With regulatory changes for reducing carbon dioxide emissions likely coming into force, many plants will revisit energy integration to reduce net heat use. One oft-mentioned idea is to add inter-reboilers to towers. This allows for using lower-temperature heat sources to provide part of the duty to the tower. The objective is to usefully recover lower-temperature heat instead of rejecting it to air or water.

Figure 1 shows a product splitter with two inter-reboilers added. The first (E32AB) used two small exchangers with the plant’s low-pressure steam — which otherwise would be vented — to provide heat to the tower. This reduced hot-oil duty demand by 3.5 million Btu/hr. The inter-reboiler worked well.

Because the first heat-integration step succeeded, the plant added a second inter-reboiler (E33) heat integrated with the bottoms stream. When attempting to start up the new exchanger, nothing happened.

Plant operations staff puzzled over this for a while but never resolved the problem before other demands took priority. The unit was allowed to sit with the exchanger piped up but not operating. After some months the curious failure of E33 to work was nearly forgotten.

As winter approached, an unexpectedly sudden storm dumped a surge of rain water on the overhead fin fans. Unit pressure widely swung. After re-establishing control, a plant operator noticed that the return line to the tower from E33, up to now cold, was hot. Investigation showed that E33 was working.

With E33 working, staff made a second curious observation — the net hot-oil duty required in E24AB had barely changed. Reduction in required duty was so small that it was nearly impossible to detect.

So now we have two mysteries: Why did E33 not work but then suddenly work? And why was no net duty saved?

To grapple with the first question, let’s more closely look at Figure 1. Notice that the E33 outlet nozzle on the return stream to the tower is 32 ft below the tower return nozzle. If at startup (or any other time) the return line to T07 is full of liquid, how does this impact the liquid in the tower side of E33? The 32 ft of liquid gives a static head of 7 psi. The higher pressure raises the bubble-point temperature of the liquid by 40°F. However, bottoms temperature is only 30°F higher than tray 50 temperature. So, the static head in the lines down to E33 prevents vaporization on the tower side of the system.

Sudden pressure drop induced by the rain storm lowered operating pressure below bubble-point pressure of the liquid in E33. Once some vaporization started, density in the return lines fell, reducing static head to E33. Thus, E33 will work once it’s been started. It just needs help starting up.

Some plants do start up inter-reboilers by suddenly dropping tower pressure. Other, more controllable alternatives, use either a temporary inert gas injection into the return line or take some distillate product that vaporizes easily to start the liquid circulating in the side reboiler (see Figure 2).