| The converter valve is the core component of HVDC transmission,and a large amount of heat is generated during the operation period of the converter valve,so the valve cooling system plays a vital role.The leakage detection of the conventional valve cooling system is mainly realized by detecting the water level value of the expansion tank.This method not only has poor accuracy but also cannot timely reflect the leakage condition.Because the water level of the expansion tank of the valve cooling system is affected by many factors,including the power of the valve group,the temperature difference between the inlet and outlet water,the temperature of the valve hall,etc.,the theoretical value of the water level of the expansion tank can also be fitted and calculated using the linear regression method and compared with the actual value.To determine the leak.Because the valve cooling system is a nonlinear system affected by many factors,the linear fitting method can not have a good detection effect on the situation where the water level fluctuates greatly.Firstly,this paper has proposed a new method of water leakage detection,using the fractal extrapolation interpolation algorithm for water leakage detection,the fractal theory has considered that the variation law within a certain period of time is similar to that of the external time period,through the historical data,we have obtained the iterative relation of the inner interval and the attractor of the iterative function.Iterated a certain number of times through a certain initial point,we obtained the mean square deviation from the attractor,and the mean square deviation is minimized when the vertical coordinate value can be used as extrapolation point function value.Then the fractal extrapolation interpolation algorithm is applied to the daily water level prediction of the power plant by utilizing the different fractal characteristics of the expansion tank water level data of the HVDC cold system.The fractal extrapolation algorithm starts from itself and looks for the internal variation of things,and extends this rule as a reference to find unknown values.It has a unique advantage for studying nonlinear systems.The system model has established and simulated by MATLAB software The experimental results showed that the method has better prediction accuracy.At the same time in order to monitor the convenience,the establishment of leak detection B / S system.On the basis of the above work,this paper has built a new detection system,the use of web-related technologies to build the system page to MVVM front-end framework Vue.js mainly SPA system,selected Mysql as the system database,the previous manual on duty records,paper archives way to automatically convert data into JSON format storage way,used Express.js based on the back-end framework Node.js build server platform,front-end page used ECharts plug-in data visualization.The system functions are: real-time inspection of data inspection,security verification login,early warning water leak detection and alarm event query capabilities.The new detection method and leak detection system provided an efficient and fast web platform for leakage detection of HVDC converter valve cooling systems,improved the accuracy of leak detection,ensuring stable operation of the converter valve,and reduced labor cost.The leak detection system has designed in this paper is accurate,reliable,smooth and safe to operate,and could provide great convenience for on-site operation and maintenance personnel. |
PDF Full Text Request