Application of Adjoint Sensitivity Analysis for Performance Enhancement of Power Plants’ Nitrogen Oxides Control Policies

Significant investments in nitrogen oxide (NOx) emission controls in the U.S. have led to a substantial reduction in emissions. However, it is unclear whether these programs have optimally reduced ozone concentrations and their corresponding health damages. Current cap-and-trade program allocates emission quotas to participants and allows the trade of quotas on a one-to-one basis. However, it does not account for spatial and temporal differences in health damage of NOx emissions. This shortcoming in the current U.S. NOx control policy is explored in this research. Spatial and temporal differences in NOx emissions can be included in policy design if emission quotas are valued differently (exchange rate policy) or if polluters pay time- or location-specific emission fees (taxation policy). The main objective of this work is to develop a decision support system model for evaluating different policies. The proposed model includes an optimization platform to predict the polluters’ behavior, and an air quality model and its adjoint (or backward) sensitivity model to calculate the derivatives of the environmental or health damage function with respect to NOx emissions used for emission differentiations. The results from a case study of U.S. power plants show that exchange rate trading outperforms current indiscriminate trading policies. These findings imply that by implementing exchange rate trading or taxation policies, current improvements in air quality could have been achieved at lower costs, or alternatively, more substantial improvements could have been reached at little to no additional costs. Furthermore, the results indicate that setting the emission fees on an hourly basis leads to a outcome comparable to setting fees based on location. Moreover, the per ton health benefit of NOx emission reductions is found to increase as emissions are reduced. This finding is particularly important from an environmental policy perspective as it impacts the optimal NOx emission reduction target. Our results also indicate that power plants in the restructured electricity market are willing to pay more for emission quotas. Uncertainties involved in the proposed model, challenges for implementation of the proposed policies, and inclusion of health impacts caused by exposure to particulate matter are main directions for future research.