ARC Collaborative Research Seminar Series
ARC seminars are free and open to the general public. Center members can download the presentation files on our password-access online portal iARC. Non-ARC attendees please email email@example.com with your requests.
Parking & directions inquires: Contact firstname.lastname@example.org) by 2:00 p.m. the day before the seminar
Remote attendance via tele/video conference: Contact William Lim (email@example.com)
Refreshments will be served 9:15-9:30am. The talks will begin at 9:30 a.m. sharp.
Event venue alternates between University of Michigan (Ann Arbor) and U.S. Army TARDEC (Warren)
February 10, Friday (9:30a.m. - 11a.m.)
University of Michigan, UM North Campus, Lurie Engineering Center. GM conference room
Thrust Area 1: Dynamics and Control of Vehicles
1. The Third Offset and the DOD Strategy in Autonomy
Dr. Bob Sadowski, Army Chief Roboticist, U.S. Army TARDEC
2. Improving Mobility through Latency Compensation in Teleoperated Ground Vehicles (project link)
Tulga Ersal (PI), Jeffrey Stein (co-PI), Yingshi Zheng, University of Michigan
Paramsothy Jayakumar, Mark Brudnak, U.S. Army TARDEC
Mitchell Rohde, Steve Rohde, Quantum Signal LLC
AbstractThis project is motivated by the Army’s need to increase the mobility performance of teleoperated unmanned ground vehicles. Communication delays make teleoperated driving very challenging especially at high speeds. This project is developing a predictor framework to compensate for the negative impact of communication delays. The proposed approach has the advantage of not requiring a dynamic model of the platform or the human operator. Pilot tests have shown the effectiveness of the approach in improving vehicle mobility with large round-trip delays of 0.9 s.
Our recent efforts aim to improve mobility of a teleoperated vehicle further by introducing a steering feedforward loop along with the predictors. This talk will first illustrate this extended framework and then focus on the experimental results with human-in-the-loop simulations, where human drivers were asked to teleoperate a simulated vehicle on a virtual track as fast as possible while keeping the vehicle in the center of the track. Results show that, compared to the difference in the performance metrics between the delayed and undelayed teleoperation scenarios, the extended framework can improve the track completion time by 48%, track keeping error by 55%, and steering control effort by 69%, with round-trip delays of 0.9 s.
March 10, Friday (9:30a.m. - 11a.m.)
University of Michigan, UM North Campus, Phoenix Memorial Lab. 2000A
Thrust Area 4: Advanced and Hybrid Powertrains -
Internal Combustion Engine & Fuels
AbstractDetailed chemical kinetic mechanisms are needed for CFD simulations used to design Diesel engine combustion systems. Despite tremendous progress, state of the art Jet-A and JP-8 mechanisms remain inaccurate within the low temperature heat release (LTHR) and negative temperature coefficient (NTC) ignition regimes relevant to Diesel ignition. This project is focused on improving these mechanisms, in particular within the Diesel relevant ignition regimes. To this end, sensitivity analysis and optimization are used to identify and then modify relevant mechanism reaction rate parameters in order to reduce ignition delay errors. Combined with the atomistic computational project of Violi focused on mechanism reaction pathway development and the motored engine ignition studies of Boehman et al., our efforts will lead to more accurate Jet-A and JP-8 chemical kinetic models needed for the development of future Army engines.
AbstractIn recent years, there has been an increasing effort to incorporate complex reaction mechanisms in simulation of reacting flows. Indeed, computational fluid dynamic calculations of reactive flows have become an important part of the design of combustion devices, such as engines. One of the key components is the accurate description of the network of reactions that can happen during combustion. Current models are developed and optimized for specific conditions, and have little chance of producing reliable extrapolations to other conditions. There are two main major problems related to the development of predictive reaction models: accuracy of rate constants and ability to obtain a complete detailed mechanism. The former is usually addressed using high-level ab initio simulations to compute reaction rates; the latter represents a big challenge and almost no literature is available on this topic. In our project, we aim at addressing this challenge and developing a novel computational procedure to identify missing reaction pathways as well as main reaction pathways for the combustion chemistry of JP-8, using atomistic simulations.
This work directly supports the effort reported by Dr. Martz, who is currently using reduced kinetic mechanisms from available detailed kinetics present in the literature, as well as the motored engine ignition studies of Dr. Boehman et al., to obtain more accurate jet fuels chemical kinetic models needed for the development of future Army engines.
Thrust Area 4: Advanced and Hybrid Powertrains - Electrification
1. Robotic Range Extender: Power and Energy Management for a Hybrid Powertrain with Quantized Power Sources (project link)
Dr. Jason Siegel, Asst. Research Scientist, Mechanical Engineering, University of Michigan
AbstractThis project addresses the need for quiet, long-life power sources for robotic vehicles which cannot be met by batteries alone (due to range), or with combustion engines (due to noise). A hybrid power source that combines a battery (BB2590) with small (245 Watt), propane-fueled solid oxide fuel cell (SOFC) is considered. A system model that takes into account degradation, which is particularly important and challenging for these systems due to the thermal stresses from the start-stop operation, was developed using physics based modeling techniques. The models are used to evaluate the impact of battery sizing and power split strategy on fuel efficiency, SOFC stack life, and battery cell life over realistic synthetic cycles. These cycles were developed using statistical models of the power measured from an instrumented PackBot, provided by TARDEC engineers. This talk will focus on the thermal model.
AbstractKnowledge of the internal temperatures of electric machines is very important since the performance of these machines, such as their torque capability and efficiency, are greatly affected by these temperatures. In our previous work an FEA-based, computationally-efficient model of thermal conduction in the electric machine components was developed. In this project, a complementary computationally-efficient model of heat convection in the air regions of an electric machine is proposed. The model is based upon the exploitation of certain properties of the heat transfer equations (i.e., conservation of mass, momentum, and energy), which we have discovered exist under certain conditions as seen in electric machines. This in turn leads to the formulation of a system identification technique, the end result being a computationally-efficient heat convection model with very high accuracy.