The Research Of Bridge Deflection Sensor Adaptive Correction Method In Bridge Health Monitoring System Generally

Bridge is one of the important transport facilities. The bridge structural healthmonitoring system attracts much attention nowadays. Sensor is the basis of the monitoringsystem，and it can directly affect the monitoring system operation working properly or not.Up to now, a lot of research on sensor fault diagnosis has been done by scholars athome and abroad. Those researches can identify the faulty sensors effectively, or it canimprove the efficiency of the monitoring system. The sensors used on bridge should betimely replaced once identified failure，but the research on the problem of replacementsensors is quite few. In order to maintain the system stability and consistency, do researchon the data of replacement sensors is very meaningful.Deflection is a important parameter of the bridge structural health monitoring system,which can show the structure performance of bridge. In order to solve the problem of datainconsistency before and after sensor been replaced，an improved deflection correction method based onWave-SVM is presented in this paper. The main content of the thesis are:(1) The paper introduces the bridge structural health monitoring system and commondeflection measurement methods at first. Then the problem of data inconsistencies before and afterdeflection sensor replaced is studied. Research shows the existing method is easily affected byexternal environment, low calculating accuracy, and does not have the disadvantages of adaptiveperformance.(2) The relativity of deflection sensor data in bridge structural health monitoring system isbeen studied. The method that using wavelet analysis in data reduction is been presented. SupportVector Machine (SVM) is also be used.(3) The paper present an adaptive correction method of deflection sensor based on Wave-SVM.The method can overcome the deficiency of the original method. By using wavelet low frequency bandand SVM, the accuracy is improved.(4) Numbers of experiments are done by using the real data of Caiyuanba Yangtze River Bridge.Results show that the mean square error of the new method is less than the existing method. The newmethod solves the problem of finding the initial state, and the precision is improved at the same time.