Analysis on oscillation in electro-hydraulic regulating system of steam turbine and fault diagnosis based on PSOBP

Abstract

The oscillation in electro-hydraulic regulating system is one of the main reasons leading to the oscillation of electric power system. In this article the influence of internal leakage and plugging of servo valve on the characteristic of the regulating system was analyzed. With the simulation in AMESIM it was found that the leakage of the power stage spool, the plugging of nozzle and orifice could both cause the oscillation of regulating system when the system oil pressure fluctuated. The fluctuation of piston displacement and oil pressure in oil servo motor in different fault state were analyzed. For preventing regulating system from oscillation it was needed to diagnose fault in the servo valve. Taking single leakage for example the change of parameters such as regulating time, steady displacement, and system frequency response, etc. were analyzed. The characteristic table of different fault was given based on simulation. In diagnosis particle swarm optimization (PSO) arithmetic was combined with BP neural network. The weight and threshold values were optimized by PSO arithmetic and five characteristics such as the piston displacement and system frequency were selected as the input of network. The results show that the PSOBP arithmetic has quick convergence speed and good generalization ability, with which the diagnosis of fault in servo valve can be done effectively.