Molten Sulphur Pump: Premature Mechanical Seal Failure

[crradovan]

Good day,

We currently have two molten sulphur pumps running with API Plan 01/62. However, mechanical seal failure is reccuring. We follow a clearly defined procedure in starting up the pump and kept all temperatures high as required.

1. Any ideas what is causing the seal failure?
2. Is the seal plan not suitable for the service?

Your inputs will be greatly appreciated.

regards,

 

[georgeverghese]

Plan 62 is not a pressurised seal flush plan, which is what you need here. Suggest one of the plan 53 arrangements. And the seal may need to be good for temperatures high enough to keep the surrounding sulphur molten. Whether it will be plan 53A, B or C is subject to disclosue of more information and the seal / pump vendor may also be able to assist. The barrier fluid may need to be kept above the melting temperature of the sulphur. Believe the melting points for sulphur are different for the 3allotropes (alpha phase, beta phase and gamma phase)-which one do you use?

 

[crradovan]

Hello, georgeverghese!

Thank you for your feedback.

The melting point of our sulphur is 116 deg. C. We are not having any problems with the secondary seal of plan 62; however, I will take heed of your suggestion. Furthermore, I am doubting the suitability of plan 01. Is it much better to install plan 02?

Thanks.

Regards,
CR Radovan

 

[georgeverghese]

Plan 53 doesnt have the process fluid as primary seal flush, since these are double seals operating with a pressurised barrier fluid, so the risk of molten sulphur cooling down on the route to cool the seal surfaces is avoided. So yes, may need to get avoid this plan 01 for this molten sulphur application.

The risk with plan 53 here is potential overcooling of the recirculating barrier fluid which may affect the molten sulphur around the seal cartridge area - this is applicable to both shutdown and normal operating modes. By the way, how do you warm up the pump casing during startup, especially around the shaft seal area?

 

[crradovan]

Hey, George.

Our pump is steam jacketed; and this includes the stuffing box.

I think the use of Plan 53B may cause complication due to the cooling configuration not unless we maintain the barrier fluid at an average temperature higher than the melting temperature. We will as the manufacturer for seals that can operate on that range.

 

[JJPellin]

I would add a steam jacket to the seal gland. I would convert from Plan 1 to Plan 2, as you suggested. Most of our sulfur pumps are vertical pit pumps with no mechanical seal. But, the one pair of sulfur pumps we have that do use mechanical seals use the arrangement I have proposed. The seal reliability is not as good as plant average. But, we have been getting acceptable seal life using this system.

Sulfur has a very narrow temperature range when it has suitable properties for lubrication of seal faces. Either too hot or too cold and the performance is poor. A good system of steam jackets with the steam conditions held a few degrees above the ideal temperature is the best way to achieve acceptable seal life. I am not sure how you would maintain this temperature using a Plan 53.

Johnny Pellin

 

[RHurst]

JJPellin has it right. Unless it's a safety regulation at your site, I almost always recommend utilizing a single seal with a plan 02/62. Dual seal plans (including those for gas seals) in hot applications such as molten sulfur tend to act as heat sinks for the mechanical seal, even at ambient temperatures. Add a steam jacket to the gland and always follow proper steam piping techniques.

With that being said and assuming your system is designed to keep the sulfur in the correct operating window (260-300F), the biggest problems with most sulfur pump applications is typically two fold:

1) There is no temperature interlock on the system. This allows operations to just switch the system on even though it may not be at temperature. Stuff breaks as the pump and seal attempt to rotate in solidified sulfur.

2) The interlock for the system is in an inappropriate location. My best recommendation for the location of a thermocouple/temperature probe would be in the seal chamber. If the seal chamber is at temperature, then the critical areas within the pump are likely close to temperature. This is your best shot at avoiding heavy damage during start-up.

 

[crradovan]

Thank you Johnny and RHurst.

We will be inquiring our manufacturer if they can provide us with the steam jacketed seal gland. Apperently our mechnical seal is, in fact, a single seal configuration. I am wrong on my previous statement of it being dual. Appologies.

With regards to the interlock system, our pumps do not have any thermocouples to monitor seal chamber temperature. What we maintain, however, is the steam inlet pressure for the pump jacket.

Do your mechanical seals have provision for installation of thermocouples? Again we will ask our manufacturer regarding this.

Regards,
CR Radovan

 

[georgeverghese]

Yes, the only way for a plan 53 to work would be if the barrier fluid were to be kept at well above 116degC. But if the plan 02 is the better way to go, okay.

Viking make pumps with electric casing heaters and temperature sensors fitted into the casing, and these include sensors picking up seal cartridge area temp also. Pump casing TSLL interlock for pump start sequence is a must as suggested.

 

[crradovan]

Hey George,

We will be considering the casing heaters and controls in our design for the resolution for this concern. Thank you very much for your inputs.

regards,
CR Radovan

 

[RHurst]

Crradovan,

To answer your question, yes, mechanical seals can be designed to include additional flush ports (NPT) so that a temperature indicator can be installed. However, I have found that you get more bang for your buck if your pump allows for an NPT connection into the stuffing box (often referred to as a cage-ring tap). If you are utilizing a standard ANSI pump for this application then there is a good chance the stuffing box design already incorporates this tap (or can be drilled). On most pump models, having a temperature indicator in the stuffing box allows you to achieve a relatively accurate reading of the bulk temperature behind the pump impeller and at the mechanical seal faces. At a minimum, these are the two areas we are most concerned about the sulfur being liquid before we turn the pump on in order to prevent failures.

As another poster mentioned, electrical tracing or the use of blankets to heat the pump is an option. For smaller pumps, often this method is very effective. For larger pumps or other designs with bulky casings, it can be difficult to dump enough heat onto the surface of the pump in order to bring the sulfur into the appropriate temperature window. This is especially true in locations where the winters can become harsh.

 

[crradovan]

Hey RHurst,

Right. We'll check for that tapping point. I never noticed one, though. Nonetheless, we can opt to install one.

Winter is not a problem for us. We don't have one. So we'll be looking into the jacketed option. We've contacted some companies that could provide us the solutions for that.

Thank you for the info!

Regards,
CR Radovan