12.1 Introduction
-
Machine tool status identification using an Electric-Load-Intelligence system.
-
Disaggregation of the energy consumption of machine tools up to the component level.
-
Comparison of the accuracies of various classifiers used in the literature.
12.2 Methodology
12.2.1 Electric-Load Intelligence Concept
12.2.2 Load Signature Acquisition and Preprocessing
12.2.3 Feature Extraction
12.2.4 Load Disaggregation
12.2.5 Conditional Inference for Detection of Material Removal
12.3 Experimental Validation of the Proposed E-LI System
12.3.1 Experimental Set-Up
12.3.2 Training Phase
12.3.3 Prediction Phase
12.4 Accuracy Analysis of the Classifiers
Classifier | Accuracy (%) | Accuracya (%) |
---|---|---|
Quadratic SVM | 92.9 | 85.9 |
Cubic SVM | 90.6 | 81.2 |
Linear SVM | 91.8 | 82.4 |
Fine Gaussian SVM | 87.1 | 82.4 |
Medium Gaussian SVM | 90.6 | 83.5 |
Coarse Gaussian SVM | 74.1 | 8.2 |
Linear discriminant | 84.7 | 83.5 |
Quadratic discriminant | 91.8 | 85.9 |
Fine k-nn | 88.2 | 78.8 |
Medium k-nn | 55.3 | 64.7 |
Weighted k-nn | 87.1 | 81.2 |
Subspace k-nn | 90.6 | 78.8 |
Subspace discriminant | 83.5 | 83.5 |
Bagged trees | 89.4 | 78.8 |
Medium tree | 64.7 | 65.9 |