A Novel Dynamic Fluid Level Prediction Method in Oil Wells based on Acoustic Signals and Time-frequency Domain Analysis

The measurement of dynamic fluid level heights in oil wells is of great significance for the monitoring, controlling, optimizing, and safety management of the oilfield production process and is a key parameter to ensure the normal operation of oil wells and improve production efficiency. Despite extensive research on methods for dynamic liquid level measurement, existing approaches have certain drawbacks. One notable limitation is the reliance on specialized equipment, leading to significant deployment costs. The acoustic-based prediction of oil well dynamic fluid level heights is widely applied due to its non-contact, high precision, and real-time capability. This paper presents a prediction method for dynamic fluid level heights in oil wells, leveraging expert models and acoustic time-frequency analysis. Unlike past work, this paper employs acoustic signals for dynamic fluid level heights measurement. In this paper, the acoustic time-frequency domain features are extracted and fused, subsequently selecting an appropriate prediction model using expert models to predict dynamic fluid level heights in oil wells. Experimental results demonstrate the method high prediction accuracy and strong robustness.