Experiences about calculating ZIP and exponential load model parameters

Abstract To reduce greenhouse emission and achieve sustainability in the electric power industry, improved efficiency and energy conservation are key to finding a sound solution. Energy conservation through voltage reduction (CVR) is one such program. To implement effective CVR programs, load models are of utmost importance in determining proper feeders and methods to implement the CVR programs. Exponential and polynomial load models stand out as the most widely adopted models. The polynomial load model, often referred to as the ZIP (constant impedance, current and power) model, is especially notable for its ability to represent the relationship between the applied voltage and power consumption. This paper sheds light on the limitations of commonly used ZIP load model estimation methods and highlights the significance of using ap-propriate measurement samples and understanding the true nature of the load. Two distinct ZIP load models, i.e., the regular ZIP (RegZIP) and the Non-traditional ZIP (NTZIP) models, have been used for ZIP parameter estimation. The RegZIP load model consumes equal or more power when voltage increases, while the NTZIP model consumes equal or less power when voltage increases. It is emphasized that the RegZIP is an existing load model and the NTZIP is an artificial load model this research created for comparison purposes. The intention of this paper is not to propose any new load model, but to investigate the implications of choosing the wrong model and the impacts of measurement errors of voltage and power data. The research has shown that the two different models can fit the same measurement data equally well for some measured data. The results have demonstrated that simply using certain methods such as the optimization method for estimating load model parameters using unfiltered real time measurements may yield misleading and unreliable results. Wrong load models will inevitably lead to wrong CVR assessment. Identifying the correct measurement samples, obtained during natural or staged events causing substantial voltage variations, is critically important in calculating load parameters and ensure the soundness of the obtained results. The accurate load parameters will be important in the assessment of CVR program effectiveness and will ensure that power system analysis applications that utilize the load models will provide reliable results.