The Application Research On Dynamic Batching System Of Humanoid Logical Predictive Controller

Through analyzing the control process of dynamic batching system of concrete mixing station, then the error and the lag appeared in the process of weigh ingredients over the idea of combining active ingredients weighed system and quantitative of ingredients. The controlled object was accused of allowing the request of the second-order systems and the second delay system, according to the PID controller parameter tuning rules of the ingredients error, using the least squares algorithms for amending the error online, the parameters estimation algorithm of least squares can compensate effectively amending the parameters errors, which improve the recognition of overall system.At the same time, simulating characteristics of human thinking features, a new type of intelligent controller is proposed:Humanoid logical predictive controller. The controller integrates the characteristics of logic controller based on pan-Boolean algebra and the model predictive controller, which is a multi-valued logic hybrid dynamic system. Because of adopting on-line rolling optimization control strategies based on a wide range of output prediction, the predictive control is input closed-loop and the output equation is very complex, its robustness and stability analysis encountered great difficulties. Taking advantage of Pan-Boolean algebra logic controller with low demand for model characteristic, using logic controller constraint forcibly to output curve, using logical reason and judgments, is carried out control decision which is able to guarantee the process in a stable and secure operation under the abnormal circumstances. To a certain extent, it solves the robust problem of predictive control.Against a large time-delay phenomena (The ratio between pure lag time of objects and time constant of objects is larger than 0.3) in the process of industrial control systems, according to MPC algorithm of predictive control, integrating fuzzy set theory of fuzzy control into the MPC algorithm, is improved a based on real-time fault-tolerant prediction algorithm of IC (RFPC algorithm). Its feature is according to forecasting model to trace accurately objective function has input, to feed back and amend every step of optimization strategy, simulating in the control experienced operators do not hesitate to cut off some important pathway when the system beyond the limit, even at risk while the system work in the normal range is controlled in accordance with experience, so that fluctuations as small as possible and has good tracing performance without error.Matlab simulation experiment suggests that the controller's control strategy can make the system run more stable and avoid ingredients precision volatility, its ingredients accuracy and speed are all required, the efficiency of the whole system has also been improved. Compared with other types of artificial intelligence controller, the controller structure is simple, physical background is explicit and mathematical concepts clear with characteristics of fast response, good robustness, small overshoot, high precision steady-state. Under the premise of ensuring stability and robustness, simultaneously solving the contradiction between performance indicators of stability, speed and accuracy, the speed and accuracy partially optimized, its control strategy close to people's mind.It will be widely used in the industrial control field.