Mahle fiber wound filter for Hot oil service catch fire

[MIANCH]

Hi All,
Couple of days ago we opened hot oil (therminol 66) filter for inspection, after opening the filter top cover smoke start producing and later on cartage color change to brown and then catch fire.
It is very un-usual to catch a fire as therminol 66 auto ignition is very high and service is very clean then how and why filter catch fire?
Can someone explain it.
thanks

 

[Latexman]

What is the filament material? Several types are available - Polypropylene, Cotton, Glass Fiber, Unbleached Cotton,
Fibrillated Polypropylene, Rayon, Nylon and Polyester. Each one has it's specific safe operating temperature range.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[EmmanuelTop]

In addition to what Latexman said, there could have been various degradation products of Therminol that have much lower flash point (Eastman say below 38 degC is not uncommon) OR degradation products that are tar-like and with significantly lower auto-ignition points, similar to heavy petroleum fractions. See also

http://emnmktassets.blob.core.windows.net/therminol/Article-Paper-Defining-Thermal-Stability.pdf

Dejan IVANOVIC
Process Engineer, MSChE

 

[MIANCH]

Its fiber glass elements.
DJ,
after reading your article, one more problem that we are facing is pump seal failure. Hence our operating temperature is 270^0C and WHRU inlet temperature is 325^0C max.
Its mean Therminol 66 degradation is started at 325^0C, while its boing point is very high.
Thanks

 

[Compositepro]

This is not at all an unusual nor an unpredictable occurrence. It is called spontaneous combustion. It happens to cotton rags soaked in oil (linseed oil is the worst). It also happens occasionally to rolls of hay. It is caused by the slow oxidation of porous material exposed to air. There is a lot of surface area for maximizing oxidation and there is very little heat loss due to the insulating effect of the porosity.

 

[EmmanuelTop]

The Therminol article says that the hot oil film temperature is 30 to 60ºF (15 to 30ºC) higher than the bulk (= measured) fluid temperature, and at certain hot spots these film temperatures are probably even higher - otherwise no degradation of hot oil would ever occur. For bulk fluid temperature in the range 325ºC, the corresponding film temperatures would be 340-355ºC, which does not mean that at some heater spots they could not be as high as 380 or 400ºC.

The pump problems are a real indicator of hot oil damage in this case - pump seal issues would point at reduced NPSH among other things, due to hot oil degradation and generated light fractions with higher vapor pressure. Coke particles are probably present as well and they also damage the seal. These are the evidences that the circulating fluid experienced a measurable extent of damage/degradation.

Dejan IVANOVIC
Process Engineer, MSChE

 

[georgeverghese]

Hot oil pumps should preferably be fitted with API plan 52 double seals. Hot oil expansion drum should be N2 blanketed.
As suggested, also check pump NPSHa for the case where hot oil vapor pressure = expansion drum operating pressure.
You've either got one or both of
(a) thermal degradation of this Therminol fluid due to excessive film temp at the WHRU hot end at low hot oil flow case and coincident max flue gas temp
(b) a leak of some flammable low boiling process fluid into hot oil through one or more leaking heat exchangers.
Check hazardous area classification drawings for all areas where this hot oil is passing through for this case where the hot oil is volatile and flammable.

 

[MIANCH]

Georgeverghese
our hot oil pump fitted with api seal plan 62+02 and expansion vessel is N2 blanketed, expansion vessel pressure is 1.0 barg and hight is 10 meter.Available pump suction pressure is 2.2 barg. Pump data is here for more detail.
Pump suction Pressure Max: 0.0 barg
Pump Minimum Pressure :-0.09 barg
NPSH available: 4.3 m
Minimum stable flow: 58.86 m3/hr.
Viscosity:0.45mm2/s
Hot oil being consumed in gas heaters and for reboiling of hydrocarbons, there is no leak of hydrocarbons in hot oil we confirmed it.
Hope you can better analyse it further and share your thoughts.

Thanks

 

[georgeverghese]

The min pump suction pressure should be higher than the max seal flush quench fluid pressure to prevent contamination of hot oil with the quench medium for this plan 62 seal?

See if the pump vendor can offer you a plan 52 or 53 alternate.

With this hot oil being flammable now, tube to tubesheet leaks on the WHRU can be very dangerous now when the GT driver is shutdown and if the hot oil is still under pressure - we have had one GT explosion in an LNG plant due to a WHRU hot oil leak.

 

[MIANCH]

Hi
our case seal flush quench fluid pressure is 4.0 barg as vendor recommended and temperature is 40^0C,its mean we need to revisit vendor for cooling pressure.
Pump model is KSB-RPH-100-360D, due you think seal plan 52 or 53 can easily be fitted.
as i wrote in my previous post that we don't have any process side leakage.
Thanks for sharing your exeprience.

 

[georgeverghese]

So there is a big risk of cooling water getting into the pump suction here.
Have my doubts if KSB can change this to to plan 52 ; no harm asking though.

 

[MIANCH]

George
Can you upload some referenced material which is stating that pump suction pressure shall be higher than quenching fluid pressure.I have read some threads today and found that water should not be used but steam or N2 can be used. Reasons are very good for steam or N2 use,then why KSB selected this plan and eagleburgman recommended for water with 45^0C and for 4 barg pressure.

 

[georgeverghese]

Mianch, dont have this reference material, but it seems obvious that there will always be some minor leak of quench fluid into the process fluid even when the shaft seal is working well.

Similarly, in a pressurised plan 52 double seal, some barrier fluid medium will always leak into process fluid, which is why the seal pot level must be checked and fluid topped up when required. When the seal has failed will topping up freqency increase.

N2 seems to be a good idea from a contamination perpective, but may be cooling of the seal parts will be very poor with gas - that may be why they chose water - they protected the seal and ignored the Therminol contamination concern?

 

[MIANCH]

George
I have some questions please answer,
1-are you operating therminol 66 as heat medium.
2-your pumps have API seal plan 52 or 53.
3-How frequently top up barrier fluid in a week or month.
4-if it is possible then upload one photo of your pump that showing seal arrangment.
Thanks

 

[georgeverghese]

1)On this last design project I worked on, we used Shell Thermia B. On another plant where I was supporting operations, it was Therminol - forget which grade ( grade 60??).
2)On this last project, we used plan 52 pressurised for these fixed speed pumps.
3)Sorry dont have info for this - quoted losses from the seal vendor sounded very low
4)Dont have this either - the seal plan included a natural convection air cooler for barrier fluid recirculation cooling since there was no cooling water on this site. The seal pot was N2 pressurised some 10psi or so above pump suction pressure to ensure positive press diff (dp = seal pot press - pump suction).
Hope this helps