Résumé
standalone microgrid in French Guiana. ANN is an artificial intelligence techniqueused to control non-linear and complex systems. ANN associated with theLevenberg–Marquardt (LM) algorithm has many advantages, such as rapiddecision-making and improved system transients. Therefore, this technique shouldbe adapted for the control of photovoltaic (PV) systems in the tropical climate ofFrench Guiana with high variation in irradiance. The microgrid is composed of a PVsource and a storage battery to supply an isolated building which is modeled by a DCload. The PV source is controlled by an ANN-based MPPT (Maximum Power PointTracking) controller. To validate our ANN-MPPT, we compared it with one of thevery popular MPPT algorithms, which is the P&O-MPPT algorithm. The comparisonresults show that our ANN-MPPT works well because it can find the maximumpower point quickly. In the case of battery control, we tested two feed-forwardbackpropagation neural network (FFBNN) configurations called method1 andmethod2 associated with the Levenberg–Marquardt (LM) algorithm. We varied thenumber of hidden layers in each of these two FFBNN configurations to obtain theoptimal number of hidden layers for each configuration which optimizes batterycontrol. Method1 is chosen because it is better than method2, in a sense that itrespects the maximum amplitude of the battery current for our application andimproves the transient regimes of this current. This best configuration (method1) is then tested with two other learning algorithms for comparison: Bayesianregularization (BR) and scaled conjugate gradient (SCG) methods. The systemperformance with LM algorithm is better than SCG and BR algorithms. LMalgorithm improves the performance of the system in transient regimes while theresults obtained with the SGG and BR algorithms are similar. Then, we focused onthe advantage of using ANN control compared to the conventional proportionalintegral control (PI control). The comparison results showed that ANN controlassociated with the LM algorithm (ANN-LM) made it possible to reduce batterycurrent peaks by 26% in transient regimes compared to conventional PI control.Finally, we present and discuss the results of our simulation obtained with theMATLAB Simulink software.