Research On GRU-based Predictive Control Of Air Conditioning Surface Cooler

With the reality of non-renewable resources,it has become an inevitable trend for the energy industry to transform in a green direction.Therefore,reducing energy consumption has become one of the important tasks for the present and the future.Adjusting the energy management of cooling and heating,such as predicting and controlling the valve opening,temperature and humidity of air conditioning surface cooler,can reduce unnecessary energy consumption and improve system operation efficiency,and also provide a theoretical and practical basis for predictive control of air conditioning surface cooler,.In recent years,many scholars have devoted themselves to studying how to predict and control refrigeration components efficiently to achieve energy saving and sustainable resource utilization.However,there are still many problems in predictive control in the field of air conditioning surface coolers,such as predictive control accuracy and stability.Therefore,one of the objectives of this research paper is to accurately predict and control air conditioning surface coolers to promote green energy development and facilitate green development.The control of surface cooler with different structures is different again.This paper adopts Transformer prediction model for the existing double-loop surface cooler to realize the control of their valves;meanwhile,this paper modifies the stratified surface cooler and uses different prediction models to realize the prediction and control of stratified surface cooler to achieve the purpose of energy saving and consumption reduction,which is another research content of this paper.This paper combines gated recurrent unit neural network(GRU)with intelligent optimization algorithm to establish a hybrid prediction optimization model based on GRU and intelligent optimization algorithm,and then transforms the surface cooler in the air conditioning system and controls the valve opening to achieve the predicted temperature and humidity output.This research can be divided into the following main contents:1.Making certain structural modifications to the evaporator component in the air conditioning system.Install an electric three-way valve after the main evaporator solenoid valve.The electric three-way valve mixes a certain amount of main evaporator return water with the chilled water from the auxiliary evaporator and continuously adjusts the ratio of chilled water entering the auxiliary evaporator and main evaporator return water based on the actual cooling load to achieve temperature control,recycling of refrigerant water to achieve energy saving.2.Predictive control,an optimization control method,is combined with GRU neural network and intelligent optimization algorithm to avoid getting trapped in local optimal solutions and sample dependency as much as possible.3.The existing water cycle optimization algorithm is modified by incorporating the migration behavior of populations in the existing biogeography algorithm into the water cycle optimization algorithm to address the non-linear problems of the surface cooler system prediction model.4.The two electric three-way valve openings of the dual-loop surface cooler have strong coupling and temporal characteristics.The Transformer model has excellent handling of temporal data and can make predictions.Therefore,this paper uses the Transformer model to predict the double valve openings of the dual-loop surface cooler to obtain a more precise amount of control.