experimental investigation to utilize low-cost sensors of early cavitation prediction in axial pump based on acoustic and vibration approaches

Cavitation, defined by the formation and collapse of vapor bubbles in a liquid, presents a significant challenge to the reliable operation of axial pumps. timely detection of cavitation is essential for preventing damage and optimizing pump performance. this research aims to evaluate detection techniques for cavitation in axial pumps through the analysis of vibration and acoustic signals. we utilized cost-effective sensors and data acquisition systems, including embedded accelerometer sensors and smartphone-based microphones, to capture these signals. our study involves a detailed analysis of vibration and acoustic data collected under various operating conditions, with a particular focus on the pump’s optimal efficiency point. by employing three directional axes for vibration data acquisition, we achieved a comprehensive examination of the cavitation phenomenon. signal processing techniques, such as feature extraction in the frequency domain, were used to identify distinct operating conditions as cavitation developed. additionally, convolutional features were applied to assess the classification accuracy when datasets were converted into spectrograms. this research includes a thorough comparison of classification algorithms and different directional axes to provide insights into the effectiveness of the detection methods. the findings demonstrate the feasibility of detecting cavitation in real operating conditions using vibration signals, while highlighting challenges associated with using low-cost commercial acoustic data for cavitation detection. this introduction sets the stage for an in-depth exploration of the methodology, results, and implications of our study on early cavitation detection in axial irrigation pumps. the analysis of acoustic and vibration signals yielded similar results in detecting cavitation. key indicators, such as peak-to-peak, rms, and variation values, were effective metrics for cavitation detection. frequency-based analysis in both the broadband range (2 khz – 10 khz) and the low-frequency range (0 – 1 khz) revealed clear trends related to cavitation presence. time-domain analysis of vibration signals proved effective for detecting and diagnosing cavitation in axial pumps. using mean and peak values for vibration and acoustic amplitude analysis in the frequency domain provided additional insights for predicting cavitation.