Th is paper modifies parameter identificat ion of singular systems with the aid of transformation of singular system to a new Strong equivalent counterpart. Singular systems should be transformed to an equivalent model in the first step of identification process. In fact choosing an appropriate equivalent singular model is of crucial importance. Inconvenient equivalent model may lead to divergence, excessive computation time and imp recise estimat ion results. Indeed a more desirable estimat

more »

... result would be attained by reducing the number of init ial conditions. Tradit ional reduction methods used before for this purpose, but they resulted low accurate estimations because important dynamics of system have been omitted wrongly using those equivalencies. In this paper, a mo re accurate equivalency transformat ion of singular systems called Strong equivalency in co mbination with the Least Square identification algorith m is performed with the aim of revising the mentioned problems. This combination of the Strong equivalency together with the identification procedure is used for the first time. Thus this new configuration imp roves not only the estimation error convergence, but also the output tracking. Performance of the proposed method is illustrated in a practical singular electric network.