OCR Text |
Show 3. NEW SYSTEM DESIGN (cont'd) Control Method (cont' d) Feedforward involves the calculation of the actual amount of NOx present (in moles) and injects the amount of ammonia to meet the setpoint outlet NOx value. This control method is based on simple inputs of fuel gas, excess oxygen and NOx concentration. The last two require that a NOx and oxygen analyzer be installed upstream of the SeR unit. The method works relatively well with the exception of the need for manual ratio adjustment if the desired NOx level in the stack is not being obtained. This adjustable ratio (initially set at 1: 1) is supplied to correct for any changes in assumptions made for the calculation blocks in the DeS. With feedback control, the NOx and oxygen is simply measured in the stack, and ammonia is continuously injected until the setpoint value is reached. Feedback is obviously a simpler control scheme except for the fact that it doesn't look at the ammonia slip. For example, in the case of a malfunctioning SeR unit, the feedback system will keep increasing ammonia until the NOx is achieved regardless of the ammonia slip. With feedforward, the operator would manually increase the ratio to maintain the outlet NOx hence becoming inherently alerted to the fact that "something is wrong". Finally, and most importantly, a typical SeR unit has an inherent lag in response to a given change in conditions. This hinders the use of feedback. For example, a change in conditions may result in the feedback control telling the ammonia to increase. As the results may not be instantaneous, the feedback control will continue to tell the ammonia to increase until it overshoots the setpoint, possibly saturating the catalyst. At this point the NOx is at a minimum, however ammonia slip is occurring. The feedback control now tells the ammonia to decrease and the same happens, it overshoots the setpoint. The system will eventually stabilize as long as the heater conditions remain fixed. If the conditions keep fluctuating, the system may not stabilize within the period necessary to stay in compliance. Feedforward on the other hand will only inject enough ammonia based upon the calculation performed, lending itself better for this type of control. Ammonia injection feedback response is currently being studied to get a better understanding of the possible ways to use it for control. If feedback can be controllable it would be the ideal replacement of the manual ratio setpoint adjustment used in feedforward. This would be the feedward with feedback reset method. 10 |