基于模糊PI双向线性预测的光伏MPPT控制

THE PV MAXIMUM POWER POINT TRACKING ALGORITHM BASED ON FUZZY PI BIDIRECTIONAL LINEAR PREDICTION METHODOLOGY

使用扰动观察法的光伏系统在初始化启动过程中或光强发生较大变化时,由于受扰动步长的限制,无法快速接近最大功率点,造成大量能量的浪费。针对该情况提出一种快速最大功率点预测算法。首先将光伏P-U曲线在0点附近Taylor展开,取线性部分作为功率函数P(u)的近似方程。由于P-U曲线存在两个0点,近似方程将从曲线的两个方向交于一点,然后近似认为P-U曲线在该点取得最大功率。最后P-U结合曲线恒压恒流区的特点,计算预测误差并给出剩余搜索范围。双向线性预测的思想用于近似估计最大功率点,此后通过模糊PI控制提高跟踪精度。此方法有效克服了扰动观察法速度与精度无法兼顾的矛盾,相对经验值预测法,该方法不依赖于统计规律,其精度受环境因素的影响较小。仿真及实验结果表明该方法具有快速、高效的优点。

When the PV system is initializing or the solar irradiation changes greatly, the traditional perturbation and observation algorithm （P＆O） cannot track the maximum power point （MPP） quickly and it may waste a lot of potential energy. In this paper, a novel prediction algorithm which can track MPP rapidly for MPP tracking （MPPT） of PV power generation is presented. First, Taylor method is used to linearize the function of P-U curve so that a bidirectional linear prediction （BLP） can be proposed for PV generation system. As there are two zero-points on P-U curve, two linear functions can be established which are used to describe the characters of P-U curve approximately. The BLP is used to estimate the MPP. It can greatly improve the MPPT response speed. Then the fuzzy PI controller is put forward to track the real MPP. Plenty of theoretical analyses indicate that the new algorithm is much faster than those traditional perturbation and observation algorithms. Comprehensive simulation and experiment results demonstrate that the new method has excellent performance, such as fast dynamic response, high efficiency, etc.