Real-World Full-Scale Applications

One of the main challenges in this type of model applications is the correct estimation of NH3 concentration at the entrance of the catalyst, which is used as an inlet parameter to the model. The problem arises from the fact that NH3 is sometimes not directly measured at catalyst entrance but rather estimated based on the instantaneous injection of urea-water solution in the exhaust stream. 

In addition, the unifonnity of NH3 concentration at catalyst entrance is not always perfect which obviously challenges the use of single-channel approaches. 

An example of model validation on standard European test cycles (ESC and ETC), performed by Tronconi et al. [29], is shown in Fig. 13.9. An excellent correlation between the NO, concentrations downstream of the catalyst is observed with an error of 3-4 % for the total NO, conversion in both cycles. 

---