Fault-tolerant Technique For Extreme-manufacturing Equipment Oriented Huge Hydraulic System

With the development of modern science and technology, industrial unit become increasingly complex, and the automatic level also become higher and higher, the thousands ton of stamp equipment, the spaceflight and the new highly energy-saving metallurgical equipment all include huge hydraulic systems which have characteristics of high pressure and high flow rate, and its control system is very complex, once this type of system takes accident, it may result in huge loss of the personnel and property, therefore, it is actually consequential to keep this modern and complicated huge hydraulic system and reliable.The structure of huge hydraulic system is very complex which exists time-variation parameters and uncertainties resulted from external disturbance, which makes that the dynamic characteristics of system very complicated, and it is hard to establish the accuracy mathematic model, so the design of controller can't adopt the traditional method based on system model. This paper take analysis on the fault characteristic of electro-fluid servo system which is applied in the huge hydraulic system, and the intelligence control based on neural network in the system is explored usefully. firstly, on the basis of analysis of system characteristics, this paper researches on neural network identification technology for establishing online dynamic model of system, which can scout the dynamic changes of the system, and the controller is designed which based on PID adaptive control strategy according to the dynamic model, with the controller which can reduce disadvantageous influence caused by nonlinearities and uncertainties of electro-fluid servo system. In the industrial control field, neural network have the ability of self-learning adaptive and nonlinear match which can offer an effective solving method for nonlinear system modeling, and fuzzy inference technology can easily build the rules base; This paper isn't simply using the advantage of nonlinear approximation ability of the Neuron Network, and also combined the fuzzy interface and de-fuzzy interface of the Fuzzy Control, which can easily build the Rule base. The black-box working mode can be avoided and intelligent control can implementation by using the Fuzzy Control.The application of traditional fault-tolerance technology in the huge hydraulic system still exist some limitations, so in this paper the author puts forward a new type of fault-tolerant method of Knowledge Modules of Hydraulic Fault,with the structure innovation of Controller, research the unit of information perception in pilot stage to aware the change of fluid information, and the information can be processed, based on the unit of information perception ,the fault-tolerance control unit can be designed. The application of knowledge modules of hydraulic fault which can solve the questions of the whole fault-tolerant in the huge hydraulic system, moreover, with it the structure of hydraulic system can be greatly simplified.The purpose of this paper is to apply the modern fault tolerant control theories to the design of huge hydraulic system in order to raise its ability which can void the equipment damage which is done by small controlling trouble and there will be no loss which is done by shutting down to repair by abrupt accident. Therefore, the fault-tolerant control research in this paper has more important realistic meaning.