Vibration-based monitoring of agro-industrial machinery using a k-Nearest Neighbors (kNN) classifier with a Harmony Search (HS) frequency selector algorithm

• DOI: 10.1016/j.compag.2023.108556
• Publication Date: 2024-01-09T12:25:17Z
• Authors: Francisco Javier ...; Víctor Martínez-M...; Ruben Ruiz-Gonzalez; Lidia Martínez-Ma...; Jaime Gomez-Gil
• Keywords: Harmony search; Smoothing; Kurtosis

Monitoring the status of rotating components is important in modern machinery. The goal this study to evaluate feasibility using a k-Nearest Neighbors (kNN) classifier combined with Harmony Search (HS) algorithm, detect operational within agricultural machines. Vibration data, source were acquired from four accelerometers located along chassis harvester. Five statuses three harvester studied: engine (low/maximum speed), thresher, and chopper (on/off balanced/unbalanced). methodology includes vibration signal acquisition, data preprocessing, smoothing, preselection frequencies, Brute Force (BF) frequency selection, classification kNN. input frequencies for chosen either BF search or HS. main results were: i) reduced training time between 92.2% 95.6%; ii) smoothing stage improved accuracy; iii) HS 82% 90% comparison BF, reaching accuracies nearly 100% five only 2 frequencies; iv) similar levels accuracy obtained when at different locations. suggested that it was feasible predict machines kNN combination preselection, algorithm. This achieved both terms computational burden, building upon previously proposed methods.