1 Abstract— A new method for determination of the steady- state equivalent circuit parameters of wound rotor induction motors using experimental starting transient measurements is presented. The algorithm data are the stator currents and volt- ages and mechanical speed. The algorithm uses the least-square method and motor dynamic equations in the synchronous refer- ence frame. Moreover, an approximation of the rotor flux that improves the accuracy of the estimation method, as well as a detailed study of errors, is included. The estimation method is applied to starting transient measurement data of a 2kW wound rotor induction motor, and parameter accuracy is verified by comparing steady-state torque- and current-slip curves calcu- lated with the estimated and laboratory measured parameters. Index Terms—Equivalent circuits, wound rotor induction mo- tor, starting transients. ARAMETER determination of induction motors is an important topic because these machines are one of the most common loads in the grid. Estimation accuracy is cru- cial because of the importance of parameters in power system dynamic behavior during faults and other perturbations. An- other important area is high performance ac drives, where parameter estimation is key in controller tuning. Several ap- proaches to induction motor parameter estimation have been presented. IEEE Std-112 [1] describes the most usual method, which is based on short-circuit and no-load tests. Other meth- ods are based on steady-state [2], frequency [3] or transient tests [4-10]. More recent and sophisticated procedures use a transient test with an extended Kalman filter [4][5]. Another possibility is to formulate a least-square minimization prob- lem [6-11] and a non-linear least-square method [12][13]. Reference [6] gives a thorough description of the least-square method and dynamic equations in the rotor reference frame and [7] uses the stator reference frame. Other examples use the synchronous reference frame [8] and [12]. This transfor- mation has the advantage that the variation of variables is smoother and their value is kept almost constant in steady state. The main issue in this paper is to develop an estimation method. One major drawback of many parameter estimation procedures is the need for large currents in the locked-rotor This research work was supported by the Spanish Ministerio de Ciencia e In- novación under Project ENE2009-10274. H. Kojooyan, Ll. Monjo, F. Córcoles and J. Pedra are with the Department of Electrical Engineering, ETSEIB-UPC, Av. Diagonal 647, 08028 Barcelona, Spain (e-mails: kojooyan@iiau.ac.ir, lluis.monjo@upc.edu, corco- les@ee.upc.edu, pedra@ee.upc.edu). test, especially in big machines. Moreover, in some industrial applications the rotor may actually not be locked. For this reason, this paper measures the direct on-line starting transi- ent, which is considered a part of motor operation. The estimation procedure is adequate for wound rotor in- duction motors and single-cage motors (modeled with the steady-state equivalent circuit in Fig. 1), where the mechani- cal transient is decoupled from the electrical transient, i.e., the time constants of the electrical and mechanical transients have different orders of magnitude. The method in this paper is based on:  Transformed equations of the induction machine in the synchronous reference frame.  Measurements of stator voltages and currents, and mechanical speed during the on-line starting transi- ent.  A smoothing process to calculate the derivatives of the above variables.  A new approximation of the rotor flux.  Validation by comparison of measured and calculat- ed steady-state torque and current-slip curves. The synchronous reference frame is useful for data prepro- cessing because of the slow variation of variables, as said be- fore. The smoothing process makes measurements free of noise, and allows calculating derivatives numerically. The key of the paper is the presentation of an approximation of the rotor flux which is usually neglected in the literature de- spite its contribution to the accuracy of estimated parameters. Finally, whereas in the literature the effectiveness of the es- timation procedures is usually determined by comparing sim- ulation parameters with estimated ones [2][3][6][8][9], stator current measurements with estimated ones [5][12], or both [7][13][14], this paper validates the accuracy of estimated parameters by comparing the steady-state torque- and current- slip curves calculated from the parameters with those meas- ured in the laboratory. Hengameh Kojooyan-Jafari, Lluís Monjo, Student Member, IEEE, Felipe Córcoles, Joaquín Pedra, Member, IEEE Parameter estimation of wound rotor induction motors from transient measurements P V s I s I r R s R r /s X sd X rd X m Fig. 1. Steady-state equivalent circuit for the single-cage model of the three- phase induction motor.