The Fault-Tolerant Control Of Water Tank

Fault-tolerant controlling is the function that when the organ or partial system gets the fault, the system still has ability to achieve basic function. It is a new approach to improve the security and dependability of system. There are two methods to achieve the fault-tolerant control, one way is hardware redundancy and the other way is software redundancy. This thesis is using software redundancy way to achieve the fault-tolerant function basing on three-tank experiment equipment. The thesis mainly contains four parts, one is two-tank system modeling, another part is water position controlling. The third part is detecting the sensor fault. The last part is fault-tolerant control. The part of two-tank system modeling particular means that constructing the model of system through the system modeling and then get discrete math model through the system identification. In the water position controlling part, there are two methods are used to control the water position, one is forecast function arithmetic and the other is PID arithmetic. In the sensor fault detecting part, firstly, establish the state observer to estimate the state, then select appropriate threshold to detect the fault. The part of fault-tolerant controlling particular means the fault decision-making after detecting the fault. That is when one sensor gets the fault, because of the relationship in two-tank, using the other water position to control this tank water position. When two sensors both get the fault, then transfer the experience data to control the water position at appointed position.