Load Recognition in Home Energy Management Systems Based on Neighborhood Components Analysis and Regularized Extreme...

Area of Science:

• Energy Management
• Artificial Intelligence
• Machine Learning

Background:

• Home Energy Management Systems (HEMS) are crucial for optimizing residential energy consumption.
• Accurate load recognition, identifying active appliances, enhances HEMS robustness but requires further exploration.
• Existing methods face challenges in classification performance, inter-class separability, and model reliability.

Purpose of the Study:

• To improve load recognition techniques for Home Energy Management Systems (HEMS).
• To enhance classification performance and model reliability in identifying household appliances.
• To explore the potential of Neighborhood Component Analysis (NCA) and Regularized Extreme Learning Machine (RELM) for advanced energy management.

Main Methods:

• Utilized Neighborhood Component Analysis (NCA) for feature extraction, focusing on improving class separability.
• Employed Regularized Extreme Learning Machine (RELM) for the classification and identification of household appliances.
• Evaluated the proposed approach against state-of-the-art methods using key performance metrics.

Main Results:

• Achieved high accuracy of 97.24% and weighted F1-Score of 97.14% in appliance identification.
• Demonstrated enhanced reliability with a Kappa index of 0.9388, surpassing competing classifiers.
• The combined NCA and RELM approach significantly improved load recognition performance.

Conclusions:

• The integration of NCA and RELM presents a pioneering and effective approach for load recognition in HEMS.
• Machine Learning techniques, specifically NCA and RELM, show significant promise for advancing energy management in residential settings.
• This research contributes to more robust and reliable HEMS through improved appliance identification capabilities.