A Data-Driven, Synthetic-Population Approach to Predict Durability Loads for Electric Vehicle Propulsion Systems

<div class="section abstract"><div class="htmlview paragraph">Designing for the durability of motor vehicles requires accounting for various stress factors, including tractive loads, electrical loads, thermal loads, and structural loads. For electric vehicle propulsion systems, it is crucial to consider not just the magnitude and repeats of these loads but also their temporal sequence throughout the vehicle’s lifespan. The order and timing of these loads influence factors such as, charge and discharge cycles or active motor heating, which ultimately impact the damage to the propulsion system components like the cell and the motor. Traditionally, lifetime loads for durability assessments are derived from a single-user load profile consisting of a set of ‘representative’ drive cycles accounting for the cumulative damage equivalent to the real-world damage covered under warranty. This profile is typically based on historical usage data, user scenarios, and industry experience, but may not capture the diverse failure modes of the different propulsion components. The uncertainty in this approach compels the designers to add more material mass, size, and capacity thereby also increasing lifecycle emissions, reducing sustainability. The advent of high-fidelity, cloud-based vehicle data offers an opportunity to create a ‘representative’ synthetic population that accurately reflects the frequency, type, and sequence of real-world driving, parking, and charging loads. In this paper, we propose a method to integrate various datasets—vehicle-measured data, survey data, and test data—to model user travel patterns, driving profiles, and associated charging behaviors comprising vehicle usage. Through this approach we generate lifetime vehicle usage patterns for a diverse set of users and identify the warranty-critical damage for each component. The warranty-critical damage for each component might occur in a different subset of users. Thus, the overall population covers the durability loads for the entire propulsion system or the vehicle. We demonstrate the application of this model to estimate durability loads for electric vehicle battery cells, highlighting its effectiveness in predicting and enhancing component longevity and sustainability.</div></div>