Network generation for Egg model

[tayloris]

This figures shows the network generated by Flownet for the Egg reservoir model.

Aditional to the connections Injector - Producer, you will connections producer-producer and also connections Injector - Injectors. I understand these connections are generated by default. However, there are connections that should no be there (in my opinion). There is a connection, for example, the one between injector 2 and injector 6 that i don't understand. Both wells are far away from each other.

It would be interesting to include in the configuration file an option to influence the creation of the network in flownet.

[wouterjdb]

This is the result of the connection being generated in 3D. It looks like the path cross one another but in fact they are above/below each other. I agree with you that it would be nice to have more flexibility in how the network is generated. I'm not sure at this point how though. Suggestions? Maybe ignoring a path that is less than x meters above/below another path if it is shorter?

However, I do not in general agree with the assumption that, based on two wells being far away, they should not be connected to one another directly. It might be less likely that they have a direct communication path, but it is the geology that dictates whether or not this is the case. We do not know that a priori and, letting the assisted history match figuring out whether or not a direct communicating path is needed, seems to be the more data-driven approach in my opinion.

If there is no further idea on how to go on with this I suggest closing this issue for now.

[wouterjdb]

As a result of #128 we now have the possibility of setting the maximum length of a connection. Maybe that is a potential solution for this problem as well?