Mechanical Seal Face Distortion

[sinbj1]

All,

I am experiencing repeated mechanical seal failures on double suction, radially split, between bearings pump. The operating conditions are listed below.

Fluid – Hydrocarbon
Operating Temperature – 485F
Suction Pressure – 30psig
Discharge Pessure – 200psi
Barrier Fluid – Royal Purple 22
Seal Pot Pressure – 100psig

The seal design was recently changed to a dual pressurized face to back rotating bellows seal. A set of seals were installed and shortly after the pump was started the thrust side seal began to leak barrier fluid in to the process. The coupling side seal never leaked. The seals were pulled and disassembled to find the thrust side seal had two wear spots that were 180o apart from one another (saddle distortion) on the inboard or primary stationary face. The seals were repaired and re-installed in the pump. Just as before, shortly after the pump was started the thrust side seal began to leak barrier fluid in to the process. The coupling side seal did not leak. Just as before, the same saddle distortion was found on inboard stationary face.

The face distortion issues have been addressed by using a stationary bellows as opposed to a rotating bellows in order to isolate the face from the gland. Any thoughts as to why the inboard or primary stationary faces experienced face distortion? Also any thoughts as to why just the thrust side seals experienced the face distortion? Any help is greatly appreciated.

 

[bk19702]

Were these high spots 180 degrees apart on the stationary face? Two high spots such as this on the stationary face are usually indicative of the ring being distorted mechanically. The high spots would correspond to disruptions in either the mating surface in the seal gland or seal chamber itself. If neither area was checked for flatness / parellelism (gland or seal chamber face), then you would most likely see the same failure mode since the variables were not changed for the thrust end seal. Switching to a stationary bellows design would require a redesign of the gland, so if the previous rotating bellows gland was to blame, it has been removed from the equation now with the alternate seal design.

 

[JJPellin]

You really didn't specify if this was a full cartridge seals and if you were replacing the entire cartridge each time. If so, it is less likely that you just had a bad gland surface. I am more interested in the design of the seal and the seal system. If this is a tandem arrangement, pressurized seal, then the inner seal has ID pressure. The configurations that I am familiar with would require the stationary face be retained with a snap ring or other method to keep it from popping out with ID pressure. If this snap ring (or other retaining method) was not providing a flat surface, then the face could be distorting against this retainer when under pressure. Did you check the faces to see if they were still distorted (optical flat) or are you basing the evaluation of the distortion purely on the noticeable wear pattern? If the faces are flat on the bench, but obviously distorted in operation (based on wear pattern), then bk19702 is probably correct. The surface adjacent to the face might not be flat and the face is distorting to conform to this surface. There is also the chance of thermal distortion. If the barrier fluid is cooled and directed at a single location on the stationary face, it could be distorting in service. You could be over-cooling the barrier fluid on one end of the pump. Are there cooling coils in the seal pots? Is the cooling water piped in series or parallel to the two pots? Does one pot run much cooler than the other? What is the face material? How is the face mounted (o-ring groove, clamp, block, etc.)? How is the face retained for reverse pressure?

As far as the difference between thrust end and radial end problems, I would have other questions. How is the pressure between the seals maintained? With double suction impellers, the seal chamber pressures should be the same (suction). But, is the barrier fluid pressure controlled to the same setting on both seals?

Johnny Pellin

 

[sinbj1]

* It is a full cartridge seal. The intial install replaced the the entire cartridge seal. The second install just repaired the first set and re-installed them.
* The inner faces do see ID pressure. The stationary face is retained with a retaining ring that is held in place with a snap ring.
* The faces were distorted under the optical flat after the fact.
* There are cooling coils on the seal pots. Each seal pot has a separate connection to the cooling water header.
* Each seal pot runs within 10 degrees of one another
* The face materials are silicon carbide vs. silicon carbide.
* The stationary face is a L-type with flexible graphite.
* Pressure between the seals is maintained with a N2 blanket applied to the seal pot (API plan 53A). The barrier setting is the same for each seal pot.

The frustrating part in all this is that the same seals were installed in the sister pump a year and a half ago and are running with no issue. This lead me to suspect something was wrong with this particular pump. The machinist swept the box face and it measured a variance of .001" perpindicularity. My thought was that when the pump was at operating temperature there was some case distortion being transmitted to the seal gland and inturn to the seal face. Specifically the case distortion is occuring on the thrust side of the pump.

Since the install of the stationary bellows design the seals are running fine. I just want to better understand why the faces were distorting in the first place.

 

[JJPellin]

I suppose it could be a problem with the pump. I have never seen a pump that distorted so badly as to permanently distort the seal face. Some pumps of this type are set up for thermal growth with alignment pins and keys along the bottom of the case or the support feet. Some use bolts with sleeves to allow for thermal growth, lock nuts or deliberate low torque to allow the case to slide under the hold down bolts. Extreme pipe strain could distort the pump. A spring can with the stops still locked. A pipe shoe that dropped off the beam. Some hot pumps use a warm up line or drilled check valve to warm the pump up before starting. Could any of these issues be messed up and causing extreme pump case distortion in service?

Johnny Pellin

 

[JJPellin]

Some pump models are more prone to extreme distortion than others. We have a few sets of Byron Jackson model DSJH (also radial-split, double suction, between bearings) that had chronic distortion with internal rubs at the wear rings and bushings. The solution involved better insulation, warm up procedures and reduced pipe strain. But, even with these, I never saw a seal face “potato chip” (our term for a saddle distortion). Some old Pacific ROV and HVC models had weak designs that could distort if not properly mounted. We are currently working on an old Pacific AC barrel pump where the case had become permanently banana shaped. We attribute it to incorrect mounting with all four hold-down bolts locked down solid. The outboard bolts are supposed to be tightened to a lower torque to accommodate thermal growth. We had a consultant once who insisted that an old cast iron case would distort under temperature and then straighten out when cold. I believe he called it thermo-plastic collapse from operation with extreme pipe strain for many years. We stress-relieved the case and corrected the pipe strain and the mysterious bearing failures stopped. I believe it was the pipe strain, not the thermo-plastic collapse. But, this cannot be proven.

Johnny Pellin

 

[flexibox]

A fair bit of discussion has already taken place, I would check the assembly of the seal. This assumes if one seal works why does the other fail? Did you check the face was square with a depth Gauge to ensure the graphoil is seated correctly. Check the anti rotation pin is not interfering with the back of the silicon face in its slot. How many drive pins are there? I do not believe that this relates to the pump. Is it a solid silicon carbide face or inserted into a metal holder? Is the silicon axially clamped?

I think its in the seal assembly somewhere together with initial lapping finish.

good to see problem solved

 

[longeron]

The other possibility is that the gland was distorted during installation. If the bolts fastening the gland to the pump were overtightened it is possible to distort the gland plate. This is fairly easy to look into- check to see if the high spots line up with the gland bolt holes. If so, this is a likely culprit.