Turkish Journal of Electrical Engineering and Computer Sciences




Identifying and controlling (ON/OFF) electrical appliance(s) from a remote location is an essential part of energy management. This motivated us to design a system that can collect the aggregate load signature from a single point, obtain the features, and finally identify the ON state of electrical appliance(s). The proposed disaggregation technique can be divided into two modules: the first part proposes an electrical installation system to disaggregate the appliance at the circuit level, whereas the second part consists of feature selection, dimension reduction, and classification algorithms. Load signatures of electrical appliances were combined with white Gaussian noise to analyze how noise affects the classification results. Amplitudes of the major eight harmonics of load signatures were selected as a feature for the classification. Various classification algorithms were applied to data to check their feasibility. The comparative evaluation showed that among the considered classifiers, the multilayer perceptron-artificial neural network (MLP-ANN) classifier leads in classification accuracy with 99.18{\%}. If the system is combined with noise, the accuracy decreases to 93.10{\%}. This paper also shows that the proposed technique reduces the space complexity and decision time of the smart meter.


Smart meter, load signature, harmonic amplitude, artificial neural network, support vector machine, Bayes classifier

First Page


Last Page