Research on Sound Signal State Diagnosis Technology of Chemical Columns Based on Support Vector Machine

DOI: https://doi.org/10.62517/jike.202604303

Author(s)

Yujie Wu1, Yike Duan2, Huarui Cai1, Wei Zhang1, Guangyan Wang1,*

Affiliation(s)

1School of Information Engineering, Tianjin University of Commerce, Tianjin, China 2School of Computer Science, Central China Normal University, Wuhan, Hubei, China *Corresponding Author

Abstract

The safety monitoring in the operation of chemical equipment is related to the national economic development and the safety of people's life and property. Existing online monitoring devices are larger in size, fewer characteristic parameters, and lower degree of intelligence. This paper takes Chemical columns equipment as the research object. Firstly, it is to collect the acoustic signals generated during the column operation process. Then the acoustic and vibration signals will serve as the input signals of a signal transmission and processing system, with the intelligent signal processing procedure of weak signal detection and enhancement, multi-variate feature extraction, multi-modal information fusion, and the online fault diagnosis and recognition by Support Vector Machine method. After 10 experimental runs, the diagnostic accuracy was measured to reach an average of 94%. The results verify the effectiveness of our proposed approach under limited data conditions.

Keywords

Chemical Process Safety; Chemical Columns; Intelligent Monitoring; Support Vector Machine; Acoustic Feature Fusion

References

[1] Zhang Z X, Wang G Y, Yang Z H. Acoustic Signal-Based Method for Recognizing Fluid Flow States in Distillation Columns. Industrial & Engineering Chemistry Research, 2022. DOI:10.1021/acs.iecr.2c02584. [2] Wang G Y, Yang Z H, Zhang Y, et al. A Preliminary Fault Detection Methodology forAbnormal Distillation Column Operations Using Acoustic Signals. Appl. Sci. 2022, 12, 12657. https://doi.org/10.3390/app122412657 [3] Zheng H, Zhang Z, Wang G, et al. State surveillance and fault diagnosis of distillation columns using residual network-based passive acoustic monitoring. Chinese Journal of Chemical Engineering, 2025, 77(1): 248-258. [4] Cortes C, Vapnik V. Support Vector Network. Machine Learning, 1995, 20(3):273-297. [5] Matthias Rychetsky, Stefan Ortmann，Manfred Glesner．Support vector approaches for engine knock detection// International Joint Conference on Neural Networks, Washington DC, 1999: 969-974. [6] Wang X F, Wang S H, Du H, et al. Fault diagnosis of chemical industry process based on FRS and SVM. Control & Decision, 2015, 30(002): 353-356. [7] Mokere R, Ghassan M, Barra I. Soil Spectroscopy Evolution: A Review of Homemade Sensors, Benchtop Systems, and Mobile Instruments Coupled with Machine Learning Algorithms in Soil Diagnosis for Precision Agriculture. Critical Reviews in Analytical Chemistry, 2010, 55(7): 21. [8] Tao Y, Yan J, Niu E,et al. An SVM-Based Anomaly Detection Method for Power System Security Analysis Using Particle Swarm Optimization and t-SNE for High-Dimensional Data Classification. Processes, 2025, 13(2). 549. [9] Qin W, Qiu Y, Feng Y .Improved Discontinuous SVM Based Fault-tolerant Control of AC/DC Converter for All-DC Offshore Wind Power System. CSEE Journal of Power and Energy Systems, 2025, 11(1):184-195.