Artificial Neural Network Maximum Power Point Tracker for Solar Electric Vehicle

TSINGHUA SCIENCE AND TECHNOLOGYISSN1007-021412/23pp204-208Volume10,Number2,April2005Artificial Neural Network Maximum Power Point Trackerfor Solar Electric VehicleTheodore Amissah OCRAN, CAO Junyi (曹军义)**,CAO Binggang (曹秉刚), SUN Xinghua (孙兴华)Research and Development Center for Electric Vehicle, Xi’an Jiaotong University, Xi’an 710049, ChinaAbstract:This paper proposes an artificial neural network maximum power point tracker (MPPT) for solar electric vehicles. The MPPT is based on a highly efficient boost converter with insulated gate bipolar transis-tor (IGBT) power switch. The reference voltage for MPPT is obtained by artificial neural network (ANN) with gradient descent momentum algorithm. The tracking algorithm changes the duty-cycle of the converter so that the PV-module voltage equals the voltage corresponding to the MPPT at any given insolation, tempera-ture, and load conditions. For fast response, the system is implemented using digital signal processor (DSP).The overall system stability is improved by including a proportional-integral-derivative (PID) controller, which is also used to match the reference and battery voltage levels. The controller, based on the information sup-plied by the ANN, generates the boost converter duty-cycle. The energy obtained is used to charge the lith-ium ion battery stack for the solar vehicle. The experimental and simulation results show that the proposed scheme is highly efficient.Key words:artificial neural network; maximum power point tracker (MPPT); photovoltaic module; digital signal processor; solar electric vehicleIntroductionPhotovoltaic (PV) generation is gaining increased im-portance as a renewable source of energy. The undesir-able rapid changes of solar power, which usually occur in a running vehicle, arise as a result of shade from buildings, large trees, and clouds in the sky. Conven-tional PV systems have difficulties in responding to rapid variations due to shade. The main drawbacks of PV systems are that the initial installation cost is con-siderably high and the energy conversion efficiency (from 12% to 29%) is relatively low. Furthermore, in most cases, PV systems require a power conditioner (dc-dc or dc-ac converter) for load interface. Therefore, the overall system cost could be reduced drastically by using highly efficient power conditioners, such as the maximum power point tracker (MPPT), to extract and maintain the peak power from the PV module even when the above-mentioned unfavorable conditions occur.Various methods of maximum power tracking have been considered in PV power applications[1-12]. Among the hill climbing methods[1-5], the perturb and observe (P&O) method tracks maximum power point (MPP) by repeatedly increasing or decreasing the output voltage at MPP of the PV module. This method requires calcu-lation of d P/d V to determine the MPP[1,2,4]. Though the method is relatively simple to implement, it cannot track the MPP when the irradiance changes rapidly. Also, the method oscillates around the MPP instead ofReceived: 2004-03-24﹡﹡To whom correspondence should be addressed.E-mail: caojy@;Tel: 86-29-87570252Theodore Amissah OCRAN et al ：Artificial Neural Network Maximum Power Point Tracker (205)directly tracking it. The incremental conductance technique (ICT) is the most accurate [6,7] among the other methods. This method gives good performance under rapidly changing conditions. However, the com-plex calculation of d I /d V and the complicated algorithm require use of a digital signal processor (DSP), which will usually increase the total system cost. The MPP tracking method using the short circuit current of the PV module utilizes the fact that the op-erating current at MPP of the PV module is linearly proportional to the short circuit current of the PV mod-ule [9]. Under rapidly changing atmospheric conditions, this method has a fast response speed of tracking the MPP, but the control circuit is complicated. The MPP tracking method using open circuit voltage of the solar panel [11] utilizes the fact that the operating voltage at MPP is almost linearly proportional to open circuit voltage at MPP of the PV module (using 76% of open circuit voltage as the MPP voltage). This method is very simple and cost-effective, but the reference volt-age does not change between samplings. MPPTs using the fuzzy logic [13,14] and the artificial neural network (ANN) [15] have been reported. These studies show that such modern control algorithms are capable of improv-ing the tracking performance as compared to the conventional methods.In this paper, we propose an ANN MPPT for solar electric vehicle PV systems. The tracking algorithm changes the duty ratio of the converter so that the PV module voltage equals the voltage corresponding to the MPP at that atmospheric condition. This adjustment is carried out by using the back-propagation ANN. The reference voltage to the MPP is obtained by an offline trained ANN. The controller generates the boost converter duty-cycle.1 PV Module CharacteristicsThe solar array characteristics significantly influencethe design of the converter and the control system, so the PV characteristics will be briefly reviewed here. The solar array is a nonlinear device and can be repre-sented as a current source model, as shown in Fig. 1. The traditional I -V characteristics of a solar array, when neglecting the internal shunt resistance, are givenby the following equation:Fig. 1 Equivalent circuit of a solar cell()o o s o g sat o o s sh exp 1V I R qI I I V I R AkT R+⎧⎫⎡⎤=−+−−⎨⎬⎢⎥⎣⎦⎩⎭ (1)where o I and are the output current and output voltage of the solar array; o V g I is the generated current under a given insolation; sat I is the reverse saturation current; q is the charge of an electron; k is the Boltzmann constant; A is the ideality factor for a P-N junction; is the array temperature; and are the intrinsic series and shunt resistances of the solar array.T s R sh R The saturation current of the solar array varies with temperature according to the following equation:3sat orr T I I T ⎡⎤=⎢⎥⎣⎦exp GO 11r qE Bk T T ⎡⎤⎛⎞−⎢⎥⎜⎢⎥⎝⎠⎣⎦⎟ (2) ()g sc 25100I I I K T λ=+−⎡⎣⎤⎦ (3)where is the reference temperature; r T or I is the saturation current at ; is the band-gap en-ergy of the semiconductor used in the solar array; B is also an ideality factor; r T GO E sc I is the short circuit current at 25°C; I K is the short-circuit current temperature coefficient and λ is the insolation in mW/cm².Equations (1)-(3) are used in the development of computer simulations for the solar array. The Matlab programming language is used. Figure 2 show the simulated current-voltage and power-voltage curves for the solar array at different insolations and different temperatures. These curves show that the output char-acteristics of the solar array are nonlinear and greatly affected by the solar radiation, temperature, and load condition. Each curve has a maximum power point (P max ), which is the optimal operating point for the ef-ficient use of the solar array.2 Artificial Neural NetworkANN technology has been successfully applied to206 Tsinghua Science and Technology , April 2005, 10(2): 204–208Fig. 2 Current-voltage and power-voltage curves for the solar array at different insolations and different temperatures (S is the solar radiation)solve very complex problems. Recently, its application in various fields is increasing rapidly [16,17].The instantaneous sum of error squares or error en-ergy at iteration is given byn 21()()2j j Cn e ε∈=∑n (4) where neuron lies in a layer to the right of neuron , and neuron lies in a layer to the right of neuronwhen neuron is a hidden unit;j i k j j ()j e n is the er-ror signal at the output of neuron for iteration ; and the set includes all the neurons in the outer layer of the network. The correction j n C ()ji w n ∆ to the synaptic weight ()ji w n is given by()()()()1ji ji j i w n w n n y n αηδ∆=∆−+ (5)where α is the momentum constant; η is the learn-ing-rate parameter of the back-propagation algorithm; is the local gradient. The error signal at the out-put is defined as()j n δ()()()j j j e n d n y n =− (6) where ()j d n is the desired response or wanted tar-gets and is the output signal. Adjustment of the weights for these layers is given by()j y n ()()()()()11ji ji ji j i w n w n w n n y n αηδ+=+−+i (7)From Eqs. (4) and (6), the mean squared error per-formance index can be rewritten as()()()()2ref A 1.2n V n V n ε=−The network training is performed repeatedly until theperformance index falls below a specified value, ideally to zero. In other words (2ref A V V ε=−)ε→0 implies 0, and then the connecting weights of the network are adjusted in such a way that the array voltage is identically equal to the maxi-mum power point voltage . At this stage the refer-ence voltage becomes equal to the maximum power point voltage .()→2ref A V V −A V mp V ref V mp V The configuration of the proposed three-layer feed-forward neural network function approximator is shown in Fig. 3. The neural network is used to obtain the voltage of the maximum power ()mp V n of the so-lar panel. The network has three layers: an input, a hidden, and an output layer. The numbers of nodes are two, four, and one in the input, the hidden, and output layers, respectively. The reference-cell open circuit voltage ()oc V n and the time parameter ()T n are supplied to the input layer of the neural network. These signals are directly passed to the nodes in the next hid-den layer. The node in the output layer provides the identified maximum power point voltage ()mp V n . The nodes in the hidden layer get signals from the inputTheodore Amissah OCRAN et al ：Artificial Neural Network Maximum Power Point Tracker (207)layer and send their output to the node in the output layer. The sigmoid activation function is utilized in the layers of the network. The training program calculates the connecting weights with the bias for the input to hidden layer mapping, the connecting weights with bias for the hidden layer tooutput layer mapping. During the training, the con-necting weights are modified recursively until the best fit is achieved for the input-output patterns in the train-ing data. The training of the net was accomplished off-line using Matlab.I {1,1}W {1}b L {2,1}W {2}b Fig. 3 Feed-forward neural network function approximator3 Experimental/Simulation ResultsA PV array used for the collection of the experimental data is JDG-M-45 (Germany) type modules. The mod-ule has a maximum power output of 45 W and a 20-V open-circuit voltage at an irradiation of 1000 W/ and a temperature. The PV module specifica-tions provided by the manufacturer in Table 1.2m 25C °Table 1 Electrical specifications for the PV module(JDG-M-45) Maximum power, P m (W) 45 Short circuit current, I sc (A) 2.93 Open circuit voltage, V oc (V) 20 Voltage at max power point, V mp (V) 17.1 Current at max power point, I mp (A) 2.64 Size (mm ×mm ×mm) 971×441×28Mass (kg)5The feed-forward back-propagation ANN, as shown in Fig. 3 was trained with values obtained from ex-perimental data of the reference cell. Gradient descent algorithm was used in training as it improves the per-formance of the ANN, reducing the total error by changing the weights along its gradient. The training parameters are as follows: learning rate parameterη=0.1; momentum factor α=0.9; number of trainingiterations=10 000; error goal=0.000 001. The conver-gence error for the training process is shown in Fig. 4.Fig. 4 Convergence error for the neural network training processThe parameters used for the training of the ANN and the output values given by the ANN after the training can be found in Table 2. Various sets of reference cell open circuit voltage V oc and a time parameter T (as shown in Table 2) are supplied as the input to the ANN. In order to validate the learning capability of the ANN, other sets of V oc different from the one in Table 2 were also supplied to the ANN, which gave out values of V mp as expected. The software Matlab was used in the training of the ANN.Table 2 The results of the trained data given by the ANNHour of the day9:00 9:30 10:00 10:30 11:00 11:30 12:00 12:30 13:00 13:3014:00 Time parameter −1 −0.9 −0.8 −0.7 −0.6−0.5−0.4−0.3−0.2−0.1 0V oc measured 0.555 0.519 0.516 0.529 0.517 0.514 0.516 0.510 0.511 0.5090.508 V mp measured0.422 0.394 0.392 0.402 0.393 0.391 0.392 0.388 0.388 0.3870.386 V mp given by ANN0.4210.3790.3880.4160.395 0.394 0.392 0.388 0.388 0.3860.385208 Tsinghua Science and Technology , April 2005, 10(2): 204–208⎤⎥⎦The weights to the hidden layer 1 from input 1 are asfollows:I 2.1706 5.3865 3.2326 5.1766{1,1}.109.830927.943597.72508.5640W −−−⎡=⎢−−⎣ The weights to the output layer 2 are I {2,1}W =[]0.20390.00650.05610.3029.−− The bias tolayer 1 is {1}[60.0887 12.996 54.4753b =−−]1.5969.− The bias to layer 2 is .{2}[0.0397]b =−4 ConclusionsAn artificial neural network MPPT for charging the battery stack of a solar (hybrid) vehicle has been pro-posed in this paper. An off-line ANN, trained using a back-propagation with gradient descent momentum al-gorithm, is utilized for online estimation of reference voltage for the feed-forward loop. Experimental data is used for the offline training of the ANN, and software Matlab is used in the training of the net. The precision of the estimation has been verified by the graph of the convergence error. The proposed method has several advantages over the conventional methods, particularly in that there is no need for voltage and current sensors, and in that it avoids a complex calculation of power. The experimental and simulation results show that the proposed scheme is highly efficient. References[1] Hua C, Lin J, Shen C. Implementation of a DSP-controlledphotovoltaic system with peak power tracking. IEEE Trans. Ind. Electron ., 1998, 45(1): 99-107.[2] Koutroulis E, Kalaitzakis K, Voulgaris N C. Developmentof a microcontroller-based photovoltaic maximum power point tracking control system. IEEE Trans. Power Elec-tron., 2001, 16(1): 46-54.[3] Kuo Y C, Liang T J, Chen J F. Novel maximum powerpoint tracking controller for photovoltaic energy conver-sion system. IEEE Trans. Ind. Electron., 2001, 48(3): 594-601.[4] Sullivan C R, Powers M J. A high-efficiency maximumpower point tracker for photovoltaic arrays in a solar-powered race vehicle. In: Proc. IEEE Power Electron. Spec. Conf., Seattle, WA, USA, 1993: 574-580.[5] Simoes M G, Franceschetti N N. A risc-microcontrollerbased photovoltaic system for illumination applications. In: Proc. IEEE Applied Power Electron. Conf., New Orleans,LA, USA, 2000: 115-1156.[6] Tse K K, Ho M T, Chung H S, Hui S Y. A novel maximumpower point tracker for PV panels using switching fre-quency modulation. IEEE Trans. Power Electron., 2002, 17(6): 980-989.[7] Brambilla A, Gambarra M, Garutti A, Ronchi F. New ap-proach to photovoltaic arrays maximum power point track-ing. In: Proc. IEEE Power Electron. Spec. Conf., Charles-ton, SC, USA, 1999: 632-637.[8] Mutoh N, Matuo T, Okada K, Sakai M. Prediction-data-based maximum power point tracking method for photo-voltaic power generation systems. In: Proc. IEEE Power Electron. Spec. Conf., Caims, Austrilia, 2002: 1489-1494. [9] Noguchi T, Togashi S, Nakamoto R. Short-current pulse-based maximum power point tracking method for multiple photovoltaic and converter module system. IEEE Trans. Ind. Electron., 2002, 49(1): 217-223.[10] Lee D Y, Noh H J, Hyun D S, Choy I. An improvedMPPT converter using current compensation method for small scaled PV-applications. In: Proc. IEEE Applied Power Electron. Conf., Dalas, TX, USA, 2002: 540-545. [11] Enslin J H R, Wolf M S, Snyman D B, Sweigers W. Inte-grated photovoltaic maximum power point tracking con-verter. IEEE Trans. Ind. Electron., 1997, 44(6): 769-773. [12] Chen Y, Smedley K, Vacher F, Brouwer J. A new maxi-mum power point tracking controller for photovoltaic power generation. In: Proc. IEEE Applied Power Electron. Conf., Miami Beach, FL, USA, 2003: 56-62.[13] Nafeh A, Fahmy F H, El-Zahab E M A. Evaluation of aproper controller performance for maximum power point tracking of a stand-alone PV system. International Journal of Numerical Modelling , 2002, 15(4): 385-398.[14] Veerachary M, Senjyu T S, Uezato K. Feedforwardmaximum power point tracking of PV systems using fuzzy controller. IEEE Trans. Aerosp. and Electron. Syst., 2002, 38(3): 969-981.[15] Hiyama T, Kouzuma S, Ortmeyer T. Evaluation of neuralnetwork based real time maximum power tracking control-ler for PV system. IEEE Trans. on Energy Conv., 1995, 10(3): 543-548.[16] Bose B K. Artificial neural network applications in powerelectronics. In: Proc. IEEE Ind. Electron. Conf., Denver, CO, USA, 2001: 1631-1638.[17] Haykin S. Neural Networks—A Comprehensive Founda-tion, 2nd Edition. New York: Prentice Hall Inc., 1999.。