About isolation valves

[diogomarin]

Here is a hydraulic model built that we insert the block valves as valves, but we have the option to transform into pipes. With this, we can open or close the valve or pipe, we have the option of adding a localized pressure drop.

A lot of work was done in CAD, QGIS and Excel to put these blocking valves in this way, as we thought about loading the base demand of the nodes.

If we configure it with a small tube run or a valve we do not interfere with the flow of water in the entire pipeline and in this way we are able to distribute the demand more adequately. And I don't need to redistribute my demand, regardless of the process adopted.

I want to make it clear that there is a significant difference in the simulation and presentation of the hydraulic model.

To explain better:

I transformed the block valves that were characterized in valves for isolation valves.

With that, I first performed the simulation with the isolation valve open and then with it closed. Follow the results in the images below.

First Simulation - Isolation Valve is Open

Second Simulation - Isolation Valve is Closed

Conclusion and suggestions:

Isolation valves are working properly, congratulations on the tool.

But to solve the total blockage of the flow in the network segment, I suggest creating nodes before and after the isolation valves and accompanied by a demand redistribution

In this way we separate the block valves from the control valves (PRV, GPV, VSP...)

[fmartine]

Hello Diogomarin,

First of all thank you for using QGISRed and for valuing the efforts we have done in the latest version 0.15 to deal with the isolation valves in the most appropriate way.

Our proposal is to differentiate regulation valves (PRV, etc.) from isolation valves. The latter are much more numerous than the former, but simpler, since it is admitted that they only have two possible states (open and closed). Moreover, in QGISRed the isolation valves also support having a head loss coefficient.

To avoid filling the hydraulic model with so many valves, which will only be used in maintenance operations (in fact traditional models ignore the isolation valves), in QGISRed they are treated as a separate theme of points, overlapped on the pipes or on the nodes of the network. We have developed algorithms to emulate the effect of the closure of these valves on the circulating flows, and it is true that for now the closure of an isolation valve causes the closure of the entire pipe, which is not always the case, although the approximation is sufficient in most cases. By the way, when isolation valves are open, QGISRed transfers the loss coefficient from the isolation valves to the pipe as a minor loss coefficient.

To take into account that the closure of a valve does not always cancel the flow in the pipe, as in the case you raise, it would be necessary to consider the isolation valves as TCV, that is, as small linear sections inserted into the pipes, which only cancel the flow in these sections.

With this aim, in version v0.16 we have planned the option of automatically integrating the selected isolation valves into the model to convert them into TCV, adding the nodes at each end, as now is available in QGISRed for service connections. In addition, it would be possible to reverse the operation and extract them to convert back into overlapped points.

In fact, in the model you show I think you have not represented all the isolation valves that exist in the network (because you would be isolating very large areas), but only those used in the sectorization, so that in version v0.16 it will be possible model the last ones as TCV and the rest as overlapped points.

Finally indicate that, if you had represented the service connections to each house, and loaded the demands to the connections, the current closure of an overlapped isolation valve, although hydraulically cancels the flow throughout the entire pipe, does not cancel the demands in those connections that can still be fed by allowing the entry of water through the ends. These demands are in fact still loaded to the end nodes when the hydraulic model is exported to EPANET.

In your case, the demands has been assigned to the network nodes by proximity (I don’t now if using the new capability included en v0.15), without associating them to services connections, and as consequence the best to accurately reproduce the flows is to proceed as you are doing, considering the isolation valves as TCV, although they can be confused with the regulating valves (PRV, etc). But consider that ignoring the services connections, you are already making an approximation.

In summary, in the next version v0.16 you will be able to decide how to consider the isolation valves, as overlapped points, with a different symbology and treatment rules, or as regular hydraulic valves typed as TCV. But currently, in v0.15 he have done our best so that overlapped valves reproduce quite accurately the behavior they would have as inserted TCV valves.

[diogomarin]

Hi Fernando,

Great that you intend to separate the control valves from the block valves.

I just wanted to portray what I had observed of isolation valves. All hydraulic models that I develop, the block valves are classified in TCV. See the image below for a model with 605 TCV valves.

And I agree with you that it greatly harms the visualization of the model, an excellent option that you are planning to have the option of transforming isolation valves (point) into TCV valves.

And yes, if I worked with connection services, I wouldn't really have this problem. But, I'm preparing a video about services connections, I intend to create another topic here on Github. I intend to show the ways I load the nodes in the hydraulic model, I currently work with 5 ways of distributing demand, it all depends on the database I receive.

I really appreciate what QGISRed has developed, it has contributed a lot to my work over the last two years. I believe that there are many tools that can improve QGISRed and I intend to continue evaluating and contributing.

[neslerel]

Thank you very much @diogomarin for your comments!

Only to clarify, the loss coefficient from the isolation valves will be transfered to the pipe as a minor loss coefficient in v0.16. In the current version (0.15) is not available yet.