Use of Pareto optimisation for tuning power system stabilizers

Abstract

Abstract. The paper presents a method for determining sets of Pareto optimal solutions (compromise sets) -parameter values of PSS3B system stabilizers working in a multi-machine power system -when optimising different multidimensional criteria. These criteria are determined for concrete disturbances when taking into account transient waveforms of the instantaneous power, angular speed and terminal voltage of generators in one, chosen generating unit or in all units of the system analysed. The application of multi-criteria methods allows taking into account the optimisation process of power system stabilizer (PSS) parameters, many sometimes contradictory requirements (criteria) without losing ability to reach the optimal solution. A choice of the compromise solution can be made by assuming the values of the weighting coefficients associated with particular components of the vector criterion and determining the equivalent, global criterion. A change of the values of those weighting coefficients in the equivalent criterion does not require, in the case of the Pareto optimization, carrying out repeated calculations.

Assessment of Polish Power System angular stability based on analysis of different disturbance waveforms

“…The objective function (7) is minimised by a hybrid algorithm [10][11][12] which is a serial combination of a genetic algorithm [12][13][14] with a gradient algorithm [12,13]. The eigenvalues with small participation factor modules in particular waveform are neglected in calculations based on this waveform.…”

“…1. The following models of the PPS generating units components were assumed for the calculations: a synchronous generator GENROU [12,15], a static [12] or electromachine [11,12,15] excitation system working in the PPS, a steam turbine IEEEG1 [11,12,15] or water turbine HYGOV [12,15] and, optionally, a power system stabilizer PSS3B [10,12,15]. For the equivalent generating units there was used the simplified model of a synchronous generator GENCLS [15].…”

Parameter polyoptimization of PSS2A power system stabilizers operating in a multi-machine power system including the uncertainty of model parameters

“…The calculations of the second stage were the starting point for the third stage. The final selection of the gains K S1 was made with use of the multi-criteria analysis [3,7,[11][12][13]. There was optimized the vector quality factor containing the same components as the additive factor minimized in the second stage.…”

“…Saturation of the magnetic circuit is taken into account in an exponential way in the model. It was assumed that the excitation systems were represented by the nonlinear model of the Polish, static excitation system [3,13]. Moreover, it was assumed that the turbines were represented by the IEEEG1 model of the steam turbine with an interstage superheater [10].…”

Material model and revealing the truth

“…High values of correlation coefficient and determination coefficient indicate the level of fitting between observed values and the values expected under the model, however, the important issue is the selection of the type of model. The issue of modeling and optimization is raised in various engineering fields of research, including civil engineering [10] but also for instance in electromagnetics [11], electromechanics [12], robotics [13], etc., where the approaches to the modeling and their objects are very different, but all of the models or optimized solutions had been elaborated to become a useful tool for prediction the performance, optimization and general design [11]. However, regardless of the field of research, the selection of the model should be done not only on the basis of the knowledge of the statistical design principles but also on the knowledge of the investigated phenomenon.…”