Research On Intelligent Control System For Papermaking Process

From ancient manual papermaking craft to modern automatic paper production technology, papermaking industry has obtained many achievements and meanwhile has met human's increasing requirements of paper product. However with the development of society, people's requirements are not only simply the quantity of paper product but also the quality and function of paper product as well as efficiency of paper production. Here the papermaking automatic control is one of the most important parts in papermaking technology, so from the angle of control how to solve and improve the quantity, quality, function and efficiency of paper product is the main direction of papermaking industry nowadays.According to the above problems, the paper takes the most important performance indexes such as basis weight, moisture content, caliper, and ash as the plants, and from the view of control, the methods are analyzed in detail to improve the control performances in the papermaking process. Although many advanced control algorithms were applied to solve the fluctuation of the paper basis weight and moisture content and other indexes as well before, the real useful and practical method is still sampling PI control which is prevailing all along. So based on the national key project"papermaking process optimal control system"awarded to Shaanxi University of Science and Technology, the paper establishes the paper machine's mathematic models, and recurring to intelligent predictive control theory, algorithm, strategy and practice, the paper machine basis weight, moisture content, caliper and ash control systems are well studied, then the intelligent optimal control structure of papermaking process is realized. Thereinto, moisture content drying curve with target function of funminimum energy consumption is proposed, meanwhile the cross-directional moisture content control with dryer surface temperature and bag zone air blowing method are analyzed. Caliper and ash control are also analyzed to abtain automatic tracking characteristics. At last with practices the CIPS is applied, and efforts to get direct guidance of engineering and practice efficiency are achieved.Firstly, according to the analysis of middle basis weight paper machine sections from headbox to reel part, the input and output characteristics of papermaking plants are obtained to establish paper machine's mathematic model. Because of paper machine plant's uncertainty, inadequacy, nonlinearity, time-delay and strong coupling, some unnecessary disturbances are ignored and approximate mathematic model can be achieved. Hence according to the technical process and signal measurement loop, multi-loop control strategies are set correspondingly.Secondly, according to the papermaking process mathematic models and control strategies, advanced control algorithm based on predictive control is applied to papermaking process control system so as to settle the complicated multi-variable and large time-delay plants. The dynamic matrix control algorithm gets the predictive model from plant's step response, its algorithm is simplicity with little calculation as well a strong anti-disturbance. For multi-variable coupling system, DMC has the latent decoupling ability. Meanwhile, combined PID with DMC, novel PIDDMC algorithm is proposed to get better dynamic performance. And global predictive control applies CARIMA model to predict within finite time domain for large time-delay plant, by successively solving Diophantine equation and adding the action of control increment in target function, optimized control law is got to obtain the purpose of future output prediction. While the model predictive function control based on CARMA selects the base function to decrease the dimension of optimal variable and develops the successively algorithm to get model output prediction, by the way the solving process of Diophantine equation is ignored, online calculating work is decreased and real time performance is improved.Thirdly, because of the characteristics of plant complication, uncertainty, nonlinearity and time-delay in papermaking process, traditional predictive control methods also have faults. So combined with the neural network which has the merits of plant's model identification and self-tuning, the integrated neural network predictive control is proposed to settle the complicated control parameters of the papermaking process. Neural network has the ability of approaching real process, and here NARMA model with BP network is used, meanwhile another neural network is used to predict real output error. By this way the control law can be directly obtained according to the plant model identification, and complicated nonlinear solution and calculation load are avoided as well. Meanwhile by the use of neural network, new neural network predictive controller can optimize the plant's control performance by the cost of moderate memory space, and the influence of disturbance and uncertainty can be minimized.Finally, with the research of above intelligent control algorithms, the technical realization and excutation of paper quality control system QCS is proposed. Here the paper basis weight, moisture content, caliper and ash control methods are analyzed. The emphasis is focus on the cross-directional control besides the basis weight and moisture content. To the moisture content, dryer surface temperature cross-direction control and bag zone air blowing control methods are analyzed, hence the DMC cross-directional moisture content control system is established. To the paper thickness control, electromagnetic heating adjustor for the controlled middle-high roller is set. To the ash control, from the view of sensor signal detection, feasibility of high accuracy control is developed. Meanwhile with the practice, multi-controller control system of papermaking process is proposed and switching method with no disturbance is realized. And filter technique is applied to solve signal disturbance for good accuracy. With above analysis, papermaking process distributed control system based on computer is proposed with hardware and software project, which is called CIPS. And with Windows xp, WinCC and control algorithm package, papermaking process control system is applied to a paper machine production line and the system can achieve optimal running states with high quantity, fine quality and low loss.