The Application Of Genetic Algorithm In Batch Reactor Fault Diagnosis

The process of intermittent reaction is an important chemical production process. The nonlinear, delay and uncertainty factors determine the complexity and risk of the process operation. The large-scale and integration development of industrial equipment promotes the importance of intermittent reaction process's fault diagnosis. This paper presents a fault diagnosis method based on improved genetic algorithm to optimize BP neural network.The neural network has non-linear approachable and self-learning ability, so it can adapt the dynamic characteristics of non-defined system. Especially BP neural network has some characteristics such as simple structure, strong robustness and high pattern recognition ability, so it has a broad range of applications in fault diagnosis. But the BP neural network has the defects that are easy to fall into the local extreme data and slow convergence speed in the training process. A global optimizing genetic algorithm is used to optimize the parameters of BP neural network and improve the defects effectively. Then this paper establishes a fault diagnosis model based on BP neural network optimized by genetic algorithm.This paper introduces the working principle, reaction characteristics and the present status of fault diagnosis of batch reactor. Then it introduces working principle and the characteristics of BP neural network, and analyses its application in fault diagnosis. According to the defects of the BP neural network in the training process, this paper puts forward the solution of using global optimization of genetic algorithm to optimize the network parameters. Then it introduces the principle and the realization process of genetic algorithm in detail. In view of the actual fault diagnosis of the batch reactor, this paper gives idea and specific steps of the genetic algorithm to optimize the BP weights and thresholds. At last this paper collects the fault data of batch reactor through the fault simulation experiments, optimizes the weights and thresholds by improved genetic algorithm and establishes the fault diagnosis system of batch reactor.This method is applied in temperature fault diagnosis of batch reactor. The simulation results show that the algorithm can improve the fault diagnosis accuracy and shorten the fault diagnosis time. The method effectively improves the accuracy and efficiency of fault diagnosis. So the method in this thesis has certain significance at theoretical study and engineering practical levels.