Flow-rate estimation in PVC and carbon-steel pipes using flow-induced vibrations and data-driven models

This study analyses the various non-intrusive methods of measuring flows within pipeline systems while considering the minimisation of pressure loss, contamination, and structural modifications. The research centres on the streams of flow-induced vibration (FIV) analysis, which has been underutilised in conjunction with data-driven modelling for analysis of standard industrial carbon-steel pipelines. An experimental study has been completed to determine flow rates within 1-inch PVC and carbon-steel pipes conveying water in a controlled laboratory environment. The experimental methodology involved the measurement of vibration responses (in the axial, lateral, and vertical directions) via a tri-axial accelerometer, which was followed by a measurement of flow-sensitive frequencies. In addition, both regression and machine learning

models were created to determine the relationship of a certain flow rate. The results showed a predominantly monotonic relationship of band-averaged vibration amplitude and flow rate with varying levels of sensitivity between. Furthermore, the application of data-driven models provided additional prediction performance and improved robustness in terms of accurately capturing the flow-induced vibrating models. The findings support the posit that combining flow-induced analysis with predictive modelling is a viable and accurate non-intrusive flow measurement method in industrial pipeline systems. For the purpose of training the model, the predictive capabilities provided by neural time-series modelling were employed.

Message: Hi FE, I need help with “Flow-rate estimation in PVC and carbon-steel pipes using flow-induced vibrations and data-driven models” in “Electrical Engineering”