State Evaluation Method of Robot Lubricating Oil Based on Support Vector Regression

Area of Science:

• Industrial Internet of Things (IIoT)
• Intelligent Manufacturing
• Predictive Maintenance

Background:

• Industrial robot maintenance decisions significantly impact operational costs, including spare parts and labor.
• The Industrial Internet of Things (IIoT) enables the collection and storage of critical robot operational data.
• Effective equipment-state-prediction models are crucial for intelligent manufacturing applications.

Purpose of the Study:

• To develop an evaluation method for predicting the state of robot lubricating oil.
• To leverage Support Vector Regression (SVR) for accurate state prediction, minimizing issues with small sample volumes.
• To reduce maintenance costs through enhanced predictive capabilities.

Main Methods:

• Utilizing IIoT technology for collecting and storing industrial robot running data.
• Extracting key features from robot operational states.
• Applying a machine learning model, specifically Support Vector Regression (SVR), based on lubricating oil element analysis.

Main Results:

• Successful prediction of robot lubricating oil state using SVR.
• Demonstrated ability to avoid structural risks associated with maintenance decisions.
• Significant cost savings in spare parts consumption, exceeding two million CNY annually.

Conclusions:

• The proposed SVR-based method effectively predicts industrial robot lubricating oil status.
• IIoT integration and machine learning enhance predictive maintenance strategies in intelligent manufacturing.
• Optimized maintenance decisions lead to substantial cost reductions in spare parts and labor.