Leakage Fault Detection Method for Axial-Piston Variable Displacement Pumps

To address the lack of health monitoring for gas turbine engine accessory components, Sentient developed a promising fault estimation algorithm for axial-piston variable displacement pumps. The key to this success was the development of a physics-based nonlinear model of the system that includes three common types of fluid leakage faults. The modeling effort and simulation results provided the thorough knowledge needed to extract features from the system responses to detect the occurrence of faults. The nominal pump and the pump with a single piston replaced by three different machined pistons were tested to emulate varying stages of piston wear. The experiments demonstrated that the model was able to capture the dynamic behavior of the actual system. By considering several time and frequency domain metrics, a fault estimation method was developed that correctly identified the presence and severity of leakage faults, while using a fraction of the memory and computational resources of other model-based methods.