Research Of ADP Strategy For Heat Supply Optimal Control Of District Heating System

The district heating system has been widely applied in our country especially in thenorth regions. Recently, it becomes a research hotspot that how to achieve the optimalcontrol goal of energy-saving and emission reduction with the precondition of meetingthe heat requirements of people. It is very difficult to use the classical control theory toachieve the overall optimization directly, because the district heating system has thecharacteristics of nonlinear, large delay, strong coupling and time-varying. An adaptivedynamic programming optimizing control method is proposed in this project, and isapplied in the district heating system.A BP neural network has been built as the short-term load forecasting model whichis used to forecast the heating load requirement of users in the next twenty-four hoursand is the basis of the optimal control as followed. The model is established by largenumbers of the offline data from the working condition and the weather forecast gainedthrough the API interface. At the same time, a graphical dynamic simulation model ofthe district heating system is built by the fluid dynamics software named Flowmaster,which is the foundation of testing the algorithm. How to solve the optimal controlproblem by quality-quantity operating approach of the total heat in the heat sources forthe district heating system is the main research content in this thesis. The mathematicaldescription of producing the total quantity of heat optimization is built and solved by theoptimal control algorithm of heuristic dynamic programming. This thesis deeply studiesan optimization control algorithm based on heuristic dynamic programming, whichincludes the model network, the action network and the critic network. The controlprogram of HDP algorithm is written on the MATLAB platform, which can obtain theoptimal set value sequence of the supply flow and the supply temperature. Finally, theco-simulation of the Flowmaster and the MATLAB is built to verify whether thealgorithm has a well control performance in the dynamic simulation model of thedistrict heating system. This thesis has verified that it is feasible to apply the HDP based optimal controlalgorithm in the district heating system. The simulation experiment results show that theheat supply can meet the demand of users following the temperature variation. It will bea beneficial trial that the adaptive dynamic programming theory is used in the feedbackcontrol engineering. The result of this research will play a guiding role for the optimalregulation of the district heating system and have a broad application prospect.