steam ejector problems

[pbrod]

dear all,
we have a vacuum problem in a dimethylphtalate/water-methanol stripping column. after a washing section the mixture with aprox 2% water/methanol is entering the stripping column at about 130°C.

a vacuum unit (Nash) should keep the column at 50mbara using 3barg steam. we have no reasons to believe that our top condensor (15°C) water is not working well.


we equipped our three stage ejector system with manometers. when working well the pressures are as following:
40mbara vacuum
120mbara between the first and second ejector. the outlet of the second ejector is to a condensor on 45°C water (shell and tube)
500mbara before the last ejector. the last ejector is condensed on a 15°C water (shell and tube).

we allmost always have a pulsating regime on the vacuum.

remarquable is that when we reduce the steam pressure from 3barg to 2barg is that the pressure before our last ejector goes down to 340mbara.
the drive steam is dry and the condensors of the ejector system is checked for fouling.
we also checked for flooding of the condensors but this was not the case.

do we need to look deeper into the ejector system?
is this a normal behaviour of our last ejector?

in advance, many thanks for the help and tips!!

very best regards,

pieter

 

[MatthewL]

Pieter,

Did you check to make sure the outlet of the condensers have a long enough vacuum leg, and that there is enough water available to seal the vacuum leg? We use a 40' (12 m) vacuum leg for our barometric, direct contact condensers on our Nash system. For the pulsation, look at your condenser cooling water temperature, we notice vacuum fluctuations that correspond to fluctuations in the cooling water temp.

Hope this helps,

Matt

Matt

Quality, quantity, cost. Pick two.

 

[pbrod]

dear Matt,

thanks for the response!

we did multiple tests with our vacuum leg. We now work with a seperate 20m leg straight down from the intermediate and end condensor. so normally this could not be te problem.

I will log for some time the water temperature today. but we are running with a seperate cooling unit that sets the water temperature on 6°C instead of 15°C and the 45°C water circuit (intermediate condensor) is now running on 28°C.

what was also noticed is that when we loose vacuum in our column 100mbara drop to 500mbara (flashing in the feed line) the first 3 meters of the vacuum pipe (50mm) inlet to the ejectors is getting hot.
can this be explained?

best regards,
pieter

 

[Compositepro]

Have you checked your ejectors for wear or deposits? When the diffuser nozzle that the steam jet flows through gets too large you will get back-flow around the steam jet when the differential pressure gets too high.

 

[MatthewL]

Pieter,

The hot inlet is probably due to the higher pressure in the column. At 100 mbara, saturated steam (the assumed condition in the column, since you are flashing the water) is 46 C, but at 500 mbara it is 81 C.

Regards,

Matt

Quality, quantity, cost. Pick two.

 

[pbrod]

we checked the ejectors for deposits and wear but could not find something (but we do not have the original orifice diameters). also the condensors were checked for fouling.

Matt, the high temperature is at the inlet of the ejector system. between our top condensor outlet there is about 20 meter pipe towards the ejectors. And it is at the ejector side of the pipe that we can notice this temperature increase. After our top condensor we can't notice this temperature increase.

i also measured the temperature drop on the intermediate condensor and this was 2°C.

anyone experience with calculating the steam nozzles, venturi nozzles?

we will do a capacity test using orifices and restart the leak tests on the column.

 

[chief]

The ejector nozzles are usually of stainless steel construction and do not generally wear very much. Check the throat body of the ejector, this may be of cast iron material and wears, reducing the vacuum efficiency over time.

Offshore Engineering&Design

 

[pbrod]

the throat body of the ejector is also stainless steel. we checked this and it looked OK. thanks for the tip!

 

[pierreick]

Is your steam ejector jacketed ? Any risk for ice formation ?
Pierreick

 

[pbrod]

last campaing we checked the convergent-divergent section and there was no risk for ice formation. now the ejector system is insulated.

 

[pbrod]

a vacuum phenomena:
- to give anyone an idea about the problem, in the attachemen there is a trend of our vacuum. the green line (1) is the vacuum. the scale is from 0mbara to 200mbara. the time axis is 40 hours. the purple line is the temperature of the 15°C cooling water circuit which cools the end condensor and the top condensor of the stripping column. the red line is the level in the bottom of our stripping tower.

it is clear that the vacuum can be stable and low for some hours, then it goes back up and later stabelises at a higher level.


our top condensor is a alpha rozenblads lamella heat exhanger. its a counter current system, oval tube, one pass.

 

[chief]

Have you considered a check valve on the final ejector stage. This is common on Steam turbine plants, and a problem with low vacuum can be traced to a faulty check.

Offshore Engineering&Design

 

[chief]

http://www.uniquesystems.com/pvs/brochures/Bulletin_PVS-80025121-ESM_-_Ejector_Maintenance.pdf

You may some pointers here,would appear you may have worn ejectors.

Offshore Engineering&Design

 

[Longinthetooth]

Echoing the comments on worn ejectors. You should check for cracks also. Even a tiny crack on the tip of the nozzle can distort the jet and cause it to lose it's efficieny. I have seen this several times.

 

[pbrod]

tomorrow we will do a pressure test on the stripping column to search for leaks.
we will also do a second check of the ejector system, wear, steam quality,...

to chief: where is the check valve installed?

is it a possibility that the gas speed in the top condensor (of the stripping column) is to low to have a good heat exchange?

 

[rmw]

you need to check were the nozzle threads into the housing of the ejector. Steam can develop a leak past the threads and that adds steam to the load and load to the condensers.

rmw

 

[chief]

The check valve is usually at the first stage suction manifold, this prevents 'leak back' and maintains the vacuum.

Offshore Engineering&Design

 

[mjpetrag]

I've had a similar problem with a binary distillation column. As we increased reflux and got more of the light component in the overheads, the vacuum will fall off. When we cut the reflux and stripped more of the heavy component out, the vacuum would drop back to set point.

Trend your all your flows (steam, feeds, etc) with the top pressure in the column. That may help us determine where the source of the problem is.

Also, do you have parallel offline ejectors? If so try swinging on to them. Also, make sure that if you have parallel ejectors that the steam, discharge & inlet process valves are blocked in to the offline one. If these are opened, the discharge of the online jet can recirc back through the offline jet and screw up the online jet.

A picture of your system with pressures and temperatures will also help us determine what is happening.

If you are checking for a leak, block in all flows to the column and let the ejectors pull down vacuum until it flatlines at the lowest pressure. Block in the ejectors and let the pressure start building. If the pressure starts coming up quicker than 4-5 mmHg/minute, you have an air leak.

-Mike