Li Fang 1,2 CEREALS 2

(1 Hebei Normal University Department of Electrical, Qinhuangdao 0666002 China Agricultural University Institute of Technology, Beijing 100083, China)

Abstract: In analyzing the characteristics of advanced control based on the summary of the development and status of grain drying process advanced control method, pointed out the drying process control problems, and proposed the development direction of the grain drying process control.

Keywords: Drying; advanced control; adaptive control; model predictive control; expert control; fuzzy control; neural network control

Grain drying The basic objective is maintaining the stability of the drying process at the lowest cost and power consumption of the drying grain get the best drying quality. Grain drying process is a typical nonlinear, multivariable, large lag, unsteady heat and mass transfer parameters associated coupling, food itself is a complex biochemical substance, in order to achieve the above objectives, the drying process must be constantly adjusting drying parameters of the dryer work process control. Automatic control of the drying process is an effective means of achieving a dryer quality, high efficiency, low cost, safe operation. Automatic control of the drying process, to achieve automatic control of grain dryer, to ensure the machine is important uniform grain moisture, dry food quality, reduce labor intensity and full drying machine production capacity. According to the National Food Authority in "" fifteen "Grain Industry Science and Technology Development Plan" [1] in the development goals, grain drying process line monitoring and automatic control has become a key issue to improve China's grain drying process efficiency and achieve " fifteen "an important way to planning. With the construction of grain depots to increase investment, food processing industry and increasing international integration, automated grain drying of grain will be added to the international flow of large market basis.

Advanced control features 1

Automatic control of grain drying process began in the 1960s. Then using feedforward control, feedback control, feedback - feedforward control and adaptive control and other traditional control methods. Traditional control theory of differential equations or transfer functions, to express knowledge and existing information systems into the drying process analytical formula. However, the use and design of the above-described method of controlling grain drying machine control system will encounter many difficulties, because: (1) grain drying process is complex, time-varying and nonlinear; (2) some drying process variables (such as grain quality and color) can not be measured directly, and some variables (such as grain moisture content) measurements may be discontinuous, inaccurate, incomplete or unreliable; process model (3) is the actual dryer approximation process and requires a lot of computing time; (4) is almost impossible to use an appropriate model to represent the drying process such as a non-linearity, hysteresis, time varying complex systems; and (5) of the controlled variable grain dryer exist between the control variables and interactions; operating conditions (6) grain dryers complex, a wide range of disturbance variables, it is difficult regulation.

Obviously, to overcome the above difficulties to the traditional method of controlling grain dryers continuous improvement, and to explore new and more effective control method. 1970s, the progress of the electronics industry, especially the development of computer technology now makes the idea of so-called advanced control can be widely disseminated. Advanced control target is to solve those using conventional control ineffective, even beyond the control of complex industrial process control problems. In recent years, modern control and artificial intelligence has made considerable development has laid a strong theoretical foundation for the implementation of advanced control systems; and computer control is distributed control system (DCS) for the popularity of computer network technology advances, compared with advanced control the application provides a powerful hardware and software platforms. In short, the need for industrial development, control theory and the development of computer and network technology, a strong impetus to the development of advanced control.

The rapid development of computer technology, artificial intelligence, control theory began to be applied in dry in machine control, significantly improved the performance of the control system was dry. Traditional control methods due to the large lag and nonlinear contact on food drying process is not suitable for controlling grain dryer. Advances in artificial intelligence technology is widely used in the field of engineering, control theory and advanced control method applied to the automatic control of grain drying process, the control method of continuous improvement, improve the control effect. After 90 years, the process has begun to control the development of intelligent, intelligent control theory combined with increasingly drying technology together, the use of artificial neural network model to simulate the drying process and control; expert system is used to predict the quality of grain, dry and process control management consulting and other aspects.

And control theory, instrumentation, computers, computer communication and network technology closely related to the advanced control system has the following characteristics:

(1) Theoretical basis of advanced control system control strategy is mainly based on the model, such as: model predictive control, these strategies take advantage of industrial process control input and output information to establish the system model, rather than relying on in-depth study of the reaction mechanism. Recently, the knowledge-based controls, such as expert control and fuzzy logic control is becoming an important development direction of advanced control.

(2) advanced control systems are commonly used to handle complex variable process control problems, such as large delay, multivariable coupling, accused of the most variable and control variables there are various constraints. Advanced control strategy is to establish a dynamic coordinated restraint in conventional single-loop control based on the control, the control system can adapt to the dynamic characteristics of the actual industrial processes and operational requirements.

(3) implement advanced control systems require high performance computers as a support platform. Due to the advanced controller controls the complexity of the algorithm and computer hardware advanced control algorithms affect two factors, complex systems are usually implemented on the host computer. With the growing development of advanced control technology and DCS functions, some basic and advanced control strategies can be implemented on a control loop DCS. The latter approach can effectively enhance the advanced control her reliability, operability and maintainability.

2 drying process control of advanced development status

Example advanced control strategy is the core of the way of advanced control systems, advanced control strategies currently a wide range of primary drying process advanced control strategies are: predictive control, fuzzy logic control, neural control, adaptive control, expert system.

2.1 model-based control

2.1.1 Adaptive Control

The basic principle of adaptive control is to control parameter variations and external disturbances drying process parameters at any time to adjust so that the dryer is in the best working condition. Adaptive control for a variety having a grain dryer, the dryer without any data on its own characteristics, environmental conditions and food situation, no special requirements, the controller in response to disturbances faster, can control the parameters of the model with the external conditions changes automatically adjust and so on. Sweden Nybrant (1985) the self-tuning technology to the cross-flow grain dryer control. Dryer exhaust temperature as an output variable, cereal grain discharge rate is used as the controlled variable, and select Auto-regressive moving average (ARMA) model of the dynamic characteristics of the performance of the cross-flow dryer. In the laboratory cross-flow drier verification testing, the standard deviation of error control during and after the 50 samples is 0.13 ℃. The results show that the adaptive controller can accurately control the temperature of the exhaust gas. Liu Jianjun [5] (2003) for HTJ-200-type drying machine study algorithm system for quantitative analysis of the sample through the online collection and intelligent optimization, the process of establishing the real-time detection of intelligent model data determined through intelligent optimization call artificial intelligence algorithm model, access control system rules, the amount of control given by the control program after D / a conversion output to the execution unit. Li Xiaobin, etc. [3] (1998) study of advanced vacuum freeze-drying equipment control systems for different demands of freeze-dried material, take two adaptive algorithms and critical ratio method DRA, self-tuning control method to solve the main controlled object control parameters - temperature lag.

2.1.2 Model Predictive Control

Process control theory research is the latest model predictive control is based on the model, and implement it in combination with optimized feedback correction control algorithm, which is particularly effective for the control of nonlinear and large time delays.

Forbes, Jacobson, Rhodes, and Sullivan [24] (1984) and dried Eltigani designed controller based model, which controls the behavior is based on a process model and a so-called fake entrance grain moisture content. Drying rate parameter updating based on the difference between the moisture content of the model predictions and the measured sensor outlet intermittently. Forbes and Eltigani different controllers that control the type of algorithm used in the process model different. Liu Qiang University of Michigan [25] (2001) proposed cross-flow dryer model predictive controller. In a simulation test Zimmerman VT-1210 Tower on the cross-flow grain drying machine, using the controller Labview can build successful operation, and the export of the corn moisture content of less than 0.7% in the control set point. Change controller entering the dryer inlet moisture content cereal considerable range, and a large step changes in air temperature can be well compensated.

Model predictive control, the more work has focused on the process of establishing and solving model and consider drying quality problems in the model. France P.Dufour [31] et al. (2003) by means of partial differential equations (PDES), to expand the model predictive control system model, so that the equation can PDES large-scale applications. They put forward a global model, designed to reduce the PDE based model optimization task solution brought online computing time. Developed a large number of applications and the actual structure of IMC combines a generic MPC framework. It uses two feedback loops in IMC- MPC structure, process performance and to correct errors based simulation model of the online optimizer caused. Denmark Helge Didriksen [29] (2002) developed a description of the quality drum dryer, the energy and momentum transfer dynamic model of a law, and applied to the dried sugar beet predictive control. The results showed that with the change in the manipulated variable and disturbance, the model has good predictive ability. Through simulation compared with the traditional model predictive control and feedback control, model predictive control showed a better performance. French I. C. Trelea, G.Trystram and F.Courtois [27] was designed in 1997 for optimizing nonlinear predictive batch drying process control algorithms on a pilot scale dryer was tested. Experiments show that the algorithm can handle significant interference and failure of the control algorithm can be easily applied to other batch process, such as freezing, sterilization or fermentation. Some scholars neural network model predictive control for process modeling. Jay [32] (1996) for the first time the neural network model predictive control for the drying process. Former French JA Hernandez-Perez et al. [33] (2004) proposed a heat and mass transfer prediction model based on artificial neural network, the model will shrink product as a function of water, applied with a hidden layer of two separate feedforward network, hidden layer with three neurons, can accurately predict the mass and heat transfer. In the data verifying apparatus, kinematics simulation and experimental tests are consistent. Online development model can be used to estimate and control the drying process.

2.2 Intelligent Control

Intelligent control is an emerging theory and technology, it is the advanced stage of development of the traditional control. This model is characterized by no closer to controlling a theory of the human brain way of thinking, is mainly used to solve the complex control system is difficult to solve using traditional methods, and the design of its controller from the shackles of the system model, algorithm simple and robust. Currently, expert control, neural control and fuzzy control, intelligent control technology is becoming an important development direction of advanced control.

2.2.1 Expert Control

Experience expert system technology and mathematical algorithms can control engineers to integrate together to maximize the use of existing knowledge, to control the effect of the traditional control methods difficult to obtain. Expert control system runs in a continuous, real-time environment, the use of real time information processing to monitor the dynamics of the system, and gives the appropriate control action. The expert system technology and food drying process control combined for food production, management and monitoring, improve production efficiency and production efficiency of food. Liu Shan [12] (2001) developed a dried food fuzzy expert control system, the simulation results were compared with the measured data, the two are basically the same. Liushu Rong [13] (2001) expert system technology combined with the drying process control, a fuzzy expert system for a high moisture grain drying process control. He Yuchun [14] (2001) by expert intelligent control the drying in the drying process parameter optimization, calculated energy consumption, efficiency, quality and benefit point in the design process of drying and drying apparatus, drying machine along the common benefit line of grain drying, the drying process so that the device is always in the best practices; at the same time, the temperature measurement and control technology and network technology, the Internet, set up a simple and effective network-based temperature control system.

2.2.2 Neural Network Control

Neural networks can provide an effective method for the modeling of complex non-linear process, which in turn can be used in the design process measurement and control systems soft on. Neural Networks in the drying process are mainly two: the drying process modeling and control.

France J.-L.Dirion (1996) [6], who developed a neural controller to adjust the semi-batch experiment reactor temperature for the formation of a basic experimental neural network learning database, the neural controller It provides a very good set-point tracking and disturbance elimination. Ancestry [9] (2000) developed a single neuron adaptive PID controller based on the design of wood drying kiln neural network model, with BP algorithm input and output characteristics are described drying kiln and model learning and training, through tests and simulation prove the conclusion meet the requirements of error indicators. Zhang Jili [10] (2003) fuzzy control and neural network technology, designed the grain drying process parameters on-line detection and intelligent predictive control system. Food exports dryer moisture range under intelligent control than smaller manual control, the former was 13.6% ~ 14.4%, which is 12.4% ~ 14.2%; exports of grain moisture fluctuation frequency intelligent control over manual control of small The former fluctuation period of about 20h, which is a period of about 8h. Wang products [11] (2003) established using improved BP network algorithm Neural Network Model drying tower through the neural network model neural network controller to achieve the arch dry grain moisture drying tower systems intelligent control, improved quality and efficiency of grain drying.

Liu Yongzhong [8] (1999) Application of artificial neural network theory to predict freeze-drying process characteristics to the drying time, sublimation share drying time, drying products productivity and sublimation interface temperature drying process parameters as the output parameters of the network model, network the mathematical model predictions and compare predicted results agree well with the calculated results. Zheng Wenli [7] (2000) on the use of artificial neural networks in the process of freeze-drying freeze-dried material weight intelligently simulate changes: for lyophilization process conditions orthogonal experiment to study, use of the network after learning of the prediction and optimization of process conditions .

2.2.3 fuzzy control

Fuzzy control is a rule-based control, direct use of linguistic control rules, which is based on knowledge of the site operator control experience or experts in the design does not need to establish accurate mathematical model of controlled object, so that the control mechanism and policy easy to accept and understand.

At present, domestic and foreign control of the drying process is the main application of fuzzy control method. Zhang Qin [15] (1994) of continuous cross-flow grain dryer were studied fuzzy control to control dryer operation by adjusting power and unloading grain auger speed heater control verification test success rate of 86.4 %. Jun-Ming Li [16] (1996) in a hot air drying column temperature as the basis, the production of corn drying in a skilled operator through observation and experience of the development of the sensory system of fuzzy control rules, the use of fuzzy control with a displacement speed of the motor regulation, and proposed cross-flow corn drier self-organizing fuzzy controller should adopt open-loop fuzzy control system to solve the corn drying process large lag. Charles Lee Tak, Li Ye Gang [17] (2001) designed a to 89c51 microcontroller core fuzzy intelligent controller on the downstream dryer through the wheat-line drying experiments demonstrate that the system response time is short, overshoot small, high control accuracy, but the entrance grain moisture fluctuations will affect the drying process.

Many domestic graduate students engaged in research work grain dryer Fuzzy Control. Fuzzy rules Mengxian Pei Northeastern University [18] (2003) in the smart modeling and intelligent control of grain drying tower, using fuzzy set theory and optimization algorithm, the grain drying system, smart model and fuzzy control systems, designed out of the system fuzzy controller. Tang Xiaojian Harbin Institute of Technology [20] (2003) Fuzzy Control of Francis Grain Dryer TS multivariable model-based simulation of the control system, and compared with traditional manual control methods and fuzzy control method. South China Agricultural University of Cao Yanming [21] (2000) against high humidity drying paddy circulation tempering process characteristics, the use of analog design method of fuzzy control our way of thinking, the development of paddy circulation dryer automatic control system. Su Yufeng Northwest Institute of Light Industry [23] (2002) using fuzzy algorithm based on practical experience of workers, freeze drying system using SCM to control, improve the degree of automation equipment.