The study, published in Discover Applied Sciences, builds on a physics-informed, reduced-order lumped-parameter thermal model of the battery pack. That model was implemented in a LabVIEW–MATLAB co-simulation platform designed to replicate electrothermal behavior with spatial and temporal precision. Rather than relying on the linear approximations used in conventional model predictive control, the NMPC approach captures the nonlinear thermodynamic interactions that arise as batteries heat, cool and exchange heat with surrounding cells under demanding driving conditions.
The reported results are striking, with an important caveat: they come entirely from simulation. Compared with conventional linear MPC, the NMPC controller reduced peak inter-cell temperature differentials by more than 60 percent and cut the required coolant air velocity by 25 to 30 percent, all without violating safety constraints. Lower coolant velocities translate directly into less energy spent on thermal management, which the authors link to improved overall energy efficiency for the vehicle.
The controller was evaluated under two driving profiles: the regional Addis Ababa–Bishoftu cycle and the standard US06 cycle used in the United States. On both, the NMPC scheme showed superior tracking accuracy and energy efficiency relative to the linear baseline. A further practical concern for any predictive controller is whether it can run fast enough to keep up with a vehicle in motion. Here, the authors report real-time computational feasibility, with each control update taking less than 150 milliseconds.
Why does the nonlinear formulation matter? That context-aware regulation is what allows the same framework to perform well across topographically and climatically diverse conditions, from highland Ethiopian routes to standardized US test cycles.
The authors are explicit about the limits of the work. The framework is described as a simulation-level proof of concept, and experimental validation is identified as essential future work to confirm real-world applicability.
Subject of Research: Nonlinear model predictive control for lithium-ion battery thermal management in electric vehicles, evaluated in simulation under the Addis Ababa–Bishoftu driving cycle
Article Title: Advanced nonlinear model predictive control for electric vehicle battery thermal management under the Addis Ababa Bishoftu driving cycle
Article References: Mamo, T., Yeneneh, K., Mosisa, D., & Tiku, G. (2026). Advanced nonlinear model predictive control for electric vehicle battery thermal management under the Addis Ababa Bishoftu driving cycle. Discover Applied Sciences. https://doi.org/10.1007/s42452-026-09652-4
Image Credits: AI Generated
DOI: 10.1007/s42452-026-09652-4
Keywords: electric vehicles, battery thermal management, model predictive control, nonlinear control, lithium-ion batteries, simulation, LabVIEW, MATLAB, driving cycle, energy efficiency, coolant air velocity, Ethiopia
News Source: Denise Maddox. (October 28, 2026). Nonlinear predictive control keeps EV batteries cool on a steep Ethiopian driving route in simulation. Scienmag.
