Hammer

Moustafa Darweesh said:

I would like to model a sudden valve closure at a pipeline end. How can I select the type of valve?

If you use a Discharge to Atmosphere (D2A) node to model closure at a pipe end, linear closure characteristics are assumed as explained in this article: Modeling Reference - Discharge To Atmosphere

If you would like to have more control over the valve closure characteristics, you will need to use a TCV node element connected to a pipe, with a D2A node at the end, set to act as an orifice (not a valve). The pipe between the TCV and the D2A would have a large diameter, short length and smooth roughness so as not to have a hydraulic impact (if the actual valve discharges directly to the atmosphere. You can find more details on the setup of the TCV and the "Valve Type" characteristics here:

Modeling Reference - Valve Closure

Valve Type field assumptions and use with a TCV

Moustafa Darweesh said:

Also, when I change the transient friction method from steady to unsteady or Unsteady - Vitkovsky, there is no different in the results.

Both of these unsteady methods are very similar and may indeed produce similar results in some models depending on the flow regime changes occurring. See more here: 

"Unsteady" vs. "Unsteady - Vitkovsky" transient friction methods in HAMMER

Moustafa Darweesh said:

Could I specify  the Vitkovsky decay coefficient to 0.039?

Not at this time. See the help topic "Unsteady or Transient Friction" which explains the assumed decay coefficient for laminar and turbulent flow.

Moustafa Darweesh said:

Could  I change any parameter in HAMMER to make the results smooth (figure 1) instead of sharp (figure 2)?

The transient results should be accurate based on the data input. If the results are not "smooth", or could be due to a data entry problem, or with the transient friction method assumptions. Please consider the above information and if the results are still not smooth for you, a copy of the model can be provided.

Thanks Jesse,

The file was attached below.

VALVE.rar

I would like my results (sharp, current study, blue line,) to be smooth and match with those shown in the figure below (red).

Hi Moustafa,

Thanks for providing a copy of the model. The reason why the valve closure transient results are "sharp" is because you have not entered a headloss coefficient in the TCV to describe the headloss characteristics as the valve is closing. So, it acts as if the valve instantly closes at the time when the pattern reaches 100% relative closure. If you add a headloss coefficient (in the initial settings section, just below your selection of "Active" as the initial status) and use one of the unsteady transient friction methods, you will see results that are more "smooth".

This exact situation is described in the following Wiki article: Gradual closure not occurring when using valve operating rule