Research On Fault Diagnosis And Fault-tolerant Control Strategy In Aluminum Electrolysis System

Aluminum electrolysis is a complex industrial process which has vulnerable to a serious of adverse factors, such as the strong electric field, magnetic field and high corrosion, etc. It makes aluminum electrolytic process becomes a complex nonlinear system which has a strong interference, a large delay, a strong coupling and differents fault types. Due to the presence of these unfavorable factors, the aluminum electrolysis production process less susceptible to supervision and control on the operation. Consequently, the effectively fault forecast about conditions of aluminum electrolytic cell has a very important meaning to reduce the energy consumption in the process of production, prolong the service life of aluminum electrolytic cell.Because in the process of aluminum electrolysis, there are lots of complex chemical and physical changes in the electrolytic cell, which has many fault types and more difficult to detect, it makes more difficult to control the system. Usually, large pre-baked aluminum electrolytic cell adopt the alumina point charging control method, without considering the operating status of each electrolytic tub, and adopts the unified control mode. It makes the control effect is not obvious and directly affect the control precision of alumina concentration.It was based on the above reasons, this thesis according to the characteristics of the aluminum electrolytic system, a kind of aluminum electrolysis fault diagnosis and fault-tolerant control strategy has been putted forward. The strategy thoughts are that:At first, to conduct a preliminary identification of fault diagnosis to the operation condition of aluminum electrolytic cells. And then, this thesis adopted different control strategies or operation mode for the different aluminum electrolytic cells state. It makes the fault diagnosis and fault-tolerant control technology for reasonable fusion introduced into the aluminum electrolysis production system.This thesis mainly to do the following works:Firstly, this thesis discussed some of the background of the aluminum electrolysis process fault diagnosis, the development of aluminum electrolysis domestic and international in this stage, and further explained the necessity for fault diagnosis.Secondly, this thesis expounded the prediction technology of fault diagnosis and fault-tolerant control technology, the main technical parameters of the aluminum electrolysis production system and the common failure types. Analyzed the shortage of the present stage in aluminum electrolytic process is controlled by a single mode, and puted forward a combined aluminum electrolytic control method based on fault diagnosis and fault-tolerant control technology.Thirdly, this thesis expounded the design method of aluminum electrolytic cell state detection system, adopted the extension neural network as the state of aluminum electrolytic cell test system, and innovatively inference the gray forecast and immune genetic algorithm into the traditional extension neural network. According to the simulation experiment results found using the improved extension neural network can improve the control superiority.Fourthly, this thesis designed an aluminium electrolytic tank fault-tolerant control system to detect the state of normal state, the anode effect, cold cell and heat cell state which using three different fault-tolerant control methods have processing function with pertinence. Finally, make the simulation research to prove that this method can improve the stationarity of alumina feeding in the fault condition.Fifthly, this thesis expounded the network control method, the control design contain hardware system and software system, in order to realize the purpose of integrate the monitoring, diagnosis, control and management.Sixthly, this thesis summarized some conclusions based on the fault diagnosis and fault-tolerant control strategy research in aluminum electrolysis system.