Decomposed Algorithms for Speed and Parameter Estimation in Induction Machines

Abstract This paper presents different decomposed algorithms for speed and parameter estimation in induction machines, and compares their performance using experimental data. It is shown that decompositions matched to the parameter coupling structure lead to the fastest convergence. Also, the conditioning of the problem is studied using componentwise condition numbers and it is shown that a significant improvement in convergence speed and a reduction of problem sensitivity are achieved by fixing the stator resistance to a prior estimate.