Study On Optimal Control Of Ground-coupled Heat Pump System Based On Artificial Algorithm

As a type of energy saving technology that uses renewable energy,Ground-Coupled Heat Pump System (GCHPS) has got more and more attentions.especiallyunder the situation that energy supply can’t meet it’s increasing requirement.GCHPS’s advantages,such as good performance in energy saving, low cost inrunning system,using renewable energy,get more attention than any timebefore.Because of it’s nonlinearity,strong coupling,and uncertainty,study on optimalcontrol of GCHPS gradual becomes the hot spot and key point in GCHPS’sresearching field. Based on the data of GCHPS in Olympic Forest Park, this thesisfocus on the energy saving optimal control of GCHPS.The control demand of GCHPS is to decrease the total energy consumption underthe condition of keeping the system safely,steadily and meet the client’srequirement.In order to reach the target of optimal control for GCHPS,firstly accurateanalysis,finding out available control variables and building the model betweenvariables and total consumption are required.Secondly,based on the model that hasbeen built,using artificial algorithms to optimal the combination of variables to findout the best combination that minimizes the total energy consumption.First,this thesis confirms the control variables and builds model between controlvariables and total energy consumption by using improved BP neural network throughreading related papers from domestic and overseas and analyizing the data formOlympic Forest Park.When the temperature is set by the client,the model calculatesthe total energy consumption based on client’s set value and the control variables.Second,improving the PSO algorithms to make better performance in runningtime,convergent effect and velocity.Using the artificial algorithms to find out the bestcombination of variables to minimize the total energy consumption based on themodel built by improved BP neural network under the condition of runningsafely,steadily and meet the client’s requirement.Third,improving the basic GA algorithms to make better performance and usedthe improved algorithm in calculating the best combination of variable based on themodel.Finding out the better way for the optimal control of GCHPS between GA andPSO algorithms through comparing.To figure out the issues of optimal control of GCHPS, this thesis analyzes the operation of GCHPS,Using improved BP neural network to build the model of controlvariables and total energy consumption and finding out the best combination byartificial algorithms.Through the work above,the predicted value of control variablesthat cause the minimum of total energy consumption can be got and it can avoid thewaste of time by using direct detecting means.