Optimization of Permanent Magnet Synchronous Generator Output Power in Wind Power Plants with ANN Back Propagation

Abstract

The focus of this research is optimizing a wind power plant using a Permanent Magnet Synchronous Generator (PMSG). The backpropagation method of the artificial neural network system was chosen to optimize the output power of the wind power generator. Based on the simulation results, the backpropagation algorithm of the artificial neural network obtains the output power based on the input variable in the form of changing wind speed. The results show that the best value is learning rate = 0.5, error = 0.0001, max. epoch= 100000, neuron hidden layer = 5. The Mean Square Error (MSE) value obtained is 0.1026 reaching the goal at epoch 14845. The reverse training regretion reaches 0.99917. The optimization results are close to the specified error, which is 0.0001, while what is obtained is 0.0145. The power generated by the wind speed is 10.7 m/s before being optimized using the back propagation neural network method worth 321 watts, while the optimized power results are 409 watts. The difference in the average target power obtained is 88 watts compared to the power of the Artificial Neural Network (ANN).