Abstract:

Improving the performance of electric machines is important for increasing energy efficiency and ensuring stable operation in broad industrial use. Applying linear regression predictors offers an effective tool for performance parameter analysis and loss of efficiency detection. Existing approaches do not precisely measure efficiency degradation due to failure to wholly account for certain impactful factors including load and temperature, leading to suboptimal performance optimization. To address such issues, in this research a methodology is introduced that uses Multivariate Linear Regression (MLR) to compute motor efficiency loss taking into consideration variations in load and temperature collectively. By jointly estimating their collective effect, the proposed method delivers more precise and dynamic predictions of performance compared to single-variable approaches. Based on this prediction, early interventions, maximized maintenance planning, and overall improved machine life and operating efficiency are realized. The findings indicate that the MLR-based predictor significantly enhances efficiency loss prediction accuracy, achieving better conformity with real performance trends, and eventually results in more intelligent, data-driven electrical machine management.