Case Study—Spiking Neural Network Hardware System for Structural Health Monitoring

Abstract

This case study provides feasibility analysis of adapting Spiking Neural Networks (SNN) based Structural Health Monitoring (SHM) system to explore low-cost solution for inspection of structural health of damaged buildings which survived after natural disaster that is, earthquakes or similar activities. Various techniques are used to detect the structural health status of a building for performance benchmarking, including different feature extraction methods and classification techniques (e.g., SNN, K-means and artificial neural network etc.). The SNN is utilized to process the sensory data generated from full-scale seven-story reinforced concrete building to verify the classification performances. Results show that the proposed SNN hardware has high classification accuracy, reliability, longevity and low hardware area overhead.

Spiking Neural Network-based Structural Health Monitoring Hardware System

“…This sensory analogue information is digitized with the help of different Analogue to Digital Converters (ADC) supported by signal processing tools. Feature extraction and identification methods relies on empirical data to differentiate between active features formed by the structural changes and momentary noise signals generated by the sensors [9]. SHM requires complex feature extraction methods i.e., orthogonal decomposition method to extract dynamic sparse features caused by the vibrational motion of the structure [9].…”

“…Feature extraction and identification methods relies on empirical data to differentiate between active features formed by the structural changes and momentary noise signals generated by the sensors [9]. SHM requires complex feature extraction methods i.e., orthogonal decomposition method to extract dynamic sparse features caused by the vibrational motion of the structure [9]. Once important featured signals are extracted from the sensory signal, a classification algorithm is used to analyze and categorize damages [25].…”

“…The advancement in high processing power on-chip processors has brought life to the hardware implementation of SNNs. A recent study has shown SNNs as an efficient and powerful classification technique for SHM [9]. Contrary to conventional artificial neurons, spiking neurons are spike-event-driven that utilize energy to process information only at the arrival of spike events at the input to neurons.…”

“…Traditional method requires personal assessments by experienced inspectors along with portable sensory equipment. These costly manual assessments can be done periodically (quarterly, biannually, or annually) as it requires dedicated high-spec equipment that further ads in costs [9]. Modern SHM techniques equipped with complex software algorithms that processes real-time structural information to automatically diagnose structural changes and damages caused by the aging, erosion, and regular load [10].…”

Damage Identification Method of Box Girder Bridges Based on Distributed Long-Gauge Strain Influence Line under Moving Load

“…Bridge structural health monitoring (SHM) has attracted the extensive attention of many researchers [1][2][3][4][5] and has been successively applied to bridge structures as an important part of disaster reduction. As a core issue in SHM, structural damage identification technology has developed rapidly.…”