Research On MPPT Strategy Of Distributed Photovoltaic Power Generation

The technology of Maximum Power Point Tracking (MPPT) is the one of the key technologies of grid-connected PV generation, which attracted lots of attention from researchers. But in the actual environment of photovoltaic power generation, the irradiance of PV array is not uniform, and the different output characteristic of PV cells have occurred from time to time, which will lead to mismatch phenomenon caused the multi peak curve of output characteristics of PV arrays. The unpredictable nonlinear output characteristic often leads that the maximum power point tracking algorithm is difficult to play a role, and the efficiency of photovoltaic power generation cannot be improved.Aiming at multi peak output characteristics of PV array, common heuristic search based multi-peak MPPT algorithms, such as Particle Swarm optimizer (PSO) and Perturb, Observe and Check algorithm (POC), can find the global maximum power point (GMPP) of multi-peak PV characteristic curve and improve efficiency under partial shading situation. However, these method have issues including sharp fluctuations, poor dynamic performance and low efficiency under unstable environments and EN50530 process. As a result, they are rarely used in practice.This paper study the multi peak characteristics of the PV output curve and the failure reason of ordinary multi peak algorithm under dynamic conditions in small and medium power photovoltaic power station. On this basis, the disadvantages of slow convergence speed of the PSO algorithm was improved, i.e., using the voltage window to reducing the searching voltage range of the algorithm, in order to enhance the speed of PSO. Then a novel dynamic multi-peak MPPT algorithm based on improved power loop is proposed in this paper. The algorithm provides a practical approach maintaining advantages of other algorithms, while providing multiple modes such as tracking, scanning and optimizing in order to reduce the number of reboot and achieve smooth tracking. This method can not only determine the global maximum power point of multi-peak characteristic, but provides superior dynamic performance. Experiments have proved that the algorithm have better performance in the EN50530 test and high efficiency. The feasibility of the proposed method is verified in laboratory and a rooftop PV power station.