Predicting Energy Consumption for Potential Effective Use in Hybrid Vehicle Powertrain Management Using Driver Prediction PDF Download

Author: Brian Magnuson
Publisher:
ISBN:
Category :
Languages : en
Pages : 409

View

Book Description
A proof-of-concept software-in-the-loop study is performed to assess the accuracy of predicted net and charge-gaining energy consumption for potential effective use in optimizing powertrain management of hybrid vehicles. With promising results of improving fuel efficiency of a thermostatic control strategy for a series, plug-ing, hybrid-electric vehicle by 8.24%, the route and speed prediction machine learning algorithms are redesigned and implemented for real- world testing in a stand-alone C++ code-base to ingest map data, learn and predict driver habits, and store driver data for fast startup and shutdown of the controller or computer used to execute the compiled algorithm. Speed prediction is performed using a multi-layer, multi-input, multi- output neural network using feed-forward prediction and gradient descent through back- propagation training. Route prediction utilizes a Hidden Markov Model with a recurrent forward algorithm for prediction and multi-dimensional hash maps to store state and state distribution constraining associations between atomic road segments and end destinations. Predicted energy is calculated using the predicted time-series speed and elevation profile over the predicted route and the road-load equation. Testing of the code-base is performed over a known road network spanning 24x35 blocks on the south hill of Spokane, Washington. A large set of training routes are traversed once to add randomness to the route prediction algorithm, and a subset of the training routes, testing routes, are traversed to assess the accuracy of the net and charge-gaining predicted energy consumption. Each test route is traveled a random number of times with varying speed conditions from traffic and pedestrians to add randomness to speed prediction. Prediction data is stored and analyzed in a post process Matlab script. The aggregated results and analysis of all traversals of all test routes reflect the performance of the Driver Prediction algorithm. The error of average energy gained through charge-gaining events is 31.3% and the error of average net energy consumed is 27.3%. The average delta and average standard deviation of the delta of predicted energy gained through charge-gaining events is 0.639 and 0.601 Wh respectively for individual time-series calculations. Similarly, the average delta and average standard deviation of the delta of the predicted net energy consumed is 0.567 and 0.580 Wh respectively for individual time-series calculations. The average delta and standard deviation of the delta of the predicted speed is 1.60 and 1.15 respectively also for the individual time-series measurements. The percentage of accuracy of route prediction is91%. Overall, test routes are traversed 151 times for a total test distance of 276.4 km.

Author: Wilson Pérez
Publisher:
ISBN:
Category : Hybrid electric vehicles
Languages : en
Pages : 0

View

Book Description
Most vehicles on the road today are conventional vehicles which require the use of nonrenewable fuels to operate. Coupled with this need is a large amount of emissions released into the atmosphere throughout the duration of every trip. To alleviate the burden this places on the environment, governments worldwide have pushed for strict mandates which aim to reduce and, eventually, eliminate the use of fossil fuels. To meet government requirements, hybrid and electric vehicles have been the focus of many car manufacturers. Advancements in vehicle technology have significantly increased the potential of hybrid vehicle technology to reduce levels of emissions and fuel consumption. Advanced energy management strategies have been developed to properly handle the power flow through the vehicle powertrain. These range from rule-based approaches to globally optimal techniques such as dynamic programming (DP). However, cost of high-power computational hardware and lack of a-priori knowledge of future road conditions poses difficult challenges for engineers attempting to implement globally optimal frameworks. A viable solution to the problem is to leverage on-board sensors present in most vehicles equipped with basic advanced driver assistance systems (ADAS) to obtain a prediction of the future road conditions. Known as look-ahead predictive EMS, this approach partially solves the lack of a-priori knowledge since a detailed view of the road ahead is available. However, uncertainty in sensors and the computational burden of processing large amounts of data creates more difficulties. This research aims to address the challenges mentioned above. A look-ahead predictive EMS is proposed which combines the use of a globally optimal approach (DP) with the equivalent consumption minimization strategy (ECMS) to obtain an optimal solution for a future prediction horizon. ECMS is highly sensitive to the equivalence factor, s, making it necessary to adapt during a trip to account for disturbances. A novel adaptation method is presented in this dissertation which uses a neural network to learn the nonlinear relationship between a speed and SOC trajectory prediction obtained from DP to estimate the corresponding s. Finally, an uncertainty analysis is performed to measure the distribution of fuel economy results over a broad range of traffic patterns. It is shown that the proposed EMS consistently improves fuel economy over the baseline strategy and is a viable option for a real-time EMS on production vehicles.

Author: Teng Liu
Publisher: Morgan & Claypool Publishers
ISBN: 1681736195
Category : Technology & Engineering
Languages : en
Pages : 99

View

Book Description
Powertrain electrification, fuel decarburization, and energy diversification are techniques that are spreading all over the world, leading to cleaner and more efficient vehicles. Hybrid electric vehicles (HEVs) are considered a promising technology today to address growing air pollution and energy deprivation. To realize these gains and still maintain good performance, it is critical for HEVs to have sophisticated energy management systems. Supervised by such a system, HEVs could operate in different modes, such as full electric mode and power split mode. Hence, researching and constructing advanced energy management strategies (EMSs) is important for HEVs performance. There are a few books about rule- and optimization-based approaches for formulating energy management systems. Most of them concern traditional techniques and their efforts focus on searching for optimal control policies offline. There is still much room to introduce learning-enabled energy management systems founded in artificial intelligence and their real-time evaluation and application. In this book, a series hybrid electric vehicle was considered as the powertrain model, to describe and analyze a reinforcement learning (RL)-enabled intelligent energy management system. The proposed system can not only integrate predictive road information but also achieve online learning and updating. Detailed powertrain modeling, predictive algorithms, and online updating technology are involved, and evaluation and verification of the presented energy management system is conducted and executed.

Author:
Publisher:
ISBN:
Category :
Languages : en
Pages :

View

Book Description
Abstract : The continued push for the reduction of energy consumption across the automotive vehicle fleet has led to widespread adoption of hybrid and plug-in hybrid electric vehicles (PHEV) by auto manufacturers. In addition, connected and automated vehicle (CAV) technologies have seen rapid development in recent years and bring with them the potential to significantly impact vehicle energy consumption. This dissertation studies predictive control methods for PHEV powertrains that are enabled by CAV technologies with the goal of reducing vehicle energy consumption. First, a real-time predictive powertrain controller for PHEV energy management is developed. This controller utilizes predictions of future vehicle velocity and power demand in order to optimize powersplit decisions of the vehicle. This predictive powertrain controller utilizes nonlinear model predictive control (NMPC) to perform this optimization while being cognizant of future vehicle behavior. Second, the developed NMPC powertrain controller is thoroughly evaluated both in simulation and real-time testing. The controller is assessed over a large number of standardized and real-world drive cycles in simulation in order to properly quantify the energy savings benefits of the controller. In addition, the NMPC powertrain controller is deployed onto a real-time rapid prototyping embedded controller installed in a test vehicle. Using this real-time testing setup, the developed NMPC powertrain controller is evaluated using on-road testing for both energy savings performance and real-time performance. Third, a real-time integrated predictive powertrain controller (IPPC) for a multi-mode PHEV is presented. Utilizing predictions of future vehicle behavior, an optimal mode path plan is computed in order to determine a mode command best suited to the future conditions. In addition, this optimal mode path planning controller is integrated with the NMPC powertrain controller to create a real-time integrated predictive powertrain controller that is capable of full supervisory control for a multi-mode PHEV. Fourth, the IPPC is evaluated in simulation testing across a range of standard and real-world drive cycles in order to quantify the energy savings of the controller. This analysis is comprised of the combined benefit of the NMPC powertrain controller and the optimal mode path planning controller. The IPPC is deployed onto a rapid prototyping embedded controller for real-time evaluation. Using the real-time implementation of the IPPC, on-road testing was performed to assess both energy benefits and real-time performance of the IPPC. Finally, as the controllers developed in this research were evaluated for a single vehicle platform, the applicability of these controllers to other platforms is discussed. Multiple cases are discussed on how both the NMPC powertrain controller and the optimal mode path planning controller can be applied to other vehicle platforms in order to broaden the scope of this research.

Author: Antonio Sciarretta
Publisher: Springer
ISBN: 3030241270
Category : Technology & Engineering
Languages : en
Pages : 294

View

Book Description
This book elaborates the science and engineering basis for energy-efficient driving in conventional and autonomous cars. After covering the physics of energy-efficient motion in conventional, hybrid, and electric powertrains, the book chiefly focuses on the energy-saving potential of connected and automated vehicles. It reveals how being connected to other vehicles and the infrastructure enables the anticipation of upcoming driving-relevant factors, e.g. hills, curves, slow traffic, state of traffic signals, and movements of nearby vehicles. In turn, automation allows vehicles to adjust their motion more precisely in anticipation of upcoming events, and to save energy. Lastly, the energy-efficient motion of connected and automated vehicles could have a harmonizing effect on mixed traffic, leading to additional energy savings for neighboring vehicles. Building on classical methods of powertrain modeling, optimization, and optimal control, the book further develops the theory of energy-efficient driving. In addition, it presents numerous theoretical and applied case studies that highlight the real-world implications of the theory developed. The book is chiefly intended for undergraduate and graduate engineering students and industry practitioners with a background in mechanical, electrical, or automotive engineering, computer science or robotics.

Author: Simona Onori
Publisher: Springer
ISBN: 1447167813
Category : Technology & Engineering
Languages : en
Pages : 121

View

Book Description
This SpringerBrief deals with the control and optimization problem in hybrid electric vehicles. Given that there are two (or more) energy sources (i.e., battery and fuel) in hybrid vehicles, it shows the reader how to implement an energy-management strategy that decides how much of the vehicle’s power is provided by each source instant by instant. Hybrid Electric Vehicles: •introduces methods for modeling energy flow in hybrid electric vehicles; •presents a standard mathematical formulation of the optimal control problem; •discusses different optimization and control strategies for energy management, integrating the most recent research results; and •carries out an overall comparison of the different control strategies presented. Chapter by chapter, a case study is thoroughly developed, providing illustrative numerical examples that show the basic principles applied to real-world situations. The brief is intended as a straightforward tool for learning quickly about state-of-the-art energy-management strategies. It is particularly well-suited to the needs of graduate students and engineers already familiar with the basics of hybrid vehicles but who wish to learn more about their control strategies.

Author: Tobias Stoll
Publisher: Springer Nature
ISBN: 3658421681
Category : Technology & Engineering
Languages : en
Pages : 245

View

Book Description
This book deals with the simulative prediction of efficiency and CO2-emissions of future powertrain systems for the year 2040. For this purpose, a suitable simulation environment is first created. This is followed by a technology extrapolation of all relevant powertrain systems, for example: combustion engines, electric drives, fuel cells as well as all relevant additional components. These components are then used to build 57 vehicle variants for the simulation. Finally, extensive simulations of the vehicle variants are carried out, evaluated and compared. Comprehensive tables of results are available for all simulated vehicle variants. The evaluations are of interest to anyone concerned with energy consumption and CO2-emissions of future vehicles.

Author: Xiangrui Zeng
Publisher:
ISBN:
Category :
Languages : en
Pages :

View

Book Description
One of the main goals of hybrid electric vehicle technology is to improve the energy efficiency. In industry and most of academic research, the powertrain control is designed and evaluated under standard driving cycles. However, the situations that a vehicle may encounter in the real world could be quite different from the standard cycles. Studies show that the human drivers have a great influence on the vehicle energy consumptions and emissions. The actual operating conditions that a vehicle faces are not only dependent on the roads and traffic, but also dependent on the drivers. A standard driving cycle can only represent the typical and averaged driving style under the typical driving scenarios, therefore the control strategies designed based on a standard driving cycle may not perform well for all different driving styles. This motivates the idea to design optimally-personalized hybrid electric vehicle control methods that can be adaptive to individual human driving styles and their driving routes. Human-subject experiments are conducted on a driving simulator to study the driving behaviors. A stochastic driver pedal model that can learn individual driver’s driving style is developed first. Then a theoretic investigation on worst-case relative cost optimal control problems, which is closely related to vehicle powertrain optimal control under real-world uncertain driving scenarios, is presented. A two-level control structure for plug-in hybrid electric vehicles is proposed, where the parameters in the lower-level controller can be on-line adjusted via optimization using historical driving data. The methods to optimize these parameters are designed for fixed-route driving first, and then extended to multi-routes driving using the idea similar to the worst-case relative cost optimal control. The performances of the two proposed methods are shown through simulations using human driving data and stochastic driver model data respectively. The energy consumption results in both situations are close to the posteriori optimal result and outperform other existing methods, which show the effectiveness of applying optimally-personalized energy management strategy on hybrid electric vehicles. Finally, a route-based global energy-optimal speed planning method is also proposed. This off-line method provides a useful tool to evaluate the potential of other speed planning methods, for either eco-driving guidance applications or future automated vehicle controls. The contributions of this dissertation include 1) a novel stochastic driver pedal behavior model which can learn independent drivers’ driving styles is created, 2) a new worst-case relative cost optimal control method is proposed, 3) a real-time implementable stochastic optimal energy management strategy for hybrid electric vehicles running on fixed routes is designed using the statistics of history driving data, 4) the fix-route strategy is extended to the multi-route situation, and 5) an off-line global energy-optimal speed planning solution for road vehicles on a given route is presented.

Author:
Publisher:
ISBN:
Category : Hybrid electric vehicles
Languages : en
Pages : 326

View

Book Description
The primary objective of this dissertation is to develop a hybrid powertrain control strategy that minimizes fuel consumption based on prior knowledge of the driven route. The primary contributions of this research work are: development of a driver classification strategy, development of a route learning and detection system to predict the driven route, and development of a global optimization based powertrain control strategy to minimize the fuel consumption for the predicted route. A driver classifier has been modeled using real time monitoring of some key dynamical parameters of a vehicle, such as, vehicle acceleration, braking, speeding index and throttle activity index. This classifier uses vehicle's powertrain signals to extract the above parameters and classifies a driver into one of three categories, aggressive, moderate and conservative using a Bayesian classifier. In the route learning technique, the learned information comprises of feedback from the vehicle's acceleration and brake sensors, steering wheel angle, driver torque demand, and data from the vehicle's lateral and longitudinal acceleration sensors. The processed route attributes, such as turns, road curvature, road grades, posted speeds and traffic lights are stored in a database in the powertrain microcontroller memory. When the vehicle is driven, the driven route is compared with the attributes stored in the memory for previous routes and identified either as a new route or one of the stored routes. When the vehicle is driven on one of the learned routes, a Hidden Markov Model based algorithm is used to predict the driven route and establish a confidence estimate for the predicted route. The hybrid powertrain is optimized using future route information. The control strategy uses real-time Fibonacci optimization to minimize gasoline fuel consumption using control actions, such as, engine on/off, transmission gear shifting, blended hybrid operation, regenerative braking and battery state of charge management strategy based on the available look-ahead information. The results indicate that substantial fuel savings can be achieved if the powertrain operation can be customized based on the driving style of the driver and advanced knowledge of the driven route.

Author:
Publisher:
ISBN:
Category :
Languages : en
Pages :

View

Book Description
Abstract : This report presents an analysis on energy consumption of a Gen II Chevrolet Volt PHEV and its energy savings potential in Real World Driving scenarios with the help of vehicle connectivity. The research on the energy consumption analysis and optimization using connectivity will focus on four main areas of contribution which includes 1.) vehicle testing on a pre-defined drive cycle and alternative routing near the Michigan Tech campus and APS research center that is a continuation of previous students' works, 2) the energy savings potential of vehicle platooning and various vehicle platoon configurations, 3) the updating of a PHEV implementation of a charge depleting-charge sustaining energy blending optimization algorithm and 4) the development of an IC Engine start-stop prediction algorithm for HEV and PHEV's using connectivity data. The first part of the report discusses the development of a Real World Drive Cycle called Reverse MTU Drive Cycle which is the successor of MTU Drive Cycle, a drive cycle previously developed local to the Michigan Technological University. The energy consumption of the PHEV on the R-MTUDC is analyzed and the baseline characteristics of the drive cycle is setup. A set of baseline drive cycle characteristics was developed and tests on the drive cycle proved that the energy consumption on the real-world drive route is consistent with variability less than 3%. The next part of the report investigates the energy savings potential of the cars when they are traveling in a platoon rather than independently. Various tests have been conducted to investigate energy savings under different platoon scenarios, like variable gap settings, variable speeds, inclusion of a vehicle with aero-modifier and effect of moving collinearly in a platoon. A platoon wide savings as high as 8.3% was achieved in the study. After that, the report discusses the on-road implementation of a Route Based Blended Mode Optimizer, in PHEVs, which comes up with an optimal control matrix using Dynamic Programming and Cost-To-Go matrix, to make use of the Hold mode capability of the Volts, to operate the cars in Charge Sustaining mode at sections of Drive Cycles where it is most efficient to be operated. Upto, 5% savings in energy was obtained using the optimizer. Some of the runs didn't provide the desired results and this is also investigated. Finally, the report presents the development of two kinds of Engine Start-Stop Optimizers, which utilizes vehicle connectivity and vehicle energy consumption model to come up with an optimal control map of regions on the predicted driving route where the engine should be turned On and Off for minimizing energy consumption in HEVs and PHEVs. The first optimizer uses vehicle and route characteristics to predict engine starts and stops and then optimizes these signals based on decisions made from energy calculations. The second optimizer uses Dynamic Programming to create a matrix of engine On and Off signals based on the route characteristics. These controllers are shown to provide energy savings as high as 8% on some routes.