Modeling Electrical Motor Dynamics using Encoder-Decoder with Recurrent Skip Connection

TL;DR

This paper introduces a novel data-driven encoder-decoder neural network with recurrent skip connections for modeling electrical motor dynamics, demonstrating effectiveness on simulated and real datasets without relying on physics-based assumptions.

Contribution

It proposes a new neural architecture and loss function for data-driven electrical motor modeling, with demonstrated domain adaptation and analysis of signal complexity effects.

Findings

Achieves good performance on high-frequency datasets

Demonstrates domain adaptation from simulated to real data

Shows impact of signal complexity on modeling accuracy

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.