Detecting Helical Gearbox Defects from Raw Vibration Signal Using Convolutional Neural Networks

Abstract

A study on the gearbox (speed reducer) defect detection models built from the raw vibration signal measured by a triaxial accelerometer and based on convolutional neural networks (CNNs) is presented. Gear faults such as localized pitting, localized wear on helical pinion tooth flanks, and lubricant low level are under observation for three rotating velocities of the actuator and three load levels at the reducer output. A deep learning approach, based on 1D-CNN or 2D-CNN, is employed to extract from the vibration image significant signal features that are used further to identify one of the four states (one normal and three defects) of the system, regardless of the selected load level or the speed. The best-performing 1D-CNN-based detection model, with a testing accuracy of 98.91%, was trained on the signals measured on the Y axis along the reducer input shaft direction. The vibration data acquired from the X and Z axes of the accelerometer proved to be less relevant in discriminating the states of the gearbox, the corresponding 1D-CNN-based models achieving 97.15% and 97% testing accuracy. The 2D-CNN-based model, built using the data from all three accelerometer axes, detects the state of the gearbox with an accuracy of 99.63%.

Helical Gearbox Defect Detection with Machine Learning Using Regular Mesh Components and Sidebands

“…The D4 state presents a localized wear on a pinion tooth flank. The D3 and D4 states are described in what follows, and in more detail in [31]. The actual speed reducer ratio ir = i in /i out is 10.421, and the relation between the input and output turning speeds (derived from the teeth numbers of the two gear pairs) is expressed in Equation ( 5).…”

“…A couple of tooth pairs are simultaneously engaged in contact at the line of action (which is tangent to both base circles of the gear pair). The deflection variation is reduced for helical gear (versus spur gear) observed in the current speed reducer, where the contact line length is larger in comparison to the spur gear [ 31 ]. More tooth pairs are simultaneously in contact in helical gears; therefore, we have a smaller deflection variation on the helical gear while the gear pair meshes.…”

“…The D4 state presents a localized wear on a pinion tooth flank. The D3 and D4 states are described in what follows, and in more detail in [ 31 ].…”

“…The main deep learning directions in machine health monitoring systems include Convolutional Neural Networks (1D or 2D CNNs), autoencoders (AEs and variants), Restricted Boltzmann Machines (RBMs and variants), and Recurrent Neural Networks (RNNs) [ 29 , 30 ]. In [ 31 ], CNN-based defect detection models use 1D and 2D grayscale images, representing the vibration image of the acquired raw data. In [ 32 ], an adaptive stacked CNN with a multi-sensor fusion algorithm fuses the diagnosis output based on vibration signals with the one based on sound signals.…”

Gearbox Condition Monitoring and Diagnosis of Unlabeled Vibration Signals Using a Supervised Learning Classifier

“…Among these, vibration signals are the most widely used because they contain a lot of information from inside the mechanical equipment. In order to monitor gearbox conditions and detect defects early, various technologies such as artificial intelligence and signal processing are being researched [4][5][6][7][8][9][10][11][12][13] It is crucial to maintain desirable performance in industrial processes where a variety of faults can occur. For most industries, FDD is an important control method because better processing performance is expected from improving the FDD capability.…”