Research On District Heating Load Forecasting And Control Based On Adaptive Genetic Artificial Neural Network

Heating load forecasting is an important condition to carry on superior regulation for district heating system. It is great significant to forecast accurately heating load for the movement management of the heating system , raising heating quality ,environment and energy saving. This paper makes a relatively thorough analysis and studies the characteristic and present condition of district hearing load forecasting. Based on the study on theory and algorithm of neural network and genetic algorithm, an adaptive genetic neural network model for heating load forecasting. The simulation results show it get high accuracy and credibility. When the neural network model is designed, in this paper the self-adapting genetic algorithm(GA) is used to optimize neural structure; In order to overcome the weaknesses of the classic BP algorithm such as slow convergence rate and the high probability to be trapping in local optimal, GA-BP combinative algorithm is designed to train the weights of neural network. In this paper, based on the theory of artificial neural network and genetic optimization algorithm, a adaptive dynamic control method is established. The simulation results show it gets rid of the weaknesses of the traditional control method ,and improve evidently the quality of district heating and save energy.