Research On Damage Identification Based On Elman Neural Network And Monitoring Data Processing

Bridge structure is an important node of modern engineering transportation systems, in order to guarantee the safe operation of the bridge, many large bridge have installed with healthy monitoring system now, but how to get useful information from large number of monitoring data importantly determine the bridge damage status base on the information become one of the highlights in the filed of agro-scientific research. This paper puts forward using the Elman neural network to discriminate the key nodes of the bridge damage, and applies this method to Shi He bridge in practical engineering. This article research mainly content as follows:(1) Selection of damage factors and damage location, using the finite element model of Shi He bridge to simulate different degree and different state damage factor as the network inputs. This article proposed the Elman neural network as the damage identification model and the BP neural network as a reference.(2) In order to improve the effect of Elman neural network damage identification avoid the location results and low damage rate due to the randomness of the initial weights of neural network.(3) In terms of monitoring data processing, pretreatment original monitoring data before analysis This paper analysis time series of Shi He bridge monitoring strain and temperature data. While it was found that the temperature and strain have periodic change trend. The cluster analysis base on strain data indication that it include two kind of clustering.This article applied Elman neural network in bridged structure damage identification, due to the special structure characteristics of Elman neural network make it has good stability,. It is concluded that this method is feasible by simulation experiment and after ant colony algorithm optimize neural net work recognition rate was improved While this method combined with Shi He bridge monitoring data, results reflect that bridge structure without exception and the data which is collected back from bridge healthy monitoring system is normally.