Optimization Of Vibration Protection For Pumped Storage Power Station

Unstable power generation and unpredictable grid connection time are the main constraints in the large-scale use of new energy power.To solve the energy dilemma,the use of energy storage to absorb new energy generation has become an important means.Photovoltaic power plants combined with pumped storage power plants and wind power plants combined with pumped storage power plants have been proven feasible to replace thermal power generation.In the context of the dual carbon strategy and with strong support from national economic policies,the pumped storage industry is booming and a large number of pumped storage power plants are about to be built and put into operation.However,the causes of mechanical oscillation of pumped storage units are diverse,the process is tedious and the mechanism is complex,and the relevant standards and regulations still need to be improved.At present,accidents related to oscillation in pumped storage power plants are frequent,especially for newly commissioned plants,and the original method alone cannot meet the requirements of high operating intensity,high capacity and high head development.Therefore,this thesis is dedicated to further improve and optimize the vibration pendulum protection of pumped storage units to ensure their safe and stable operation.This study summarizes the directions of vibration pendulum protection optimization,including redundant configuration,working condition differentiation,vibration protection measurement point optimization,logic optimization,vibration protection fixed value optimization,vibration protection frequency selection optimization and shaft system space state protection optimization,by investigating the design of vibration pendulum protection configuration in domestic pumped storage power plants and the technological change variation occurred during the actual application.In addition,we deeply analyze the causes of accidents or failures of pumped storage units and propose vibration suppression and noise reduction techniques,such as setting elastic supports and vibration isolators to weaken the transmission of vibration energy or even block the transmission process.This thesis proposes the use of feedforward neural network(FNN)to set the alarm value of oscillation protection.To address the problem of insufficient accuracy of FNN,we improve the neural network by using the Golden Jackal Optimization Algorithm(GJO)and obtain the feedforward neural network(GJO-FNN)based on the Golden Jackal Optimization Algorithm for setting the alarm value of the oscillation protection.At the same time,the corresponding model was developed and simulated for the operating conditions of the pumped storage unit.The results verify the improvement of FNN prediction accuracy by the Golden Jackal algorithm and demonstrate that the feedforward neural network based on the Golden Jackal optimization can be used to predict the oscillation of pumped storage units.Based on this,the prediction value plus the maximum error value is used to set the alarm value for oscillation protection.This can improve the sensitivity of oscillation protection,prevent the expansion of accidents.This thesis analyzes the actual accident cases of vibration and swing of pumped storage power stations in China by GJO-FNN,and the results prove that the method has remarkable effect on foreign objects in waterways,improper setting of fixed values and operation and maintenance inappropriate problems,and can meet the four characteristics of relay protection requirements.Meanwhile some suggestions are put forward to prevent and treat the abnormal vibration and swing during operation,and ensure the safe and stable operation of pumped storage units and power grids.