The Design And Analysis Of Variable-Frequency Constant-Pressure Water-Supply System Based On Neural Network

With the rapid development of social economy and seriously lack of energy resource,it demands the better quality and reliability of water supply system.With the help of advance techniques of automation,control and communication,it is inevitable tendency to design water supply system which has high function and saves energy well.At the same time this system can adapt different water supply fields.First,this thesis summarys current development situation and application of controltheory in water supply system in China and abroad,then analyzes characteristics ofwater supply system with variable-frequency and principle of frequency control.It expatiates on energy conservation theory of frequency control,and then it takes user side as pressure control point because the system is more reliable by it.Second,after doing some research,fuzzy-PID is determined to be taken because ofits superiority.Nowadays,most of fuzzy-PID controllers are designed with initial parameters by neural network,using fuzzy control rules by expertise.But the system's overshoot is big,and adjusting time is long using PID with initial parameters by neural network.And it'ssubjective to build control rules based on expertise.Therefore,the paper improves fuzzy-PID in these two aspects.Then it introduces principle of neural network,and sets initial parameters of PID by neural network.The result indicates that overshoot is reduced and adjusting time is curtailed.Then,with this group of parameters as initial parameters,and fuzzy control rules using expertise,a fuzzy-PID controller is designed preliminary.