qgisred / QGISRed

GNU General Public License v2.0
23 stars 8 forks source link

Request, demand allocation consider different diurnals #83

Open Lolaento opened 2 years ago

Lolaento commented 2 years ago

Hello again,

I've noticed that with the Punctual consumption, demand allocation it doesn't take into consideration the pattern of the individual meters. Right now when using the punctual consumption it sums the demand from individual meters onto the closest junction. Useful, unless the meters have a different pattern (could be commercial and residential), which case atm I am fixing them manually after importing the demands.

Will something like this be implemented in the future?

Lolaento

fmartine commented 2 years ago

Hi @Lolaento,

QGISRed is thought to be used in different ways concerning demand allocation. If you have only average demands as punctual consumptions, you can use the current option to allocate the demands by adding them at nodes. For version 0.16 we are developing a new tool to assign demand patterns automatically to nodes by sectors, or just to the selected nodes. So if you use the label to classify the nodes, you can use it to select the nodes before assigning the corresponding pattern. But if the classification is done at the level of punctual consumption, you can still separate them by type o consumption in different shapes. Then first allocate one type of demand, assign a demand pattern to the nodes with non zero demands, export the scenario to a file and delete all demands to build another scenario. Following that you can export different files, one for each type of demand, including in each file the nodal base demands and IDpattern for the corresponding nodes. Finally you can import all files additively as multiple demands for the shared nodes. All these options will be available in the next version 0.16

But this could be cumbersome and will be justified only if you are going to do it just once. Another option we are also developing for version 0.16 is to convert automatically punctual consumptions in service connections. Now each connection can have its own base demand and ID pattern, and they are allocated to nodes automatically as multiple demands (one per connection linked to the node) each time a simulation is launched. But if there are thousands of connections, you will have also thousands of IDpatterns, which is inefficient. So you will have the option to aggregate demands and patterns to get a unique equivalent pattern per node, even differentiating connections with patterns from those without. But in this case you still will have one IDpattern per node which can be excessive, so another option is leave pattern blank and finally assign patterns by sectors, or separate demands just by type and then assign a unique pattern per type of demand.

Finally, how do you get a different pattern for each punctual consumption? Are patterns statics o dynamics? If you have access to Automated Meter Readings (AMR), then patterns are changing every day. The QGISRed solution will come in future versions to address this challenge, but current version already offer an algorithm to do it in the menu option Load Meter Readings.