Performance evaluation of the ZIP model-phaselet frame approach for identifying appliances in residential loads

This paper presents the analysis and development of a new approach to monitor and update the ON-OFF status of appliances in residential loads (RSLs). The proposed approach is structured to employ power meter readings P to determine values for the magnitude |S| and phase θ of the apparent power. The value of |S|, associated with a value of P, is determined using Newton iterations, where a value of θ is calculated using 6 phaselet frames during each iteration. Once the iterations converge, the values of P and θ are used to construct the ZIP model (polynomial model) for the RSL, from which P is provided. The constructed ZIP model provides values for the constants Kpf and Kqf that relate the change in frequency to the active and reactive power demands of the modeled load. The obtained values of Kpf and Kqf are compared to standardized values that are defined for each appliance in a RSL. The ZIP model-phaselet frame approach is implemented as an algorithm for monitoring the ON-OFF status of appliances in RSLs. The algorithm for the proposed approach is developed without a need to collect data for training. Test results show simple implementation, good accuracy, and insensitivity to variations in energy demands.