The Unknown Abnormal Condition Monitoring Method for Pumped-Storage Hydroelectricity

Area of Science:

• Energy storage systems
• Hydroelectric power generation
• Condition monitoring and fault detection

Background:

• Pumped-storage hydroelectricity (PSH) facilities store energy as gravitational potential energy.
• Condition monitoring of PSH, particularly its submerged and rotating hydropower turbines, is complex.
• Existing methods often require expert intervention for detecting new abnormal conditions.

Purpose of the Study:

• To develop an automated method for detecting novel abnormal conditions in PSH.
• To enable automatic updating and optimization of existing classification models for new faults.
• To reduce reliance on expert knowledge in PSH condition monitoring.

Main Methods:

• Implementation of a novel automated detection system for PSH.
• Automatic updating and optimization of abnormal condition classification models.
• Evaluation using sensor data from an on-site PSH facility.

Main Results:

• The proposed method achieved 85.89% accuracy in detecting new abnormal PSH conditions.
• Detection of new conditions required fewer than three data points.
• Average classification accuracy for identified conditions reached 99.73%.

Conclusions:

• The developed automated method effectively detects and classifies new abnormal conditions in PSH.
• The system demonstrates high accuracy and efficiency, even with limited data.
• This approach offers a significant advancement in PSH condition monitoring and maintenance.