Adaptive super-twisting observer for fault reconstruction in electro-hydraulic systems

Area of Science:

• Control Systems Engineering
• Robotics
• Fault Diagnosis

Background:

• Classic sliding mode observers suffer from chattering, conservative gains, and strict conditions on fault/uncertainty distribution.
• Electro-hydraulic servo systems (EHSS) require robust fault reconstruction for reliable operation.
• Unknown bounded perturbations pose significant challenges to observer performance.

Purpose of the Study:

• To develop an adaptive-gain super-twisting sliding mode observer for fault reconstruction in EHSS.
• To overcome limitations of traditional sliding mode observers, including chattering and gain conservatism.
• To relax stringent conditions on fault and uncertainty distribution.

Main Methods:

• An adaptive-gain super-twisting sliding mode observer design.
• Utilizing the equivalent output-error-injection feature of sliding mode techniques.
• Implementing a fault reconstruction strategy based on the observer.

Main Results:

• The proposed observer effectively reconstructs faults in EHSS under bounded perturbations with unknown bounds.
• The adaptive gain law eliminates observer gain overestimation.
• Chattering effects are significantly attenuated.
• Experimental validation confirms precise fault reconstruction capabilities.

Conclusions:

• The adaptive super-twisting sliding mode observer offers a robust and effective solution for fault reconstruction in EHSS.
• This approach enhances system reliability by addressing limitations of conventional observers.
• The method demonstrates superior performance in handling uncertainties and reducing chattering.