自回归滑动平均模型风速预测最大风能追踪策略研究

Research of a new MPPT strategy based on ARMA model wind prediction

变速恒频风力发电机组在额定风速以下的最大风能追踪(Maximum Power point Tracking,MPPT)效果对于机组的效率有很大影响。现有的最大风能追踪策略不论是功率控制模式还是转速控制模式都是无风速测量下的最大风能追踪策略。究其原因,就在于风速无法精确测量。引入时间序列法中的自回归滑动平均模型(ARMA)对风速进行超前一步预测。根据该预测风速的大小来确定下一时刻最优功率点搜索的起始风机转速,再利用变步长转速扰动的最大风能追踪策略(爬山法)找到最优功率点。仿真表明,时间序列法对风速具有较好的预测效果,有效地缩小了最优功率点的搜索区间,缩短了搜索时间,提高了机组的运行效率。

The maximum power point tracking （MPPT） strategy has great impact on the efficiency of the Variable-Speed Wind Turbine Generator （WTG） when the wind speed is below the rated speed. Now the MPPT has two control methods that are power control method and speed control method. All of them had not used the wind speed information because the wind speed measured is usually inaccurate. In this paper, wind prediction method based on ARMA model was introduced. The previous wind speed information was used to predict the wind speed of next time. The predictive speed obtained by ARMA model was used to calculate the initial WTG rotor speed of the MPPT searching process. Then rotor speed disturbance was add to the initial speed and finally to complete the MPPT prrocess. Simulating results show that the ARMA wind speed prediction method can effectively reduce the MPPT searching area and search time. The effectiveness of the WTG is also improved.