Development and Validation of a Range Prediction Model for New Energy Vehicles

Abstract

Given the significant discrepancy between the official driving range of new energy vehicles and users' real-world experience, this paper analyzes the range attainment rate by considering three key factors: ambient temperature, driving speed, and air-conditioning operation. To address this issue, this study collects and preprocesses real-world test samples from media platforms such as Autohome and Dongchedi, combines them with official test reports from the China Automotive Technology and Research Center (CATARC), and constructs a range prediction model based on multiple linear regression. The empirical results reveal that ambient temperature acts as the dominant factor influencing driving range variation, contributing to nearly 40% of total range attenuation, while high-speed driving and air-conditioning operation also significantly increase vehicle energy consumption. Across several test scenarios, the coefficient of determination (R²) remains at 0.84, and the prediction residuals fall within a 15% error band. These results show that the proposed model can partially correct the prediction bias in traditional NEDC/CLTC standards under extreme operating conditions. It also provides a quantitative basis for driving decisions under complex conditions and may improve users' confidence in trip planning.