Chandrasekar, A., Sengupta, S., Hingane, S., Gururaja, C. et al., "Comparative Analysis of Model Predictive Control (MPC) and Conventional Control in Supervisory Controller of a Retrofit HEV," SAE Technical Paper 2017-26-0093, 2017, doi:10.4271/2017-26-0093.
The constant pressure on reducing fuel consumption and emissions in cost sensitive automotive markets has brought focus on retrofit HEV solutions. Through retrofit mechanism, existing conventional (solely engine powered) vehicles can be quickly converted into HEVs. However, the retrofit HEV, although cost effective, poses challenges in developing strategies to control the motor for a given fuel economy, emissions, drivability, battery life requirements and driver inputs. These challenges are primarily due to non-availability of calibration data from the OEMs. This paper focuses on the benefits and challenges with design, tuning and performance of MPC based supervisory controller against a conventional one for a retrofit HEV, using practical data in simulation environment. The inherent characteristics of MPC will lead to the choice of best possible inputs, while respecting the constraints. Hence, MPC is used to control the motor torque in such a manner so as to achieve maximum possible utilization of motor (reducing engine usage) while respecting the safety critical constraints of battery viz., SOC, current, voltage, etc. This is realized by framing the dynamical structure of the retrofit HEV into an MPC problem by identifying the manipulated variables, controlled variables, states, cost functions, constraints and desired set-points. Finally, different modes of HEV operation viz., acceleration/deceleration, battery charge/discharge, motor/regenerate are analyzed for both supervisory controllers to comparatively assess their performance for same drive cycles. Such kind of study is expected to give insight into the suitable choice of supervisory control strategy for retrofit HEV under various circumstances.