The Control Strategy Research Of Air Handling Unit In Central Air Conditioner

With the development of society and advances in technology, people are required toimprove the material life. And then central air conditioning has been accessed to millionsof households. Demand growth is an inevitable trend. Due to technical limitations, familieswith central air-conditioning system is neither economic nor reality, but now thedevelopment of home central air conditioning are quite mature technology.Firstly, on the basis of reading many foreign and domestic references, it summarizesthe development and status of the AHU (Air Handling Unit) in central air conditioner. Itintroduces the predictive control method of the central air-conditioning system and theapplication of neural network model identification algorithm.Secondly, the structure characteristics and the working principle of the central airconditioner and AHU are introduced. Cold or heat source is produced by the host machine.And then the source is sent to the respective ends by connecting pipes. Above all itachieved the purpose of heating and cooling. Only open the central air conditioning whichroom has people we can save electricity consumption, that we can save energy much morethan before.Then, for the characteristics that nonlinearity, large time delay, multi disturbance havea serious impact on the AHU. The controller is designed using predictive control method,which is more commonly used in the field of industrial process. The neural networkidentification method is adopted. The neural network for the establishment of a non-linearsystem prediction model is effective. The neural network model is used as a forecast modelin predictive control.Lastly, the experiment can do under laboratory experimental environment, withAdvantech IPC control simulation platform. The dates of room temperature, humidity, airsupply and the damper valve opening degree electrical signal are collected. The controlledobject system is identified by the neural network model, which has strong disturbanceability. Then according to the optimal control, partial derivatives are obtained. Using them can get the more appropriate and more energy-saving control output variables. Finally,according to the recursive iterative methods, the system predicts the future value of thecontrol output and the predicted output. It is necessary to simulate and analyze usingMATLAB/Simulink. Compared with the classical control strategies, the predictive controlmethod based on neural network model has higher precision, more robust, better controleffect.