| Solar heating system collector efficiency and operating conditions are affected by outdoor weather conditions,which are highly random and fluctuating.The existing fixed control strategy represented by temperature difference control is difficult to achieve the goal of real-time adjustment of solar heating system operation strategy based on outdoor weather conditions,resulting in low solar collector efficiency and high auxiliary heat source output.At the same time,the negative feedback control strategy represented by PID ignores the large time lag characteristics of the solar heating system heat transfer process,which makes it difficult to achieve optimal control of the solar heating system and causes problems such as large fluctuations in the indoor thermal environment and energy waste.Based on this,this study proposes a solar heating system day-ahead control strategy based on model predictive control,which can solve the problem of random fluctuations of the solar heating system and realize real-time regulation of the solar heating system based on meteorological conditions,and at the same time,solve the problem of difficulty in achieving optimal control due to the time lag problem of the solar heat transfer process.First,this study uses a sequence-to-sequence long and short term memory learning model to establish a prediction model for key operating parameters such as outdoor temperature,solar radiation,and building load of the solar heating system,and obtains the values of key operating parameters for the coming day;brings the key operating parameter prediction model into the solar heating system operation regulation model to obtain a solar heating system day-ahead prediction control model;uses NSGA-III to solve the model and compare it with the traditional fixed and negative feedback control strategy.The main conclusions obtained are as follows:(1)Using the sequence-to-sequence long short-term memory learning(seq2seq-LSTM)model,a solar heating system key operating parameter prediction model was developed to predict and obtain the future day key operating parameter values for Lasa as an example.The results show that the average R~2,RMSE,and MAE of the seq2seq-LSTM model are 0.97473,0.57°C,and 0.43°C,respectively,when predicting the outdoor dry bulb temperature,and the average R~2,RMSE,and MAE are 0.84404,104.07 W/m2,and 90.86 W/m2 when predicting the solar radiation.compared with the LSTM direct prediction model and LSTM recursive prediction model,the seq2seq-LSTM model has larger R~2,smaller RMSE,MAE,and The prediction results are more accurate.(2)Combining the prediction model of key operating parameters of solar heating system and the operation regulation model,a control strategy and method of solar heating system before the day based on model prediction was established.The results show that compared with the traditional solar heating system,the average collector efficiency of the solar heating system based on the model prediction is increased by8.9%,the collector heat is increased by 9%,the heat loss of the water tank is reduced by4%,and the indoor temperature fluctuation is smaller with an average temperature difference of 0.54℃,the temperature fluctuation is reduced by 12.05%,and the indoor thermal environment is more comfortable.(3)Finally,this study compared the pump consumption,auxiliary heat source energy consumption and total energy consumption under the model prediction-based day-ahead control strategy with the traditional control strategy.The results show that the solar heating system based on the model prediction can more effectively enhance the collector system heat collection and make full use of the heat storage tank heat,resulting in a higher water flow rate in the collector cycle and end circulation system,which increases the collector pump energy consumption by 12.5%and the end heating pump energy consumption by 112.3%;however,due to the more efficient use of the heat storage system,the auxiliary heat source turn-on frequency and system energy consumption are significantly reduced.The energy consumption of auxiliary heat source is reduced by 94.1%;the total energy consumption of the system is reduced by 34.3%. |
