Intelligent Realtime Control to Reduce Ammonia Slip
Publication Date: 2005-Aug-30
The IP.com Prior Art Database
INTELLIGENT REALTIME CONTROL TO REDUCE AMMONIA SLIP
This invention describes a method to reduce ammonia slip using an intelligent real-time control scheme.
Ammonia (NH3) is used as a reagent to convert oxides of Nitrogen (NOx) in combustion flue gases by Selective Catalytic Reduction (SCR) process. Figure 1 shows a typical SCR process. The injection of NH3 must be quite precise to yield a desired NH3/NOx mole ratio. Excess ammonia, in addition to waste of the reagent, has detrimental effect due to fouling of the air preheater surfaces caused by the formation of ammonium sulfate or bisulfate, deteriorates the quality of fly ash making it unsuitable for use, and increases the stack opacity above the regulated limits. This process is used for systems using solid (coal, biomass, pet coke, etc.) as well as liquid fuels (oil, etc.). However, for coal-fired utility boilers SCR process is specifically useful due to comparatively large amounts of NOx produced.
The coal feeding system in the utility boilers typically does not lead to a uniform feed rate and causes surges of high and low coal concentrations in the primary air stream. In addition, utility boilers usually have multiple burners ranging in number from a few to over a hundred. With the existing technology, it is quite difficult to feed all the burners with same quantity of coal. As a result the amounts of NOx generated fluctuates significantly both as a function of time and over the cross-section of the boiler. The load of these utility boilers also varies according to demand and usually a change in load is accommodated by taking a few burners in or out of service. This also leads to a transient behavior of NOx emissions especially during the step change.
The control of ammonia injection poses difficulty due to long lag/response times needed for NOx measurements. A number of solutions have been proposed both based upon control schemes as well as on simulation based schemes (Boyden et. al., 2003) but ammonia injection in proper quantity at the right time and at the right location still remains a challenge.
BRIEF DESCRIPTION OF THE FIGURES
Figure 1 shows a schematic flow diagram of SCR process.
Figure 2 is a schematic of the invention.
Figure 3 shows a Neural Network using temperature measurements from video camera to predict NOx concentrations in real time.
Figure 4a shows a NOx mapable grid for a four-port ammonia injector.
Figure 4b shows a NOx mapable grid for an eight-port ammonia injector.
Figure 5 shows the use of ratio control to calculate ammonia manipulations.
Figure 6 shows feedback and ratio controller to improve SCR performance.
This innovation proposes use of a video sensor in conjunction with a controller in the area of artificial intelligence, such as Neural Networks (NN), coupled with feed forward control that will yield a quick response and precise ammonia injection when and where it is required. The NN will be populate...