Thrust 1 - Efficiency
Title
1A.1: Integrated Algorithms for Optimal Energy Use in Mobile Fluid Power Systems
Project Leaders
Prof. Kim Stelson (UMN) and Prof. Andrew Alleyne (UIUC)
Statement of Project Goals
The goal of this project is to identify means of regulating power generation and distribution in mobile fluid power systems that maximize the overall system efficiency and demonstrate them on center test beds. From our previous work in the study of energy management strategies (EMS) we have concluded that there is no single strategy which is optimal for all applications [2, 6, 8]. Therefore, we propose to develop a toolbox of EMS design methods and decision algorithms which will identify the best design method for a chosen application. These algorithms will select the optimal design from the EMS toolbox based on a number system attributes such as knowledge of duty cycle, ability to store energy, and problem constraints. In this way, we plan to improve the energy efficiency of mobile fluid power applications without any loss in their performance. The first center test bed that we are targeting is the passenger vehicle where our goal is to demonstrate a 100% improvement in fuel economy over non-hybrid vehicles.
Project's Role in Support of the Strategic Plan
This project will attack the energy management barrier at a very fundamental level. It will result in the development on an energy management toolbox and decision algorithms which together can be used to choose the design method best suited for a chosen application. Such a toolbox is critical to doubling the efficiency of fluid power within the hydraulic hybrid passenger vehicle and the general class of mobile fluid power systems (large trucks, buses, construction equipment, etc).
Figure 1: General framework for a fluid power energy management system.
Description and Explanation of Research Approach
This project explores how to achieve optimal energy usage in the general class of fluid power systems by (a) understanding when to use available power sources (b) achieving smooth transitions between different modes of operation. Deterministic dynamic programming can be used to find the optimal behavior for an assumed drive cycle but cannot be used in real time. To design real time implementable energy management strategies that address both barriers, three methods are being studied: rule based, stochastic dynamic programming, and model predictive control. These methods have been identified as being most successful for hybrid electric vehicles [1, 4, 5, 7] and all three are being studied in the context of mobile fluid power systems because each method has unique advantages/disadvantages. This study will include HIL testing using an augmented earthmoving vehicle powertrain simulator (A-EVPS) that includes energy storage as well as the hydraulic hybrid passenger vehicle test bed. This will lead to the development of an EMS design toolbox which will be used in conjunction with decision algorithms to choose the design method best suited for a given application.
Figure 2. Earthmoving Vehicle Powertrain Simulator
To design the decision algorithms it will be necessary to identify the system attributes which correlate to EMS design strengths and weaknesses. Two potential attributes are constraints and knowledge of the duty cycle. By imposing constraints on the optimization problem, one is able to capture many real world considerations, such as limitations on the accumulator's state of charge, engine on/off cycling, and emissions. Introducing these considerations as constraints enables the EMS to intelligently utilize the DOF's within the powertrain to meet these as well as performance and efficiency demands. Similarly, knowledge of the duty cycle is a key attribute which differentiates the EMS design methods as well as mobile fluid power applications. Systems which follow a prescribed duty cycle, such as construction equipment, could be effectively controlled using a rule-based strategy in which some knowledge of the duty cycle is assumed when constructing the EMS. However, applications which do not follow prescribed duty cycles, such as passenger and rescue vehicles, would be best suited to stochastic dynamic programming and model predictive control where little assumption of the duty cycle is needed. All of this information will be incorporated into the decision algorithm.
Figure 3: Decision algorithm
With the integration of multiple research efforts within the CCEFP, we are able to study optimization of the powertrain architecture, multiple EMS design methods, and conduct simultaneous hardware-in-the-loop studies (using TB3 and the A-EVPS), enabling us to take a much broader view of the energy management problem. Other work on hydraulic hybrid powertrains has focused on larger vehicles, using a parallel or series architecture, and optimized over a fixed drive cycle [3, 9]. Two of the major gaps that we plan to address are taking advantage of the hydromechanical powertrain architecture and developing an EMS toolbox that integrates the advantages of the rule-based, SDP, and MPC approaches. The hydromechanical configuration is more complex, but offers the potential to for greater efficiency and performance. Also, while other researchers have focused on establishing the merit of a single strategy we recognize that there is no single EMS design method which will give optimal performance for all applications. Rather, each design approach has its own unique advantages/disadvantages and a decision algorithm is needed to identify the optimal design approach.
References
[1] Borhan, H.A., et al. Predictive energy management of a power-split hybrid electric vehicle. 2009. Piscataway, NJ, USA: IEEE.
[2] Deppen, T.O., A.G. Alleyne, K.A. Stelson, J.J. Meyer. Predictive Energy Management for Parallel Hydraulic Hybrid Passenger Vehicle. Proc. of the ASME Dynamic Systems and Control Conference. DSCC 2010.
[3] Filipi, Z., et al. Combined optimization of design and power management of the hydraulic propulsion system for the 6 x 6 medium truck. International Journal of Heavy Vehicle Systems, 2004, 11(3-4):372-402.
[4] Johannesson, L., et al. Assessing the potential of predictive control for hybrid vehicle powertrains using stochastic dynamic programming. 2007. IEEE Transactions on Intelligent Transportation Systems, 8(1): 71-83.
[5] Lin, C.C., H. Peng, and J.W. Grizzle. A stochastic control strategy for hybrid electric vehicles. in Proceedings of the American Control Conference. 2004.
[6] Meyer, J.J., et al. Energy Management Strategy for a Hydraulic Hybrid Vehicle Using Stochastic Dynamic Programming. in Proceedings of the 6th FPNI PhD Symposium. 2010.
[7] Sciarretta, A. and L. Guzzella, Control of hybrid electric vehicles. IEEE Control Systems Magazine, 2007. 27(2): p. 60-70.
[8] Stelson, K.A., et al. Optimization of a passenger hydraulic hybrid vehicle to improve fuel economy. in Proceedings of the 7th JFPS Symposium on Fluid Power. 2008.
[9] Wu, B., et al., Optimal power management for a hydraulic hybrid delivery truck. Vehicle System Dynamics, 2004. 42(1-2): p. 23-40.