Towards Faster E-Motor Deployment: A Fleet Digital Twin Approach for Predicting Efficiency Under Unseen Load Profiles

Abstract

The electric motor is the central component of the electric vehicle (EV) propulsion system, responsible for the energy conversion process that drives vehicle motion. Researchers involved in motor development often need to investigate the efficiency of electric motors under new load profiles prior to field deployment. However, experimental investigations during the product development phase are time-consuming and costly. This paper proposes a novel approach called Fleet Digital Twin (FDT) to predict the efficiency of electric motors under new load profiles, ones that have not yet been tested on the physical assets. The contribution of this paper is to consider several physical setups to create the FDT. To achieve this goal, five experimental setups of electrical drivetrains, each consisting of permanent magnet synchronous motors (PMSMs), were manufactured and deployed at different locations. The data from all setups was collected on a centralized cloud platform. Data from four of the setups was used to develop the FDT, while data from the fifth setup was reserved exclusively for validation. The FDT parameters were derived through an optimization study using time-domain data from the measured load profiles. Unlike a traditional digital twin based on a single prototype, this FDT is developed based on multiple field-deployed products, capturing critical variations arising from manufacturing tolerances, component aging, and differing operating conditions. Results confirm that the proposed approach accurately assesses motor efficiency under new load profiles, supporting informed and rapid deployment decisions across fleets of electric vehicles.