3D FEA Based Squirrel Cage Rotor Model for Design Tradeoffs and Performance Analysis Md Ashfanoor Kabir * , Rajib Mikail , Steven Englebretson and Iqbal Husain Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, NC, USA ABB US Corporate Research Center, Raleigh, NC, USA Email: * mkabir@ncsu.edu AbstractAn accurate rotor resistance estimation model of squirrel cage induction motors (SCIMs) is developed in 3D FEA. 2D transient analysis was utilized for excitations in the 3D model to improve its accuracy over previous 2D and analytical methods. Rated and starting performance from the FEA model match with the nominal and locked- rotor performance of a 3-phase, 460 V, 1 hp test machine. A modified ring model has been proposed and machine torque-slip characteristics and nominal performance have been analyzed. The effect of slot opening and 4 classes of SCIM bar geometry have been investigated to analyze their relative performance. Finally, four different ring and bar combinations are suggested, with the modified rotor structure presenting gain in starting and rated performance compared to the test machine. Results present the design tradeoffs and performance analysis, first for a 1 hp SCIM and then extended for a higher power (10 hp) machine. Keywordsinduction motor; 3D FEA; starting performance; loss model; end-ring resistance; IM design classes; performance analysis; torque-slip characteristics, locked-rotor I. INTRODUCTION The major portion of the world’s electrical energy is being generated or utilized by electric machines, and induction machines (IM) are the largest share-holder in this field for electrical loads. The high cost of permanent magnets (PMs) and robustness and simplicity of induction machines have expanded the research focus on this field [1], [2]. To improve SCIM performance, use of 3D finite element analysis (FEA) can aid development of accurate models [3]. Despite numerous investigations on rotor bar design of SCIM [4]-[8], few studies can be found on accurate end-ring models [9]-[11]. In this research, an accurate rotor resistance model which incorporates the end-ring resistance, ௥௜௡ is developed in 3D FEA. One of the advantages of the proposed model over other reported methods is that the 3D cage model incorporates detailed cage geometry variation along axial and radial directions. The proposed model also imposes precise current density distribution from 2D transient field analysis by segmenting the bar-ring contact surface to improve accuracy in resistance estimation. The model is able to account for non-uniform current in rotor end rings as well as predict the effect of variation of ring geometry parameters on the end ring and total rotor resistance. Based on the developed model in FEA, sensitivity of ௥௜௡ variation with ring geometry is studied and the value of ௥௜௡ is estimated. Next, the effect of bar-slot openings and different bar geometries on torque-slip characteristics and rated and starting performances are investigated. Finally, four different combinations of bar and end-ring design are proposed, and their design trade-offs are discussed. Detailed analysis on a 1 hp test machine is presented, and results are extended for a 10 hp machine for generalization. II. DEVELOPING THE RESISTANCE MODEL An accurate squirrel cage (SC) model is necessary to properly estimate the end-ring resistance. At first, the joule loss and current in the ring were determined using analytical models. Three different analytical models [10], [12], [13] were utilized. The first one is the ideal loss estimation model assuming a sheet conductor of uniform thickness and sinusoidal current distribution replacing the bars. Losses in both end-rings and the peak current around each pole of the machine can be expressed as in (1) [13]. ௥௜௡−௟௢௦௦ = ݌ ߩ ܬ ߨ ሺͳሻ ܫ ௣௘௔௞−௥௜௡ =ℎ ܬ ݋ ݏ ߨ  =ℎ ܬ ߨ ሺʹሻ Here, ܬ is the peak linear current density of the uniform sinusoidal sheet approximating the rotor bars; is distance along periphery from pole mid; is the thickness or height of the ring; is pole pitch; ݌ is number of poles; is the ring cross sectional area; and ߩ is the resistivity of the ring material. The second method, based on Alger [12], estimates the ring resistance by considering the effective radius that carries the current at the end-ring outer diameter. Assuming uniform current distribution through the ring cross-section, the expression for joule loss in both rings for this case is given in (3), where is the ratio between ring inner and outer diameter. ௥௜௡−௟௢௦௦ = ߩ ߨ ʹ ܫ ௣௘௔௞−௥௜௡ × ͳ ͳ− ሺ͵ሻ 978-1-4799-6735-3/15/$31.00 ©2015 IEEE 2696