Dryer Vessel Desiccant Charge.

[MacMcMacmac]

Hi folks. I recently opened up our desiccant dryer to investigate why the pressure drop across the beds has risen dramatically, and why I am finding a lot of dust in the after filters. I found that previous maintenance crews had installed a flat baffle plate on the top of the desiccant charge to help prevent a crater forming directly below the inlet nozzle. This plate was attached to a frame resembling a wagon wheel with 4 spokes. Within the spokes, sections of screen keeps the desiccant charge from stirring due to the airflow, while a layer of 5/8" catalyst support balls on top of the screen helps to break up the turbulent flow and air jet before it goes into the bed. It seems like one of the beds has experienced a backflow energetic enough to lift the bed. This has tilted the frame and it has settled back into the desiccant on an angle. Naturally, this has led to the air flow being directed to a restricted area of the bed, stirring up the desiccant, resulting in pulverized beads and dust formation. The catalyst support is nowhere to be seen. I think I will be forced to remove this frame and plate, as it is bent and the screens have ruptured. I was wondering if I would incur a backpressure penalty if I topped off the desiccant bed with a 32" deep layer of catalyst support, bringing it up almost flush with the 18" inlet nozzle. I was thinking about fabricating a perforated cone which would be inserted into the bed support to help the air jet penetrate and spread out through the be support. The top of the vessel would restrict the movement of the bed support and we would no longer need the screens and baffle plate, which was apparently added by our staff after the fact. It has worked well, and is still working well in the other tower, but the wrecked one is impossible to repair or replace by me hanging upside down into the vessel through an 18" opening. I realize the thorough repair would be to drain and sift the desiccant, then drop myself inside the tank and do the appropriate servicing, but time and budget are the restraining factors at the moment. FWIW I have never seen a dryer tower with so much dead space above the desiccant. I do not know why this was done, I can only think it was to provide a pulsation buffer for the incoming air when the compressor first came online.

 

[Montemayor]

Mac:

I don’t know the criticality of the dryers performance in your plant, but I sense an urgency over this issue. And it is, in my opinion, an urgent issue to resolve ASAP and with a final, successful solution.

Your adsorber unit seems to be a pretty large one if your inlet nozzle is an 18” size. This just makes the resolution that much more important. Unfortunately your description is hard to understand, and you haven’t furnished any flow diagram (to reveal the type and direction of flows, regeneration, bed numbers, etc.) nor have you given us fabrication drawings of the vessels. This leaves a lot of questions up in the air.

However, one basic issue that stands out is: you are suffering from some back-flow (probably up-flow) effect. Whether this is from regeneration stream or from the feed stream is unknown – since there is no flow diagram. Regardless, it is important to identify what, specifically could have caused the bed(s) to lift. For that, you should do a thorough operational analysis of the dryer unit. You should compile a complete (maybe a week’s operation) evaluation of how the unit is being operated throughout its NEMA cycle everyday. This may require some personal time at the unit – especially during the bed switching times – to identify flow and pressure peaks and surges.

The next thing to resolve is to install a “fix”. I have done this on several units in the field and the type and details of a fix depends on what you find as the cause, the internal superficial velocities, the potential magnitude of the bed lifts, and other flow surges that may occur. The basic fix is to always maintain your bed in a stable, fixed position. To do that, the original designer and specifier for the unit should have rigidly specified DOWNFLOW for the incoming humid feed air through the adsorption beds. This is always the preferred way to avoid future problems. However, the top of the bed should also be maintained in a sustained and stable location. For that, you require a mechanical method to “tie-down” the bed – even as it is inherently becoming lower and lower with time due to settling and attrition of the adsorbent through use and cycling. The type of mechanical restraint and hold-down device depends on the size, type and details of the vessel’s internal fabrication. To discuss this, a fabrication drawing is needed.

If, for example, your feed is fed in a down-flow direction and your regeneration is done in an up-flow direction, then the culprit for the bed lifting is the regeneration stream. When I have used humid feed gas as the regeneration fluid, I have always employed BOTH feed and regeneration fluids flowing in a con-current, down-flow direction through the adsorbent beds and never having to suffer any bed lifting.

So I believe that to recommend some experienced remedies to this problem, a lot of Basic Data is needed about the application and the way it is operated.

 

[Thealanator]

There are a few questions that come to mind:

Did the sudden pressure rise occur after any type of maintenance or upset?

When you do maintenance do you de-pressurize both towers?

Have you seen much dust coming from the atmospheric exhaust?

Have any vessel reliefs tripped recently?

How long has desiccant been in the tower?

 

[MacMcMacmac]

Hi guys. Sorry for the lack of info. Flow is indeed downward through the bed. Regeneration of the bed is upward, but flow is generated by a 15hp Roots blower which is encapsulated in its own pressure vessel and driven externally through a coupling and sealed at the input shaft with a mechanical seal. It could not by itself generate the pressure or flow to lift an 8500lb desiccant bed which is restrained at the top by screens and a frame which is interference fit to the inner wall of the vessel.

Switching towers is a manual operation on this dryer, and the operator has to trip a pressure equalizing valve before switching towers, otherwise, the higher pressure tower may violently feed the offline tower when the 10" gate valves change positions from offline to online and vice versa. Alternately, the main 8" pipe leading to the storage tanks may be at a higher pressure then the dryer when it is brought online in the morning and rapidly backfeed the dryer. The dryer is isolated from the entire system when it is shut down; all 10" inlet and outlet valves on both towers close. If there is sufficient pressure drop in the dryer overnight, the difference can be fairly high.

I think I shall propose that a new operating procedure be developed whereby the dryer and the pipe between it and the non-return valve on the main storage tanks is drained nightly to prevent such an upset in the future. There are hand valves on the outlet header which could be automated to dump the dryer nightly when the control mains are opened.

 

[MacMcMacmac]

And for thealanator:



Did the sudden pressure rise occur after any type of maintenance or upset? See the post above. I suspect someone switched beds without equalizing first.

When you do maintenance do you de-pressurize both towers? Yes.

Have you seen much dust coming from the atmospheric exhaust? None.

Have any vessel reliefs tripped recently?

Yes. This was on the regeneration pipe circuit which is held to 280psi or less due to the rating on the heater vessels being 300psi MAWP. Someone closed the vent valve from the pressure regulator trying to retain full pressure in the tanks overnight and forgot to open it again. I then was asked to order a and install a new safety valve to replace the "faulty" one. A 300psi safety valve pops at 300psi, so naturally it must be faulty...

How long has desiccant been in the tower? The last notes I have on any desiccant maintenance shows 1962. The dryers were opened up and screens renewed in 1985. The dust problem has come on fairly recently. I have worked in this plant for 4 years now and pressure drop across the filters has been consistent at 3psi up until a month ago. Curiously, this is about when a lot of cross training was begun, and "green" staff were learning the ropes.

 

[Thealanator]

It appears that your problem stems from a sudden de-pressurization.
Changing the relief valve is probably not a bad idea. If the relief valve saw the towers at the time of its trip, you can imagine what high-velocity dust and desiccant particles would do to the seats, soft goods, etc. Rebuilding/testing the removed unit will give you a good spare.

If this dryer is shut down nightly you might also look at any inlet headers between the dryer and compressor. If inlet pressure is low (compressor off) and a tower has pressure, you will get a backflow when you open the tower inlet valve.