Font Size: a A A

Research On Boiler Drum Water Level Control Based On Neural Network

Posted on:2019-06-22Degree:MasterType:Thesis
Country:ChinaCandidate:L Q FengFull Text:PDF
GTID:2382330563490611Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Boiler is a very important power output device in industrial production,thermal power generation and other industries.The stable and safe operation of boiler is premise for stable production of all industries.Therefore,in order to ensure stability and efficiency of various industries,safety and effective control of boiler operation is very necessary.Effective control of drum water level is the primary reference for ensuring normal operation of industrial boilers.Commonly used PID control method in industry is highly dependent on exact model of controlled quantity,and has a poor adaptability to nonlinear system,and its parameters will not change once it is set,so it is not flexible enough to deal with many kinds of disturbances in the bad conditions.So the effective control of water level of drum appears to be very important.In this paper,mathematical model of water level of drum is built in boiler's principle.Trend of water level under disturbance of water flow,steam flow alone,under common disturbance is analyzed.Secondly,based on three impulse(water level,feed water flow and steam flow),genetic algorithm and extended calman filter theory are introduced,and a neural network controller is built based on neural network.It is used to replace commonly used PID controller in boiler water level control,to realize effective control of nonlinear and unstable controlled object of boiler drum water level,and to achieve satisfactory energy saving effect and better control quality.Finally,through MATLAB Simulink simulation module to build framework of three impulses of drum water level,and apply neural network controller designed by project and conventional PID controller to this framework.Through simulation,control effect of two are compared.Simulation results show that when adding steam disturbance and feed water disturbance,neural network controller can overcome disturbance and recover to steady state quickly,and it is better than conventional control to overcome disturbance.Feasibility of the controller is also illustrated by simulation of the controller's advantages,such as reasonable structure,strong anti-interference ability and so on.
Keywords/Search Tags:neural network, genetic algorithm, extended Calman filter, three impulses
PDF Full Text Request
Related items