Reinforced Concrete Bridge Structural Health Classification using Hybrid Principal Component Analysis and Artificial Neural Network Approach with Wireless Sensor Network

Structural health monitoring (SHM) and evaluation of bridges isinteresting and radically important with the exponentially increasing demand for safety and its functionality in relation with economic, industrial and civilbenefts. In bridge health evaluation systems, the modal parameters can only access the required accuracy after multiple and looping experiments due tovarying ambient parameters. The paper proposed an approach of bridge structural health evaluation and classification method based on the combined multivariate linear principal component analysis (PCA) and multilayer artificial neural network (ANN) which is used to compensate temperature on vibration data and classify bridge health respectively. Bridge health can be classifed as good, needs rehabilitation and critical. A wireless sensor network (WSN)composed of six autonomous motes is installed along the deck of the reinforced concrete laboratory bridge test platform to acquire vibration, and bridge andenvironment temperature data. The critical points of the bridge whereinmaximum defections occur are determined using moment diagram. Nondestructive testing (NDT) using controlled vibration testing (CVT) and platform physical alteration technique is implemented with formulated damage test cases to provide different excitations on the bridge. Peak picking (PP) algorithm was used to mitigate data congestion. The optimization technique of scaled conjugate gradient (SCG) is the algorithm used to train the three-layer feedforward ANN and sigmoidal function was used as the activation type for output neurons. There is 0.0038881 cross-entropy (CE) performance and 99.8% accuracy during network training. Satisfactory tested neural network CEperformance of 0.00636106 and 98.4% accuracy in classifying bridge health is obtained.