An Orthogonal Matching Pursuit based signal compression and reconstruction approach for electromechanical admittance based structural health monitoring

Abstract

Abstract Signal compression and reconstruction were critical for damage detection in engineering structural health monitoring (SHM), on account of large amounts of sensor data collected and processed in signal acquisition system. A signal compression and recovery approach for damage detection using piezoelectric ceramic transducer (PZT) in structural monitoring system was proposed in this article. The basis of this approach was to first perform a linear projection of the transmitted data x in the monitoring system onto y by a random matrix and subsequently to feedback the data y to the receiving system. An algorithm of sparsity-adaptive Orthogonal Matching Pursuit (OMP) modified via optimal parameter analysis was explored to improve both the recovery effect and the compression ratio (CR) of compressed sensing (CS) in data processing stage. A statistical index was then introduced to identify the vector characteristics. The proposed method was sufficiently validated with the electromechanical admittance (EMA) data collected in an experiment for local damage detection on a simply-supported steel beam, and further applied to a long-time health monitoring of full-scaled shield tunnel segment structure. Qualitative and quantitative comparisons between the reconstructed and the original signals in structural damage detection under multiple conditions indicated that the proposed compression and recovery approach was of high accuracy and robustness to immune from the sensor conditions and temperature/environment impact, thus providing promising assistance to the impedance/admittance based SHM practice.