Flammable mixture inside pump 1

[Tom_H]

A centrifugal pump is used for liquid transfer (filling) of a flammable and somewhat volatile liquid from a truck to a vessel. If the pump deadheads due to it continuing running despite the truck container being empty, is there then any risk of a flammable atmosphere occuring inside the pump? I.e. due to the pump sucking in air which then mixes with the gas at the discharge side of the pump (in the piping) and then possibly igniting due to mechanical ignition sources inside the running pump? Seems pretty far-fetched, but how is this potential risk normally handled?

 

[Artisi]

If the tanker is empty and the pump is allowed to run, it won't be running dead headed, it will be off-prime and churning whatever is in the casing and heating up the product - it could reach its boiling point - can't really comment on the likely outcome beyond this point.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[ashtree]

You could put some sort of no flow or excessive pressure detection on the pump which will cause it to shutdown and protect against dead heading.

Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[JJPellin]

We have two services that have a similar risk. We have used various methods to mitigate the risk. The first service that has similarities is our oily water sewer. Sumps accumulate various liquids which are automatically pumped out based on sump level. If the level instrument were to fail, the pump could continue running with the sump empty. We protect against the potential process safety event using three methods. The first one is not relevant to your example. We use multiple, redundant level instruments with a mid-select logic. In case this fails, we shut the pumps down on low flow. Lastly, we use bushings in the pump made of a material designed to survive extended dry-run conditions.

The other service is nearly identical to yours. We unload tanker loads of a volatile chemical. We are required, by contract to completely empty the tanker. We park the tanker on a pad with a slight slope so that the unloading hose is at the low point at the very back. The unloading pump is a seal-less design, fully lined with Teflon ®. The pump will shut down on low flow or low motor amps.


Johnny Pellin

 

[JJPellin]

If you are actually concerned about deadhead operation, some of the same safeguards can still be used. A low amp trip can be set to trigger under deadhead conditions. A low flow trip is also still applicable. Under deadhead conditions, there is less risk of air being drawn in to create a combustible mixture. But, wear rings and bushings designed to survive dry running could stabilize the rotating parts and reduce the chance of a bearing and seal failure.

Johnny Pellin

 

[ScottyUK]

As a word of caution, using a 'low amp trip' is an awful solution if you're trying to protect against low flow. Most motors draw 1/3 to 1/2 of their rated full-load current as magnetising current under open shaft conditions, and the change in current magnitude between, say, 25% load and 5% load can be much smaller than you might imagine. A power measuring relay which detects when the delivered power to the load has dropped below a certain threshold, or a power factor relay which detects the change in the phase of the line current relative to the voltage as the motor moves from a laden to an unladen state are both far better at accurately detecting an under-load condition. We use a lot of power measuring relays which the process engineers insist on calling 'low amp trips'.

In my opinion measuring the electrical power into the motor, even if we were to use a revenue-class instrument, is still a rather poor analog of process conditions in the pump: at best it forms a weak secondary layer of protection to back up the primary layers of detecting the flow and measuring pump bowl temperature.

 

[georgeverghese]

Looks like this setup is "dead in the water" even before the ship sets sail - how do you justify the presence of air in the vapor space of a truck filled with a flammable liquid in the first place? The pump pulling in air with this flammable liquid just makes things much worse. It should only be some inert gas in the vapor space.
A flammable liquid has a closed cup flash point temperature which is lower than the max normal ambient temp.

 

[Tom_H]

Thanks for all the responses. In response to the last post, the reason that air could be sucked in in our particular case is that a hatch is opened in the truck (container) in order to prevent underpressure. Thus, when all the liquid is pumped out, air could be pumped in until the pump is shut down. Is this reasonable? Perhaps not - I am not an expert in unloading operations.

 

[LittleInch]

I'm with Artisi on this one.

It's not dead heading - that's when you close the pump discharge valve when it's still running - it's called dry running.

Now in reality pumps designed to pump liquids tend to be really bad at pumping gases or mixtures of gases. Therefore you won't really suck in large quantities of air before the pump just spins around not pumping anything in a mixture of liquid and air. If you're pumping into a tank or vessel where liquid level is higher than the pump then there should be a NRV to prevent back flow, but the pump will have liquid in front of it.

You really shouldn't be designing to operate that way, but I understand that's sometimes what happens. For this scenario I would always prefer to dump the truck into a buried vessel / tank of some size open to atmosphere via a tall vent and pump from there and control flow on level. Then when the flow gets too low use a small stripper pump to pump out the dregs.

Anyway, even if you do get some sort of explosive atmosphere in the pipe, the ability of something inside the pump to actually ignite that mixture seems to me to be so low as to be negligible. I will stand corrected, but I've never heard of an internal pipe explosion caused in this way. So to me you're trying to solve a problem which doesn't exist.

Inert gas fill of a tanker being offloaded is the ideal scenario, but I appreciate it sometimes doesn't always happen that way. Certainly in petrol filling stations they normally just seem to clip the hatches open and then drain away....


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

If this fluid is indeed flammable at operating temperature range, then the ISOtainer should be inert gas blanketed during unloading ops. Hook up a forward sensing self regulated PCV to feed inert gas and maintain ISOtainer pressure during unloading, in addition to maintaining electrical grounding for the ISOtainer. Else you would be taking on more risk for no benefit, given the frequency of this operation, the fact that these operations are manned and are sequential in nature, the inventory in the ISOtainer and the high likelihood that there may ignition sources some where in the vicinity. If your Company management refuses to minimise risk ( for little effort for this case ) and make things safe for plant operators and truck drivers, then this high risk operation should be acknowledged and recorded in the Plant Major Hazards List as Corporate responsibility with accountability - very likely your Company's insurer's capital assets premium will be higher if they get a whiff of this.

 

[chicopee]

If you observe propane transfer, you will notice two lines one fro liquid delivery to the storage and the other for vapor transfer between storage tank and bobtail. No air gets in the system, perhaps a vapor return line in your case would be exemplary.