Dynamic equivalent model of Distribution Network Cell using Prony analysis and Nonlinear least square optimization

Abstract

The paper presents initial results in development of dynamic equivalent model of Distributed Network Cell (DNC) comprising various types of loads and distributed energy resources. The method used for development of the model is based on Prony analysis and Nonlinear least square optimization. The dynamic equivalent model is developed in MATLAB based on simulated measurement data of DNC. The model is given in the form of an equivalent second order transfer function that can be used in dynamic stability studies. The model estimation procedure is evaluated on a case study and the initial simulation results show very good performance of the proposed dynamic equivalent model. Index Terms-Dynamic equivalent, Prony analysis, Nonlinear least square optimization. § · § ¸ © ¹ © Z · ¹ (5) The accuracy of the estimation procedure was checked by calculating the Root Mean Square Error (RMSE) values. The RMSE is a frequently used measure of the differences between values predicted by a model or an estimator, and the Samila Mat Zali received her B.Eng (1999) and M.Sc. (2002) degrees from National University of Malaysia (Malaysia), both in Electrical, Electronic and System Engineering,. Currently, she is working towards her Ph.D degree in

Black‐box dynamic equivalent model for microgrids using measurement data

“…Prony analysis provides an efficient way to fit a dynamic response with a sum of damped sinusoids [19], [20], [23]. However, for each individual disturbance in a MG configuration (different disturbance amplitude or pre-disturbance steady state condition of the MG) a new set of Prony terms must be extracted, resulting in a huge amount of data and considerable computational burden.…”

“…Furthermore, modelling restrictions due to the network complexity, generation mix and control strategies are overridden by the black-box approach, ensuring the flexibility and generalized form of the developed models. The black-box parameters are derived using system identification techniques such as sub-space methods [17], [18] and Prony analysis [19], [20]. In most cases simulation results are used for the model parameter identification and validation, whereas in some cases field measurements from conventional, extended transmission networks [19].…”

Measurement‐based analysis of the dynamic performance of microgrids using system identification techniques

“…For the analysis of the MGs dynamics, black-or grey-box modelling techniques are used in [23]- [27] to develop equivalent This paper is a postprint of a paper submitted to and accepted for publication in IET Generation, Transmission and Distribution, and is subject to Institution of Engineering and Technology Copyright. The copy of record is available at IET Digital Library, or http://dx.doi.org/10.1049/ietgtd.…”

“…The model parameters are directly extracted from measurements without prior knowledge of the exact MG features and structure. However, system identification methods are applied in MGs only on simulation data [23]- [27]. Therefore, the need to ensure the validity of the results from application of system identification methods on real measurement data from MGs is crucial.…”

“…Therefore, nonlinear least square optimization is additionally applied in discrete time form on the measured responses to identify with higher accuracy the unknown parameters of (1) [23], [29]. The proposed procedure is a two-step approach, where first the initial values of the set of parameters , , , ]…”