Research On Forecasting Models Of PV Power Plants Based On Output Prediction And Market Model Based On Typical Weather

With the rapid development of photovoltaic power generation,the market pricing reform of photovoltaic power generation is imperative,but the current on-grid price of photovoltaic power generation market still depends on national policy subsidies,resulting in high power generation costs.In order to further activate the photovoltaic power generation market,this paper establishes a day-ahead market scheduling model based on the output forecast of photovoltaic power generation,and discusses the pricing model under the typical weather model.Forecasting the PV output in the future is the basis for the participation of PV power plants in the market.Therefore,this paper first introduces the categories and methods of PV output prediction,and focuses on BP,ELMAN and PSO-ELMAN neural networks.By comparison,the PSO-ELMAN neural network was selected to predict the future PV output in four typical weather conditions.In view of the random nature of PV output,and referring to the bidding mechanism of thermal power grid-connected,this paper establishes the electricity price model of PV power plant participating in the market.Then,using the fruit fly algorithm,combined with the data of the output prediction and the electricity price model of the day-ahead market,the relevant simulation analysis was carried out.That is,the comparison of the purchase cost of small-sized power supply in four typical weather conditions and PV subsidy price reversed under the premise of equal total cost.The analysis of the example shows that compared with the fixed PV subsidy price,the day-ahead market output declaration mode under four typical weathers makes the system purchase cost show different degrees of decline.Reversed PV subsidy price when total cost is equal,it shows different degrees of r:ise,comparing with the fixed subsidy price.This paper establishes the day-ahead market output declaration model,and carries out relevant simulation research based on the measured data of the experimental platform,hoping to provide a reference for the formulation of follow-up policies.