Maximum power point tracking algorithm of improved perturbation and observation method

The invention discloses a maximum power point tracking algorithm of an improved perturbation and observation method. The method comprises the following steps: collecting a photovoltaic output voltage U and a photovoltaic output current I; taking the output voltage and the output current as the input of a gradient algorithm, and taking the optimal step length alpha of the perturbation and observation method as the output of the gradient algorithm; substituting the optimal step length alpha into a perturbation and observation method, performing a perturbation and observation method flow, and searching a maximum power point; and repeating the steps until the observation is finished. According to the gradient algorithm-based optimization perturbation and observation method, a plurality of experience problems exist in a fixed step length set according to an empirical value, the problems caused by the experience can be optimized by carrying out self-adaption on the step length in the gradient algorithm-based optimi