Blocked in line heater coil 1

[SNORGY]

Hi all...

I have an indirect fired line heater with a 50/50 EG/Water bath maintained at 90 C. Gas enters the line heater preheat (first pass) coil at a known wellhead pressure, temperature and flow rate. Under a certain condition, the wellhead ESDV and the angle choke on the coil outlet close simultaneously, thereby blocking in the gas.

With "U" and "A" both known or reasonably estimated, I can use NTU effectiveness to get the process outlet temperature. What I am interested in, though, is an accurate determination of the actual trapped mass of gas in the coil at the time it is blocked in. The ensuing effect of heat input on the gas follows an isochoric process, and the intent is to determine the requirements, if any, for thermal relief. Coarsely, I can estimate the average density (and thus mass) using some form of temperature averaging, or I can use the end states and estimate the average density using some form of density averaging (arithmetic or "2/3 Rule" or otherwise), or I can do something else.

What would be the best averaging method for me to arrive at the most accurate estimate of the trapped mass, or is the arithmetic average density between the two end states close enough?

 

[BigInch]

Assume, temporarily, linear pressure and temperature profiles, "the 1/2 rule", and take the average and calculate trapped mass.
Now try the "2/3" rule, although it is not = 2/3.

If there is any significant difference, more accuracy will be needed.

If you need more accuracy, you might need the exact pressure and temperature profiles (typically not linear) at the moment the gas is locked in. Integrate the {Γ]F(p&t) from inlet to outlet. Depending on the pressure, compressibility may need to be considered as well.

Unless your coil is a mile long, I suspect that the "1/2" rule could be used.

Independent events are seldomly independent.

 

[SNORGY]

Thanks, BigInch.

It is a high pressure coil, just over 5000 psig, in fact.

I tried averaging based on determining an arithmetic mean temperature and calculating the corresponding density. I also tried averaging the arithmetic mean densities and back-calculating the temperature corresponding to that. The difference in densities amounted to about 2% throughout the range I looked at. The cubic EOS that I used was Peng Robinson, since that's what the process simulation used, although I probably could have used R-K, SRK or VDW to get a realistic Z. There is also some free water and condensate in the gas stream, which my single phase based approach does not account for. I characterized the gas as a 20.3 MW natural gas, assumed a corresponding critical pressure and temerature, took a guess at the acentric factor, estimated the vapor pressure at 0.7 x Tc using Lee-Kesler, based my compressibility and density determinations on P-R, and then checked the Z-factors against the GPSA charts. The densities appear to be accurate (well, as accurate as they can be, given the assumptions I made) at the end states. I extracted the OHTC (U-value) for the coil from the process simulation for the line heater to spare me the effort of manually determining the film coefficients. Knowing the coil dimensions, I determined a "UA", and with all that in place, I used NTU-effectiveness to get the temperature rise under flowing conditions, whereupon I then blocked in the line heater and assumed isochoric heat input to establish the pressure-temperature profile up to the relief valve set pressure.

The conclusion is that, for a cold start-up gas case at high wellhead pressure, either a thermal relief is required or the bath temperature needs to be controlled to a fairly low setting until the pressure in the well decays to the point where the blocked-in pressure rise will not exceed the piping or coil design limits.

The added "oh, by the way" in all of this is that it has been stated that a sour flare is not wanted and neither are relief valves going to atmosphere wanted, which is somewhat inconvenient since the gas is sour. I am looking at tying the outlet of the thermal relief valve into the 300# (PN-50) pipeline system downstream of the facility, but unless it is taken right to the pipeline outside of the lease edge valve, there may not be enough of a buffer volume in the piping to accommodate the incremental mass of gas that would need to be transferred across the relief valve. This situation limits the line pack pressure against which the facility can operate.

One proposal under consideration is a gas charged surge pot (like a pulsation dampener) in lieu of the thermal PSV, intended to absorb the volume of gas that would otherwise need to be transferred across the relief valve. I think the idea has some merit, but the idea of a pressurized balloon at 0.7 x 5000 psig (typical gas precharge for, say, a pulsation dampener) inside a vessel on top of a pipe in the middle of a -45 C blizzard is like replacing an unwanted but familiar safeguard with something that would leave a lot of people feeling skeptical.

Have you faced this sort of thing before?

 

[BigInch]

No. I've always run to a high stack flare. In lieu of that, second choice would be an unloaded surge chamber, similar to the type used in liquid lines, basically just the relief valve connected to another pressure vessel or maybe a dead end pipeline type holder, but of larger volume so that it will hold whatever the relief valve let's go, at under the 300# (or whatever) pressure limits.

Independent events are seldomly independent.

 

[SNORGY]

Kind of what I thought. Whatever buffer volume needs to accommodate an incremental 9 kg of gas. The trouble with a surge chamber is that I'd have to rationalize it as falling outside of the realm of ASME...otherwise, I'd need a relief valve on *it*, too.

Thanks, BigInch.

 

[BigInch]

What about a pipe drip-type.

Independent events are seldomly independent.

 

[SNORGY]

Thanks.

After some analysis and discussions, and a lot of calculation iterations, we decided to limit the line heater bath temperature to a point where a thermal relief scenario was difficult to achieve, and if it was achieved, the mass of gas transferred would not unacceptably pressurize the lower pressure piping. In combination with other process safeguards, this seemed to be the optimum solution.

 

[BigInch]

I like solutions that eliminate the problem completely! Good job.

Independent events are seldomly independent.