Regression Models to Predict Coastdown Road Load for Various Vehicle Types
| Main Author: | |
|---|---|
| Language: | English |
| Published: |
The Ohio State University / OhioLINK
2020
|
| Subjects: | |
| Online Access: | http://rave.ohiolink.edu/etdc/view?acc_num=osu1595265184541326 |
| id |
ndltd-OhioLink-oai-etd.ohiolink.edu-osu1595265184541326 |
|---|---|
| record_format |
oai_dc |
| collection |
NDLTD |
| language |
English |
| sources |
NDLTD |
| topic |
Automotive Engineering Mechanical Engineering Statistics coastdown testing fuel economy certification chassis dynamometer testing road load statistical modeling regression exploratory data analysis kernel density estimation wind tunnel testing aerodynamic drag |
| spellingShingle |
Automotive Engineering Mechanical Engineering Statistics coastdown testing fuel economy certification chassis dynamometer testing road load statistical modeling regression exploratory data analysis kernel density estimation wind tunnel testing aerodynamic drag Singh, Yuvraj Regression Models to Predict Coastdown Road Load for Various Vehicle Types |
| author |
Singh, Yuvraj |
| author_facet |
Singh, Yuvraj |
| author_sort |
Singh, Yuvraj |
| title |
Regression Models to Predict Coastdown Road Load for Various Vehicle Types |
| title_short |
Regression Models to Predict Coastdown Road Load for Various Vehicle Types |
| title_full |
Regression Models to Predict Coastdown Road Load for Various Vehicle Types |
| title_fullStr |
Regression Models to Predict Coastdown Road Load for Various Vehicle Types |
| title_full_unstemmed |
Regression Models to Predict Coastdown Road Load for Various Vehicle Types |
| title_sort |
regression models to predict coastdown road load for various vehicle types |
| publisher |
The Ohio State University / OhioLINK |
| publishDate |
2020 |
| url |
http://rave.ohiolink.edu/etdc/view?acc_num=osu1595265184541326 |
| work_keys_str_mv |
AT singhyuvraj regressionmodelstopredictcoastdownroadloadforvariousvehicletypes |
| _version_ |
1719457599912411136 |
| spelling |
ndltd-OhioLink-oai-etd.ohiolink.edu-osu15952651845413262021-08-03T07:15:42Z Regression Models to Predict Coastdown Road Load for Various Vehicle Types Singh, Yuvraj Automotive Engineering Mechanical Engineering Statistics coastdown testing fuel economy certification chassis dynamometer testing road load statistical modeling regression exploratory data analysis kernel density estimation wind tunnel testing aerodynamic drag The fuel economy label (window sticker) is used by every vehicle manufacturer in the United States to report fuel economy for two purposes. First, the values reported on the sticker are certified by the United States Environmental Protection Agency (EPA) and are used for certifying emissions regulations like the Corporate Average Fuel Economy (CAFE). Second, the fuel economy numbers are used by consumers to compare competing vehicles in the marketplace. These fuel economy numbers are generated through a process that involves standardized testing on a chassis dynamometer using standard drive cycles. As a result, the test requires an accurate replication of the resistive forces that a vehicle experiences in the real-world, which requires an accurate estimation of road load applied by the road and the surroundings opposing the vehicle motion. The estimation also depends on the type (aerodynamic shape, drivetrain configuration, etc.) of vehicle being tested. To get a description of road load that is as close as possible to reality, several noise factors and residuals need to be estimated as well, which forms the bulk of this thesis.Vehicle coastdown method is widely used to determine road load coefficients for testing vehicles on a chassis dynamometer for fuel economy certification. However, apart from being a time-consuming procedure for each variant in a mass production vehicle lineup, the repeatability of track coastdown testing procedure is sensitive to environmental conditions, the track surface condition as well as on the type of vehicle being tested (for example, SUVs, sedans, hybrid vehicles, manual transmissions, etc.). As a result, several attempts have been made to accurately model the coastdown road load parametrically. This thesis explores various ways in which such parametric models can be obtained and methods to minimize risks related to overfitting and collinearity of variables. Since, the vehicle’s road load is dependent on several physical phenomena, various lab testing methods are employed to determine the road load components (rolling resistance, aerodynamic drag and driveline parasitic loss) individually to come up with a model for road load force that can be applied on the chassis dynamometer. The model obtained from lab tests is defined as a second order polynomial in vehicle speed using the linear least squares curve fit procedure specified by SAE standards J1263, J2263 and J2264. Lab testing is carried out using various equipment under a well-defined testing procedure. Wind tunnels are used for aerodynamic drag testing, parasitic measurements are performed by disconnecting various powertrain components iteratively and rolling resistance measurement is performed for the tires using appropriate tire drum apparatus. However, this model developed solely through lab tests ignores certain noise factors that cause a significant deviation (residual) between the track coastdown force and the model developed in the lab by adding all road load components.This thesis explores the various regression modeling techniques and recommended practices for estimating residual force and the associated residual horsepower. Additionally, different models for various clusters of vehicles based on certain characteristics have been explored. 2020 English text The Ohio State University / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=osu1595265184541326 http://rave.ohiolink.edu/etdc/view?acc_num=osu1595265184541326 restricted--full text unavailable until 2022-08-24 This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws. |