With the increasing penetration of solar electricity in residential, institutional and commercial centres around the globe, the phenomenon of partial shading (PS) in Photovoltaic (PV) power generation is gaining attention. Under Partial shading condition (PSC), cells that are shaded tends to have an equivalent current with cells that are unshaded in series-connection, due to this, the shaded cells operates in reverse bias and consequently becomes load and consumes the generated power. This causes a serious problem known as hotspot. This is characterized by the presence of excessive heat which consequently reduces the total generated power. Recently, researchers use the technique of bypass diodes across the PV cells so that the problem of partial shading can be reduced, but this solution taken alone, has made the nonlinearity and complexity of the system to increase. The shaded cells generate multiple peaks with only one global peak. Conventional Maximum Power Point Tracking (MPPT) algorithms do not differentiates the global peak from local peaks which may end up tracking local peak as global peak, this results in serious power loss. This paper seeks to solve this problem by modelling a PV system under PSC and through the application of Improved Harmony Search algorithm (IHSA) and variable step Perturb & Observe (P&O) to track the global peak instead of local peaks. Simulation was done in MATLAB/Simulink 2018a environment, and the results under standard test condition (STC) and PSC showed that the proposed IHSA had an improvement of 25%, 3.17% and 2.27%, 3.07% and 2.21%, 3.26% and 2.26% when compared with the improved particle swarm optimization (IPSO) under STC and PSC conditions respectively, which had a better advantage of minimizing power oscillation and improving the efficiency of the system, improved MPPT tracking, reduced error and a better tracking efficiency in both conditions.
 Keywords: MPPT, photovoltaic system, partial shading, tracking efficiency, Harmony search algorithm
Read full abstract