An improved technique for load identification in residential buildings

In this paper, a new approach about non-intrusive load identification technologies in household is developed. To identify the electrical equipment connected to the house, a load signature must be analyzed. These signatures are measurable parameters of the load, like power level, duty cycle and rate harmonic distortion, that allows the identification of the load connected to the network. There are three kinds of electrical loads, classified according to their power (one state of power, multiple states of power and infinite states of power) and most of the techniques in the literature cannot identify all of them. This paper presents an improvement to these techniques with an algorithm that can identify all types of electrical loads. As the purpose of proving the effectiveness of the technique, a comparison between a method that does not identify all types of load and the one proposed in this paper is made and validated.