3 Problems Process Control Research

3.1 drying technology and control technology is not fully integrated

Drying process is a typical multivariable, large inertia, highly nonlinear complex systems, in line with the establishment of an ideal mathematical model of the actual drying process is very difficult; and modeling to spend a lot of energy, and sometimes even impossible. Usually facilitate research to simplify the modeling conditions, simplified equation does not accurately reflect the drying process, simplify the often result in errors. Some models, such as heat and mass transfer model, controlled drying process optimization, fuzzy control and intelligent control type, etc., are not perfect place. While drying theory limited diffusion theory circles, did not find the material itself characteristic function, which is to establish an accurate model brings difficulty. Even some of the drying process can establish a precise mathematical model, its structure is often very complex and difficult to design and implement effective control. The current study basically stuck on a one-dimensional mathematical model based control, control is often only a particular parameter, the control effect is not ideal, but can not complete the intelligent control of multiple targets. Not a good mathematical model in the implementation of control have to seek other indirect methods, which affected the accuracy and effectiveness of control to some extent, drying technology and control technology in combination was good enough to make the dryer control to play dryer highest performance, improve product quality and the role of not fully reflected.

3.2 less drying process control method and control Effect

3.2.1 Process Control control variable less

Drying process in a conventional single-variable control system technology, the target control is mainly limited to the smooth operation of a particular variable or several variables, to ensure smooth production and fewer accidents. With the growing trend of large-scale grain drying industry, integrated, continuous and complicated process control of a higher quality requirements, a good control system not only to protect the security and stability of the entire production system to meet certain constraints, and should bring certain economic and social benefits. In grain drying, when once a hot air drying section temperature and humidity changes, not only have a direct impact on the drying section temperature and moisture content of grain, but also indirectly affect to the next exit and the drying tower temperature and moisture content of food. If the grain discharge motor speed is slowed down or speed up, not only drying grain moisture vary tower exports, within every period of drying temperature and moisture food segment will change accordingly. In this series of complex process of change, inevitably accompanied by delay, coupling, and a series of non-linear time-varying process. If only the error and error rate of change as the controlled variable input control system, when the increase in the system's internal or external interference, it is difficult to ensure its control. Classic fuzzy control system is often simplified as a research question univariate single input single output fuzzy controller has great limitations in the application, enter the controller only charged with bias and deviation change variables, essentially the equivalent of a variable single input parameter PD regulator. Thus, the complexity of the drying process determines the control amount and the control amount is more than one, there is a complex relationship between influence each other, each is the best value of the control amount of mutual restraint factors also exist, it is difficult finding the optimal control scheme .

3.2.2 Application of advanced control methods focus on small and single

Although decades began to explore how intelligent control applied to the drying process, but the design method on grain drying advanced control system, little studied, but also focus on research in some way more. National Food Authority "fifth" period to spend a lot of money for solving the grain drying process Moisture online testing and automatic control, combined with some grain research and development work carried out some projects, but most units are designed fuzzy control method. Browse domestic dissertations can see, more is the use of neural network mathematical model of the drying tower, fuzzy thinking on the performance of the dryer and comprehensive evaluation of the dryer design optimization; not a single application of the model to predict control reported. Although the advanced control method has many advantages, but there are also single method of one kind or another inadequate. Fuzzy control is based on the experience of skilled operators only on the basis of need through self-learning system, constantly revised parameters in order to gradually approach the target value. The water on dried food factors, not easy to find a skilled operator empirical parameters, without the use of a more accurate reflection of the Mathematical Model of dryer control amount for automatic control design, it is difficult to ensure food quality after drying. While the adaptive control can solve the problem to a certain degree of uncertainty, but the algorithm is complex, computationally intensive, and the process is not dynamic modeling and disturbance of poor adaptability, robustness issues to be further addressed, the application is limited. Development of expert systems-friendly graphical interface is based on the development direction of the drying process control, but as a result of a search problem solving long time expert system for online control ability is relatively poor. In neural network modeling forms, BP algorithm based network has a long training time, and often disadvantage not converge; radial basis function approximation although the drying process can greatly improve the speed of convergence, and the network can converge to the global minimum , but it is difficult to determine the center coordinates. Most of the existing nonlinear model predictive control method can only be used for the slower process control, real-time requirements for higher unfavorable drying process control. Thus, a single application of a control strategy is not necessarily best to take advantage of process control.

Detection control more than 3.3, the moisture sensor accuracy and stability is not high

Grain drying parameters of measurement and control instruments directly related to the quality and efficiency of drying. Domestic grain dryer automatic control applications much, some drying machine is equipped with digital display of air temperature and over-temperature alarm and grain discharge rate of the display device, but can not automatically controlled. Domestic grain moisture tester for grain moisture measurement and display of simple, there is no grain drying equipment with real-time, on-line control systems can not achieve automatic control of grain drying process. Grain moisture testing is difficult to achieve fast online measurement, the current domestic use due to the drying equipment Dynamic Process Water and detected none of the stereotypes, can not achieve automatic control of grain drying process. Line moisture sensor test test accuracy and stability issues together have not been well addressed, not really mature stage really reliable detection, affecting the accuracy of the process approach.

4 Development Direction

4.1 improve the drying process model

Continue in-depth study of the drying process the internal law of heat and mass transfer material; to establish a mathematical model to accurately reflect the state of the drying process will help improve the automatic control of the drying process. At the same time, you can create intelligent model of the drying process, with intelligence model to replace the mathematical model, intelligent control system can approximate a real system and its effective control. As with the use of artificial neural network technology to create a mathematical model, artificial neural network technology is capable of more independent variables mapped to multiple dependent variables, it is particularly suited to complex grain drying process.

More than 4.2 kinds of Controlling method of penetration

Single use some advanced control technology is difficult to give full play to advantages, is an inevitable trend of mutual penetration of various control strategies, complementarity, mutual economic advantage, combined into a composite control strategy. Hybrid control strategy combining multiple control strategies to overcome the lack of a separate policy, but also has excellent characteristics, can better meet the requirements of different applications, is the future direction of development. Studies have shown that, instead of using a neural network fuzzy inference method will enable online control expert system greatly improved; neural network expert system artificial neural networks and expert systems together for problem solving is a useful attempt; neural network combined with the traditional control theory the control system with a considerable degree of intelligence. Thus, the hybrid control strategy will encourage research and mathematical simulation laboratory research stage neural network control system for actually stay in control. Fuzzy PID control was combined fuzzy variable structure control, adaptive fuzzy control, fuzzy predictive control, fuzzy neural network control, fuzzy control, expert control complex emerging, I believe there will be greater development and wide application.

4.3 depth study of the control strategy

A drying process of the system has been unable to adopt a single quantitative mathematical model based on the traditional control theory and control technology, the need for further development of advanced process control systems, advanced process control study law, and existing control theory and method to process migration and transformation control areas, these areas are also more and more attention control field. To further strengthen the control theory, such as the three major mechanism of predictive control: prediction model, the feedback correction method for solving the optimization strategy efforts, all to go to research and breakthroughs; drying process control in the urgent need to develop a good real-time model predictive control method, under the premise of quality assurance dried to make online computing time reduction; focus on cross-disciplinary research, learn from other effective control method to solve the existing problems of process control, continuous improvement, development and innovation prior drying process control algorithm ; further improve the reliability of automatic drying quality control system, the establishment of an adaptive control algorithm capabilities.