KNN Methods with Varied K, Distance and Training Data to Disaggregate NILM with Similar Load Characteristic

Non-Intrusive Load Monitoring (NILM) enables detection of appliances which are ON or OFF even the characteristics for each equipment installed in homes, industries, laboratories, etc. by disaggregating the total electrical consumption at the central Power panel. The K-NN method is one of the most simple and commonly used machine learning methods for classifying with good performance and competing with even complex methods. In this paper the K nearest neighbor (KNN) method is performed on NILM AMPds data which having distinctive similar load characteristic between different appliances, with 9 different distances, 7 types of total training data (10% -70%) and performed for k (1-25) for best result, then an accuracy performance comparison for disaggregation on 100% data and cross-validation (10%-80%) data also performance comparison of disaggregation on data which feature real power only, compared with data which feature having additional reactive power data, have done. From the test and research results it was found that by adding reactive power data, the disaggregation results on NILM data which having distinctive similar load characteristic between different appliances with KNN method were more than 20% accurate. It up to 95.06% accuracy on 70% training data, while for disaggregation on data that test data were completely different from the training data, disaggregation with 20% training data provides better performance in terms of accuracy as well as process speed.