Evaluation of energy saving potentials, costs and uncertainties in the chemical industry in Germany

In 2014, 19.3% of Germany's industrial final energy consumption could be allocated to the chemical industry. Energy efficiency measures with focus on the chemical industry could thus contribute to reaching the German goal of reducing greenhouse gas emissions. To achieve this goal, energy planners and industries alike require an overview of the existing energy efficiency measures, their technical potential as well as the costs for realizing this potential. Energy efficiency opportunities are commonly presented in marginal cost curves, which rank these measures according to specific implementation costs. Existing analyses, however, do not take uncertainties in costs and potentials sufficiently into account. The aim of this paper is to create a marginal cost curve of energy efficiency measures for the chemical industry in Germany, while quantifying the uncertainties of the results and identifying the most influential input parameters. The identification of energy efficiency measures and the quantification of the associated technical potentials and costs were identified based on literature data and own assessments. Based on these findings a cost curve was created for the current technical potential. This potential was found to be 24.4 PJ per year, of which 23 PJ had negative lifetime costs. To investigate the uncertainties of these results, Monte Carlo simulations were performed to quantify the standard deviations of the implementation potential and costs. Furthermore, a sensitivity analysis, based on Morris Screening and linear regression, was conducted in order to identify the most influential model input parameters. With the applied approach, it was shown that uncertainties have a non-negligible impact on the final energy saving potential and costs, as well as the shape of marginal cost curves. The standard deviation of the energy saving potential was found to be 3.1 PJ. Furthermore, it is possible to systematically prioritise efforts in refining data.