The Application And Research Of Alkali Recovery Evaporation Section Based On Neural Network

With the development of our national pulp and paper industry, the scale isgetting more and more extensive. The amount of water consumption and wasteliquor in pulp and paper making are constantly increasing. Followed with this, theuntreated waste water is released arbitrarily, so the water pollution is becomingvery serious. Because of alkaline pulp gets a large proportion in our nationalchemical pulp production, black liquor is the major polluting sources in alkalinepulp making, solving the black liquor is the focus of water pollution prevention.Practice proves that alkali recovery is an effective way to deal with black liquor.Evaporation section in alkali recovery is a complicated industrial process, socustom control method can't get a satisfactory performance. With the developmentof process control automation level, some advanced intelligent control techniqueshave been successfully used in paper making system. The neural network is animportant branch of intelligent control which has self-learning, self-adapting andnonlinear approximation ability, how to use it on evaporation section of alkalirecovery is the content of this paper.In this paper, the control system of alkali recovery evaporation is researched.The process of evaporation section is introduced. By comparing and analyzing thecontrol method which has been used in black liquor treatment process, therequirements of evaporation section control and the problems of the traditionalcontrol method are understand better. To lay a better theoretical foundation forcontrol system design, PID control method, neural network control algorithms andtheir combination are introduced.The level of black liquor is an important parameter in the alkali recoveryevaporation, as a controlled unit, it has characteristics of nonlinear, large delay andtime-varying. PID neural network is a new type of neural network, in whichtraditional PID combined with neural network. Because of the virtues of thePIDNN, PIDNN control method is used for the black liquor level. The simulationresults in Matlab environment shows that the PIDNN has good adaptability andstrong robustness, compared with conventional PID and PID control based on BPneural network, the performance of black liquor control was greatly improved. As aresult, PIDNN achieve effective control of the black liquor level.The key problem of alkali recovery evaporation control system is the black liquor concentration control. Taking into account the black liquor concentration isa complex controlled unit, detailed analysis is made to the relationship among theparameters in black liquor Evaporation. Decoupling control for black liquorevaporation is presented that significantly improves the effect of the concentrationof control, and has certain application value. Through integrate PID algorithm andneural network algorithm, PID control based on RBF neural network is designed, inwhich the efficiency of black liquor concentration and flow rate for the controlledvariable, into the efficiency of steam pressure and feed efficiency of black liquorflow for the operating variables.Effectiveness of the method is proved bysimulation the black liquor in Matlab software. It increases the effect ofconcentration of control, so has a potential application value.After concluding this paper, the shortcomings of the study are pointed out andsome suggestions of application need to be further addressed are given, in order toprepare the application of neural networks in practice.