A Fault Diagnosis Framework for Centrifugal Pumps by Scalogram-Based Imaging and Deep Learning

Centrifugal pumps are the most vital part of any process industry. A fault in centrifugal pump can affect imperative industrial processes. To ensure reliable operation pump, this paper proposes a novel automated health state diagnosis framework for that combines signal to time-frequency imaging technique and an Adaptive Deep Convolution Neural Network model (ADCNN). First, vibration signals corresponding different conditions acquired. Vibration obtained from carry great deal information generally, statistical features extracted retain meaningful information. However, these either insensitive weak incipient faults or unsuitable tracking severe faults, thus, decreasing classification accuracy. tackle problem, is applied signals. For purpose, Continuous Wavelet Transform (CWT) decompose over scales extract both time frequency domains. The CWT form two-dimensional images commonly referred as scalograms. scalograms then converted into grayscale (SGI). Over past few decades, CNN models have been established effective practice pattern recognition. Consequently, CWTSGIs finally provided inputs proposed ADCNN architecture achieve feature extraction faults. performance diagnostic (CWTSGI + ADCNN) validated with dataset collected testbed specifically designed diagnosis. experimental results suggest based on CWTSGI outperformed existing methods average improvement 4.7 – 15.6%.