8 replies to this topic

[ayan_dg]

Dear all,

I am calculating minimum design metal temperature of ASME vessel made of LTCS which is storing dry natural gas (HC dew point is -14deg C) . The minimum temperature at site for the last 50 years is 4 deg C.
The design pressure of the vessel is 14 barg , the maximum operating pressure is 9 barg.

The vessel is having PSV but no auto blowdown valve though it has manual valve to depressurize to flare.

My question is as it has no auto blowdown valve should I consider isentropic depressurization by Hysys dynamic depressurization utility to arrive at MDMT or should I consider a simple isenthalpic depressurization from design pressure to atmospheric pressure at 4 deg C .
The isenthalpic depressurization will obviously give a higher MDMT value.

Please advise.

[paulhorth]

Ayan,

Since this vessel has a design pressure above 6.9 barg I would have expected that it would have an auto blowdown valve. I hope it is provided with water deluge at least. Can you tell us how large it is?
I suggest that you carry out a depressuring using HYSYS starting from the high pressure trip setting and the minimum ambient temperature of 4 C. Use an orifice size defined from the manual blowdown valve. Use 100% isentropic efficiency for the first run. If the minimum temperature is above -46 deg C then you are OK. If not, you can then try 50% efficiency to check how sensitive the result is.

Paul

[fallah]

The vessel is having PSV but no auto blowdown valve though it has manual valve to depressurize to flare.

ayan,

Seems the manual valve to be for maintenance cases...

The reason of not having BDV may be: The vessel isn't located in a fire zone and the fire case is not applicabla for that. Is it right?

Did you use the curves UCS-66 of ASME to calculate MDMT?

IMO, because you didn't use BDV in your vessel you can not use isentropic depressurization by Hysys dynamic depressurization utility to arrive at MDMT.

Anyway, in PSV sizing we normally ignore the frictional losses in the PSV nozzle and consider ISENTHALPIC process.

Fallah

[ayan_dg]

The vessel is in an area where there are no chance of having a pool fire. Also the design pressure of the vessel is less than 17 barg.
The manual valve will be used to depressurize the vessel during maintenance.
I believe I need to depressuring calculation through that 1 inch globe valve to arrive at MDMT .

[kkala]

Though not familiar with HYSYS and its depressurizing utility, following notes may be useful (subject to criticism).
1. It is given that depressurization will be implemented through a (not specific) manual valve. This is judged realizable in the mentioned case, see http://www.cheresources.com/invision/topic/14892-pcv-on-spheres/, mainly post No 27, para 3.2.
2. Throttling through a valve is not a reversible process (*). It is improbable for the downstream gas to revert to its previous higher pressure without spending energy on it. Downstream entropy (per unit mass) will increase. Enthalpy upstream and downstream valve will be same.
Note (*): See books of physics in the entropy chapter, Schaum's Theory and problems of Thermodynamics (throttling processes), M Zemansky's Heat and thermodynamics (throttling process). According to the latter, "initial and final enthalpies are equal" is the precise expression, since one cannot speak of the enthalpy of a system passing through such non equilibrium states.
3. For the flows in pipes upstream or downstream the valve, one can consider isentropic flow, or with some efficiency, according to above posts. Probably HYSYS depressurization does consider equal enthalpies upstream / downstream the valve, known by those using it (if not known, it has to be investigated).
Isentropic flow along whole piping (including valve) does not seem appropriate.

Edited by kkala, 27 May 2012 - 03:06 PM.

[paulhorth]

Kkala,

Please allow me to clarify what could be a misunderstanding here.

During a depressurisation of a vessel (with no inflow) there are THREE expansion processes to consider. Each process affects the gas temperature.

• The gas contained in the vessel, upstream of the blowdown valve, expands as its pressure falls. The mass decreases while the volume remains constant.
  
• The gas which passes through the valve into the discharge piping expands from the upstream pressure (which is reducing) to the downstream pressure (which is approximately constant and determined by the flare system).
  
• The gas downstream of the valve undergoes further expansion as its pressure falls along the line due to friiction losses

Process (2), through the valve, is indeed isenthalpic, which I think is what you have said.
Process (3), the continued expansion along the piping from the valve, will be something between adiabatic (no heat pickup from the pipe) and isothermal. Not exactly isenthalpic, because of work done against friction, and some heat pickup.
Process (1) is the one which the original poster is asking about and which can be modelled in Hysys. If there is no heat input from the vessel wall or from liquid, the expansion inside the drum is in principle isentropic. This process results in a lower temperature than isenthalpic expansion and so the correct modelling is important in determining the minimum gas temperature. However, due to turbulence, heat pickup, and other irreversible effects, the real expansion is not completely isentropic. Hysys recognises this by allowing the user to enter an "isentropic efficiency". 100% means completely isentropic, while 0% means isenthalpic. Hysys also attempts to calculate heat transfer from the vessel wall, and thus to calculate the metal temperature, though I am not convinced by the method used.

The isentropic nature of this expansion (with no feed) was recognised in the 1980s, I devised a way to model it using a programmable calculator before the simulation programs offered this feature.

If there is liquid in the drum, the expansion is going to be close to isenthalpic because of the flash vapour from the liquid.

And, of course, in a fire case depressurisation there will be more heat pickup from the vessel wall.

Paul

[vikramltv]

Ayan_dg

Your understanding is correct.

[kkala]

Sincere thanks, Paul, for the care and the useful clarifications given. Process 1 per post No 6, concerning the content of the vessel * (being gas 100%), can be considered as isentropic, as it has been clarified.
Is calculated minimum gas temperature (in the vessel) considered as minimum design metal temperature (MDMT) in the mentioned adiabatic process? Or vessel MDMT is usually taken a bit higher, considering heat input from ambient?
Advise from anybody would be welcomed.


* Vessel not under fire.

Edited by kkala, 04 June 2012 - 02:16 AM.

[paulhorth]

Kkala,
The problem of determining the gas and metal temperatures during blowdown is complex and difficult to summarise in a short post here.
There are two main reasons why the actual minimum metal temperature will be significantly higher than the minimum gas temperature. First, the heat transfer from vessel wall to gas will be poor, because the velocity of gas is low and the heat transfer will be by natural convection. This means that the wall temperature has to be significantly higher than the gas for heat transfer to occur. Second, the heat capacity of the metal is usually a lot more than that of the gas, so that the temperature change for a given quantity of heat transferred will be larger in the gas than in the metal. Both these effects mean that the bulk metal temperature will stay well above the bulk gas temperature.
However, local cooling at the nozzle where the gas leaves the vessel can be much greater, because the gas velocity here is high leading to high heat transfer. The minimum design metal temperature for the vessel has to cover this local condition, and this can result in a MDMT which is close to the gas temperature. The blowdown nozzle and local area can be protected from cold gas by fitting a thermal sleeve. This is a thin sleeve in stainless steel fitted inside the nozzle, extending into the vessel, which shields the nozzle from the cold gas flowing out. This feature can allow a low MDMT to be avoided.

The heat transfer from vessel to gas has been the subject of research, and Imperial College, London, have produced a software package which will model accurately the transient temperatures of gas and metal during a depressuring process, based on experimental measurements. However, using their package can be expensive.I will admit to being out of date on the latest information in this field.

Many oil companies now have a recommended procedure for approaching the problem. I would suggest starting using th approach in my earlier post to estimate the most conservative minimum gas temperature, then consider the implications for the MDMT and proceed from there.

One more point - you can forget heat input to the vessel from ambient air, this will be negligible (except in the case of a fire, obviously!)

Paul

Edited by paulhorth, 07 June 2012 - 10:25 AM.