A self-sensing framework for weak fault detection of planetary gearbox

Planetary gearboxes serve as fundamental components within modern electro-mechanical assemblies. Prior research has shown that monitoring these systems ensures operational reliability and prevents catastrophic mechanical failure. Traditional diagnostic techniques typically necessitate the installation of external vibration or acoustic sensors to monitor structural health. These supplementary sensing devices introduce significant financial burdens and complicate the physical architecture of the machinery. High costs associated with sensor procurement and maintenance often limit the widespread adoption of advanced diagnostic protocols in industrial settings. Engineers frequently encounter difficulties when attempting to integrate multiple external transducers into compact or high-speed mechanical environments. This absence of evidence motivated the development of alternative monitoring strategies that utilize internal system data rather than external hardware.

This research develops a self-sensing framework designed to identify subtle mechanical impulses within planetary gearboxes without external instrumentation. The investigation focuses on extracting diagnostic information directly from the existing motor servo control systems. Researchers aimed to construct a comprehensive hole control model representing the motor-driven planetary gearbox assembly. The study evaluates the feasibility of using intrinsic electrical signals to replace traditional physical transducers for fault identification. Engineers sought to create an adaptive signal processing method capable of isolating weak fault signatures from complex background noise. The project intends to provide a cost-effective solution for real-time health monitoring in sophisticated electro-mechanical equipment. This objective includes validating the approach across multiple gearbox components to ensure broad diagnostic applicability.

The investigative process began by capturing preliminary electrical data from the integrated servo control systems. A mathematical hole control model was established to simulate the interaction between the driving motor and the planetary gearbox. The team implemented a multi-signal assisting Adaptive Time Synchronous Averaging (ATSA) technique to process the measured servo control signals. This analytical framework incorporates a Periodic Enhanced Gini (PEG) index to facilitate the adaptive selection of optimal processing parameters. Experimental validation involved testing the framework against weak faults occurring in three distinct components of the gearbox assembly. Each component was analyzed separately to determine the sensitivity of the self-sensing approach to localized mechanical degradations. Data acquisition focused on identifying specific impulse patterns within the current and voltage fluctuations of the servo drive.

The proposed self-sensing framework successfully identified weak fault impulses across all three tested planetary gearbox components. Analysis of the motor servo control signals provided sufficient resolution to detect mechanical anomalies without supplementary hardware. The multi-signal assisting Adaptive Time Synchronous Averaging (ATSA) method effectively isolated fault-related transients from the baseline operational noise. Utilizing the Periodic Enhanced Gini (PEG) index allowed for the precise optimization of parameters required for signal enhancement. Experimental data confirmed that the hole control model accurately reflects the electromechanical coupling necessary for fault propagation. The framework demonstrated consistent performance in detecting subtle impulses that are typically obscured in raw electrical data. Results indicated that the adaptive averaging technique significantly improves the signal-to-noise ratio of internal diagnostic data.

Integrating diagnostic capabilities into existing servo control systems offers a viable alternative to sensor-heavy monitoring configurations. This self-sensing approach reduces the economic barriers to implementing comprehensive fault detection in industrial planetary gearboxes. The methodology provides a scalable scheme for monitoring weak mechanical faults in diverse electro-mechanic equipment. Future applications may leverage these intrinsic signals to enhance the autonomous health management of complex robotic and manufacturing systems. The study establishes a foundation for developing sensorless diagnostic tools that maintain high sensitivity to early-stage mechanical wear. These findings suggest that internal motor data contains rich diagnostic information previously overlooked in traditional maintenance paradigms. Implementation of this framework could lead to more resilient and cost-efficient industrial automation technologies.