Advanced control of grain drying process (2)

3.1 the drying technology and control technology are not fully combined

The drying process is a typical multivariable, large inertia and highly nonlinear complex system. It is difficult to establish an ideal mathematical model of drying process, and it takes a lot of energy to establish the model, sometimes even impossible. In order to facilitate the research, the modeling conditions should be simplified. The simplified equation can not reflect the drying process correctly, and the simplification often brings errors. Some models, such as heat and mass transfer model, optimal control model of drying process, fuzzy control and intelligent control model, have some shortcomings. At the same time, the drying theory research is limited in the circle of diffusion theory, and the characteristic function of material itself is not found, which also brings difficulties to the establishment of ^ - model. Even if the mathematical model of drying process can be established, its structure is often complex, and it is difficult to design and realize effective control. The current research basically stays on the control based on one-dimensional mathematical model, often only control a specific parameter, the control effect is not ideal, and can not complete the multi-objective intelligent control. Without a good mathematical model, we have to seek other indirect methods in the implementation of control, which affects the accuracy and effect of control to a certain extent. The combination of drying technology research and control technology research is not good enough, which makes dryer control not fully reflect the role of high efficiency of dryer and improvement of product quality.

3.2 the research on drying process control method and control effect is less

3.2.1 less control variables in process control

The drying process control system is based on the conventional single variable technology, and the control goal is mainly limited to the stable operation of one or several variables, so as to ensure stable production and less accidents. With the grain drying industry becoming large-scale, integrated, continuous and complex, higher requirements are put forward for the quality of process control. A good control system should not only protect the stability of the system and the safety of the whole production, meet certain constraints, but also bring certain economic and social benefits. In grain drying, once the temperature and humidity of the hot air in a certain drying section changes, it will not only have a direct impact on the temperature and moisture content of the grain in the drying section, but also indirectly affect the temperature and moisture content of the next section and even the outlet of the drying tower. If the rotation speed of grain discharging motor slows down or speeds up, not only the moisture content of grain at the outlet of drying tower will change, but also the temperature and moisture of grain in each drying section will change accordingly. In this series of complex change process, there will be time-delay, coupling, time-varying and a series of nonlinear processes. If only the deviation and variation rate of the controlled variables are taken as the input of the control system, when the internal or external interference of the system increases, it is difficult to ensure the control effect. In the classical fuzzy control system, the research problem is often simplified as single input single output single variable fuzzy controller, which has great limitations in application. The input of the controller is only the deviation and variation of the controlled variable, which is essentially equivalent to a single input PD regulator with variable parameters. Therefore, the complexity of drying process determines that there are more than one controlled quantity and controlled quantity, and there are complex influence relations between them. The optimal value of each controlled quantity will also have mutual restriction factors, so it is difficult to find an optimal control scheme.

3.2.2 the application of advanced control is few and the method is centralized and single

Although it has been explored how to apply intelligent control in drying process for decades, there is little research on the design method of grain drying advanced control system, and there are more researches focused on a certain method. During the Tenth Five Year Plan period, the State Grain Administration spent a lot of money to solve the on-line test and automatic control of moisture content in the drying process of grain, and carried out some research and development work combined with some grain depots, but most design units adopted fuzzy control method. It can also be seen from the domestic dissertations that most of them use neural network to establish mathematical model of drying tower, use fuzzy thought to comprehensively evaluate the performance of dryer and optimize the design of dryer; there is no report on the application of model predictive control. Although advanced control method has many advantages, single method also has some disadvantages. Fuzzy control is based on the experience of skilled operators. It needs self-learning of the system and constant correction of parameters to gradually approach the target value. However, there are many factors affecting the moisture content of grain during drying, so it is not easy to find the experienced parameters of skilled operators, and it is difficult to ensure the quality of grain after drying without using the mathematical model method which accurately reflects the control quantity of dryer. Although the adaptive control can solve the uncertain problems to some extent, the algorithm is complex, the calculation is large, and the adaptability to unmodeled dynamics and disturbances is poor. The problem of system robustness needs to be further solved and its application is limited. Developing a friendly graphical interface based ^ system is one of the development directions of drying process control. However, due to the long search time for problem solving, ^ system has poor ability in online control. In the form of neural network modeling, the neural network based on BP algorithm has the disadvantages of long training time and non convergence. Although the radial basis function approximation drying process can greatly improve the convergence speed and make the network converge to the global small value, it is difficult to determine the center coordinate. Most of the existing nonlinear model predictive control methods can only be used for slow process control, which is disadvantageous to the drying process control with high real-time requirements. Therefore, the single application of a certain control strategy can not give full play to the advantages of process control.

3.3 the detection is more than the control, and the accuracy and stability of the moisture sensor are not high

The detection and control instruments of grain drying parameters are directly related to the quality and economic benefits of drying. The application of automatic control of domestic grain dryers is few. Some dryers are equipped with digital display of air temperature, over temperature alarm and grain discharge speed display devices, but they can not be automatically controlled. Because of the simple measurement and display of grain moisture by domestic grain moisture detector, there is no real-time and on-line control system matched with grain drying equipment, which can not realize the automatic control of grain drying process. It is difficult to realize on-line and rapid measurement of grain moisture content. At present, the drying equipment used in China can not realize the automatic control of grain drying process because there is no fixed dynamic process moisture detection method. The accuracy and stability of the on-line moisture measurement sensor are not well solved, and there is no mature stage to truly reliable detection stage, which affects the accuracy of the process method.
4 development direction

4.1 improvement of drying process model

Further study on the heat and mass transfer law in the drying process and establish a mathematical model which can reflect the state of drying process is helpful to improve the automatic control of drying process. At the same time, the intelligent model of drying process can be established, and the intelligent model can replace the mathematical model. The intelligent control system can approach the real system and control it effectively. If the mathematical model is established by using artificial neural network technology, the artificial neural network technology can map multiple independent variables to multiple dependent variables, so it is particularly suitable for complex grain drying process.

4.2 combination of multiple control methods

It is difficult to give full play to the advantages by using a single advanced control technology. An inevitable trend is that various control strategies penetrate each other, learn from each other, and benefit from each other to form a compound control strategy. The combination of multiple control strategies overcomes the shortcomings of single strategy, and has better characteristics, which can better meet the requirements of different applications, and is the future development direction. The research shows that the on-line control ability of the system will be greatly improved by using neural network instead of fuzzy mathematics reasoning method; the neural network system combining artificial neural network and ^ system is a beneficial attempt to solve the problem; the combination of neural network and traditional control theory makes the control system have a considerable degree of intelligence. Therefore, the compound control strategy will promote the research of neural network control which stays in the stage of mathematical simulation and laboratory research to be used in practical system control. Fuzzy PID explicit and close control, fuzzy variable structure control, adaptive fuzzy control, fuzzy predictive control, fuzzy neural network control, fuzzy control and so on are emerging, which are believed to have greater development and wide application.

4.3 in depth study of control strategy

The traditional control theory and technology based on quantitative mathematical model can not be used in the design of drying process system. It is necessary to further develop advanced process control system, study advanced process control law, and transplant and transform the existing control theory and method to the field of process control. These aspects are also paid more and more attention by the control field. Further strengthen the control theory research, such as the three mechanisms of predictive control In the drying process control, it is urgent to develop a model predictive control method with good real-time performance to reduce the on-line calculation time under the premise of ensuring the drying quality; it pays attention to the interdisciplinary research and uses other effective control methods for reference to solve the existing process control problems To solve these problems, the existing drying process control algorithm should be improved, developed and innovated; the reliability of the automatic drying quality control system should be further improved, and the control algorithm with adaptive ability should be established.