Load signatures improvement through the determination of a spectral distribution coefficient for load identification

In this paper a novel and simple methodology for developing distinct load signatures is proposed. The analysis relies on the exhaustive utilization of the information embedded in the harmonic behavior of a load, towards the formulation of an appropriate data form that could describe the behavior of a Low Voltage (LV) load in a unique and representative way. Based only on the current magnitude during one period of the steady state, a special coefficient is formulated under a simple procedure determining the spectral distribution of the current. A load identification algorithm is also developed and presented in order to examine the robustness and effectiveness of the resulting load signatures. The results are very promising since they indicate that uniqueness in load signatures is indeed added by the proposed technique and hence, the improvement of the effectiveness of the signatures could contribute in more efficient Nonintrusive Load Monitoring (NILM) algorithms.