Method For Underdetermined Fault Diagnosis With The Prior Knowledge Of Gas Turbine

Gas path diagnosis is an important research area of information and intelligence of gas turbine. On the one hand,gas turbines are more and more widely applied in a vast field because of the gas turbine is a high efficiency, safety and environmental protection, stable and reliable power equipment.Thus it is necessary to strengthen state monitoring and fault diagnosis to avoid accidents; on the other hand, as one of the main fault type of gas turbine,gas path faults have the characteristics of harsh conditions, monitoring difficult and lack of information. In the event of a gas path fault, or the performance decreases, or it affects the reliability and safety, or it causes accidents. Therefore, it is quite important to do gas path fault diagnosis theory and application research. The main research content of this paper is the problem of the fault diagnosis of gas turbine under the conditio n of insufficient information. The key point is the research on the diagnosis method of the prior knowledge of gas path fault and its application.Firstly, the fault model of the turbine is established and the typical gas path fault simulation analysis is carried out. The model is consistent with the actual gas turbine operation data, and the fault setting meets the typical gas path fault characteristics. The deviation of the measurement parameters in typical gas path fault and variable working conditions is analyzed, and the prior knowledge of the gas path fault is formed.Secondly, the fault of the gas turbine is analyzed and the diagnosis method is studied. Through the analysis of the linear gas path established gas path fault equation, through discussing constrained least squares, solving method based on 1-norm constraint and based on 2-norm constraint, the constraint rules do not conform to the actual gas path fault situation and makes the diagnosis effect is poor. A priori knowledge of the typical gas path fault is integrated into the genetic algorithm to solve the problem of the gas path fault. The experimental results show the feasibility of the method.Then, we solve the underdetermined fault diagnosis problem with the prior knowledge of gas turbine. Firstly, the fault of gas path fault is isolated using the deviation direction and size of the measuring parameters, and then the corresponding constraint is used to identify the fault. The di agnosis results show that the method can diagnose gas path fault well.Finally, the practical application of gas path fault diagnosis is discussed. In view of the actual situation, load, atmospheric environment and sensor fault, the influence factors can be separated from the gas path by the expansion of the influence coefficient matrix. Finally, the effect of measurement noise is discussed.