ENERGY EFFICIENCY MONITORING AND INDEX DECOMPOSITION ANALYSIS LIU NA NATIONAL UNIVERSITY OF SINGAPORE 2006 ENERGY EFFICIENCY MONITORING AND INDEX DECOMPOSITION ANALYSIS LIU NA (Master of Management Science & Engineering, Tsinghua University, Beijing, China) A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF INDUSTRIAL AND SYSTEMS ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE 2006 ACKNOWLEDGEMENTS I would like to express my deepest appreciation to Professor Ang Beng Wah, my supervisor His serious attitude in guiding student’s research and amiable personality make him a respected supervisor Prof Ang sets a perfect example for me via his sagacity, diligence, gentleness, and patience The experience of being his research student is happy and enriching I would also like to express my gratitude to Jessica Palmer of the Office of Energy Efficiency (OEE) of Canada and Robert Tromop of the Energy Efficiency and Conservation Authority (EECA) of New Zealand for providing data and helpful suggestions for my research During the past four years in the National University of Singapore (NUS), I appreciated the academic atmosphere of the university very much The devoted professors and students, the comprehensive collections in NUS library, and numerous academic activities were a great help to my research In particular, I owe my thanks to all the other members of the Department of Industrial and Systems Engineering I have learnt a lot through coursework, seminars, being a tutor, and discussions with laboratory mates All these activities have made my stay in the department enjoyable and memorable Last but not the least, I would like to thank my husband for his continuous support and encouragement, my dearest daughter who makes my life full of expectations, and her four grand-parents for their wholehearted help LIU NA i TABLE OF CONTENTS ACKNOWLEDGEMENTS i TABLE OF CONTENTS iii SUMMARY .ix LIST OF TABLES xi LIST OF FIGURES .xv LIST OF ABBREVIATIONS xix LIST OF NOTATIONS xxi CHAPTER : INTRODUCTION 1.1 Motivations of Energy Efficiency Monitoring 1.2 Definition of Energy Efficiency .3 1.3 Evolution of Energy Efficiency Indicators .4 1.4 Issues Related to Energy Efficiency Monitoring 1.5 Scope and Structure of the Thesis CHAPTER : AGGREGATE ENERGY EFFICIENCY INDICATORS .11 2.1 Introduction 11 2.2 Data and Statistical Analysis 13 2.3 Commercial Energy Consumption Per Capita versus Income Per Capita 16 2.4 Substitution of Commercial Energy for Non-Commercial Energy 18 2.5 Aggregate Energy Intensities versus Income Per Capita 19 2.6 Cross-Country Energy Elasticity versus Income Per Capita 21 2.7 Conclusion 23 CHAPTER : INDEX DECOMPOSITION ANALYSIS AND KEY FACTORS 25 3.1 Introduction 25 iii 3.2 Index Decomposition Analysis 26 3.3 Review of Country Practices 30 3.3.1 Australia 31 3.3.2 Canada 31 3.3.3 New Zealand 33 3.3.4 USA 33 3.3.5 APEC 34 3.3.6 European SAVE project 35 3.3.7 IEA 36 3.4 Empirical Study of the Impacts of Key Factors in IDA 37 3.4.1 Data source and scope 37 3.4.2 Impacts of IDA method 39 3.4.3 Impacts of activity indicator 42 3.4.4 Impacts of sector disaggregation 44 3.5 Conclusion and Problems that Need Further Investigation 45 CHAPTER : FACTORS SHAPING CHANGES IN INDUSTRIAL ENERGY USE 49 4.1 Introduction 49 4.2 Scope and Information Sources 50 4.3 Complications in Comparing Different Information Sources 57 4.4 Trends in Industrial Energy Decomposition Analysis 59 4.5 Results Presentation 66 4.6 A Multi-Country Analysis 71 4.7 Analysis for Selected Countries 74 4.7.1 United States 78 4.7.2 Canada 78 iv 4.7.3 Australia .79 4.7.4 Japan .80 4.7.5 South Korea 81 4.7.6 China 81 4.7.7 Other countries .82 4.8 Conclusion and Further Discussion 82 CHAPTER : HANDLING ZERO AND NEGATIVE VALUES IN THE LMDI APPROACH 85 5.1 Introduction 85 5.2 Zero and Negative Value Problems in LMDI Approach 88 5.3 Small Value Strategy for LMDI with Zero Values 89 5.3.1 Case study 91 5.3.2 Case study 95 5.3.3 Application of SV strategy 99 5.4 Analytical Limits for LMDI with Zero Values 100 5.4.1 Analytical Limits for various cases with zero values .100 5.4.2 Comparisons between cases with zeros 103 5.4.3 Application of analytical limit strategy 105 5.5 Negative Values 107 5.5.1 Procedure for handling negative values .108 5.5.2 Case study 113 5.6 Conclusion 114 CHAPTER : COMPOSITE ENERGY EFFICIENCY INDEX AND CONSISTENCY IN AGGREGATION 115 6.1 Introduction 115 6.2 Composite Energy Efficiency Index 116 6.3 Review of Aggregation Methods 117 v 6.4 Consistency in Aggregation 120 6.4.1 Type of consistency in aggregation 121 6.4.2 Advantages of consistency in aggregation 122 6.4.3 Partial fulfillment of consistency in aggregation 124 6.5 Empirical Study 127 6.6 A Framework for Economy-Wide Energy Efficiency Study 133 6.7 Conclusion 135 CHAPTER : COMPARING DECOMPOSITION METHODS: AN AHP ANALYSIS 137 7.1 Introduction 137 7.2 Alternatives 141 7.3 Criteria of Comparison 143 7.3.1 Factor reversal 144 7.3.2 Time reversal 145 7.3.3 Proportionality 146 7.3.4 Additive/Multiplicative 147 7.3.5 Aggregation 148 7.3.6 Special value robustness 148 7.3.7 Ease of computation 150 7.3.8 Transparency 150 7.3.9 Ease of formulation 151 7.3.10 Extensibility 151 7.4 Methodology and Tool 154 7.4.1 AHP and Expert Choice 154 7.4.2 The comparison procedure 156 7.4.3 Pairwise comparisons 159 vi Appendix G Analytical Limits of LMDI I and Their Speed of Convergence This is exactly the analytical limit of ∆C emf ,ij Therefore for this variable, the AL strategy converge to AL instantly, no matter how large or how small δ is For the other factors in the same sub-category, ⎛ QT lim L(C , C ) ln⎜ ⎜Q δ →0 ⎝ = lim δ ij T ij lim ∆Cact ,ij δ →0 lim δ δ →0 ⎞ ⎟ ⎟ ⎠ δ →0 ⎛ QT ln⎜ ln(QT S I M δ ) − ln(Q Si0 I i0 ⋅ M ij ⋅ δ ) ⎜ Q ⎝ = lim Q S I M δ − Q Si0 I i0 M ij δ T T T T i i ij T T T i i ij δ δ →0 ⎛ QT ln⎜ = lim ⎜Q δ →0 ln(Q T S I M ) − ln(Q S I ⋅ M ) i i ij ⎝ Q S I M − Q Si0 I i0 M ij T ⎛ QT = k ⋅ ln⎜ ⎜Q ⎝ where k = T T T i i ij T T T i i ij ⎞ ⎟ ⎟ ⎠ ⎞ ⎟ ⎟ ⎠ ⎞ ⎟ ⎟ ⎠ T QT SiT I iT M ij − Q Si0 I i0 M ij ⎛ QT ⎞ and ln⎜ ⎟ are both normal numbers T ⎜Q ⎟ ln(Q T SiT I iT M ij ) − ln(Q S i0 I i0 M ij ) ⎝ ⎠ Similarly, lim ∆C str ,ij δ →0 lim δ δ →0 ⎛ ST = k ⋅ ln⎜ i0 ⎜S ⎝ i ⎞ lim ∆C int,ij ⎛ IT ⎟ , δ →0 = k ⋅ ln⎜ i0 ⎟ ⎜I lim δ ⎠ ⎝ i δ →0 T ⎛ M ij ⎞ lim ∆C mix ,ij δ →0 ⎟, = k ⋅ ln⎜ ⎟ ⎜M lim δ ⎠ ⎝ ij δ →0 ⎞ ⎟ ⎟ ⎠ From the above four equations, we can see, for Type I zero-change, the comparison of speed of convergence of SV and lim δ = results in a real number, this δ →0 means the convergence speed of SV strategy is comparable to that of lim δ = When δ →0 the zeros values are replaced by a number which is at least two orders of magnitude smaller than the smallest positive numbers in the dataset, the calculation results from SV strategy would be of little difference from those of AL strategy 227 Appendix G Analytical Limits of LMDI I and Their Speed of Convergence For Type II change, M ij changes from to a positive number and all the other variables are positive For the factor of ∆C mix ,ij with zero changes, the AL strategy T T gives the limit of C ij − C ij = C ij To compare the speed of convergence of SV strategy and that of lim δ = , we have the following calculation: δ →0 T ⎛ M ij ⎜ lim L(C , C ) ln T ⎜M0 lim ∆C act ,ij − C ij δ →0 ⎝ ij δ →0 = lim δ lim δ T ij ij δ →0 ⎞ T ⎟ − C ij ⎟ ⎠ δ →0 T ⎛ M ij ⎜ ln 0 T T ln(Q T S iT I iT M ij U ij ) − ln(Q S i0 I i0 M ij δ ) ⎜ M ij ⎝ = lim Q S I M U − Q S i0 I i0 M ij δ T T i T i T ij T ij δ δ →0 ⎛ QT ln⎜ ⎜Q δ → ln(Q T S T I T M T U T ) − ln(Q S I M δ ) i i ij ij i i ij ⎝ Q S I M U / δ − Q S i0 I i0 M ij T = lim = lim δ →0 = lim δ →0 ⎞ T ⎟ − C ij ⎟ ⎠ (Q T (ln(Q T δ T i T i T ij T ij T T S iT I iT M ij U ij / δ − Q S i0 I i0 M ij ) ⎞ ⎟ ⎟ ⎠ ' δ T T S iT I iT M ij U ij ) − ln(Q S i0 I i0 M ij δ ) ) ' δ =∞ For the other factors, we have the following results: ⎛ QT T lim L(Cij , Cij ) ln⎜ ⎜Q δ →0 ⎝ = lim δ lim ∆Cact ,ij δ →0 lim δ δ →0 ⎞ ⎟ ⎟ ⎠ δ →0 ⎛ QT ln⎜ 0 ln(QT S I M δ ) − ln(Q Si0 I i0 ⋅ M ij ⋅ δ ) ⎜ Q ⎝ = lim T Q S I M U ij − Q Si0 I i0 M ij δ T T T T i i ij T T T i i ij δ δ →0 ⎛ QT ln⎜ ⎜Q δ →0 ln(Q T S I M U ) − ln(Q S I M 0δ ) i i ij ⎝ Q S I M U / δ − Q Si0 I i0 M ij T = lim = lim δ →0 = lim δ →0 228 (Q (ln(Q δ T T =∞ T T i i T T i i T ij T ij T ij T ij T T SiT I iT M ij U ij / δ − Q Si0 I i0 M ij ) ' δ T T SiT I iT M ij U ij ) − ln(Q Si0 I i0 M ij δ ) ) ' δ ⎞ ⎟ ⎟ ⎠ ⎞ ⎟ ⎟ ⎠ Appendix G Analytical Limits of LMDI I and Their Speed of Convergence Similar conclusion is applicable to lim ∆C str ,ij , lim ∆C int,ij , and lim ∆C emf ,ij This δ →0 δ →0 δ →0 means, the speed of convergence for all the factors with Type II zero-change is slower than the speed of convergence of lim δ = Therefore it is too slow and unacceptable δ →0 229 Appendix H Proof of Consistency in Aggregation APPENDIX H PROOF OF CONSISTENCY IN AGGREGATION For an aggregation analysis that can be performed through either one-step or multi-step procedure, consistency in the results is a desirable attribute We shall prove how the decomposition methods listed in Appendix C fulfill this property The proof is based on the work by de Waziers (2004) To illustrate, we use an aggregation structure of two levels: (a) the sectoral level (subscript i) refers to the first level of disaggregation which has a total of p sectors and (b) the sub-sectoral level (subscript j) refers to the second level of disaggregation which has total of li sub-sectors under sector i For each method, we first present the formula for one-step analysis, then present the formulae for the first step in two-step analysis If the aggregation function for the second step in two-step analysis makes the final formula identical to those of one-step analysis, this method is said to be consistent The Additive Methods The decomposition formula for the additive approach is given by: ∆X tot = X T − X = ∆X x1 + ∆X x2 + + ∆X xr + ∆X rsd where ∆X x1 , ∆X x and ∆X xr are explaining factors to the change of X , and ∆X rsd is the residual term for those unexplained factors 231 Appendix H Proof of Consistency in Aggregation In the Laspeyres index, ∆X x z from one-step analysis is given by: [ li p 0 one ∆X x z − step = ∑∑ x10,ij x 2,ij ( x T,ij − x z ,ij ) x r0,ij z ] i =1 j =1 When we calculate the same effect in two steps, we first calculate the value for each sector: li [ 0 two ∆X x z ,i− step (1) = ∑ x10,ij x2,ij ( xT,lj − xz ,ij ) xr0,ij z ] j =1 and then calculate the national aggregate indicator: p [ ] p li [ ] 0 two two one ∆X x z − step ( ) = ∑ ∆X x z ,i− step (1) = ∑∑ x10,ij x2,ij ( xT,ij − xz ,lj ) xr0,ij = ∆X x z − step z i =1 i =1 j =1 Thus, the Laspeyres index is consistent in aggregation in type of A2 in the additive p two two one form Similarly, using ∆X x z − step ( ) = ∑ ∆X x z ,i− step (1) = ∆X x z − step , we conclude that i =1 Paasche, M-E, S/S, AMDI, LMDI I, MRCI and Stuvel are all consistent in type of A2, i.e the aggregation function and the decomposition function are different, but the final result is consistent However, for LMDI II method, the situation is different In the one-step procedure p li ⎛ xT ⎞ one one ∆X x z − step = ∑∑ wij − step ln⎜ z ,ij ⎟ , the weight factor is given by the following ⎜ x0 ⎟ i =1 j =1 ⎝ z ,ij ⎠ formulae which involves an impact of a macro parameter: 232 Appendix H Proof of Consistency in Aggregation one wij −step T ⎛ X ij X ij ⎞ ⎜ 0, T⎟ L⎜ ⎟ ⎝X X ⎠ = p li ⎛ X0 XT L⎜ ij , ij ∑∑ ⎜ X X T i =1 j =1 ⎝ ⎞ ⎟ ⎟ ⎠ L( X , X T ) In the first step of the two-step procedure, this weight factor becomes: ⎛ X0 XT ⎞ L⎜ ij , ij ⎟ ⎜ X0 XT ⎟ i i − two step (1) wij = l ⎝ ⎠ L(X0, XT ) T i ⎛ Xij Xij ⎞ ⎜ , ⎟ ∑LL⎜ X0 XT ⎟ j =1 ⎝ i i ⎠ No linear function allows the same results in the single-step and the two-step procedure Consequently, LMDI II is not consistent in aggregation from a theoretical point of view The Multiplicative Methods The formula for the multiplicative decomposition is given by: r Dtot = X / X = ∏ D x z ⋅ Drsd T z =1 where D x1 , D x and D xr are explaining factors to the change of X , and Drsd is the residual term for those unexplained factors The Laspeyres index The basic formula for the Laspeyres method in the multiplicative approach is given by: 233 Appendix H Proof of Consistency in Aggregation ∑∑ (x li p D one − step xz = x 2,ij x T,ij x r0,ij z 1,ij i =1 j =1 ∑∑ (x p li 1,ij i =1 j =1 x 2,ij x r0,ij ) ) In the first of the two-step procedure, ∑ (x li D two − step (1) x z ,i = 1,ij j =1 x 2,ij x T,ij x r0,ij z ∑ (x li j =1 1,ij x 2,ij x r0,ij ) ) Thinking about a bridge to connect the above two equations, we notice that ∑ (x 1,ij ) x 2,ij x r0,ij represents the value of the aggregate indicator at the sectoral level, j and ∑∑ (x ) 0 1,ij ,ij i x xr0,ij represents the value of the same indicator at the macro level j Consequently, we can aggregate the sectoral indices as follows: p two D xz ,i − step ( 2) = ∑ i =1 X i0 two − step (1) one D x ,i = D x z ,i − step X0 z The aggregation function is a weight-based function of the sectoral indices Using this aggregation function allows the method to give consistent results However the residual terms calculated through this aggregation function are not consistent Hence, this method has consistency in aggregation of type B, i.e results for each factor are consistent, but the results for the composite index are not consistent because residual term is not consistent 234 Appendix H Proof of Consistency in Aggregation LMDI I The proof of the consistency in aggregation for LMDI I is given by Ang and Liu (2001) In the single-step procedure the formula is given by: D one − step xz ⎡ p li one − step ⎛ x T,ij z = exp ⎢∑∑ wij * ln⎜ ⎜x ⎢ i =1 j =1 ⎝ z ,ij ⎣ ome − step ij where w = ( T L X ij , X ij ⎞⎤ ⎟⎥ ⎟⎥ ⎠⎦ ) T L( X , X ) In the first step of the two-step procedure, the formula is: D two − step (1) x z ,i ⎡ li two − step (1) ⎛ x T,ij z = exp ⎢∑ wij * ln⎜ ⎜x ⎢ j =1 ⎝ z ,ij ⎣ two − step (1) ij where w = ⎞⎤ ⎟⎥ ⎟⎥ ⎠⎦ T L( X ij , X ij ) ( L X i0 , X iT ) Then we calculate the aggregate index using the same formula as the one at the sectoral level: ⎤ ⎡ p two two two D xz − step ( 2) = exp ⎢∑ wij − step ( ) * ln D x z ,i − step (1) ⎥ ⎦ ⎣ i =1 ( two where wij − step ( ) = ) L(X i0 , X iT ) L(X , X T ) Therefore 235 Appendix H Proof of Consistency in Aggregation D two − step ( ) xz ⎡ p ⎛ ⎡ li two ⎛ xT two ⎢∑ wij − step ( ) * ln⎜ exp ⎢∑ wij − step (1) * ln⎜ z ,ij = exp ⎜ x0 ⎜ ⎢ i =1 ⎢ j =1 ⎝ z ,ij ⎣ ⎝ ⎣ ⎡ p li one ⎛ x T,ij z = exp ⎢∑∑ wij − step * ln⎜ ⎜x ⎢ i =1 j =1 ⎝ z ,ij ⎣ ⎞⎤ ⎞⎤ ⎟⎥ ⎟⎥ ⎟⎥ ⎟⎥ ⎠⎦ ⎠⎦ ⎞⎤ one ⎟⎥ = D x − step z ⎟⎥ ⎠⎦ Consequently, LMDI I has consistency in aggregation of type A1, i.e the aggregation function and the decomposition function have the same formula, and the final results are consistent LMDI II Inspired by the proof for Laspeyres method, we create an aggregation function for LMDI II to make the results consistent For one-step analysis, we have: ⎡ p li one ⎛ x T,ij z one D xz − step = exp ⎢∑∑ wij − step * ln⎜ ⎜x ⎢ i =1 j =1 ⎝ z ,ij ⎣ ⎞⎤ ⎟⎥ , ⎟⎥ ⎠⎦ T ⎛ X ij X ij ⎞ L⎜ , T ⎟ ∑ ⎜X X ⎟ j =1 ⎝ ⎠ = T p li ⎛ X ij X ij ⎞ ∑∑ L⎜ X , X T ⎟ ⎜ ⎟ i =1 j =1 ⎝ ⎠ li one where wij − step For two-step analysis, we have D one where wij − step 236 T ⎛ X ij X ij ⎞ ⎟ ⎜ , L ⎜ X0 XT ⎟ i ⎠ ⎝ i = T li ⎛ X ij X ij ⎞ ⎜ ∑ L⎜ X , X T ⎟ ⎟ j =1 ⎝ i i ⎠ two − step (1) x z ,i ⎡ li two − step (1) ⎛ x T,ij z = exp ⎢∑ wij * ln⎜ ⎜x ⎢ j =1 ⎝ z ,ij ⎣ ⎞⎤ ⎟⎥ ⎟⎥ ⎠⎦ Appendix H Proof of Consistency in Aggregation If we choose the following aggregation function, the results will be same for onestep and two-step analysis: two D xz − step ( 2) T li ⎡ ⎛ X ij X ij ⎞ ⎢ ∑ L⎜ , T ⎟ ⎢ p j =1 ⎜ X X ⎟ ⎝ ⎠ * ln D two − step (1) = exp ⎢∑ x z ,i T p li ⎛ X ij X ij ⎞ ⎢ i =1 ⎟ L⎜ , ⎢ ∑∑ ⎜ X X T ⎟ i =1 j =1 ⎝ ⎠ ⎣ ( ⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ ) Other methods Regarding to Stuvel index method, van Yzeren (1958) proved it is consistent in type A1 in multi-step analysis Vartia (1976) also points out that Stuvel index is consistent in type of A1 The proof is omitted here as it is quite complicated For other multiplicative methods, it is clearly shown in de Waziers (2004) that the aggregation functions are either non-linear (for example Paasche method), or complicated (for example Fisher Ideal), or non-exist at the moment (for example M-E and AMDI) So we roughly consider them inconsistent in aggregation In conclusion, all the additive methods except LMDI II are consistent in aggregation of type A2 The reason for this particularity of the additive methods is that each change is mathematically independent of other changes at the same or higher level of aggregation; consequently, summing them allows the obtaining of changes at the higher level of aggregation As for the multiplicative methods, LMDI is the only method consistent of type A1 For the other methods, it may also be possible to develop some aggregation functions to make the results consistent, but these techniques are not so neat, sometimes rather cumbersome 237 Appendix I Energy Consumption and Activity Data for US Economy APPENDIX I ENERGY CONSUMPTION AND ACTIVITY DATA FOR US ECONOMY Table I-1 Energy consumption and activity for US economy, 1985 and 2000 1985 Level Residential (Million Households) Level 2000 Intensity (Energy/Activity) Energy (TBtu) Activity Intensity (Energy/Activity) 1882.32 9.01 208.94 2051.36 9.57 214.32 306.91 1.85 166.08 474.99 2.77 171.29 Mobile Home 94.04 0.69 135.71 96.24 0.64 149.32 Multi-Family (2-4 units) 460.02 3.17 145.21 388.95 2.73 142.64 Multi-Family (>4 units) 418.41 3.73 112.21 399.37 4.21 94.77 Single-Family Detached 2930.54 13.89 210.94 3806.76 16.39 232.23 Single-Family Attached 167.41 0.89 187.89 323.05 1.78 181.75 Mobile Home 187.40 1.12 167.03 216.96 1.17 186.22 Multi-Family (2-4 units) 427.42 2.71 157.95 353.61 2.06 171.87 Multi-Family (>4 units) 340.84 2.97 114.76 267.78 2.78 96.45 Single-Family Detached 3837.85 20.53 186.97 5609.49 25.03 224.08 Single-Family Attached South Activity Single-Family Attached Midwest Energy (TBtu) Single-Family Detached Northeast Level 4/5 220.16 1.22 180.33 488.69 2.83 172.80 Mobile Home 302.40 2.24 134.76 597.73 3.20 186.93 Multi-Family (2-4 units) 226.70 1.85 122.69 312.82 2.21 141.80 Multi-Family (>4 units) 431.19 3.70 116.67 517.67 4.54 114.12 239 Appendix I Energy Consumption and Activity Data for US Economy 1985 Level Level 2000 Activity Intensity (Energy/Activity) Energy (TBtu) Activity Intensity (Energy/Activity) 1740.64 10.06 172.94 2188.90 12.43 176.09 Single-Family Attached 63.12 0.49 127.56 268.26 2.27 118.23 Mobile Home 141.27 0.99 142.69 277.55 1.69 164.66 Multi-Family (2-4 units) 233.41 2.18 107.16 130.86 1.39 94.27 Multi-Family (>4 units) 337.33 3.50 96.44 393.48 5.03 78.20 11470.87 58.31 196.72 17196.31 75.62 227.40 Food Mfg 863.50 88567.00 0.01 1014.20 101145.00 0.01 Beverage and Tobacco Product Mfg 147.60 49414.00 0.00 96.10 23224.00 0.00 Textile Mills 195.60 12897.00 0.02 106.60 15682.00 0.01 Textile Product Mills 53.80 6945.00 0.01 23.70 8444.00 0.00 Apparel Mfg 31.40 24939.00 0.00 50.10 22451.00 0.00 Leather and Allied Product Mfg 11.80 5135.00 0.00 8.60 3871.00 0.00 Wood Product Mfg 382.20 41739.00 0.01 322.90 44130.00 0.01 Paper Mfg 2351.50 48144.00 0.05 2620.80 49954.00 0.05 76.60 100101.00 0.00 94.00 86635.00 0.00 Petroleum and Coal Products Mfg 1878.60 27276.00 0.07 1739.80 25498.00 0.07 Chemical Mfg 2346.80 98839.00 0.02 3548.80 184192.00 0.02 Plastics and Rubber Products Mfg 223.10 27287.00 0.01 331.00 59814.00 0.01 Nonmetallic Mineral Product Mfg 896.50 26141.00 0.03 983.10 39699.00 0.02 Primary Metal Mfg 1558.20 40437.00 0.04 1953.90 57397.00 0.03 Fabricated Metal Product Mfg 309.00 74453.00 0.00 420.70 99623.00 0.00 Machinery Mfg 243.00 64677.00 0.00 200.00 235313.00 0.00 Computer & Electronic Product Mfg Commercial (Billion SF) Industry Manufacturing (GDP 1996$) Energy (TBtu) Single-Family Detached West Level 4/5 131.10 36250.00 0.00 203.40 133820.00 0.00 Printing and Related Activities 240 Appendix I Energy Consumption and Activity Data for US Economy 1985 Level Level 2000 Level 4/5 Energy (TBtu) Activity Intensity (Energy/Activity) Energy (TBtu) Activity Intensity (Energy/Activity) Electrical & Component Mfg 117.10 32330.00 0.00 145.60 208682.00 0.00 Transportation Equipment Mfg 329.40 163926.00 0.00 489.30 172087.00 0.00 Furniture and Related Product Mfg 49.20 18787.00 0.00 96.20 24428.00 0.00 Miscellaneous Mfg 80.00 62244.00 0.00 86.00 61712.00 0.00 5683.01 471686.82 0.01 4972.21 646600.00 Non-manufacturing Automobiles 2577.21 3.28 3160.14 720.49 4.39 6129.78 1478.43 4.15 147.41 114.96 1.28 190.23 144.55 1.32 Scheduled carriers 1283.62 351.07 3.66 1824.29 708.42 2.58 General aviation 135.67 12.30 11.03 165.58 15.20 10.89 Commuter rail 10.69 6.53 1.64 13.79 9.40 1.47 9.99 10.43 0.96 12.11 13.84 0.87 Light rail 0.53 0.35 1.52 1.58 1.36 1.17 Intercity rail 10.01 4.79 2.09 11.48 5.50 2.09 Highway 2632.48 746.32 3.53 4468.00 1285.50 3.48 Rail 414.40 876.98 0.47 487.66 1465.96 0.33 Air 330.82 9.05 36.56 778.71 30.22 25.77 Waterborne 1007.21 964.91 1.04 1185.18 763.42 1.55 Pipeline Total 8444.06 Heavy rail Freight (Kton-miles) 3.79 Busses Transport 2188.83 Light-duty trucks Passenger (Billion Passengermiles) 8288.48 517.37 219.54 2.36 659.55 299.57 2.20 62128.09 80239.84 241 ... aggregate energy intensity change using decomposition analysis, and the related application and technical issues 24 Chapter Index Decomposition Analysis and Key Factors CHAPTER : INDEX DECOMPOSITION ANALYSIS. .. Related to Energy Efficiency Monitoring For each of the three phases of energy efficiency monitoring, namely energy- GDP ratio, IDA performed at sectoral level, and composite energy efficiency index, ... evolvement of energy efficiency monitoring and some methodological issues of decomposition and aggregation analysis applied to energy Regarding the evolvement of energy efficiency monitoring, the