Learining in carbon accounting and energy efficiency

Abstract

In the context of energy use and greenhouse gas emissions, the manufacturing industry plays a dual role, namely (i) as consumer of energy and (ii) as producer of energy-consuming technologies. Although the manufacturing industry has been subject to extensive research, knowledge gaps still exist regarding (i) a detailed and accurate accounting of the use of fossil fuels for non-energy purposes (e.g., as feedstock in the chemical industry) and the resulting emissions as well as (ii) with respect to the rate at which cost decline occurs for novel and efficient energy demand technologies. In the first part of this thesis, we develop bottom-up models that quantify non-energy use of fossil fuels and related emissions at various levels of detail. We quantify in Chapter 2 worldwide non-energy use in the year 2000 to be 20 2 EJ, thereby accounting for 5% of the global total primary energy supply. In Chapter 3, we estimate that in the year 2000, non-energy use leads on a global scale to emissions of 700 90 Mt CO2. The forty industrialized Annex I countries account for 51% (360 50 Mt CO2) of this total, where as the remaining non-Annex I countries account for 49% (340 70 Mt CO2). Chapter 4 shows that yearly non-energy use emissions in Germany increased on average by 1.8% per year between 1990 and 2003. Our non-energy use emission estimates help closing major gaps in the German GHG inventory. We conclude that our bottom-up models allow a more detailed, complete, and accurate accounting of non-energy use emissions in comparison to existing methodologies. To address the second research question, we apply the experience curve approach to model prices (as proxy for actual production costs) of energy demand technologies as power-law function of cumulative production. We quantify the change in productions costs by so-called learning rates (LR), which represent the percentage of cost decline with each doubling of cumulative production. In Chapter 5, we estimate that prices of condensing gas combi boilers decline at a learning rate of 14 1%. For large appliances, we find in Chapter 6 a robust long-term price decline for wet and cold appliances at learning rates of, on average, 29 8% and 9 4%, respectively. These results are in line with the findings of our literature review (Chapter 7), where we identify that learning rates of energy demand technologies are approximately normally distributed around a mean of 18 9%. By modelling the specific energy consumption of large appliances as function of cumulative production, we extend the conventional experience curve approach and quantify technological learning from a broader perspective. The resulting energy experience curves indicate a robust decline in the specific energy consumption of wet appliances (LR of 18 3% to 35 3%) and cold appliances (LR of 13 3% to 17 2%). We conclude that our results indicate that technological learning helps to achieve a more efficient use of energy at declining costs.