎｓｉｔｙｆａｃｔｏｒｓｕｎｄｅｒｄｉｆｆｅｒｅｎｔｃｏｎｆｉｎｉｎｇｐｒｅｓｓｕｒｅｓ图１３　不同围压下的无量纲Ｔ应力Ｆｉｇ１３　ＤｉｍｅｎｓｉｏｎｌｅｓｓＴｓｔｒｅｓｓｕｎｄｅｒｄｉｆｆｅｒｅｎｔｃｏｎｆｉｎｉｎｇｐｒｅｓｓｕｒｅｓ由图１２～１３可以看出，随着围压的增大，Ⅰ型应力强度因子逐渐减小，代表裂纹面压缩程度提高，而Ⅱ型应力强度因子也逐渐减小，但围压对两者影响很小，而Ｔ应力会随着围压的增大而增大．４　结　论本文利用ＡＢＡＱＵＳ软件针对平台巴西圆盘模型进行了参数化二次开发，对试件裂纹扩展进行了数值验证，研究了围压对试件的影响，并提取了不同情况下裂纹尖端奇异参数即应力强度因子和Ｔ应力．结果表明，试件的断裂韧度随围压增大而增大，在裂纹倾角为０°和９０°时，试件为纯Ⅰ型开裂，且Ⅰ型应力强度因子随裂纹倾角增大而减小．当裂纹倾角为３０°时，试件为纯Ⅱ型开裂，且Ⅱ型应力强度因子随裂纹倾角增大呈现先增大后减小的趋势．在试件首先达到纯Ⅰ型开裂和纯Ⅱ型开裂情况时，Ｔ应力均为负值，且Ｔ应力随裂纹倾角增大而增大．随着围压增大，Ⅰ型和Ⅱ型应力强度因子逐渐减小但所受影响程度很小，而Ｔ应力增大．参考文献（Ｒｅｆｅｒｅｎｃｅｓ）：［１］贺晶晶，师俊平．冻融循环作用下砂岩三点弯曲断裂性能试验及其破坏形态研究［Ｊ］．岩石力学与工程学报，２０１７，３６（１２）：２９１７－２９２５．（ＨＥＪｉｎｇｊｉｎｇ，ＳＨＩＪｕｎｐｉｎｇ．Ｆｒａｃｔｕｒｉｎｇｂｅｈａｖｉｏｒａｎｄｆａｉｌｕｒｅｐａｔｔｅｒｎｏｆｓａｎｄｓｔｏｎｅｉｎｔｈｒｅｅｐｏｉｎｔｂｅｎｄｉｎｇｔｅｓｔｕｎｄｅｒｆｒｅｅｚｉｎｇｔｈａｗｉｎｇｃｙｃｌｅｓ［Ｊ］．ＣｈｉｎｅｓｅＪｏｕｒｎａｌｏｆＲｏｃｋＭｅｃｈａｎｉｃｓａｎｄＥｎｇｉｎｅｅｒｉｎｇ，２０１７，３６（１２）：２９１７－２９２５．）［２］朱思尘，李江腾．干燥和饱水状态下含层理构造板岩巴西劈裂实验能量研究［Ｊ］．中南大学学报（自然科学版），２０１８，４９（８）：２０２４－２０３０．（ＺＨＵＳｉｃｈｅｎ，ＬＩＪｉａｎｇｔｅｎｇ．ＥｎｅｒｇｙｒｅｓｅａｒｃｈｏｎｓｌａｔｅｓｗｉｔｈｂｅｄｄｉｎｇｓｔｒｕｃｔｕｒｅｕｎｄｅｒＢｒａｚｉｌｉａｎｓｐｌｉｔｔｉｎｇｔｅｓｔｓｉｎｄｒｙａｎｄｓａｔｕｒａｔｅｄｃｏｎｄｉｔｉｏｎ［Ｊ］．ＪｏｕｒｎａｌｏｆＣｅｎｔｒａｌＳｏｕｔｈＵｎｉｖｅｒｓｉｔｙ（ＳｃｉｅｎｃｅａｎｄＴｅｃｈｎｏｌｏｇｙ），２０１８，４９（８）：２０２４－２０３０．）［３］吴秋红，赵伏军，李夕兵，等．径向压缩下圆环砂岩样的力学特性研究［Ｊ］．岩土力学，２０１８，３９（１１）：３９６９－３９７５．（ＷＵＱｉｕｈｏｎｇ，ＺＨＡＯＦｕｊｕｎ，ＬＩＸｉｂｉｎｇ，ｅｔａｌ．Ｍｅｃｈａｎｉｃａｌｐｒｏｐｅｒｔｉｅｓｏｆｒｉｎｇｓｐｅｃｉｍｅｎｓｏｆｓａｎｄｓｔｏｎｅｓｕｂｊｅｃｔｅｄｔｏｄｉａｍｅｔｒａｌｃｏｍｐｒｅｓｓｉｏｎ［Ｊ］．ＲｏｃｋａｎｄＳｏｉｌＭｅｃｈａｎｉｃｓ，２０１８，３９（１１）：３９６９－３９７５．）［４］王辉，李勇，曹树刚，等．含预制裂隙黑色页岩裂纹扩展过程及宏观破坏模式巴西劈裂试验研究［Ｊ］．岩石力学与工程学报，２０２０，３９（５）：９１２－９２６．（ＷＡＮＧＨｕｉ，ＬＩＹｏｎｇ，ＣＡＯＳｈｕｇａｎｇ，ｅｔａｌ．Ｂｒａｚｉｌｉａｎｓｐｌｉｔｔｉｎｇｔｅｓｔｓｔｕｄｙｏｎｃｒａｃｋｐｒｏｐａｇａｔｉｏｎｐｒｏｃｅｓｓａｎｄｍａｃｒｏｓｃｏｐｉｃｆａｉｌｕｒｅｍｏｄｅｏｆｐｒｅｃｒａｃｋｅｄｂｌａｃｋｓｈａｌｅ［Ｊ］．ＣｈｉｎｅｓｅＪｏｕｒｎａｌｏｆＲｏｃｋＭｅｃｈａｎｉｃｓａｎｄＥｎｇｉｎｅｅｒｉｎｇ，２０２０，３９（５）：９１２－９２６．）［５］周北明，张明明，高原．中心直裂纹圆盘无量纲应力强度因子获取精度研究［Ｊ］．人民长江，２０１７，４８（２０）：１０１－１０６．（ＺＨＯＵＢｅｉｍｉｎｇ，ＺＨＡＮＧＭｉｎｇｍｉｎｇ，ＧＡＯＹｕａｎ．ＡｎａｌｙｓｉｓｏｎａｃｃｕｒａｃｙｏｆｄｉｍｅｎｓｉｏｎｌｅｓｓｓｔｒｅｓｓｉｎｔｅｎｓｉｔｙｆａｃｔｏｒｏｆｃｅｎｔｒａｌｓｔｒａｉｇｈｔｔｈｒｏｕｇｈＢｒａｚｉｌｉａｎｄｉｓｃ［Ｊ］．ＹａｎｇｔｚｅＲｉｖｅｒ，２０１７，４８（２０）：１０１－１０６．）［６］ＡｌＳｈａｙｅａＮＡ．ＣｒａｃｋｐｒｏｐａｇａｔｉｏｎｔｒａｊｅｃｔｏｒｉｅｓｆｏｒｒｏｃｋｓｕｎｄｅｒｍｉｘｅｄｍｏｄｅⅠ Ⅱｆｒａｃｔｕｒｅ［Ｊ］．ＥｎｇｉｎｅｅｒｉｎｇＧｅｏｌｏｇｙ，２００５，８１（１）：８４－９７．［７］ＳａｒｆａｒａｚｉＶ，ＨａｅｒｉＨ，ＦａｔｅｈｉＭ．ＦｒａｃｔｕｒｅｍｅｃｈａｎｉｓｍｏｆＢｒａｚｉｌｉａｎｄｉｓｃｓｗｉｔｈｍｕｌｔｉｐｌｅｐａｒａｌｌｅｌｎｏｔｃｈｅｓｕｓｉｎｇＰＦＣ２Ｄ［Ｊ］．ＰｅｒｉｏｄｉｃａＰｏｌｙｔｅｃｈｎｉｃａＣｉｖｉｌＥｎｇｉｎｅｅｒｉｎｇ，２０１７，６１（４）：６５３－６６３．［８］栗青，常兆荣，高阳，等．隐晶质玄武岩的应变演化规律与强度破坏特征［Ｊ］．沈阳工业大学学报，２０２２，４４（１）：１０９－１１５．（ＬＩＱｉｎｇ，ＣＨＡＮＧＺｈａｏｒｏｎｇ，ＧＡＯＹａｎｇ，ｅｔａｌ．Ｓｔｒａｉｎｅｖｏｌｕｔｉｏｎｒｕｌｅａｎｄｓｔｒｅｎｇｔｈｆａｉｌｕｒｅｃｈａｒａｃｔｅｒｉｓｔｉｃｓｏｆａｐｈａｎｉｔｉｃｂａｓａｌｔ［Ｊ］．ＪｏｕｒｎａｌｏｆＳｈｅｎｙａｎｇＵｎｉｖｅｒｓｉｔｙｏｆＴｅｃｈｎｏｌｏｇｙ，２０２２，４４（１）：１０９－１１５．）［９］彭凡，董世明．一类非均布载荷下中心裂纹圆盘Ｔ应力分析［Ｊ］．应用数学和力学，２０１８，３９（７）：７６６－７７５．（ＰＥＮＧＦａｎ，ＤＯＮＧＳｈｉｍｉｎｇ．ＴｓｔｒｅｓｓｉｎａｃｅｎｔｒａｌｌｙｃｒａｃｋｅｄＢｒａｚｉｌｉａｎｄｉｓｋｕｎｄｅｒｎｏｎｕｎｉｆｏｒｍｐｒｅｓｓｕｒｅｌｏａｄ１１１第１期刘钧玉，等：基于ＡＢＡＱＵＳ二次开发的巴西圆盘断裂机理［Ｊ］．ＡｐｐｌｉｅｄＭａｔｈｅｍａｔｉｃｓａｎｄＭｅｃｈａｎｉｃｓ，２０１８，３９（７）：７６６－７７５．）［１０］孙欣，朱哲明，谢凌志，等．基于ＳＥＮＤＢ试样的砂岩复合脆性断裂行为研究［Ｊ］．岩石力学与工程学报，２０１７，３６（１２）：２８８４－２８９４．（ＳＵＮＸｉｎ，ＺＨＵＺｈｅｍｉｎｇ，ＸＩＥＬｉｎｇｚｈｉ，ｅｔａｌ．ＩｎｖｅｓｔｉｇａｔｉｏｎｏｎｍｉｘｅｄｍｏｄｅｆｒａｃｔｕｒｅｂｅｈａｖｉｏｒｏｆｓａｎｄｓｔｏｎｅｕｓｉｎｇａＳＥＮＤＢｓｐｅｃｉｍｅｎ［Ｊ］．ＣｈｉｎｅｓｅＪｏｕｒｎａｌｏｆＲｏｃｋＭｅｃｈａｎｉｃｓａｎｄＥｎｇｉｎｅｅｒｉｎｇ，２０１７，３６（１２）：２８８４－２８９４．）［１１］赵彦琳，范勇，朱哲明，等．Ｔ应力对闭合裂纹断裂行为的理论和实验研究［Ｊ］．岩石力学与工程学报，２０１８，３７（６）：１３４０－１３４９．（ＺＨＡＯＹａｎｌｉｎ，ＦＡＮＹｏｎｇ，ＺＨＵＺｈｅｍｉｎｇ，ｅｔａｌ．ＡｎａｌｙｔｉｃａｌａｎｄｅｘｐｅｒｉｍｅｎｔａｌｓｔｕｄｙｏｎｔｈｅｅｆｆｅｃｔｏｆＴｓｔｒｅｓｓｏｎｂｅｈａｖｉｏｒｏｆｃｌｏｓｅｄｃｒａｃｋｓ［Ｊ］．ＣｈｉｎｅｓｅＪｏｕｒｎａｌｏｆＲｏｃｋＭｅｃｈａｎｉｃｓａｎｄＥｎｇｉｎｅｅｒｉｎｇ，２０１８，３７（６）：１３４０－１３４９．）［１２］ＢｅｌｙｔｓｃｈｋｏＴ，ＢｌａｃｋＴ．Ｅｌａｓｔｉｃｃｒａｃｋｇｒｏｗｔｈｉｎｆｉｎｉｔｅｅｌｅｍｅｎｔｓｗｉｔｈｍｉｎｉｍａｌｒｅｍｅｓｈｉｎｇ［Ｊ］．ＩｎｔｅｒｎａｔｉｏｎａｌＪｏｕｒｎａｌｆｏｒＮｕｍｅｒｉｃａｌＭｅｔｈｏｄｓｉｎＥｎｇｉｎｅｅｒｉｎｇ，１９９９，４５（５）：６０１－６２０．［１３］刘红岩．考虑Ｔ应力的岩石压剪裂纹起裂机理［Ｊ］．岩土工程学报，２０１９，４１（７）：１２９６－１３０２．（ＬＩＵＨｏｎｇｙａｎ．ＩｎｉｔｉａｔｉｏｎｍｅｃｈａｎｉｓｍｏｆｃｒａｃｋｓｏｆｒｏｃｋｉｎｃｏｍｐｒｅｓｓｉｏｎａｎｄｓｈｅａｒｃｏｎｓｉｄｅｒｉｎｇＴｓｔｒｅｓｓ［Ｊ］．ＣｈｉｎｅｓｅＪｏｕｒｎａｌｏｆＧｅｏｔｅｃｈｎｉｃａｌＥｎｇｉｎｅｅｒｉｎｇ，２０１９，４１（７）：１２９６－１３０２．）［１４］曹金凤，王旭春，孔亮．Ｐｙｔｈｏｎ语言在Ａｂａｑｕｓ中的应用［Ｍ］．北京：机械工业出版社，２０１１．（ＣＡＯＪｉｎｆｅｎｇ，ＷＡＮＧＸｕｃｈｕｎ，ＫＯＮＧＬｉａｎｇ．ＡｐｐｌｉｃａｔｉｏｎｏｆＰｙｔｈｏｎｌａｎｇｕａｇｅｉｎＡＢＡＱＵＳ［Ｍ］．Ｂｅｉｊｉｎｇ：ＣｈｉｎａＭａｃｈｉｎｅＰｒｅｓｓ，２０１１．）［１５］黄志刚，孙洪祥，黎然．基于ＧＭＴＳ准则的岩石复合型断裂机理研究［Ｊ］．力学季刊，２０１８，３９（３）：６３８－６４４．（ＨＵＡＮＧＺｈｉｇａｎｇ，ＳＵＮＨｏｎｇｘｉａｎｇ，ＬＩＲａｎ．Ｉｎｖｅｓｔｉｇａｔｉｏｎｏｆｍｉｘｅｄｆｒａｃｔｕｒｅｂｅｈａｖｉｏｒｓｏｆｒｏｃｋｂｙｔｈｅｇｅｎｅｒａｌｉｚｅｄｍａｘｉｍｕｍｃｉｒｃｕｍｆｅｒｅｎｔｉａｌｓｔｒｅｓｓｃｒｉｔｅｒｉｏｎ［Ｊ］．ＣｈｉｎｅｓｅＱｕａｒｔｅｒｌｙｏｆＭｅｃｈａｎｉｃｓ，２０１８，３９（３）：６３８－６４４．）（责任编辑：钟　媛　英文审校：尹淑英）２１１沈　阳　工　业　大　学　学　报第４５卷。

spectresimulation介绍

Outputs/To be plotted/selected on schematic
schematic子菜单用来在电路原理图上选取要显示的波形（点击连线选取节点电压，点击元件端点选取
节点电流），这个菜单比较常用
2021/4/2
30
第30页/共88页
其它有关的菜单项（3）
Outputs/Setup
2021/4/2
25
第25页/共88页
dc（直流分析）
dc（直流分析）可以在直流条件下对
temperature，Design Variable，Component
Parameter，Model Parameter进行扫描仿真
举例:对温度的扫描（测量温度系数）
电路随电源电压变化的变化曲线等
2021/4/2
9
第9页/共88页
File菜单
在File菜单下，主要的菜单项有New、Open、Exit等
New菜单项的子菜单下有 Library、Cell view两项。 Library项打开New Library窗口，Cell view项打开Create New File窗口。
Open菜单项打开相应的 Open File窗口。
11
第11页/共88页
Tools菜单
在Tools菜单下，比较常用的菜单项有
Library Manager
Library Path Editor
Technology File Manager
Library Manager项打开的是库管理器。在窗口的各部分中，分别显示的是Library、Category、Cell、 View相应的内容。
2021/4/2
15
第15页/共88页

ROHM 自动氢氧化錫 DC DC 電源轉換器模擬用戶指南说明书

User ’s GuideROHM Solution Simulator3.5V to 40V Input, 2A Single 2.2MHz Buck DC/DC Converter for AutomotiveBD9P235MUF-C / Frequency ResponseThis circuit simulates the frequency response of BD9P235MUF-C. You can customize the simulation conditions by changing the parameters of components highlighted in blue.General CautionsCaution 1: The values from the simulation results are not guaranteed. Please use these results as a guide for your design. Caution 2: These model characteristics are specifically at Ta=25°C. Thus, the simulation result with temperature variancesmay significantly differ from the result with the one done at actual application board (actual measurement).Caution 3: Please refer to the datasheet for details of the technical information.Caution 4: The characteristics may change depending on the actual board design and ROHM strongly recommend todouble check those characteristics with actual board where the chips will be mounted on.1 Simulation SchematicFigure 1. Simulation Circuit2 How to simulateThe simulation settings, such as frequency range or convergence options, are configurable from the ‘Simulation Settings’ shown in Figure 2, and Table 1 shows the default setup of the simulation.In case of simulation convergence issue, you can changeadvanced options to solve.The parameters V_VIN, V_VOUT and I_IOUT are defined in the ‘Manual Options’.Figure 2. Simulation Settings and executionTable 1. Simulation settings default setupParameters Default Note Simulation Type Frequency-Domain (Do not change Simulation Type)Start Frequency 100 Hz Simulate the frequency response for the frequency range from 100Hz to 1MHz. End Frequency 1.0e6 Hz Advanced options BalancedManual Options“.param V_VIN=12 V_VOUT=3.3 I_IOUT=1.0”See “Simulation Condition” for detailsSimulation Settings Simulate VOUTIOUTVIN3 Simulation Conditions3.1 How to define V IN, V OUT and I OUTThese parameters are used to setup the simulation conditions and BD9P235MUF-C_Average model parameters, therefore these are defined in the Manual Options as the common variables.Table 2 shows the default value of V IN, V OUT and I OUT. Those values are defined and can be set in the ‘Manual Options’ text box from Simulation Settings as s hown in Figure 3.The output voltage of VBAT and the load resistance RL are automatically set according to those parameters.Table 2. Simulation ConditionsParameters Variable Name Default Value Units DescriptionsV IN V_VIN 12 VV OUT V_VOUT 3.3 VI OUT I_IOUT 1.0 ASet V_VIN, V_VOUT and I_IOUTFigure 3. Definition of V IN, V OUT and I OUT3.2 Resistive Load RLRL is the resistive load and its resistance is determined from V OUT and I OUT. The resistance value is defined as the equation below.Table 3. Resistive loadInstance Name Default Value UnitRL {V_VOUT/I_IOUT} ohm4 BD9P235MUF-C_Average modelThe simulation model in this circuit is designed for frequency response, and the functions not related to frequency response are not implemented.Table 4. BD9P235MUF-C_Average model terminals used for frequency responseTerminals DescriptionPVIN, VIN Power supply inputEN Enable inputPGND Power groundOCP_SEL Over current selector inputSW Switching nodeGND GroundTable 5. BD9P235MUF-C_Average model terminals NOT used for frequency responseTerminals DescriptionBST Input is ignore (no switching operation in this model)MODE Input is ignore (no switching operation in this model)SSCG Input is ignore (no switching operation in this model)RESET The function is not implementedVOUT_DIS Input is ignore (no switching operation in this model)VOUT_SNS Function not implementedVCC_EX Function not implementedVREG Function not implemented(Note 1) This model is not compatible with the influence of ambient temperature.(Note 2) This model is not compatible with the external synchronization function.(Note 3) Use the simulation results only as a design guide and the data reported herein is not a guaranteed value.4.1 BD9P235MUF-C Simulation Model ParametersBD9P235MUF-C_Average model has its parameters shown in Table 6. All the parameters are pre-defined and fixed in the simulation. V_VIN is substituted to VIN_VOLTAGE as shown in Table 6.Table 6. Parameter ListParameters Values DescriptionVIN_VOLTAGE {V_VIN} VIN voltageFigure 4. Property Editor of BD9P235MUF-C_Average model5 Peripheral ComponentsTo set parameters of components, open ‘property’ by double click or right click on a component. You can input a value toa property text box if available. Please refer to the hands-on manual for more details.5.1 Bill of MaterialTable 7 shows the list of components used in the simulation schematic. Each of the capacitor and inductor has the parameters of equivalent circuit shown below. The default value of equivalent components are set to zero except for the parallel resistance of L1. You can modify the values of each component.Table 7. List of components used in the simulation circuitType Instance Name Default Value UnitsCapacitor CIN1 0.1 µFCIN2 4.7 µFCREG 1.0 µFCOUT1 22 µFCOUT2 22 µFInductor L1 3.3 µH5.2 Capacitor Equivalent Circuits(a) Property editor (b) Equivalent circuitFigure 5. Capacitor property editor and equivalent circuit5.3 Inductor Equivalent Circuits(a) Property editor (b) Equivalent circuitFigure 6. Inductor property editor and equivalent circuitThe default value of PAR_RES is 6.6kohm.(Note 5) These parameters can take any positive value or zero in simulation but it does not guarantee the operation of the IC in any condition. Refer to the datasheet to determine adequate value of parameters.6 Open Loop Transfer Function (OLTF) MonitorOLTF1 is the insert model to measure AC open loop transfer function and is inserted to acquire the gain and phase output. To monitor the gain and phase from OLTF1, select probe items ‘dbMag’ for gain and ‘phase’ for phase plot, respectively from ‘property’ of OLTF1.Figure 7. Probe Items of OLTF17 Link to the product information and tools7.1 Product webpage link:https:///products/power-management/switching-regulators/integrated-fet/buck-converters-synchronous/bd9p235muf-c-product7.2 Related documentsThe application notes are available from ‘Documentation’ tab of the product page.7.3 Design assist tools a re available from ‘Tools’ tab of the product page.The Circuit constant calculation sheet is useful for Febiding the application circuit constants.NoticeROHM Customer Support System/contact/Thank you for your accessing to ROHM product informations.More detail product informations and catalogs are available, please contact us.N o t e sThe information contained herein is subject to change without notice.Before you use our Products, please contact our sales representative and verify the latest specifica-tions :Although ROHM is continuously working to improve product reliability and quality, semicon-ductors can break down and malfunction due to various factors.Therefore, in order to prevent personal injury or fire arising from failure, please take safety measures such as complying with the derating characteristics, implementing redundant and fire prevention designs, and utilizing backups and fail-safe procedures. ROHM shall have no responsibility for any damages arising out of the use of our Poducts beyond the rating specified by ROHM.Examples of application circuits, circuit constants and any other information contained herein areprovided only to illustrate the standard usage and operations of the Products. The peripheral conditions must be taken into account when designing circuits for mass production.The technical information specified herein is intended only to show the typical functions of andexamples of application circuits for the Products. ROHM does not grant you, explicitly or implicitly, any license to use or exercise intellectual property or other rights held by ROHM or any other parties. ROHM shall have no responsibility whatsoever for any dispute arising out of the use of such technical information.The Products specified in this document are not designed to be radiation tolerant.For use of our Products in applications requiring a high degree of reliability (as exemplifiedbelow), please contact and consult with a ROHM representative : transportation equipment (i.e. cars, ships, trains), primary communication equipment, traffic lights, fire/crime prevention, safety equipment, medical systems, servers, solar cells, and power transmission systems.Do not use our Products in applications requiring extremely high reliability, such as aerospaceequipment, nuclear power control systems, and submarine repeaters.ROHM shall have no responsibility for any damages or injury arising from non-compliance withthe recommended usage conditions and specifications contained herein.ROHM has used reasonable care to ensur e the accuracy of the information contained in thisdocument. However, ROHM does not warrants that such information is error-free, and ROHM shall have no responsibility for any damages arising from any inaccuracy or misprint of such information.Please use the Products in accordance with any applicable environmental laws and regulations,such as the RoHS Directive. For more details, including RoHS compatibility, please contact a ROHM sales office. ROHM shall have no responsibility for any damages or losses resulting non-compliance with any applicable laws or regulations.W hen providing our Products and technologies contained in this document to other countries,you must abide by the procedures and provisions stipulated in all applicable export laws and regulations, including without limitation the US Export Administration Regulations and the Foreign Exchange and Foreign Trade Act.This document, in part or in whole, may not be reprinted or reproduced without prior consent ofROHM.1) 2)3)4)5)6)7)8)9)10)11)12)13)。

COMSOL入门讲座

propagation in microwave components and photonic devices.
From Modeling to Analysis
Geometry Modeling and CAD Tools Visualization and Selection Tools Modeling Physics and Equations Meshing Solving the Model Postprocessing and Visualization
Scattered-Field Formulation VS full-wave formulation Generating a perfect plane wave Incident field direction and phase Set PML by Manual Symmetric Boundary Condition
?scatteredfieldformulationvsfullwaveformulationgeneratingaperfectplanewaveincidentfielddirectionandphasesetpmlbymanualsymmetricboundarycondition?????metalslit?complexepsilonsettingusingmanualpmltosimulateinfinitestructuredomainplotparameterscrosssectionplotparametersdefinesourcetype????modelingphysicsandequations?metalslitarray?periodicboundarycondition?superlens?pointsourcesettingb?wavelengthscanninggtiisppwid?braggratinginsppwaveguide?waveguidehbendwithsparametersscatteredfieldformulationfullwaveformulation?differentapplicationmodessctevstesctmvstm?purpose

1、下载文档前请自行甄别文档内容的完整性，平台不提供额外的编辑、内容补充、找答案等附加服务。
2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
3、如文档侵犯您的权益，请联系客服反馈,我们会尽快为您处理(人工客服工作时间：9:00-18:30)。

Separation and Puriﬁcation Technology 73 (2010) 100–105Contents lists available at ScienceDirectSeparation and PuriﬁcationTechnologyj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /s e p p urSimulation of a modiﬁed cyclone separator with a novel exhaustJihui Chen ∗,Xu LiuUtilities Department,CISDI Engineering Co.Ltd.,Chongqing 400013,PR Chinaa r t i c l e i n f o Article history:Received 23December 2009Received in revised form 10March 2010Accepted 11March 2010Keywords:Cyclone separator Separation efﬁciency Pressure drop Exhausta b s t r a c tSimulations on the separation efﬁciency and the pressure drop of a modiﬁed cyclone separator with a novel exhaust under different insert depths (250mm,350mm,450mm,and 550mm),sloping orienta-tions (0◦,90◦,180◦,and 270◦),contracting angles (15◦,30◦,and 45◦)and sloping angles (5◦,10◦,and 15◦)were performed by computational ﬂuid dynamics technique in this study.Results indicate that despite the pressure drop increases to some extent,the separation efﬁciency of such separators fundamentally exceeds that of the traditional linear-pipe-shaped ones due to the conﬁguration of the novel exhaust can be made to harmonize with the ﬂow ﬁeld within the separator.Both the separation efﬁciency and the pressure drop change with the sloping orientation,and they have the same changing rule,the max-imum at 90◦and the minimum at 270◦.The separation efﬁciency increases ﬁrstly and then decreases with increasing insert depth of the novel exhaust,and the maximum appears when the insert depth is 450mm.The separation efﬁciency of the modiﬁed separator is about 4.6–7.9%higher than that of the traditional one under different sloping angles and contracting angles.Simulation results agree relatively well with the previous experimental results.Besides,based on an overall consideration of separation efﬁciency and pressure drop,we think that it is the most economic condition when the contracting angle is 30◦.However,the contracting angle of 45◦is still more suitable when the key problem is the separation efﬁciency despite its largest resistance loss.© 2010 Elsevier B.V. All rights reserved.1.IntroductionCirculating ﬂuidized bed (CFB)combusting technology has developed fast during the last three decades and is widely used in the world due to its good characteristics such as satisfactory desul-furization efﬁciency,low NOx emission and suitability for different types of fuels [1].As a key equipment for CFB boilers,gas–solid cyclone separator,which helps the circulation of the solids in the furnace,has strong effects on the combustion efﬁciency,the circu-lation rate,the desulfurization efﬁciency,and so on [2].As we know,the size of cyclone separators increases rapidly with increasing the capacity of CFB boilers.However,large cyclone not only results in some manufacturing,installation and operation problems,but also has lower separation efﬁciency [3].In recent years,experimental and theoretical investigations have been conducted in order to achieve the less size and higher separation efﬁciency of the gas–solid cyclone separators.The France’s Stein Corporation developed a cyclone separator with an offset mounted exhaust [4].Germany Siemens Corporation developed a cyclone separator with guide vanes at the top of the∗Corresponding author.Tel.:+862363547546;fax:+862363548888.E-mail address:chenjihui zl@ (J.Chen).separator [5].A novel double-inlet square cyclone separator with two furnaces developed by Clean Coal Combustion Laboratory of Chongqing University is very suitable to the scale-up of the CFB boilers [6].A type of square cyclone separator with downward exhaust was developed and granted a Chinese patent [7].Its sep-aration efﬁciency was shown as good as that of the traditional cyclone of circular cross-section separator and its particle cut-diameter is about 15␮m.Moreover,researchers have carried out many investigations on the effect of cylinder height and diameter,the cone opening size,the exhaust diameter,length,insert depth,offset and turbulence intensity and boundary layer on the separa-tor performance [8–17].These results lay important foundations for optimizing and scale-up of cyclone separators.As we know,however,the inside ﬂow ﬁeld of a single-inlet cyclone separator has not axial symmetry.The tangential velocity and the radial velocity at different angles in cyclone separator are also not uniform [18,19].Thus,the shape of exhaust will have an inﬂuence on the ﬂow and on the separation efﬁciency and pressure drop of cyclone separator.In the previous work,experimental investigations have been carried out on the inﬂuence of the bottom-contracted and edge-sloped exhaust on the separation efﬁciency and the pressure drop of a cyclone separator under different insert depths and sloping orientations in a visual cold large-scale CFB setup [2].1383-5866/$–see front matter © 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.seppur.2010.03.007J.Chen,X.Liu /Separation and Puriﬁcation Technology 73 (2010) 100–105101Fig.1.Schematics of the novelexhaust.Fig.2.The geometric models of two cyclones.In order to validate the experimental results and comprehen-sively examine the performance of the modiﬁed cyclone separator with the novel exhaust,simulations were performed by computa-tional ﬂuid dynamics technique in this study.2.Task formulation and numerical methodIn the present work,the effect of a novel exhaust on the separa-tion efﬁciency and pressure drop of a modiﬁed cyclone separator under different exhaust insert depths (250mm,350mm,450mm,and 550mm),sloping orientations (0◦,90◦,180◦,and 270◦),con-tracting angles (15◦,30◦,and 45◦)and sloping angles (5◦,10◦,and 15◦)will be studied in a CFB system.Besides,all the results will be compared with that of the traditional cyclone separator.2.1.Geometric modelIt should be pointed out that the sloping orientation represents the angle between the two directions of the sloped edge and the inlet stream.And the angle is described in clockwise when looking down the exhaust.The modiﬁed exhaust and the description of contracting angle and sloping angle of the exhaust can be seen in Fig.1.In addition,the three-dimensional geometric models of two cyclones are shown in Fig.2.2.2.Mathematic model(1)Governing equations for continuous phaseContinuity equation:∂ ∂t +∂∂x i ( v i )=0(1)Momentum equations:∂∂t ( i )+∂∂x j ( i j )=−∂p ∂x j +∂∂x j∂ i∂x i +∂ j ∂x j−23ıij∂ i ∂x i−∂( i j )∂x ji(2)Reynolds stress model:∂( v i v j)∂t+∂( v k v i v j)∂x k=D ij +p ij + ij −εij(3)wherep ij =−v i v k∂v j ∂x k+v j v k∂v i∂x k,D ij =∂∂x kt k ∂v i vj ∂x k,p ij =−v i v k∂v j∂x k +v j vk∂v i∂x k, ij = ij 1+ ij 2, ij 1=−C 1εkv i v j−23ıij k,ij 2=−C 2p ij −23ıij G k,G k =−2 v i v k∂v i∂x k,and εij =23ıij ε∂∂t ( k )+∂∂x j ( i k )=−∂∂x j+tk∂k ∂x j+12(P ij+G ij )− ε(4)102J.Chen,X.Liu/Separation and Puriﬁcation Technology73 (2010) 100–105Table1Values of the coefﬁcients in Eq.(9).Item k1k2k3Re p124.00.00.0≤1.0222.730.09 3.690.1–1.0329.17−3.89 1.22 1.0–10446.5−116.670.61710–100598.33−27780.364100–1000∂∂t ( ε)+∂∂x j( iε)=−∂∂x j+tε∂ε∂x j+εk12Cε1P ij−Cε2 ε(5)where Cε1=1.44,Cε2=1.92, k=0.82, ε=1.0,C1=1.8,C2=0.6,and i,j=x,y,z.(2)Equations for discrete particlesm p dw pdt=F D+F G+F C(6)where F D=(1/ p)(C D Re p/24)m g(v g−v p),F C= P d3v2t/6r,and F G= P d3g/6are drag force,centrifugal force,and gravitational force,respectivelyp= p d2p18(7)Re= p d pv g−v p(8)where m p,d p, p,and p are the mass,diameter,density and velocity of particles;v t and v g are tangential velocity of particles and air-ﬂow velocity;C D and are the resistance coefﬁcient and viscosity coefﬁcient,respectively.The resistance coefﬁcient can be described asC D=k1Re p+k2Re2p+k3(9)The coefﬁcients are shown in Table1.2.3.Simulation conditionsFloater is used as simulating material and it has a density of 620kg/m3.The particles in this simulations have four sizes(0.3mm, 0.5mm,0.7mm,and0.9mm),and each of these are5000.The amount of the particles separated by the cyclone separator divided by the total amount of the original particles is the separation efﬁ-ciency.The simulating conditions are shown as follows:(1)the exit boundary is deﬁned pressure exit and the exit pressure is0Pa;(2) the boundary of the particles capturing exit is deﬁned as capturing wall of particle phase;(3)the rigid wall boundary and the par-ticle phase are respectively deﬁned as no slip boundary condition and reﬂection;(4)considering the coupling between gas phase and solid phase.Note that there has not continuity of the separation efﬁciency because only the above typical particles with four sizes were tracked.However,it still has not any effect on the comparison of the separationefﬁciency between two different cyclone separators under the same operation conditions.The pressure drops of the two separators are all obtained by subtracting the average pressure value at the cross-section of the separator entrance from that of the exhaust outlet.Fig.3.Tangential velocity in the contraction port zone of the modiﬁed cyclone. 3.Simulation results and analysis3.1.Theﬂowﬁelds and pressureﬁelds of two different cyclone separatorsparison of theﬂowﬁeldsThe distributions of the tangential velocity and the axial veloc-ity of the modiﬁed cyclone separator are shown in Figs.3and4, respectively.The results show that the tangential velocity and the axial veloc-ity around the bottom of the novel exhaust have not axial symmetry due to the contraction and sloping of the exhaust.Fig.3also indi-cates that the tangential velocity at the bottom of the exhaust (X=0.05m)increases with the height,and the position of the max-imum approaches gradually to the geometric centre of the exhaust. Besides,from Fig.4we can also see that the maximal value of the up-ﬂows of the axial velocity deviates from the geometric centre of the exhaust.And the down-ﬂows between the positivehalf and the negative half of the x-axis are all asymmetric.By contrast,the distributions of the tangential velocity and the axial velocity of the traditional linear-pipe-shaped cyclone separa-tor are shown in Figs.5and6,respectively.Fig.4.Axial velocity in the contraction port zone of the modiﬁed cyclone.J.Chen,X.Liu /Separation and Puriﬁcation Technology 73 (2010) 100–105103Fig.5.Tangential velocity at the exhaust’s bottom zone of the traditional cyclone.It can be found by comparing Fig.3with Fig.5that the axial symmetry of the tangential velocity of the modiﬁed cyclone sep-arator is not so good as that of the traditional cyclone separator due to the contraction and sloping of the exhaust.Besides,the tangential velocities of two different separators have remarkable differences in distribution regularities.Fig.5shows that the tan-gential velocity of the outside quasi-free vortex of the traditional separator ﬁrst increases gradually then decreases slowly along the radial direction.On the contrary,the tangential velocity oftheFig.6.Axial velocity at the exhaust’s bottom zone of the traditional cyclone.modiﬁed separator changes markedly,especially around the zone where the tangential velocity becomes maximal (Fig.3).In addition,because the sloped port is at the positive direction of x-axis and the boundary layer effect in this region has been greatly changed,there does not appear a phenomenon that the tangential velocity equals to zero.At the same time,the position of the maximum of the tan-gential velocity moves to the geometric centre of the exhaust.As a result,the particles inside the quasi-forced vortex will be carried into the outside quasi-free vortex,and then the separation process goeson.Fig.7.The pressure ﬁelds of two different cyclones.104J.Chen,X.Liu/Separation and Puriﬁcation Technology73 (2010) 100–105Table2Comparison of separation efﬁciency and pressure drop between two different cyclone separators.Insert depth(h/mm) (˛=90◦,ˇ=15◦,=10◦)Sloping orientation(˛/◦)(insert depth450mm,ˇ=15◦,=10◦)Sloping angle( /◦)(insert depth450mm,ˇ=15◦,˛=90◦)Contracting angle(ˇ/◦)(insert depth450mm,=10◦,˛=90◦)The traditionalcyclone(insertdepth=450mm)2503504505500◦90◦180◦270◦5◦10◦15◦15◦30◦45◦Actual valueÁ(%)−3.5+2.3+5.5+5.3+1.8+5.5+3.2−1.6+4.8+5.5+5+5.5+7+7.591%P(%)+4.3+9+10.2+17.9+4.7+10.2+3.5−2.7+14+10.2−3.2+10.2+31.3+1861980PaOn the other hand,there are some differences on the axial veloc-ity between the two different cyclone separators(Figs.4and6). However,the main difference is reﬂected in the position’s change of the maximum of the axial velocity.As shown in Fig.6,the central position of the axial velocity of the traditional separator changes continually.So there will be much more secondary vortexes inside theﬂowﬁeld in this kind of separator.These secondary vortexes will effect the movement of the up-ﬂows and the down-ﬂows, resulting in a worse separation.Therefore,if the sloping orienta-tion of the modiﬁed exhaust can be harmonize with the innerﬂow ﬁeld of separator,the newﬂowﬁeld will be more favorable for separation process.Both the previous experimental results[2]and the simulation results showed in Table2in this paper show that the sloping orientation has a remarkable effect on separation efﬁciency,the maximum at90◦and the minimum at270◦.Simulation results agree relatively well with experimental results.parison of the pressureﬁeldsFig.7illustrates the pressure distributions of the two different cyclone separators.From Fig.7we can see that the symmetry of the pressureﬁeld of the traditional linear-pipe-shaped separator is better than that of the modiﬁed separator due to the contraction and sloping of the exhaust.Besides,the pressure drop of the modiﬁed separator is higher than that of the traditional separator under the same operation con-ditions.The major reason lies in that the conﬁguration change of the exhaust’s bottom will decrease the totalﬂow area of the exhaust, resulting in the increase of theﬂuid velocity.Consequently,the resistance loss(pressure drop)increases.The simulation result of the pressure drop is10.6%and the previous experimental result is about10%[2].They show good agreement.3.2.The separation efﬁciencies and pressure drops of twodifferent cyclone separatorsIt should be pointed out that the simulation results of theﬂow ﬁelds and the pressureﬁelds obtained here are similar to those in Section3.1.This paper will not give a full illustration of what have been presented in the previous discussion so as not to fall into unnecessary repetition.However,all the simulation results of sep-aration efﬁciency and pressure drop have been listed in Table2. Note that the separation efﬁciency and the pressure drop of the traditional linear-pipe-shaped separator are regarded as basic val-ues in the table.And all the results of the modiﬁed separator are the relative values which are gained by subtracting the calculation values of the modiﬁed separator from the basic values.From Table2we canﬁnd that the separation efﬁciencies of the modiﬁed separator fundamentally exceed that of the traditional separator,while the pressure drops of the modiﬁed separator are usually higher than that of the traditional one.The effects of sloping angles and contracting angles on the sepa-ration efﬁciency are all positive.As shown in Table2,the separation efﬁciencies increase about4.6–7.9%under different sloping angles and contracting angles.When the contracting angle is45◦,the sepa-ration efﬁciency has a largest rise(7.9%),whereas the pressure drop increases remarkably(3062Pa).The main reason is that the total ﬂow area of the exhaust decreases with the increase of the con-tracting angle,results in reducing the probability of short-circuit of particles.But,at the same time,theﬂuid velocity will also increase and result in the increase of the resistance loss.By contrast,there is a higher increase of the separation efﬁciency(7.3%)with a lower rise of the pressure drop(551Pa)when the contracting angle is 30◦.Therefore,it is the most economic condition.However,when the key problem is the separation efﬁciency,the exhaust with a contracting angle of45◦is clearly more suitable.On the other hand,Table2also indicates that the inﬂuence of exhaust insert depth and sloping orientation on the separation efﬁ-ciency of the modiﬁed separator is more complex.The separation efﬁciency increasesﬁrstly and then decreases with the increase of the insert depth,and the maximum at450mm.The separation efﬁ-ciency becomes maximal when the sloping orientation is90◦and the minimum at270◦.In addition,due to there is more friction loss when the insert depth increases,the pressure drop increases slightly with increasing the insert depth of the exhaust.Besides,the pressure drop also changes with the orientation of the sloped edge, and the change rule is similar to that of the separation efﬁciency.4.ConclusionsSimulations on the separation efﬁciency and pressure drop of a modiﬁed cyclone separator with a novel exhaust under different exhaust insert depths,sloping orientations,contracting angles and sloping angles were performed in this study,and some signiﬁcant results were summarized as follows:(1)Despite that the pressure drop of the modiﬁed cyclone sep-arator increases to some extent,the separation efﬁciency of such separators fundamentally exceeds that of the traditional linear-pipe-shaped ones due to the conﬁguration of the novel exhaust can be made to harmonize with the innerﬂowﬁeld of the modiﬁed separator.(2)Both the separation efﬁciency and the pressure drop changewith the sloping orientation,and they have the same rule of change,the maximum at90◦and the minimum at270◦.The calculating results agree relatively well with the previous experimental results.(3)Separation efﬁciency increasesﬁrstly and then decreases withincreasing the insert depth of the novel exhaust,and the max-imum appears when the value of the insert depth is450mm.The simulation results and the previous experimental results also show reasonably good agreement.(4)Separation efﬁciencies increase about4.6–7.9%under differ-ent sloping angles and contracting angles.Based on an overall consideration of separation efﬁciency and pressure drop,we think that it is the most economic condition when the contract-ing angle is30◦.However,the contracting angle of45◦is still more suitable when the key problem is the separation efﬁciency despite its largest resistance loss.J.Chen,X.Liu/Separation and Puriﬁcation Technology73 (2010) 100–105105Nomenclatureh exhaust insert depth(mm)u cyclone inlet stream velocity(m/s)P pressure drop(Pa)Greek letters˛sloping orientation(◦)ˇcontracting angle(◦)sloping angle(◦)Áseparation efﬁciency(%)References[1]P.A.Basu,S.Fraser,Circulating Fluidized Bed Boiler—Design and Operations,Butterwoths-Heinemann-Reed Publishing,USA,1991.[2]J.H.Chen,X.F.Lu,H.Z.Liu,C.Y.Yang,Effect of the Bottom-contracted and Edge-sloped Exhaust on the Cyclone Separator Performance,Chem.Eng.J.129(2007) 85–90.[3]K.F.Cen,X.D.Li,Y.X.Li,J.H.Yan,Y.L.Shen,S.R.Liang,M.J.Ni,Experimentalstudy ofﬁnned tubes impact gas–solid separator for CFB boilers,Chem.Eng.J.66(1997)159–169.[4]L.Ireneusz,S.Joachim,K.Frank,K.Stanislaw,Operational experiences with highefﬁciency cyclones:Comparison between boiler A and B in the Zeran Power Plant-Warsaw,Poland,in:Proceedings of the17th International Conference on Fluidized Bed Combustion,Jacksonville,United States,May18–21,2003.[5]U.Muschelknautz, E.Muschelknautz,Special Design of Inserts and ShortEntrance Ducts to Recirculating Cyclones.Circulating Fluidized Bed Technology, Science Press,Beijing,PR China,1997,pp.597–602.[6]Y.J.Huang,X.F.Lu,H.Z.Liu,R.S.Amano,Investigation on the structure andarrangement of a large-scale circulatingﬂuidized bed boiler,in:Proceedings of the19th International Conference on Fluidized Bed Combustion,Vienna, Austria,May21–24,2006.[7]K.F.Cen,et al.,Square-shaped heat exchanging cyclone separator withdownward-exhaust,China Patent,No.942,232,2006,76.3.[8]C.J.Stairmand,The design and performance of cyclone separators,Trans.Inst.Chem.Eng.29(1951)356–383.[9]J.Dirgo,D.Leith,Performance of theoretically optimized cyclones,Filtr.Separat.22(1985)199–225.[10]H.S.Bryant,R.W.Silverman,F.A.Zenz,How dust in gas affects cyclone pressuredrop,Hydrocarbon Process.62(1983)87–90.[11]R.B.Xiang,S.H.Park,K.W.Lee,Effects of cone dimension on cyclone perfor-mance,Aerosol Sci.32(2001)549–561.[12]W.B.Smith,R.R.Wilson,D.B.Harris,Aﬁve-stage cyclone system for in situsampling,Environ.Sci.Technol.13(1979)1387–1392.[13]J.Dirgo,D.Leith,Cyclone collection efﬁciency:comparison of experimen-tal results with theoretical predictions,Aerosol Sci.Technol.4(1985)401–411.[14]H.Buttner,Size separation of particles from aerosol samples using impactorsand cyclones,Part.Part.Syst.Char.5(1988)87–93.[15]J.C.Kim,K.W.Lee,Experimental study of particle collection by small cyclones,Aerosol Sci.Technol.12(1990)1003–1015.[16]D.L.Ioza,D.Leith,The logistic function and cyclone fractional efﬁciency,AerosolSci.Technol.12(1990)598–606.[17]X.D.Li,J.H.Yan,Y.C.Cao,M.J.Ni,K.F.Cen,Numerical simulation of the effects ofturbulence intensity and boundary layer on separation efﬁciency in a cyclone separator,Chem.Eng.J.195(2003)235–2400.[18]J.J.Derksen,S.Sundaresan,H.E.A.Van Den Akker,Simulation of mass-loadingeffects in gas–solid cyclone separators,Powder Technol.163(2006)59–68.[19]R.Zhang,P.Basu,Simple model for prediction of solid collection efﬁciency of agas–solid separator,Powder Technol.147(2004)86–93.。