Font Size: a A A

Simulation Research On Active Disturbance Rejection Control Of Superheated Steam Temperature Based On Neural Network

Posted on:2018-11-06Degree:MasterType:Thesis
Country:ChinaCandidate:Y J GuoFull Text:PDF
GTID:2322330515957572Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Superheated steam temperature is an important control parameter of coal-fired power unit.The steam temperature being too high or too low will affect the safety and economy of the unit operation.Generally speaking,when the unit is in normal working condition,the steady state error of the actual operating value and the expected value of the superheated steam temperature should be controlled within the range of +5?.At present,the most commonly used method of superheated steam temperature control is to adjust the water-spray flow to each-stage superheater.Because the superheated steam temperature system has the characteristics of nonlinearity,large delay,and large inertia and so on,the cascade PID control is adopted for the water spraying control of superheated steam temperature.When the unit is involved in deep peak load regulation,due to the change in steam temperature characteristics,the PID parameters have to be re-optimized and re-adjusted,inappropriate retuning will seriously affect the quality of steam temperature control,and the tuning process is time-consuming and difficult to realize online.With the development of advanced control technology in recent years,the neural network and active disturbance rejection control is more and more widely applied in thermal power plants.According to the characteristics of superheated steam temperature system of supercritical boiler,this paper combines the neural network inverse characteristic modeling and the linear active disturbance rejection control,and a method of active disturbance rejection control for superheated steam temperature based on neural network inverse model is proposed.On the basis of detailed analysis of supercritical boiler steam temperature characteristics and the influence factors,a neural network inverse model was established for the superheated steam temperature system of a 600 MW supercritical unit,in order to achieve the purpose of pseudo linearization treatment of superheated steam temperature system.On the basis of this,active disturbance rejection controller is designed to control the pseudo-linearized superheated steam temperature system.Based on MATLAB platform,real-time superheated steam temperature control program is developed by combined neural network inverse model with linear active disturbance rejection control algorithm.Real-time control simulation experiments were carried out with full-scope simulator.Experimental results show that,compared with the original cascade PID control system,the active disturbance rejection controller which is combined with the neural network inverse model significantly reduces the overshoot of the system,greatly reduces the convergence time of the adjusting process,and improves the control quality and robustness of superheated steam temperature.
Keywords/Search Tags:supercritical boiler, superheated steam temperature, neural network inverse model, linear active disturbance rejection control, simulation experiments
PDF Full Text Request
Related items