Direct Normal Irradiance Prediction And Dynamic Modeling And Control Of Solar Tower Thermal Power Generation System

Solar tower thermal power generation is an important way to solve the conflict between environment and energy.Since the direct normal irradiance(the energy source of the solar tower thermal power generation system)is obviously affected by cloud and weather conditions,the solar tower thermal power generation system is in the process of continuous dynamic adjustment,which brings difficulties to the steady operation and control of the system.In addition,the solar tower thermal power generation system is a complex thermal system with strong coupling and multi-variable.Therefore,it is of great significance to study the direct normal irradiance prediction and the dynamic modeling and control of the solar tower thermal power generation system.Based on the theory of energy flow and the theory of thermoelectricity,the dynamic model for the air absorber,heat transfer pipeline,heat storage tank and steam generator of the solar tower thermal power generation system are established from the perspective of temperature-driven heat transport,and the correctness of the dynamic model is verified by comparing with the actual data from other papers.On this basis,the influence of different parameters and boundary conditions of each equipment on the heat transfer process is analyzed.The results of simulation show that the dynamic model can accurately describe the characteristics of each device,and lay a good foundation for further research on the control scheme of solar tower thermal power generation system.Combining lifelong learning theory with neural network,a short-term direct normal irradiance prediction model based on dynamic expandable neural network is established.The historical irradiance data and weather information of one of the data observation stations are divided into different training tasks according to the natural months.Training and testing for different tasks based on the short-term direct normal irradiance prediction model,and analyzing the changes of the network structure and the prediction accuracy of the model during the training process to verify the dynamic scalability of the short-term direct normal irradiance prediction model.By migrating the trained prediction model to the other data observation sites,and analyzing the changes in the network structure and prediction accuracy during the model migration process to verify the mobility of the short-term direct normal irradiance prediction model.Compared with other mature algorithms,the superiority of the proposed prediction algorithm is proved.Based on the dynamic model of solar tower thermal power generation system and direct normal irradiance prediction model,taking the direct irradiance prediction value as interference signal,the system’s air circulation(the air flow in the air heat sink and the heat storage tank)as the controlled volume,and studying the control of outlet temperature of the air absorber and the outlet temperature of the steam generator,the classical optimal preview control algorithm is adopted.Simulation results show that the control scheme can overcome the influence of direct irradiance fluctuation on the solar tower thermal power generation system.