LIST OF FIGURES |
v |
LIST OF TABLES |
vii |
ACRONYMS AND ABBREVIATIONS |
ix |
ACKNOWLEDGEMENTS |
xi |
EXECUTIVE SUMMARY |
xiii |
|
1 |
INTRODUCTION |
1 |
|
1.1 |
BACKGROUND |
1 |
1.2 |
BASIC TERMINOLOGY |
2 |
1.3 |
ORGANIZATION OF REPORT |
3 |
2 |
METHODOLOGY AND DATA |
5 |
|
2.1 |
DIESEL FUEL AND EMISSIONS DATABASES |
5 |
2.2 |
EPA UNIFIED MODEL |
7 |
3 |
ISSUES WITH STEPWISE REGRESSION |
9 |
|
3.1 |
RELATIONSHIPS AMONG FUEL PROPERTIES |
9 |
3.2 |
THE CONSEQUENCE OF CORRELATIONS |
10 |
3.3 |
THE EFFECT OF ALIASING ON REGRESSION COEFFICIENTS |
12 |
3.4 |
THE EFFECT OF ALIASING ON PREDICTIVE POWER |
15 |
3.5 |
COMMENTS ON THE EPA UNIFIED EMISSION MODELS |
20 |
3.6 |
SUMMARY |
21 |
4 |
APPLICATION OF PCR+ TO DIESEL EMISSIONS |
23 |
|
4.1 |
STATISTICAL BACKGROUND |
23 |
4.2 |
EIGENVECTOR REPRESENTATION OF FUELS |
24 |
4.3 |
A PCR+ MODEL OF EMISSIONS |
26 |
4.4 |
DIESEL FUEL EFFECTS ON EMISSIONS |
28 |
4.5 |
PARTITIONING OF THE MODEL SUMS OF SQUARES |
34 |
4.6 |
ISSUES OF STATISTICAL BIAS |
37 |
5 |
DIESEL FUEL EIGENVECTORS |
39 |
|
5.1 |
EIGENFUEL STRUCTURE OF COMMERCIAL DIESEL FUELS |
39 |
5.2 |
EIGENFUEL STRUCTURE OF EXPERIMENTAL FUELS |
41 |
5.3 |
DISCUSSION AND SUMMARY |
45 |
6 |
USE OF EIGENFUELS IN EXPERIMENT DESIGN |
47 |
|
6.1 |
METHODOLOGICAL PRINCIPLES OF EXPERIMENT DESIGN |
47 |
6.2 |
DEMONSTRATION AND APPLICATION OF PRINCIPLES |
48 |
6.3 |
DISCUSSION AND SUMMARY |
51 |
7 |
EIGENFUELS IN FUELS RESEARCH |
53 |
|
7.1 |
DESIGN OF MONTE CARLO SIMULATIONS |
53 |
7.2 |
ESTIMATING THE PROPERTIES OF ALTERNATIVE FUELS |
56 |
8 |
EIGENFUELS IN FUELS REFORMULATION |
59 |
|
8.1 |
WHAT HAPPENS WHEN EIGENFUELS ARE CHANGED? |
59 |
8.2 |
A FRAMEWORK FOR FUEL REFORMULATION |
62 |
9 |
REFERENCES |
65 |
|
APPENDIX A: SUPPORTING DATA |
67 |
APPENDIX B: THE MULTIPLICITY OF MULTIPLE REGRESSION |
77 |
APPENDIX C: CALCULATION OF THE ALIAS MATRIX |
105 |
APPENDIX D: PARTITIONING THE MODEL SUM OF SQUARES |
113 |
APPENDIX E: EXPERIMENT DESIGN AND DATA ANALYSIS |
129 |
APPENDIX F: EIGENVECTORS: THEIR ROLE IN EMISSIONS |
145 |