Distributed Energy System Optimization Based On Photovoltaic Power Prediction

As a widely distributed intermittent renewable energy source,solar power has been widely used in recent years in the construction of centralized power stations and distributed energy systems.Aiming at the characteristics of unstable photovoltaic output power and the requirements of the national standard for ultra-short-term power prediction,an ultra-short-term prediction model based on the random forest algorithm and long-short-term memory network algorithm was proposed.Firstly,the main influencing factors of photovoltaic power and the characteristics of series-parallel photovoltaic arrays were analyzed theoretically.Then,actual operating data of large-scale photovoltaic power stations in Inner Mongolia were collected.After completing the data preprocessing work,based on the numerical weather prediction(NWP)data,more useful input features were constructed through feature engineering.And the importance of different features was ranked using the random forest algorithm.Besides,the effect of the number of features on prediction accuracy was studied.Secondly,by improving the long short-term memory network algorithm,a multi-step prediction model of "sequence to sequence" was designed to use the historical data of the past12 hours to predict the output power of the next 4 hours.Finally,the accuracy of the prediction model was verified by actual onsite data.The evaluation indicators are the root mean square error(RMSE)between the predicted power and the real output power,and the monthly qualified rate(QR).The results show that the RMSE of the composite prediction model is8.81%,and the QR is 97.67%.The prediction effect is stronger than the existing prediction model of the power station and has a strong application value.Besides,the effect of the number of input features on the prediction accuracy for the same prediction model was studied.In this example,when the number of input features is 6,the optimal prediction effect can be achieved.In addition,the application of photovoltaic power prediction in the distributed energy system scenario was investigated.By coupling photovoltaic power generation systems,battery energy storage systems with existing combined cooling,heating and power system,the distributed energy system was established.The energy supply relationship and mathematical model of the distributed energy system(DES),combined cooling,heating and power system(CCHP)and separation production system(SP)were established.Based on the evaluation indicators of energy,environment,economy,the particle swarm algorithm was used to determine the optimal configuration combination of different systems under different operation strategy.The results show that compared to the existing energy supply system,the distributed energy system can effectively utilize photovoltaic power and significantly reduce the consumption of fossil energy and carbon dioxide emissions,but has only a slight decline in economic performance.Therefore,it has broad application prospects.