Search Results

A Machine Learning-Genetic Algorithm (ML-GA) approach was developed to virtually discover optimum designs using training data generated from multi-dimensional simulations. Machine learning (ML) presents a pathway to transform complex physical processes that occur in a combustion engine into compact informational processes. In the present work, a total of over 2000 sector-mesh computational fluid dynamics (CFD) simulations of a heavy-duty engine were performed. These were run concurrently on a supercomputer to reduce overall turnaround time. The engine being optimized was run on a low-octane (RON70) gasoline fuel under partially premixed compression ignition (PPCI) mode. A total of nine input parameters were varied, and the CFD simulation cases were generated by randomly sampling points from this nine-dimensional input space. These input parameters included fuel injection strategy, injector design, and various in-cylinder flow and thermodynamic conditions at intake valve closure (IVC).

This paper presents a unified modeling environment to simulate vehicle driving and powertrain operations within the context of the surrounding environment, including interactions between vehicles and between vehicles and the road. The goal of this framework is to facilitate the analysis of the energy impacts of vehicle connectivity and automation, as well as the development of eco-driving algorithms. Connectivity and automation indeed provide the potential to use information about the environment and future driving to minimize energy consumption. To achieve this goal, the designers of eco-driving control strategies need to simulate a wide range of driving situations, including the interactions with other vehicles and the infrastructure in a closed-loop fashion.

As connectivity and sensing technologies become more mature, automated vehicles can predict future driving situations and utilize this information to drive more energy-efficiently than human-driven vehicles. However, future information beyond the limited connectivity and sensing range is difficult to predict and utilize, limiting the energy-saving potential of energy-efficient driving. Thus, we combine a conventional speed optimization planner, developed in our previous work, and reinforcement learning to propose a real-time intelligent speed optimization planner for connected and automated vehicles. We briefly summarize the conventional speed optimization planner with limited information, based on closed-form energy-optimal solutions, and present its multiple parameters that determine reference speed trajectories.

As a result of increasingly stringent regulations and higher customer expectations, auto manufacturers have been considering numerous technology options to improve vehicle fuel economy. Transmissions have been shown to be one of the most cost-effective technologies for improving fuel economy. Over the past couple of years, transmissions have significantly evolved and impacted both performance and fuel efficiency. This study validates the shifting control of advanced automatic transmission technologies in vehicle systems by using Argonne National Laboratory's model-based vehicle simulation tool, Autonomie. Different midsize vehicles, including several with automatic transmission (6-speeds, 7-speeds, and 8-speeds), were tested at Argonne's Advanced Powertrain Research Facility (APRF). For the vehicles, a novel process was used to import test data.

Interest in operational cost reduction is driving engine manufacturers to consider low-cost fuel substitution in heavy-duty diesel engines. These dual-fuel (DF) engines could be operated either in diesel-only mode or operated with premixed natural gas (NG) ignited by a pilot flame of compression-ignited direct-injected diesel fuel. Under certain conditions, dual-fuel operation can result in increased cycle-to-cycle variability (CCV) during combustion. CFD can greatly help in understanding and identifying critical parameters influencing CCV. Innovative modelling techniques and large computing resources are needed to investigate the factors affecting CCV in dual-fuel engines. This paper discusses the use of the High Performance Computing resource Titan, at Oak Ridge National Laboratory, to investigate CCV of a dual-fuel engine.

Traditional autonomous vehicle perception subsystems that use onboard sensors have the drawbacks of high computational load and data duplication. Infrastructure-based sensors, which can provide high quality information without the computational burden and data duplication, are an alternative to traditional autonomous vehicle perception subsystems. However, these technologies are still in the early stages of development and have not been extensively evaluated for lane detection system performance. Therefore, there is a lack of quantitative data on their performance relative to traditional perception methods, especially during hazardous scenarios, such as lane line occlusion, sensor failure, and environmental obstructions.

Building a vehicle model with sufficient accuracy for fuel economy analysis is a time-consuming process, even with the modern-day simulation tools. Obtaining the right kind of data for modeling a vehicle can itself be challenging, given that while OEMs advertise the power and torque capability of their engines, the efficiency data for the components or the control algorithms are not usually made available for independent verification. The U.S. Department of Energy (DOE) funds the testing of vehicles at Argonne National Laboratory, and the test data are publicly available. Argonne is also the premier DOE laboratory for the modeling and simulation of vehicles. By combining the resources and expertise with available data, a process has been created to automatically develop a model for any conventional vehicle that is tested at Argonne. This paper explains the process of analyzing the publicly available test data and computing the parameters of various components from the analysis.

The Prius - a power-split hybrid electric vehicle from Toyota - has become synonymous with the word “Hybrid.” As of October 2010, two million of these vehicles had been sold worldwide, including one million vehicles purchased in the United States. In 2004, the second generation of the vehicle, the Prius MY04, enhanced the performance of the components with advanced technologies, such as a new magnetic array in the rotors. However, the third generation of the vehicle, the Prius MY10, features a remarkable change of the configuration - an additional reduction gear has been added between the motor and the output of the transmission [1]. In addition, a change in the energy management strategy has been found by analyzing the results of a number of tests performed at Argonne National Laboratory's Advanced Powertrain Research Facility (ARRF).

Over the past couple of years, numerous Hybrid Electric Vehicle (HEV) powertrain configurations have been introduced into the marketplace. Currently, the dominant architecture is the power-split configuration, notably the input splits from Toyota Motor Sales and Ford Motor Company. This paper compares two vehicle-level control strategies that have been developed to minimize fuel consumption while maintaining acceptable performance and drive quality. The first control is rules based and was developed on the basis of test data from the Toyota Prius as provided by Argonne National Laboratory's (Argonne's) Advanced Powertrain Research Facility. The second control is based on an instantaneous optimization developed to minimize the system losses at every sample time. This paper describes the algorithms of each control and compares vehicle fuel economy (FE) on several drive cycles.

Eco-Approach and Departure (EAD) has been considered as a promising eco-driving strategy for vehicles traveling in an urban environment, where information such as signal phase and timing (SPaT) and geometric intersection description is well utilized to guide vehicles passing through intersections in the most energy-efficient manner. Previous studies formulated the optimal trajectory planning problem as finding the shortest path on a graphical model. While this method is effective in terms of energy saving, its computation efficiency can be further enhanced by adopting machine learning techniques. In this paper, we propose an innovative deep learning-based queue-aware eco-approach and departure (DLQ-EAD) system for a plug-in hybrid electric bus (PHEB), which is able to provide an online optimal trajectory for the vehicle considering both the downstream traffic condition (i.e. traffic lights, queues) and the vehicle powertrain efficiency.

This paper describes a configuration and controller, designed using Autonomie,1 for dual-motor battery electric vehicle (BEV) heavy-duty trucks. Based on the literature and current market research, this model was designed with two electric motors, one on the front axle and the other on the rear axle. A rule-based control algorithm was designed for the new dual-motor BEV, based on the model, and the control parameters were optimized by using a genetic algorithm (GA). The model was simulated in diverse driving cycles and gradeability tests. The results show both a good following of the desired cycle and achievement of truck gradeability performance requirements. The simulation results were compared with those of a single-motor BEV and showed reduced energy consumption with the high-efficiency operation of the two motors.

In an earlier paper, the authors described how Model-Based System Engineering could be utilized to provide a virtual Hardware-in-the-Loop simulation capability, which creates a framework for the development of virtual ECU software by providing a platform upon which embedded control algorithms may be developed, tested, updated, and validated. The development of virtual ECU software is increasingly valuable in automotive control system engineering because vehicle systems are becoming more complex and tightly integrated, which requires that interactions between subsystems be evaluated during the design process. Variational analysis and robustness studies are also important and become more difficult to perform with real hardware as system complexity increases. The methodology described in this paper permits algorithm development to be performed prior to the availability of vehicle and control system hardware by providing what is essentially a virtual integration vehicle.

Knock integrals and corresponding ignition delay (τ) correlations are often used in model-based control algorithms in order to predict ignition timing for kinetically modulated combustion regimes such as HCCI and PCCI. They can also be used to estimate knock-inception during conventional SI operation. The purpose of this study is to investigate the performance of various τ correlations proposed in the literature, including those developed based on fundamental data from shock tubes and rapid compression machines, those based on predictions from isochoric simulations using detailed chemical kinetic mechanisms, and those deduced from data of operating engines. A 0D engine simulation framework is used to compare the correlation performance where evaluations are based on the temperatures required at intake valve closure (TIVC) in order to achieve a fixed CA50 point over a range of conditions.

In recent studies, optimal control has shown promise as a strategy for enhancing the energy efficiency of connected autonomous vehicles. To maximize optimization performance, it is important to accurately predict constraints, especially separation from a vehicle in front. This paper proposes a novel prediction method for forecasting the trajectory of the nearest preceding car. The proposed predictor is designed to produce short to medium-length speed trajectories using a locally weighted polynomial regression algorithm. The polynomial coefficients are trained by using two types of information: (1) vehicle-to-vehicle (V2V) messages transmitted by multiple preceding vehicles and (2) vehicle-to-infrastructure (V2I) information broadcast by roadside equipment. The predictor’s performance was tested in a multi-vehicle traffic simulation platform, RoadRunner, previously developed by Argonne National Laboratory.

This paper introduces an eco-driving highway cruising algorithm based on optimal control theory that is applied to a conventionally-powered connected and automated vehicle. Thanks to connectivity to the cloud and/or to infrastructure, speed limit and slope along the future route can be known with accuracy. This can in turn be used to compute the control variable trajectory that will minimize energy consumption without significantly impacting travel time. Automated driving is necessary to the implementation of this concept, because the chosen control variables (e.g., torque and gear) impact vehicle speed. An optimal control problem is built up where quadratic models are used for the powertrain. The optimization is solved by applying Pontryagin’s minimum principle, which reduces the problem to the minimization of a cost function with parameters called co-states.

Extreme Fast Charging (XFC) infrastructure is crucial for an increase in electric vehicle (EV) adoption. However, an unmanaged implementation may lead to negative grid impacts and huge power costs. This paper presents an optimal energy management strategy to utilize grid-connected Energy Storage Systems (ESS) integrated with XFC stations to mitigate these grid impacts and peak demand charges. To achieve this goal, an algorithm that controls the charge and discharge of ESS based on an optimal power threshold is developed. The optimal power threshold is determined to carry out maximum peak shaving for given battery size and SOC constraints.

Model-based control system design improves quality, shortens development time, lowers engineering cost, and reduces rework. Evaluating a control system's performance, functionality, and robustness in a simulation environment avoids the time and expense of developing hardware and software for each design iteration. Simulating the performance of a design can be straightforward (though sometimes tedious, depending on the complexity of the system being developed) with mathematical models for the hardware components of the system (plant models) and control algorithms for embedded controllers. This paper describes a software tool and a methodology that not only allows a complete system simulation to be performed early in the product design cycle, but also greatly facilitates the construction of the model by automatically connecting the components and subsystems that comprise it.

While spark-ignition (SI) engine technology is aggressively moving towards challenging (dilute and boosted) combustion regimes, advanced ignition technologies generating non-equilibrium types of plasma are being considered by the automotive industry as a potential replacement for the conventional spark-plug technology. However, there are currently no models that can describe the low-temperature plasma (LTP) ignition process in computational fluid dynamics (CFD) codes that are typically used in the multi-dimensional engine modeling community. A key question for the engine modelers that are trying to describe the non-equilibrium ignition physics concerns the plasma characteristics. A key challenge is also represented by the plasma formation timescale (nanoseconds) that can hardly be resolved within a full engine cycle simulation.

We report experimental observations of cyclic combustion variability during the transition between propagating flame combustion and homogeneous charge compression ignition (HCCI) in a single-cylinder, stoichiometrically fueled, spark-assisted gasoline engine. The level of internal EGR was controlled with variable valve actuation (VVA), and HCCI combustion was achieved at high levels of internal EGR using the VVA system. Spark-ignition was used for conventional combustion and was optionally available during HCCI. The transition region between purely propagating combustion and HCCI was mapped at multiple engine speeds and loads by incrementally adjusting the internal EGR level and capturing data for 2800 sequential cycles. These measurements revealed a complex sequence of high COV, cyclic combustion variations when operating between the propagating flame and HCCI limits.

Modern vehicles use automated driving assistance systems (ADAS) products to automate certain aspects of driving, which improves operational safety. In the U.S. in 2020, 38,824 fatalities occurred due to automotive accidents, and typically about 25% of these are associated with inclement weather. ADAS features have been shown to reduce potential collisions by up to 21%, thus reducing overall accidents. But ADAS typically utilize camera sensors that rely on lane visibility and the absence of obstructions in order to function, rendering them ineffective in inclement weather. To address this research gap, we propose a new technique to estimate snow coverage so that existing and new ADAS features can be used during inclement weather. In this study, we use a single camera sensor and historical weather data to estimate snow coverage on the road. Camera data was collected over 6 miles of arterial roadways in Kalamazoo, MI.