Hybrid method of DFT and constructive backpropagation for energy monitoring

This study present new technique to identify similar signal of appliance such as the same brand and the specification of appliance. The proposed method was developed in the study was converting signal RMS current of appliance from time domain to frequency domain. To distinguish similar signals fourier transformation was conducted to result low frequency which represent of appliance. And then Constructive Bcak Propagation-Neural Network (CBP-NN) was employed for decision making of identification. The CBP-NN was designed using five input and two output. The experiment result shown that the method working properly with very small, i.e., error 0. 0659% for case 1 and 0. 2797% for case 2.