Early detection and classification of internal leakage in hydraulic actuator using SVM

What is it about?

Internal leakage in hydraulic actuators, a common issue in hydraulic machines, adversely affects dynamic performance and energy efficiency. Detecting this leakage early is challenging as it becomes apparent only when extreme and the actuator stops responding to commands. This paper proposes a practical method for early detection of internal leakage in the boom actuator of mobile hydraulic machines. By analyzing work-cycle data with minimal hardware, a Support Vector Machine (SVM) classifier is trained and validated using pressure and boom angle displacement signals. Employing an event-based feature extraction method and binary Particle Swarm Optimization, the classifier achieves over 95% accuracy for timely preventive maintenance.

Featured Image

Photo by Mohamed Nohassi on Unsplash

Photo by Mohamed Nohassi on Unsplash

Why is it important?

This study introduces a novel method for detecting and classifying cross-port leakage in the boom actuator of hydraulic excavators. The approach can be applied to other load handling machines as well. Early detection of internal leakage prevents further damage to system components. Actuator chamber pressure and boom angle displacement signals are utilized, and statistical and system-specific features are extracted using an event-based method. Optimal feature selection is achieved through discrete particle swarm optimization. An SVM classifier with a cubic kernel function achieves a high classification accuracy of 97.5%, making it reliable for monitoring actuator internal leakage.

Perspectives