Optimization of a retrofit urea-SCR system.

摘要

Oxides of nitrogen (NOx) emissions from legacy diesel engines are often many times over currently mandated standards, contributing significantly to degradation of air quality and negative environmental impact. A retrofit urea-SCR (selective catalytic reduction) aftertreatment system offers a viable solution for reducing NOx emissions from older diesel engines. A stand-alone retrofit urea-SCR aftertreatment system was previously developed by West Virignia University (WVU) engineers, implementing a pre-SCR NOx sensor, open-loop feed-forward control, and stoichiometric NOx reduction logic. During experimental testing at WVU, the urea-SCR system demonstrated NOx reductions of 2% to 53%, depending on the test cycle. In order to optimize the system, this dissertation considered additional control configurations. To evaluate the emissions performance of each control strategy, a neural network heavy-duty diesel engine model was developed along with separate four-state chemical and thermal SCR catalyst models. Each model component was validated with experimental data recorded from the WVU Engine and Emissions Research Laboratory (EERL). The following control configurations were considered: (1) pre-SCR NOx sensor, open-loop feed-forward control, (2) post-SCR NOx sensor, closed-loop feed-back proportional-derivative (PD) control, (3) pre- and post-SCR NOx sensors, closed-loop feed-back proportional-integral-derivative (PID) control, (4) pre-SCR NOx sensor, model-based control.;The evaluation process considered differences between a highly instrumented and highly engineered system. Emissions performance was evaluated over two transient on-road test cycles (FTP, ACES HHDDT_S) and one steady-state marine test cycle (ICOMIA E5), implying broad applicability of the aftertreatment system. The evaluation process was characterized by overall NOx reduction percentage, maximum ammonia slip in parts per million (ppm), and average ammonia slip (ppm). The complexity of the sensor configuration and control strategy calibration were evaluated, as well as how adaptable a given configuration was to variations in engine behavior and sensor measurement accuracy. Finally, total cost was compared between each control configuration, considering system capital, maintenance, operation, control strategy engineering, and system calibration. A final cost per ton of NOx reduced was presented for each control configuration, assuming a six year operational cycle in marine and on-road applications. Based on the collective emissions, complexity, and cost analyses, a configuration implementing pre- and post-SCR NOx sensors and closed-loop PID control was identified as optimal for a retrofit application. Model results demonstrated NOx reductions of 44%, 53%, and 47% over FTP, ICOMIA, and ACES High-Speed Cruise (HHDDT_S) cycles, respectively. The total annual NOx reduction cost was