The Study Of Fault Diagnosis System Based On Immune Population Network Algorithm

As one of the highly parallel and self-adapted information study systems, Biology immune system can automatically recognize and eliminate antigen invaded into the body. This system which keeps the stability of inner space also has the ability of study, memory and self-adaptation. The Artificial Immune System (AIS) illumined by the biology immune system is new intelligent research field after Neural Network, Fuzzy System and Evolution Computing in the searching field of modern information. This method is used in the aspect of Automation, Information Security, Data Mining, and Pattern Recognition.Fault Diagnosis is integration subject developed from 60 or 70 years in the last century. With the development of Technology, Facility becomes more complex and difficult of fault diagnosis augment obviously. Therefore, we should research and explore practical fault diagnosis technology, not only for the need of modern product and facility automation, but also for the subject of maintenance and management of modern mechanism facility operating.In this paper, we start from introducing the history of biology immune system, and then introduce some basic concepts such as antigen, antibody, lymphocyte and basic structure composed by immune apparatus, immune cells, immune molecule and basic mechanism on immune tolerance, immune response. We also summarize the further mechanism which concludes immune recognize, immune memory, adaptation and so on.Then, we propose an immune adjusting algorithm based on immune adjusting mechanism. This algorithm could improve the speed of immune system response and maintain the stability of immune system. The main work of this algorithm is dealing with the uncertain status existed in inner and outer of system to reduce warp between practicing operating and expectation operating. Latter, we combine the algorithm with Fuzzy PID Controller to improve the performance of regular PID control system. And we apply this controller to the superheated steam temperature control system under 4 typical loads of a certain supercritical 600 MW units in fossil-fired power plant and Two-degree-freedom control system as a practical example. Simulation studies show that the performance of the new Fuzzy-Immune PID control system is superior to regular PID cascade control system and Two-degree-freedom PID control system. It has some excellent features such as strong robustness, stability and fast response velocity and so on.At last, we propose Immune Population Network Algorithm after analyzing population immune algorithm and network immune algorithm. This algorithm could execute multi-point parallel search from local to global search field. It has higher global and local optimization search ability for it using the new kind of search mechanism. We prove the effect of this algorithm in the optimization search by using testing functions. At the same time, we apply this algorithm to the fault diagnosis system trial of rotating machinery and power plant dust-making system. The result testified this algorithm is useful in fault diagnosis system.