Evaluation of prediction error effects in wind energy-based electric vehicle charging

This paper first presents a battery operation scheduler for the sake of practical integration of wind energy generation and electric vehicle charging, and then measures its performance mainly focusing on the effect of wind speed prediction errors. The operation scheduler decides whether to charge or discharge a station battery on each time slot based on current wind speed reading and next speed prediction. Its control logic straightforwardly activates generation facilities according to the minimum wind speed for energy generation and the current battery capacity. Next-hour wind speed is predicted by an artificial neural network trained by a series of hour-by-hour speed records. The performance measurement results obtained from simulation show that the depletion ratio is affected by 6.8 % and the energy loss by 3.5 %. This result is valid for the whole given parameter range except only a few cases. Moreover, judging from the observation that the largest renewable energy loss is just 0.9 %, the battery management scheme overcomes the misprediction effect by adaptively compensating for the generation loss on each time slot.