|
|
When scenarios were first introduced the way they were implemented didn't seem intuitive. With increasing interest in such as adaptive pathways the number of scenarios can rapidly increase because each scenario is its own network.
This proposal is about changing the approach to scenarios so that elements can be maintained as separate entities (layers) so they can be brought together at simulation time in different combinations.
If you have a base network and want to look at proposed development in 10 years, 20 years and 30 years time the existing scenario system requires that you have four networks. If you then want to improve the network in the 20 year model it adds another scenario and more again if you want to consider several options. Factor in that development within a catchment won't happen in the order you might think initially and there is spatial variation to consider. Every combination is a different network.
It is very difficult to keep track of the (say) 20 year development proposals if a developer or local government policy chooses to progress part of the 20 year plan in the 10 year plan and defers part of the 10 year plan. That's a lot of scenarios to change.
With both time and spatial variations it is quite easy to see that the number of scenarios needed quickly spirals out of control.
If scenarios are changed from being whole networks to only containing changes, the 10 year development proposals could be maintained as an entity. It wouldn't be run on its own but would be added as 'layer' on to the base at simulation time. Similarly the 20 year development would be the additional development to add on top of the 10 year. As development plans change some areas might be moved from the 20 year layer to the 10 year. And as development is built-out it could be moved into the base model. maintaining the 10 year and 20 development layers would be much easier.
When putting together options for (say) flooding each option would be a separate 'layer' to run with the base (and other layers, such as the 30 year development proposals).
Setting-up a simulation would start by picking the base and then choosing which modifications to make to the base by overlaying the required layers. So this could be Base+ 20 year development+ Option A.
Option A might consist of building some new sewers, upsizing others and abandoning others so it would need to be validated against the base to ensure that it is a functioning network. Similarly a 20 year development plan would need to connect to the existing system so could include a notional network which itself could be modified in time.
It follows that versioning for the base would be separate from each 'layer'. Each layer would have its own versioning. When a simulation is run it would be clear as to which version of the base and which version of each layer was used. Solution options may go through several versions but the base would see far fewer changes.
