複雜遮陰條件下光伏發電系統最大功率點跟蹤優化算法

A New Optimization MPPT Algorithm for Photovoltaic Power Systems Under Complex Shading Condition

在光伏發電系統最大功率點跟蹤（maximum power point tracking, MPPT）中，由於旁路二極體而導致光伏陣列的P-U曲線出現多個功率極值點，傳統的MPPT方法會無法追蹤全域最大功率點。為了解決這個問題已經提出了諸如確定性粒子群算法（deterministic particle swarm optimization, DPSO）等群體智慧優化算法。然而，在均勻光照條件下，DPSO算法將搜索整個電壓範圍，從而產生較大的功率震盪。本文提出一種結合爬山法和局部學習粒子群算法的新的全域最大功率點跟蹤方法，通過計算I-U曲線上短路電流和最大功率點電流處的輻照度來實現遮陰情況的檢測，從而有效地區分均勻光照和局部陰影的變化。除此之外，在粒子群算法中加入局部學習策略，提高了算法的求解精度。模擬結果證明了所提方法可以快速跟蹤外部環境變化，並減小系統在最大功率點附近的震盪。

In the maximum power point tracking (MPPT) of photovoltaic (PV) power systems, multiple power extreme points occur in the P-U curve of the PV array due to the bypass diodes. The conventional MPPT algorithm will fail and can’t track the global maximum power point. To solve this problem, a swarm intelligence optimization algorithm, such as deterministic particle swarm optimization (DPSO), has been proposed. However, under the condition of uniform irradiance, the entire range of voltage will be searched by the DPSO algorithm, which may cause excessive power oscillations. The paper puts forward a new global maximum power point tracking technique combing the hill climbing and local-learning particle swarm optimization algorithm. The method is adapted to accomplish the detection, achieved by calculating the short-circuit current on the I-U curve and intensity of irradiance at the maximum power point current. In this way, the variance between uniform irradiance and partial shading conditions can be effectively distinguished. In addition, the accuracy of the algorithm can be enhanced by adding local learning to the algorithm of particle swarm optimization. The experimental results show that the proposed technique can quickly track the changes of external environment and reduce the systematic oscillation near the maximum power point.