The paper focuses on a power-split architecture in which the internal combustion engine (ICE) and electric motors can, either alone or together, power the vehicle while the battery pack may be charged simultaneously through the ICE.

Broadly, existing EMS for PHEVs are either rule-based or optimization-based.

• Rule-based EMS are fundamental control schemes operating on a set of predefined rules without prior knowledge of the trip. The control decisions are made according to the current vehicle states and power demand only. Such strategies are easily implemented but the resultant operations may be far from being optimal due to not considering future traffic conditions.

• Optimization-based EMS aim at optimizing a predefined cost function according to the driving conditions and behaviors. The cost function may include a variety of vehicle performance metrics, such as fuel consumption and tailpipe emissions.

However, notes the UCR team, most existing PHEV EMS have one or more of the following limitations:

1. Lack of adaptability to real-time information, such as traffic and road grade. This applies to rule-based EMS the parameters or criteria of which have been pre-tuned to favor certain conditions (e.g., specific driving cycles and route elevation profiles). Most EMS that are based on global optimization off-line assume that the future driving condition is known;only a few studies have focused on the development of on-line EMS for PHEVs.

2. Dependence on accurate (or predicted) trip information that is usually unknown a priori. Many of the existing EMS require at a minimum the trip duration as known or predicted information prior to the trip. Furthermore, it is reported that the performance of EMS is largely dependent on the time span of the trip [20]. There are very few studies analyzing the impacts of trip duration on the performance of EMS for PHEVs.

3. Emphasis on a single trip level optimization without considering opportunistic charging between trips. The most critical feature that differentiates PHEVs from conventional HEVs is that PHEVs’ batteries can be charged by plugging into an electrical outlet. Most of the existing EMS are designed to work on a trip-by-trip basis. However, taking into account inter-trip charging information can significantly improve the fuel economy of PHEVs.

To address these limitations, we herein propose a generic framework of on-line EMS for PHEVs that uses an evolutionary algorithm (EA) to optimize vehicle fuel economy in real time. For the purpose of on-line implementation, the optimization is conducted on a sliding time window basis rather than on an entire trip basis. Meanwhile, two types of state-of-charge (SOC) control strategies (i.e., SOC reference control and self-adaptive control), which govern the utilization of vehicle battery power to achieve optimal fuel efficiency for the vehicle without the knowledge of trip duration, are proposed within the framework and compared with conventional binary control strategies.

The major contributions of this paper include: 1) development of a generic framework of on-line EMS for PHEVs; 2) exclusion of trip duration as required information for PHEVs’ energy management; 3) quantification of the performance of the proposed EMS with respect to different trip durations; and 4) consideration of the impacts due to inter-trip charging opportunities.

—Qi et al.

Flow chart of the proposed on-line EMS. Click to enlarge.

The framework consists of information acquisition (from external sources); prediction; optimization; and powersplit control. The entire trip is divided into segments or time horizons. Both the prediction and the control horizons keep moving forward (in a rolling horizon style) while the system is operating. More specifically, the prediction model is used to predict the power demand at each sampling step (i.e., each second) in the prediction horizon. Then, the optimal ICE power supply for each second during the prediction horizon is calculated with this predicted information.

Control of the vehicle’s SOC is formulated as a combinatory optimization problem that can be efficiently solved by the estimation distribution algorithm (EDA). In each control horizon, the pre-calculated optimal control decisions are inputted into the powertrain control system (e.g., electronic control unit (ECU)) at the required sampling frequency.

Using real-world data to evaluate the strategy, they found that the self-adaptive control strategy used in the proposed system statistically outperforms the conventional binary control strategy with an average of 10.7% fuel savings without considering charging opportunity and 31.5% fuel savings when considering charging opportunity.

In reality, drivers may switch routes, traffic can be unpredictable, and road conditions may change, meaning that the EMS must source that information in real-time. By mathematically modeling the energy saving processes that occur in nature, scientists have created algorithms that can be used to solve optimization problems in engineering. We combined this approach with connected vehicle technology to achieve energy savings of more than 30%. We achieved this by considering the charging opportunities during the trip—something that is not possible with existing EMS.

—Xuewei Qi, lead author

The current paper builds on previous work by the team showing that individual vehicles can learn how to save fuel from their own historical driving records. Together with the application of evolutionary algorithms, vehicles will not only learn and optimize their own energy efficiency, but will also share their knowledge with other vehicles in the same traffic network through connected vehicle technology.

Even more importantly, the PHEV energy management system will no longer be a static device--it will actively evolve and improve for its entire life cycle. Our goal is to revolutionize the PHEV EMS to achieve even greater fuel savings and emission reductions.

—Xuewei Qi

This project was supported in part by the National Center for Sustainable Transportation. The UCR Office of Technology Commercialization has filed patents for the inventions above.

Resources

• X. Qi; G. Wu; K. Boriboonsomsin; M. J. Barth, “Development and Evaluation of an Evolutionary Algorithm-Based Online Energy Management System for Plug-In Hybrid Electric Vehicles,” IEEE Transactions on Intelligent Transportation Systems doi: 10.1109/TITS.2016.2633542