A Multilevel Isolation Forrest and Convolutional Neural Network Algorithm for Impact Characterization on Composite Structures

Abstract

In this paper, a Deep Learning approach is proposed to classify impact data based on the type of impact (Hard or Soft Impacts), via obtaining voltage signals from Piezo-Electric sensors, mounted on a composite panel. The data is processed further to be classified based on their energy, location and material. Minimalistic and Automated feature extraction and selection is achieved via a deep learning algorithm. Convolutional Neural Networks (CNN) are employed to extract and select important features from the voltage data. Once features are selected the impacts, are classified based on either, Hard Impacts (simulated from steel impactors in a lab setting), Soft Impacts (simulated from silicon impactors in a lab setting) and their corresponding location and energy levels. Furthermore, in order to use the right data for training they are obtained from the signals as anomalies via Isolation Forests (IF) to speed up the process. Using this approach Hard and Soft Impacts, their corresponding locations and respective energies are identified with high accuracy.

Feature Selection Algorithm Based on CFS Algorithm Emphasizing Data Discrimination

“…The algorithm principle of the isolated forest is: continuously cutting the data set until every data becomes an isolated point. In both low and high dimensions, abnormal points are more likely to be isolated, while normal data needs to be cut more times to become isolated due to its dense and close distribution [22] . By comparing the path length from the root node when each point is isolated with the standard value, we can judge which points are outliers, namely outliers and fault points.…”

Design and Analysis of a Novel Piezoceramic Stack-based Smart Aggregate

“…In SHM of civil infrastructure, notable research based on advanced algorithms, such as damage index-based damage assessment [8,9], image-based damage locating [10] and artificial intelligence algorithms [11], have been reported to provide solutions to real-time monitoring and early warning of civil infrastructure [12,13] by using various sensors, such as strain gauges [14,15], piezoelectric transduces [16,17], optical fiber sensors [18,19], etc. With advantages of low cost, wide bandwidth [20] and quick time response [21,22], apart from in SHM of civil infrastructure, piezoelectric transducers have also been widely used in other fields, including aerospace [23][24][25][26], transportation [27,28] and energy [29,30], and they have displayed good applicability in both passive and active situations [31,32], such as vibration sensing [33], acoustic emission [34], active sensing and electromechanical impedance-based damage detection [35][36][37].…”