Waterloo team develops adaptive anti-idling regenerative power system for service vehicles
19 June 2017
A team at the University of Waterloo in Canada has developed a novel adaptive auxiliary battery-powered anti-idling system for service vehicles (e.g. delivery trucks or public buses); the energy storage system in the regenerative auxiliary power system (RAPS) is able to electrify the vehicle auxiliary systems.
Service vehicles generally have pre-determined routes, making it feasible to use a model predictive control strategy to improve fuel economy. However, the team noted, the mass/load of such service vehicles is time-varying during a drive cycle. The researchers developed an adaptive model predictive controller to account for this variation. A vehicle can save up to 7% fuel by using anti-idling system and the adaptive model predictive controller, compared to a prescient model predictive controller. A paper on their work is published in the journal Energy.
The team used a Kalman Filter based algorithm to identify the vehicle mass online, and designed a predictive power management controller for the anti-idling system.
The contributions of this study can be summarized as follows: firstly, most of the existing MPCs [model predictive controls] focused on the passenger vehicles, whose loads do not vary that much. Unlike those of the passenger vehicles, the mass of service vehicles can vary up to 500% from the fully loaded to the unloaded situations. As the vehicle mass is closely related to FC [fuel consumption], it is urgent to consider the time-varying mass into the PMSs [power management strategies]. Thus, an online KF-based [Kalman Filter] identification algorithm for the vehicle mass is integrated into the PMS to make an adaptive MPC, which seldom exists in the current literature.
Then, unlike the fact that the auxiliary power is treated as a constant or even neglected in the current PMSs developed for HEVs, this paper takes the A/C-R system as an example, whose power consumption is varying with the ambient temperature and vehicle operating conditions. This varying auxiliary power is integrated into the MPC to make the prediction and action more accurately. According to authors’ knowledge, none of the literature has done this.
In addition, almost all the MPC used for PMSs in HEVs do not put forward a way to treat the uncertainties, noises, and disturbances, which are usually fully ignored in the forecast or prediction process. The current robust MPC algorithms that consider the presence of uncertainties are mainly using the min-max approach or worst-case method, where the objective function is calculated over the worst cases. Nominal controllers that ignore the effects of uncertainties would produce a poor performance if applied in real plants, and such robust controllers present too conservative control laws. Thanks to the characteristic of service vehicles that the pre-defined route is repeatedly driven. The rough or the average drive cycle can be obtained by the data collected on each repetition. As mentioned before, a prediction method with a large step size is used in the proposed MPC to alleviate disturbances or uncertainties and reduce computational efforts.
At last, the adaptive MPC is verified by the prescient MPC to show its advantages over the conventional MPC and its performance approximates to that of the prescient MPC.
—Huang et al.
The researchers said that their proposed approach is independent of the powertrain topology such that it is able to be directly extended to other types of hybrid electric vehicles.
Yanjun Huang, Soheil Mohagheghi Fard, Milad Khazraee, Hong Wang, Amir Khajepour (2017) “An adaptive model predictive controller for a novel battery-powered anti-idling system of service vehicles,” Energy, Volume 127, Pages 318-327 doi: 10.1016/j.energy.2017.03.119