Short-term prediction of PV output based on weather classification and SSA-ELM

In this paper, according to the power output characteristics of distributed photovoltaic users, the SSA-ELM (Sparrow Search Algorithm - Extreme Learning Machine) model based on weather type division is proposed for photovoltaic power day ahead prediction. Because the solar panel power generation sequence of photovoltaic users contains high frequency fluctuations, in this paper we use the power sequence convergence effect to make cluster prediction on all photovoltaic panels to reduce the randomness of distributed photovoltaic. The prediction accuracy is further improved by dividing weather types. The historical data of distributed PV users in a region of Gansu province is used for modeling verification, and the results show that the prediction error of the proposed method is lower. In bad weather, the root mean square error is at least 0.02 less than the comparison model, and the average annual accuracy rate is 93.2%, which proves the applicability of the proposed method in different output types.

Automated summary

In this paper, a SSA-ELM model based on weather type division is proposed for photovoltaic power day ahead prediction, taking into account the high frequency fluctuations in the solar panel power generation sequence of photovoltaic users. By using the power sequence convergence effect, cluster prediction is made on all photovoltaic panels to reduce the randomness of distributed photovoltaic, and the accuracy is further improved by dividing weather types. The proposed method is validated using historical data of distributed PV users in a region of Gansu province, showing lower prediction error compared to a comparison model, with a minimum root mean square error of 0.02 in bad weather and an average annual accuracy rate of 93.2%, proving its applicability in different output types.