D. K. Chaturvedi, et. al. International Journal of Engineering Research and Applications www.ijera.com ISSN: 2248-9622, Vol. 11, Issue 5, (Series-IV) May 2021, pp. 43-49 www.ijera.com DOI: 10.9790/9622-1105044349 43 | Page Development of Integrated Softcomputing Approach for Stator Resistance Estimation of Three Phase Induction Motor D. K. Chaturvedi*, Mayank Pratap Singh*, O.P. Malik** ABSTRACT In this paper, an integrated Quantum inspired GA (QGA) based generalized neural network (QGA-GNN) has been developed. The QGA-GNN is used for estimation of stator resistance of a 5hp Three phase Induction Motor (3Φ I.M.) under different healthy and unhealthy working conditi ons. The simulation model is used to collect the set of data for estimating stator winding resistance under healthy and faulty (i.e. 10%, 20%, 30% or 40% short circuited) conditions. The motor current and motor speed are considered as input and stator resistance as output of the proposed technique. The results obtained from QGA-GNN are compared with the ANN and GNN. QGA-GNN is giving good results under different working conditions. It is found that the training epochs required in ANN is about 50,000, in GNN - about 400 epochs and in QGA-GNN it is negligible. The superiority in terms of RMSE of QGA-GNN is 0.001 in comparison to ANN which is 0.012. Keywords — Stator Resistance Estimation, Three Phase Induction Motor, Artificial Neural Network, Genetic Approach, Quantum Genetic Approach, Soft Computing techniques. --------------------------------------------------------------------------------------------------------------------------------------- Date of Submission: 12-05-2021 Date of Acceptance: 25-05-2021 --------------------------------------------------------------------------------------------------------------------------------------- I. INTRODUCTION Three phase induction motor (3Φ I.M.) stator resistance non-linearly changes with operating conditions, like weather conditions, magnetic, electrical and mechanical couplings, etc. The precise estimation of stator resistance is necessary for good modeling and control of 3Φ I.M. Stator resistance of a 3Φ I.M. are normally obtained by performing various tests, such as, dc test, zero-load ac test, rated shaft load and stationary rotor test for balanced operating conditions [1]. In the case of an unbalanced 3Φ I.M., like damaged rotor cage or aluminium-bars, stator resistance cannot be exact [2-4]. Some investigators have used 3Φ I.M. current frequency spectra for finding damaged bars in 3Φ I.M. [5]. Firstly, these methods are off line method for parameter estimation and secondly, they are not fault tolerant or any learning mechanism. Therefore, soft computing approaches are used. Soft computing approaches [6], namely, ANN and its variants [7.8], Fuzzy system [9], neural fuzzy system [10], wavelet [11] and GA [12] have been used parameter identification and condition monitoring. ___________________ * Dept. of Electrical Engineering, Dayalbagh Educational Institute, Dayalbagh, Agra, U.P., India, Email: dkc.foe@gmail.com ** Dept. of Electrical and Computer Engineering, University of Calgary, AB, CANADA, email: maliko@ucalgary.ca ANN can handle large blocks of information at a time because of its parallel processing capability. Hence, it is an effective approach for 3Φ I.M. stator resistance evaluation. But, there is no specific ANN structure and configuration for a given problem. Also it is not known that which type of neuron will be good for given problem. Hence, many neuron structures have been developed such as summation neuron, product neuron or combinations of these structures. To overcome these problems GNN was developed by Chaturvedi et.al. [13-17]. But the training issues of GNN also remain same as of ANN i.e. sufficient and good data for training, stuck in local minima if backprop training algorithm is used. To overcome these problems and accurately estimate the 3Φ I.M. stator resistance QGA-GNN is proposed in this paper and the results are compared with ANN and GNN. This paper is divided into five sections. The first section deals with the introduction of the paper and second section introduces the soft computing and Quantum computing and development of QGA-GNN. The third section describes the methodology for estimating the stator resistance of 3Φ Induction Motor. The results and discussion are mentioned in section fourth. Finally, the paper is concluded in section fifth and the references are given. RESEARCH ARTICLE OPEN ACCESS