Research Of The Central Air-conditioning Temperature Control Based On Neural Network

with the development of intelligent buildings, the control algorithm and precision of the control system is gradually improved. After the 80's years the theory of artificial neural network(ANN)sprang up that is a new subject based on the structure and function of brain, which can be applied to the artificial intelligence, mode identification, simulation of biology and advanced control etc. It is attatched importance that artificial neural network is applied in the heating, ventilation and air conditioning field. The PID controller has many advantage such as simple structure, parameter of certain physical meaning, good dynamic and static characteristic and so on. The conventional PID is not also ignored. The combination of the broad-using PID controller and ANN is a way of intelligent control.Genetic algorithm (GA) is a kind of optimizational search method based on natural selection and the principles of genetics and genetic. It screens the individuals in accordance with the value function and a series of genetic manipulation, So the individuals with high adaptive value have been retained form the new group, and the new group fitness of the individuals can rise up to meet the conditions of a certain limit. At this point, the individuals with highest value of the group are the optimal solution to optimize the parameters. Because of the unique working principle, genetic algorithms can optimize and search on the overall and complexity space. Genetic algorithm used in neural network can optimize the artificial neural network (ANN) structure and learn the weights of ANN. In this paper,genetic algorithm was used to determine the right of the initial threshold of BP neural network to determine a better search space and to replace the initial threshold of randomly selection,then BP neural network can training, learning and search out the optimal solution or near optimal solutionin the solution space .Aimed at the temperature control of air conditioning system, this paper simulates and studies through the experiment on the algorithm of neural network PID control, and bring forward a mean that the best control combines neural network PID, which is adaptive to the control of temperature in the air conditioning system. Firstly, the digital simulation and analysis is to the neural network PID control; secondly through the MATLAB, the temperature control of air conditioning system is simulated and the combination of the best control and neural network PID control is put forward; at last the practicality and validity of neural network PID controller is proved by the experiment that the water temperature is controlled.