Modeling Electrical Motor Dynamics using Encoder-Decoder with Recurrent Skip Connection
Abstract
Electrical motors are the most important source of mechanical energy in the industrial world. Their modeling traditionally relies on a physics-based approach, which aims at taking their complex internal dynamics into account. In this paper, we explore the feasibility of modeling the dynamics of an electrical motor by following a data-driven approach, which uses only its inputs and outputs and does not make any assumption on its internal behaviour. We propose a novel encoder-decoder architecture which benefits from recurrent skip connections. We also propose a novel loss function that takes into account the complexity of electrical motor quantities and helps in avoiding model bias. We show that the proposed architecture can achieve a good learning performance on our high-frequency high-variance datasets. Two datasets are considered: the first one is generated using a simulator based on the physics of an induction motor and the second one is recorded from an industrial electrical motor. We benchmark our solution using variants of traditional neural networks like feedforward, convolutional, and recurrent networks. We evaluate various design choices of our architecture and compare it to the baselines. We show the domain adaptation capability of our model to learn dynamics just from simulated data by testing it on the raw sensor data. We finally show the effect of signal complexity on the proposed method ability to model temporal dynamics.

Related Publications
- Sagar Verma, Nicolas Henwood, Marc Castella, Francois Malrait, and Jean-Christophe Pesquet, Modeling Electrical Motor Dynamics using Encoder-Decoder with Recurrent Skip Connection, Association for the Advancement of Artificial Intelligence (AAAI 2020), New York, NY, USA, 2020. [PDF, Poster]
- Sagar Verma, Nicolas Henwood, Marc Castella, Al Kassem Jebai, and Jean-Christophe Pesquet, Neural Networks based Speed-Torque Estimators for Induction Motors and Performance Metrics, Annual Conference of the IEEE Industrial Electronics Society (IECON 2020). [PDF, PPT, Video]
- Sagar Verma, A Survey on Machine Learning Applied to Dynamic Physical Systems, ArXiV Preprint. [PDF]
Bibtex
If you use this work or dataset, please cite :
@inproceedings{verma2020motorbench, title={Neural Networks based Speed-Torque Estimators for Induction Motors and Performance Metrics}, author={Verma, Sagar and Henwood, Nicolas and Castella, Marc and Jebai, Al Kassem and Pesquet, Jean-Christophe}, booktitle={Annual Conference of the IEEE Industrial Electronics Society (IECON 2020)}, pages={6}, year={2020} } @inproceedings{verma2020motor, title={Modeling Electrical Motor Dynamics using Encoder-Decoder with Recurrent Skip Connection}, author={Verma, Sagar and Henwood, Nicolas and Castella, Marc and Malrait, Francois and Pesquet, Jean-Christophe}, booktitle={Association for the Advancement of Artificial Intelligence (AAAI 2020), New York, NY, USA, 2020}, pages={8}, year={2020} } @inproceedings{Verma2018ASO, title={A Survey on Machine Learning Applied to Dynamic Physical Systems}, author={S. Verma}, archivePrefix = {arXiv}, eprint = {2009.09719}, year={2018} }
Dataset
The recorded time-series data from electrical motor can be downloaded from here: Dataset
Code
Simulation code and model implementations are available at: Code