Accelerating performance evaluation of deep-bar induction machines from parameter identification

Abstract

Deep rotor bar induction machines have good starting characteristics as compared with cylindrical bar squirrel-cage configurations. High starting torque is achieved by the speed-dependent secondary resistance. The variation of secondary resistance and leakage inductance determines the machine starting and acceleration performance. Conventional measurement methods can only provide nominal parameters under blocked rotor and no-load test conditions. This makes details of the dynamic parameters and an evaluation of the acceleration performance from standstill ill-defined. High performance control of the machines requires a knowledge of real-time machine parameters as well. To determine the starting characteristics of deep bar induction machines, or to operate the machine at high performance, a knowledge of the dynamically-varying parameters is needed. This paper describes the application of a new algorithm identifying the variable parameters of deep rotor bar induction machines, namely secondary resistance and secondary leakage inductance, based on the easily measurable terminal conditions of the machine, namely voltage, current and rotational speed. The algorithm is simply an algebraic operation, which is always stable and easily realizable, and no convergence problems arise. Practical application of this method to an evaluation of the starting and accelerating performance of a 4-pole, 60 Hz GE (General Electric, Canada) 10000 hp deep bar induction motor demonstrates the utility of the new parameter identification method.