Leak Detection in Slow Oscillation High-Pressure Fluid-Filled Circuits

This paper presents a fully automatic leak detection system for slow oscillation high-pressure fluid-filled (HPFF) circuits. It also presents the concept of detecting leaks in such circuits, utilizing only a limited number of sensors. It will describe algorithms built upon this concept, including two different approaches to the construction of the decision module, crucial for detection capabilities. The proposed solution utilizes adaptive digital filters for modeling the circuit behavior and a Bayesian network coupled with a software leak simulator in the decision stage. These tests were performed using data obtained from a real-life circuit, including real-leak data. This paper may enable practical application of high-precision low-cost leak detection solutions for a wide range of HPFF circuits.