shell side of HX

[reactorshell]

i have a vertical shell and tube HX with a hot oil heating medium flowing on the shell side downwards. The HX was quite poorly designed, with baffles cuts of <10% and relatively small baffle spacing. My suspicion is that the shell side is not properly filled with the heating medium during operation.

Has anyone ever experienced the problem of a shell side of a HX not being fully filled with fluid, resulting in a reduced heat transfer to the tubes. How can I test out my hypothesis then?

 

[25362]

Do you suspect gases dissolved in the hot oil becoming released and collecting at exchanger shell top section ?

 

[srfish]

Since vertical heat exchangers are notorious for trapping non-condensibles under the upper tubesheet, have you tried venting ? hopefully there is a vent there.

Regards
Dale

 

[reactorshell]

My exact suspicion as well!!! But how would the gases get dissolved into the hot oil? that would be really difficult to prove?

There was a venting exercise done in the plant a few years ago. apparently, when the HX was vented, the heat transfer seems to have improved, indicating a better heat transfer to the tube side. However, once venting stopped, the temperature of the tube side fluid decreased again, indicating a lower heat transfer coefficient. This probably suggested that there must be continous venting to ensure a proper heat transfer within the HX. But this could also result in the possibility of loss of the hot oil from the vent? I was going to propose another venting exercise anyway. I will see what's the result this time round.

srfish,

What do you mean by non-condensibles? Do you have experience of this phenomenon? Or where can I get information on this?

Thanks everyone for giving your suggestions. Really appreciate these kind efforts!!

 

[quark]

It is always a standard practice to let in liquids from the bottom to avoid gas obstructing the passage of liquid. Main problem in your case seems to be vapor as suggested by others.

Does your return piping go upwards after heat exchanger? If so don't worry about partial filling. Secondly check for the quality of hot oil. In every sense providing venting arrangement is always better.

 

[reactorshell]

quark,

unfortunately, when the HX was originally designed for, the piping was such that the hot oil flows from the top down. In additionm the return piping does not go upwards after the heat exchanger and therefore we might expect a low pressure zone and therefore accumulation of vapours on the shell side. Venting therefore seems like the most feasible solution, short of makor hardware and piping changes. But is venting the only solution to the problem?

 

[25362]

The hot oil flow should be controlled under back-pressure from a CV downstream the H/E. This back-pressure would help in keeping the dissolved gases from disengaging, although turbulence may be a factor. Venting, if at all, would have to be carefully done to avoid safety problems and oil losses.

Uncondensibles and other gases can originate from progressive oil degradation (ie, thermal cracking). The balance tank may absorb inert gas from the blanket or air if vented to the atmosphere. In long-time use gases may accumulate up to 10% w/w.

Thermal decomposition of mineral oils is usually accompanied by an increase in the hot oil viscosity. What is the type of hot oil fluid used ? In oxidative degradation solid particles may form, it would be practical to filter a side stream continuously with fiberglass cartridges provided the particles are not too fine. Check acetone insolubles (as char and loosened scale) to keep the oil out of erosive conditions. Higher TAN numbers may indicate oxidation or hydrolysis of the oil and possible corrosiveness.

In any case samples should be representative: taken from points of good fluid circulation, neither from unflushed sight glasses nor from leaks. Good luck !

 

[Montemayor]

reactorshell:

Quark is exactly correct. What you describe is a classic example of how NOT to install a vertical S&T exchanger. The introduction of all liquid fluids into the bottom nozzle of heat exchangers and promoting a positive, upward flow that expels any non-condensables is a basic requirement of the process that all too often goes unheeded by the inexperienced and naive engineers who probably slept through Heat Transfer 101 in college. The basic requirements for establishing a decent Overall Heat Transfer Coefficient (&quot;U&quot is that of establishing the proper film coefficients on BOTH sides of the heat transfer surface. This means that if you are basing your self on a liquid film coefficient, then it is your responsibility to ensure that a LIQUID phase is present there 100% of the time. Any gas film that contaminates the liquid surface will corrupt the heat transfer rate down to practically nothing. Gas heat transfer coefficients are relatively insulators, rather than conductors - in fact, static gas is the ingredient that gives insulation material (i.e., Polyurethane) its excellent insulation propterty: zilch heat conduction. I am reciting basic heat transfer theory merely to reinforce the basic requirements of a heat exchanger. You must not allow the prescence of inerts. What you are experiencing is not a phenomena. It is a direct result of a bad installation and the bad results are to be expected - they don't come as a surprise. The solution to this problem is to repipe the exchanger in the proper, logical manner: shell flow ingressing from the bottom and out at the top. I have seen this mistake done in the field with rather large exchangers (>230 ft2) and resolved the issue as I've described. The very un-orthodox baffle window cut (10%) is another amateurish mistake in my opinion. A cut of 20-25% is the norm and I've never heard or seen any justification for using such a ridiculously small cut. This, and your close baffle spacing, can only lead to increased turbulence and pressure drop which may influence the release of dissolved gases. It certainly doesn't help.

Take a serious look at your tube-side fluid as well. How is that piped? Don't forget: A vertical unit doesn't only give you shell-side piping constraints; it also applies to the tube-side. In other words, if you have a multiple pass tube-side unit, you are confronting the same gas accumulation problem whether you know it or not. Think about it. The geometry of the apparatus leaves you no alternative but to accept the inevitable traping of any gas in the tube side when you are making multiple vertical passes.

I personally have never accepted the idea of having to continuously &quot;vent&quot; a vertical unit. I have never seen any application where this is done successfully or cost effectively. It's easy to say that you are going to vent, but how do you propose to do this? Continuously? Every so often, manually? From an operational point of view the second alternative is a pain in the backside and I don't trust a fellow human to do this without fail. The first alternative involves equipment and instrumentation suitable to the specific unit and operation. I've found that it's much simpler and effective to relocate the unit in a horizontal configuration and be done with it.

This is an application in heat transfer that unfortunately is seldom discussed - primarily because it doesn't have a clear-cut resolution nor explanation for logically expecting such a geometric orientation to work. People just simply expect the application to work without applying engineering analysis to it. I wish you luck.

 

[25362]

All that has been said concerning vertical liquid-liquid exchangers, gas trapping, and the difficulties attached to their venting are entirely right. We still don't know what HT fluid is being used and whether the tube-side stream is boiling or not.

Speaking of gases, a fresh (thermally sound) mineral oil may have a vapour pressure of 14.7 psi at 600^oF!
Besides, lights or gases contained in the hot oil should be ventable from the system expansion tank or from system vents.

If you have gases as a result of thermal degradation of the hot hydrocarbon circulating oil, one should expect solids depositing on colder surfaces impairing the HT. That's one reason to check a sample with acetone in the lab. Besides, the HT fluid viscosity may change with time.

Of all thermo-physical properties of a liquid HT fluid, viscosity is the one affecting HE performances the most. HTC are inversely proportional to the viscosity to the 0.4 power. Viscosity may change in one single exchanger in as many as four times (as from 5 to 20 cP), thus the HTC can drop by a factor of 1.7 specially if the fluid is an aromatic-rich mineral oil.

One more point: when generalizing on gases having lower-than-liquid HT coefficients, one must remember that hydrogen gas films may show equal or better convection HT coefficients than liquid hydrocarbons especially at the temperatures (300^oC or more) encountered in hydrotreaters.

Hydrogen has a thermal conductivity more than 6.5 times higher than for air and 2.5 times higher than for liquid HC oils. Therefore, hydrogen-bearing liquid streams don't necessarily suffer from low HT convection coefficients at high temperatures. As a curiosity: air and nitrogen are also sometimes used as HT fluids, but they need extreme pressures to achieve good HT.

Aside from changing the HE altogether as suggested, it can be summarized:

1. Don't thermally crack the hot oil in the heater.
2. Vent safely uncondensibles and lights formed through the system expansion tank and other (&quot;relatively cold&quot vents in the circuit.
3. Filter a by-pass stream &quot;in line&quot;.
4. Check the hot oil fluid in the lab regularly, taking representative samples.
5. Keep a back-pressure on the vertical &quot;consumer&quot; to avoid, as much as possible, the release of lights or uncondensibles.
6. At every overall maintenance shut down, clean the system from coke and deposits.
7. When starting up the vertical unit, and other parts of the hot oil circuit, vent any air, and drain any water, collected using cold oil flushing.
8. Start hot oil circulation gradually to minimize mechanical thermal stresses.
9. Remove leaks.
10. Follow safe &quot;emergency&quot; or &quot;planned&quot; shut-down procedures.

I hope I've been of help. Good luck !

 

[25362]

In addition to wahtwas said before, to release lights from the returning hot oil thru your expansion balance tank, one needs either a separate line discharging in to the tank (under an inert gas blanket) or a large balance line. In the first case, the circulating pump receives its feed mainly from this tank.

If, on the other hand, the circulating pump is connected to the tank only as a make-up provision, this line would have to be sufficiently large to allow for the disengaging of lights coming with the returning hot oil, while liquids move down to the pump as make-up (for example, to cover leaks and oil losses).

If this is the situation, and the make-up line is not large enough, gases and vapours may stay with the circulating hot oil, and collect at high points in the circuit.