What is the typical reflux ratio for regeneration systems?

Is the 4% base for Yu Gong the sum of the flow rates of the concentrated lean liquid and the semi-lean liquid?

We don’t have semi-poor liquid. I first heard about 4% in 1998 from Shenglian Design Institute; later, a company named KBC came to our plant to assist us with energy-saving optimizations. The people who came were foreigners, and our reflux ratio was 4.5% (with centralized amine solution regeneration and sulfur-containing exhaust gases). Another regeneration system in our plant had a reflux ratio of around 3.5%. At that time, it was said that our reflux ratio was too high, while the other system was considered more appropriate. The actual reflux ratio should be determined based on the quality of desulfurization; in practice, it often drops to around 2.5%, while currently it is about 3%. As for the steam consumption per unit, since this parameter leaves a particularly strong impression on me, in most cases it is 70 kg or 80 kg per ton of amine solution; the lowest value recorded was 60 kg, while the highest values reached 100 kg or 110 kg. This figure also depends on the operating cycle of the plant.

Oh, according to Yu Gong’s calculation! Here, the reflux ratio is around 2-3%, and it’s not continuous either; in other words, this reflux is controlled based on the liquid level in the reflux tank. To maintain that liquid level, there is often almost no flow at all. (The numerical values are available, but I personally think they are false values, as I checked the situation on site), so I’m wondering whether it would be possible to implement intermittent operation – that is, to set a range for the liquid level in the reflux tank; the pump should start running when the level exceeds a certain value, and stop running when it falls below another value.

During operation in winter, to ensure the pump keeps running without freezing, the return line of the pump is kept open to maintain a stable liquid level in the return tank.

Thanks for the reminder! But the backflow volume here is indeed very small; even when the return line is blocked by a flow-limiting orifice, there is basically no backflow, so we simply stopped using it!