Don’t skimp on back pressure.
In today’s world of advanced HRSG design for fast start, efficiency, clean emissions an old nemesis has re-emerged; gas side pressure drop. For decades the effect of too much pressure drop on the gas turbine has been known. Not only can too much delta P thru the exhaust system cause them to go offline but power production also suffers. HRSG manufacturers always are always limited, by customer guarantees, in the amount of back pressure they are allowed to impose on the turbine.
The new problem is that in some of the advanced designs being implemented there tends not to be enough back pressure provided on the system. In the conservative realm of engineering all parties involved in a power plant design tend to over estimate how much pressure drop is taken by their equipment in Evaporator sections, pollution catalysts, silencers and economizers.
SCR and CO catalyst manufacturers typically operate 25-40% below their guaranteed pressure drop numbers. And in recent years they have developed ultra-low pressure drop module designs. These are great for retrofits into existing sites, but as we are seeing new installations seem to be suffering if not applied properly.
Even those designs utilizing CFD models on the gas side flow are missing the mark by as much as 40%. I know that computers don’t lie, but the old adage “bad stuff in equates to bad stuff out” seems to apply. Two recent HRSG installations were examined for failing emission guarantees. During evaluation it was found that the back pressure on one unit had a guarantee of 12″ WC back pressure, but at full load was only taking 6.9″. The other had an 8″ guarantee and was operating at 5.3″.
In the new world of emission requirement of 2 ppm NOx and 2ppm ammonia slip the effects of not enough back pressure become very evident. Velocity profiles are critical in SCR catalyst performance and all guarantees have a velocity caveat associated with the the design. In fluid dynamics class engineers become familiar with the convex profile of flow thru a pipe, friction along the walls slows the flow near the walls. In the case of HRSG duct work there is often not enough pressure drop thru the components in the gas path to ‘flatten’ this profile. The result is high velocities in the center and bad emissions in the stack.
What’s worse is almost every HRSG design has one 90 degree turn in its duct as it turns and goes up the stack. In a low DP situation this results in chocking at the top of the turn and acceleration in the bottom of the turn. Stack height can also magnify this effect because of stack draft.
In one of the evaluated installations mentioned above a CFD was performed on the design. It predicted that the velocity profile at the point of ammonia injection was +/- 1% RMS of the mean. At the SCR catalyst face the profile had degraded to +/- 13%. This should have raised a red flag as the profile became worse in a straight empty duct. The +/- 13% was within the catalyst manufacturers requirement, so it was assumed to be acceptable.
Do to the lower than anticipated pressure drop thru the HRSG the field measurements taken revealed a +/- 17% maldistribution of the velocity at the catalyst face. Field data also showed that the velocity profile downstream of the catalyst exceeded +/- 20%. Both these profiles were outside the +/- 15% required by the catalyst manufacturer.
The fix? Add the missing back pressure using perforated plates; and/or adding turning vanes in the turn. Both expensive field modifications.
Good practice going forward is to use expected pressure drop values in the design rather than guaranteed values. Ask vendors and engineers to provide both expected and guaranteed pressure drop values. Potentially this approach can actually save money on the equipment as likely the duct cross section size can be decreased slightly and denser tube arrangements can be used. It will most certainly save some headaches in the field.
Pollution abatement is a reality of this age. No one really wants to buy any as it has a negative impact on the bottom line. Paying (pun intended) close attention to its implementation can avoid adding insult to injury with unnecessary field issues. Watch the Delta P, and keep the customer happy.
