Fault Diagnosis And Performance Analysis Of Turbine Thermal System

As an important system, the operation of the regenerative cycle system will have a direct impact on the safety and economy of the entire plant in a large number of thermal power units. At present, much successful work has been done in this field, but most of it focuses on turbine vibration faults diagnosis based on shafting supervision.It is noticed that faults that occur to the system of thermal power units are also in a quite high proportion, but less successful onsite applications of fault diagnosis system to the thermal system has been found at home and abroad. Based on the theory of artificial intelligence, there is a combination of neural networks fuzzy theory and expert system in the regenerative cycle system, especially in the high pressure heater system being investigated.It will realize the optimization of unit operation.This paper analyzes the failure mechanism of the high pressure heater system and summarizes the typical characteristics of the fault knowledge based on the existing research results combining with field experience .In this paper, a new neural network training method for BP network called"learning rate self-adaptive adjustment based on constant error correcting rate"is put forward which greatly shortens the network convergence time. Two-tier pattern recognition methods of high pressure heater fault diagnosis system are put forward, and error analysis is used to optimize operation of generating units on the basis of the establishment of the performance model. The performance analysis and fault diagnosis system of the thermal system is designed and developed with the function of real-time data collection, computing performance, condition conitoring, fault diagnosis, mass storage and statistical analysis.