30 replies to this topic

[barbara_s]

Dear All

I'm simulating the depressuring using Hysys.
I defined heat flux parameter -> Detailed -> Convection and I shoul like to know the difference between Fixed U and Updated U, in the cold blowdown this parameter is very important!!!!

thks in advance!

[Zauberberg]

The difference is as the name says: fixed U-value is kept constant throughout depressurization process, while continually updated U-value is being calculated based on actual fluid conditions. You can find some more information in the Aspen article attached below.

My experience with Aspen's heat transfer capabilities is not satisfactory. If I were at your place, I would try to estimate the U-value based on initial conditions and then keep it constant throughout the blowdown process. It will give more reliable values in any case.

Best regards,

Attached Files

•  Depressurisation - A Practical Guide.PDF   317.26KB   86 downloads

[barbara_s]

Thks for your reply!

At this moment I have U updated and I have the results better than with U fixed, but I dont understand why...
Anyway, if I have to estimate the U-valve, I can use the tool in Hysys (Estimate coeff now)????

thks

Barbara

[Zauberberg]

How do you know that the results with updated U are "better" than with the fixed U-value? What does the word "better" mean?

You are somewhere very close to natural convection inside the vessel, and it is fairly easy to get the U-value based on published correlations. You can do it manually and see what kind of changes in the inside film coefficient occur by changing fluid properties during blowdown - consider initial and final stage at least to get a picture. Outside conditions and the vessel wall material/thickness are the same, of course.

[barbara_s]

thks I will estimate the U value!!!! And I will simulate my case!

Best regards

Barbara

[Zauberberg]

In such case I agree, -14 degC and -66 degC is quite significant difference, calling for different materials of construction. Doing it manually could clear some doubts.

I have become suspicious by time regarding Hysys' temperature profile calculation during blowdown. It's never that much low in reality, but we have to be persistent and consider all possible scenarios - especially when ligh Hydrocarbon liquid is present in the vessel at the beginning of blowdown process. In such cases, very low temperature may occur as the blowdown progresses towards the end.

Good luck,

[barbara_s]

I don't have a good experiece with depressuring simulation...
And I had, I have and I will have many problems!!!!
Now I'm estimating the transfer coeff...

thks for your help!!!

Barbara

[Zauberberg]

This sounds as if the army of demons is right behind your back

Attack the problem from a relaxed standpoint. The worst thing that can happen is that you would not be able to solve it. And it's not the end of the world. In the meantime, while working on this subject, you will learn plenty of other engineering things.

Of course, we'll be glad to assist you if you come up with any particular question.

[barbara_s]

If you agree I can send you a file with the simulation...

[Zauberberg]

Yes, you can upload the file (but ZIP it first, otherwise it may not work), and also give some background on the subject.

In addition, there is a good reference with respect to correlations for natural/forced convection available at ChE website, and you should look at these as well: http://www.cheresour.../convection.pdf

[barbara_s]

Simulation Data:
- Vessel diameter 1880 mm
- Vessel heigth 7320 mm

- P_iniatial= P_design = 74 barg(Fire Case)
- T_initial= T_operat (Fire Case)
- P_iniatial= P_eq (T= 11°C as required by client) = 67.5 barg (Cold Case)
- T_iniatial= T_environm (T= 11°C as required by client) (Cold Case)
- flare header backpressure = Patm (ar required by client) (cold / fire case)

thks!!!!!

Attached Files

•  2106-V-101_6_Average_mod.zip   1.08MB   42 downloads

[Zauberberg]

Barbara,

I promise I'll have a look into your simulation file but it will probably take some time. Meanwhile, what I'd suggest you to do, is to try to calculate heat transfer coefficients assuming natural convection. Consider initial (high pressure) conditions, and final (low pressure) conditions, because both liquid and vapor phase properties are significantly different at the beginning and at the end of blowdown process. Different thermodynamic/transport properties will affect heat transfer coefficients calculated for both stages. Also, do the calculation work for both phases inside the vessel, vapor and liquid. For heat transfer correlations refer to the document I have attached earlier in this thread.

One interesting thing from the simulation file which doesn't seem to be logical, is the total mass of liquid phase inside the vessel (start/end of blowdown). If you create a strip chart you can see that the mass of liquid increases by time, which is kind of contradictory - especially if we remember that the vessel is depressurized down to ambient pressure, and to a fairly high temperature (-5 degC).

In any case, try to start your manual calculation work in order to estimate U-value (we are looking for ballpark figures only), and let's see what is the difference between initial and final conditions. If the difference is significant, we should start analyzing temperature profile and liquid vaporization rate during blowdown; if the difference is small, we will investigate where does the huge temperature drop come from.

[barbara_s]

Barbara,

I promise I'll have a look into your simulation file but it will probably take some time. Meanwhile, what I'd suggest you to do, is to try to calculate heat transfer coefficients assuming natural convection. Consider initial (high pressure) conditions, and final (low pressure) conditions, because both liquid and vapor phase properties are significantly different at the beginning and at the end of blowdown process. Different thermodynamic/transport properties will affect heat transfer coefficients calculated for both stages. Also, do the calculation work for both phases inside the vessel, vapor and liquid. For heat transfer correlations refer to the document I have attached earlier in this thread.


-ok

One interesting thing from the simulation file which doesn't seem to be logical, is the total mass of liquid phase inside the vessel (start/end of blowdown). If you create a strip chart you can see that the mass of liquid increases by time, which is kind of contradictory - especially if we remember that the vessel is depressurized down to ambient pressure, and to a fairly high temperature (-5 degC).

-I haven't saw that the liquid increases inside the vessel, I suppose that this mass increase is due to the low temperature that we have inside the vessel (about -56°C)...


In any case, try to start your manual calculation work in order to estimate U-value (we are looking for ballpark figures only), and let's see what is the difference between initial and final conditions. If the difference is significant, we should start analyzing temperature profile and liquid vaporization rate during blowdown; if the difference is small, we will investigate where does the huge temperature drop come from.

I will estimate the U-value with:

- Nu= a(Gr*Pr)^m
- Nu= h*d/k

where

a= 0.59 and m= 1/4 when (10^4<Gr*Pr<10^9)
a= 0.13 and m= 1/3 when (Gr*Pr>10^9)

Ref. Perry "Chemical Engineering Handbook"

thks for your help!

[barbara_s]

***

[barbara_s]

Doubt: to estimate the U-value I'm using:

Fire Case:
t=0 (time)
- T_initial= T_op
- P_iniatial= P_design

======> U-value Vapor = 95.5 kJ/s m^2 °C
======> U-value Liquid = 641.7 kJ/s m^2 °C


t=t_final
- T_final= ????
- P_final= 0.5*P_design

Cold Case
t=0 (time) in my case T_iniatial= Texternal = 11°C = T_inside vessel (to estimate Gr number I set Delta_T (Tinside vessel - Twall) = 0.1°C)

- T_initial= 11°C (as required by client)
- P_iniatial= P_eq (T=11°C) about 67.5 barg

======> U-value Vapor = 96.8 kJ/s m^2 °C
======> U-value Liquid = 186.4 kJ/s m^2 °C

t=t_final
- T_final= ???? (I have suppose a value????)
- P_final= P_atm (1 bara)

[Zauberberg]

Barbara,

Check your calculation procedure (or check the units you used) as it seems that the values you have obtained are excessively high. I wouldn't expect natural convection heat transfer coefficient to exceed 30-50 W/m2C in the best case. The values you are getting are much, much higher (>10,000 W/m2C).

As for the final temperature in both cases, use the values calculated by Hysys. As said earlier, we are looking for the ballpark figures and I believe that +/- 10degC difference in final temperature (as assumed) will not affect the results up to that extent to be considered significant.

[barbara_s]

Sorry there was a mistake...

Fire Case:
t=0 (time)
- T_initial= T_op
- P_iniatial= P_design

======> U-value Vapor = 95.5 kJ/s m^2 °C kJ/h m^2 °C = 26.5 W/m^2 °C
======> U-value Liquid = 641.7 kJ/s m^2 °C kJ/h m^2 °C = 178.2 W/m^2 °C



Cold Case
t=0 (time) in my case T_iniatial= Texternal = 11°C = T_inside vessel (to estimate Gr number I set Delta_T (Tinside vessel - Twall) = 0.1°C)

- T_initial= 11°C (as required by client)
- P_iniatial= P_eq (T=11°C) about 67.5 barg

======> U-value Vapor = 96.8 kJ/s m^2 °C kJ/h m^2 °C = 27.8 W/m^2 °C
======> U-value Liquid = 186.4 kJ/s m^2 °C kJ/h m^2 °C = 51.8 W/m^2 °C

[Zauberberg]

That seems to be a good result. I've got quite similar figures as well.

Now, what we should do is to perform the same sort of calculations for the final vessel conditions - final pressure and temperature in both cases. As we agreed, use the final temperature as calculated by Hysys when "continually updated U" option is chosen.

The idea is to get the picture how does the U-value really changes during blowdown, and whether it has significant effect on the temperature profile. If we see that, for example, the final U-values are much less than the initial ones, we can start thinking in the direction that using a fixed U-value is probably not the best option.

Next, we should analyze what Hysys does in the U-value calculation, and plot our results against the U-value calculated by Hysys during blowdown process.

It seems like we are on the right track, do you agree?

[barbara_s]

Now I have simulate the same case with fixed-U (the U-values estimates by me) and updated-U and the results of min Temperature of fluid and wall are similar!

[Zauberberg]

You have perform a blowdown study based on the U-value calculated for initial conditions in the vessel. However, bear in mind that we need to check the U-value at final conditions and, if that figure shows much different than the U-value at initial conditions, we know that we cannot use a fixed U-value, whatever it is.

So, proceed with calculating U-value at the final conditions for both cases (cold/warm blowdown) and see how much they differ from the U-value calculated for initial conditions.

[barbara_s]

You have perform a blowdown study based on the U-value calculated for initial conditions in the vessel. However, bear in mind that we need to check the U-value at final conditions and, if that figure shows much different than the U-value at initial conditions, we know that we cannot use a fixed U-value, whatever it is.

So, proceed with calculating U-value at the final conditions for both cases (cold/warm blowdown) and see how much they differ from the U-value calculated for initial conditions.

I don't know where I can see the properties of liquid and vapor at the final condition...

[Zauberberg]

Use the approximation, i.e. assuming that the composition of initial vapor stream is identical to the final vapor stream. This will not introduce significant error as you have Methane-rich gas. Composition does not really change that much in order to affect gas/vapor properties.

To do that, just add another blank material stream in Hysys flowsheet, and copy the composition from initial vapor stream. Then use final pressure and temperature values from previous blowdown simulations, get the properties from gas stream and calculate the U-value with those parameters. Let's see what will come up.

[barbara_s]

Use the approximation, i.e. assuming that the composition of initial vapor stream is identical to the final vapor stream. This will not introduce significant error as you have Methane-rich gas. Composition does not really change that much in order to affect gas/vapor properties.

To do that, just add another blank material stream in Hysys flowsheet, and copy the composition from initial vapor stream. Then use final pressure and temperature values from previous blowdown simulations, get the properties from gas stream and calculate the U-value with those parameters. Let's see what will come up.

Yes, I did It, but at the final condition I don't have the liquid properties...
and I don't know wich properties I have to take into account (density, heat capacity, viscosity, etc)...

[Zauberberg]

The same thermodynamic/transport properties you have used for calculating heat transfer coefficient for natural convection at the initial process conditions.

My guess is that there will be no liquid phase inside the vessel at the end of blowdown process, so let's continue the calculation work for vapor phase only.

[barbara_s]

The same thermodynamic/transport properties you have used for calculating heat transfer coefficient for natural convection at the initial process conditions.

My guess is that there will be no liquid phase inside the vessel at the end of blowdown process, so let's continue the calculation work for vapor phase only.

ok I'm running the simulation with h_vap= 13.1 kJ/h m^2 °C, h_liq =0...