Abstract
The traditional nonintrusive load identification (NILD) algorithms result in high computational costs based on classification models. And they cannot accurately identify loads with similar voltage-current trajectories that are frequently encountered in practical applications. Aiming at these deficiencies, an NILD algorithm is proposed based on combined weighting-the technique for order preference by similarity to ideal solution (TOPSIS) of feature fusion. The feature is fused by one image feature and eight numerical features with the combination weighting method. The weights are calculated by combining the principal component analysis, entropy, and the criteria importance through intercriteria correlation weighting methods to improve the utilization of features. The similarity between these nine features of a load and the corresponding features of other loads is calculated by the TOPSIS algorithm. Similarity analysis is used to determine whether or not a load is a "known load," and to obtain an identification result. If it is an "unknown load", the result can be obtained by dynamically updating the database and identifying it again. The results obtained indicate that this proposed algorithm can significantly improve the accuracy of load identification while reducing the computational costs.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call