4 replies to this topic

[_luis]

Dear Experts,

 

I'm a junior process engineer doing relief calculations for an MTBE vaporizer. Here's the background of the process:

 

Subcooled MTBE is pumped to a heat exchanger (shell side, DP = 17.5 barg) where it undergoes heating, vaporization, and superheating all in one shell. This is done using a 25 barg steam (tube side, DP = 30 barg). The superheated MTBE goes to a reactor section (DP = 17.5 barg) afterwards to undergo a vapor-phase reaction converting it to Isobutylene. Afterwards, it is cooled using a series of coolers (DP = 17.5 barg). A PSV protects the vaporizer with a set pressure of 17.5 barg. The normal operating pressure in the vaporizer is 10 barg. (SEE ATTACHED SKETCH)  MTBE VAPORIZER.PNG   46.33KB   3 downloads

 

Here are my two main problematic scenarios (apart from common ones like fire):

(a) blocked discharge - 40 tons/hr

( tube rupture - 120 tons/hr

 

The relief load calculated is based on the assumption that all of the heat that can go in the exchanger is taken to vaporize a potentially infinite liquid pool on the exchanger (since the feed pump can supply enough pressure to deliver MTBE during relief conditions).

The concern is that we are limited to only 14 tons/hr. For this specific project, we take credit for SIS action and we intend to use a high pressure trip that closes the steam supply via SIS. This reduces the heat input to a minimal amount (only a small flow from the leakage through the steam valves) .

 

Unfortunately, when we generated a pressure vs time analysis, it takes only 5 seconds to reach the relief pressure. The fastest response time we have available is 10 seconds (from detection to valve closure).

 

Now, I'm considering the use of a rupture disc which would be piped towards an upstream tower (volume = 165m³) with a design pressure of 6 barg. I am concerned however, about how much depressurization can I really get away with.

Any ideas/comments/suggestions? I would appreciate them

 

[Bobby Strain]

The rate you indicate for tube rupture looks to be quite more than I would expect. And I don't think you told the whole story. For example, what is the relieving material for the tube rupture case? Maybe you should show us your tube rupture calculations so we might comment.

 

Bobby

[_luis]

Bobby

 

For blocked discharge: Total heat input is equivalent to: Q=UAdelT. U is clean coefficient, A is total area, and delT is the maximum temperature difference between supply steam and inlet temperature. This total heat input is used to vaporize the MTBE with latent heat of around 270 kJ/kg at the relief pressure.

For the tube rupture case: A separate relief is calculated considering only steam as the relieving material. A secondary case is also considered where the heat input coming from this steam entry is used to vaporize the pool of MTBE. Because of the higher steam entry (thus, higher heat input), the vaporization is larger.

Tube rupture flow calculation is based on two-hole orifice (critical flow calculation based on API).
 

[Bobby Strain]

I checked steam flow through ruptured 3/4 inch OD tube (both ends) and the flow is 7.5 tonnes/hr. So maybe you want to have someone check your calculations.

 

Bobby

Edited by Bobby Strain, 25 April 2014 - 12:59 PM.

[_luis]

Bobby

7.5 tonnes/hr is correct - assuming that we are only relieving steam. However in this case, we needed to consider the scenario of the heat of the steam being transferred to the MTBE pool. In which case, this 7.5 tonnes/hr of 25 barg steam can deliver around 5500 kW of heat -- thus vaporizing around 120 tons/hr (latent heat is only 165 kJ/kg -- please disregard the 270kJ/kg I mentioned earlier).

Anyway, so far we have increased the design pressure of the shell such that it can accommodate the time difference of the steam valve closure and the flare limit.