Continuous monitoring of the performance of a NOx reduction system using electrochemical sensors

Abstract
The efficiency of a NO_x removal system is continuously measured using a TüVapproved analyser coupled to a high-temperature sampling system. Both NO and NO₂ are measured directly using electrochemical sensors without the need to convert NO₂ to NO. It is important to measure both components separately since the ratio of NO to NO₂ in stack gas is much smaller after NO_x removal. The sampling system has been designed to give fast response and to minimise the possibility of sample loss. Identical analysers are installed both before and after the NO_x removal system so as to reduce the likelihood of systematic errors.

Introduction
Stringent emissions legislation throughout the world requires operators of many combustion processes to monitor continuously the concentrations of a range of stack gases . Among the most important of these are NO and NO₂ (collectively NO_x). The very low levels of NO_x emissions now required cannot always be achieved by the use of low-NOx combustion methods alone and so some form of scrubbing must be used to reduce NO_x levels. Selective catalytic reduction (SCR) is widely used in Europe and is gaining increasing acceptance in the USA. This method uses ammonia injection in the presence of a suitable catalyst to convert the NO_xto N₂ and H₂O. Depending on the application and the amount of ammonia slip that can be tolerated, NO_x reductions between 80% and 90% can be achieved (1).

A number of methods are available for measuring NO and NO₂, the most common being the chemiluminescent NO detector (2) which can be coupled to a thermochemical NO_x converter to give a total NOx measurement. The typical efficiency of the NO₂ to NO conversion process is 90 to 95% when the converter is new, declining to 80% or less as the converter ages. The resulting error in total NO_x is small where the NO/NO₂ ratio is large, as in a traditional combustion application without deNOx. However where the ratio is close to unity, as is the case after an SCR process, the error becomes unacceptably large and concern has been expressed (3) that NOx converters cannot meet the accuracy requirements for low-level NO_x measurement.

For this reason, it is preferable to use a method which measures both NO and NO₂ directly. Optical methods using ultraviolet spectroscopy (4) have proved successful but are rather expensive. In-situ infrared measurements of NO and NO₂ are made difficult by the strong interference from water vapour at the principal absorption wavelength (5) whilst extractive IR measurements can suffer from baseline drift.