Refinement Overview 🌟

[xldeltares]

Objective:
convert and calibrate a hybridurb model from a Delft3DFM 1D model. Test case Eindhoven.

About Delft3DFM model:
model topology: rr + dflowfm
model results: rainfall + flood on manholes

About hybridurb:
graph model: distributed physical network + semi-distributed hydrological network.
regression model: static features (sewer physical parameters) + dynamic feature (hydrological process) --> probability of flooding

About test case Eindhoven:
sewer network + hydraulic structures (pumps and weirs)
Memo on hybridurb.docx
rionet design rainfall
Rainfall_NL_RIONED.xlsx

Steps

1. convert delft3dfm to geometry files --> get graph topology and physical params (pipe size, roughness, capacity, bedlevel etc)
2. conduct graph optimization --> get flow path and hydrological params (tc, ua, e) to run hydrological processes
3. Run Eindhoven and graph model for multiple rainfall scenarios --> get calibration dataset
4. calibrate regression model --> get complete hybridurb model

MileStones

1. Graph representation of Eindhoven Model

• Convert dflowfm model files to geometry files #11
• Get sewer parameters from delft3dfm files: pipe/pump at edges #21
• Get sewer parameters from delft3dfm files: manholes/weirs at nodes #24
• Construct the graph model from geometry files 👍 #14
• Add edge attributes required for computing maximum capacity, and compute maximum capacity #35
• Flow Path delineation for hydrological params 👍 #25
• network optimization #36 (joined development with hydromt_delft3dfm)
• Construct function to create 1D grid from graph and create graph from 1D grid. #23
  [optional]
• Convert rr files to geometry file #12
• Get rainfall runoff parameters from rr files: e.g. at nodes. #22
• Construct RR info to graph 👍 #26

2. Run model (rainfall runoff and/or runoff routing) For Eindhoven model

• Add routing model 👍 #38 (vision Allow 2D model output as input (thinking about how to separate rr from sewer) #30)
  [optional]
• Convert rr files to general settings #18
• Construct RR module based on NWRW 👍 #17

3. Calibration of regression module for Eindhoven model

• Prepare workflow to generate Delft3D FM model scenarios #16
• Prepare workflow to generate graph model scenarios #31
• Prepare calibration dataset (Delft3D FM) #13
• Prepare calibration dataset (graph) #32
• Calibrate regression model 👍 #15 (vision change regression model to neural network #28 & Allow use of state for operational use. e.g. water level #29 )

5. Hybrid model for Eindhoven

• Run hybridurb Eindhoven model 👍 #27

Planning
https://github.com/users/xldeltares/projects/1/views/3?filterQuery=

Attachement (flow chart)