Delft3D 4 Suite vs. Delft3D FM Suite (D-Morphology)
Modeling the evolution of
tidal inlet systems
Abdel Nnafie
Luitze Perk
Abdel Nnafie
Luitze Perk
Test case: Tidal inlets
Source: EMODnet, 2014
𝐖𝐚𝐝𝐝𝐞𝐧 𝐒𝐞𝐚
𝐍𝐨𝐫𝐭𝐡 𝐒𝐞𝐚
𝐕𝐥𝐢𝐞𝐥𝐚𝐧𝐝
𝐓𝐞𝐱𝐞𝐥
𝐓𝐞𝐫𝐬𝐜𝐡𝐞𝐥𝐥𝐢𝐧𝐠
𝐀𝐦𝐞𝐥𝐚𝐧𝐝 𝐏𝐢𝐧𝐤𝐠𝐚𝐭
𝐋𝐚𝐮𝐰𝐞𝐫𝐬
𝐒𝐜𝐡𝐢𝐥𝐝
𝐓𝐞𝐫𝐬𝐜𝐡𝐞𝐥𝐥𝐢𝐧𝐠, 𝐳𝐨𝐨𝐦𝐞𝐝 𝐢𝐧
Nether-
lands
North
Sea
Objective
Verification of Delft3D FM relative to Delft3D 4
Test case: Tidal inlet-basin system
Models should be able to simulate the ebb-tidal delta+
observed channel-shoal network in tidal basin
Model description
Hydrodynamics: Shallow water equations
➢
𝜕𝐷
𝜕𝑡
+ 𝛻 ∙ 𝐷 Ԧ𝑣 = 0
➢
𝜕𝑣
𝜕𝑡
+ 𝛻 ∙ Ԧ𝑣 Ԧ𝑣 + 𝑓𝑒 𝑧 × Ԧ𝑣 = −𝑔𝛻η-
𝑔 𝑢2+𝑣2
𝐶2 𝐷
Ԧ𝑣+
1
𝐷
(𝛻 ∙ 𝐷υ𝛻) Ԧ𝑣
• Morphodynamics:
➢Van Rijn (1993), bedload + suspended load transport
➢𝐵𝑒𝑑 𝑠𝑙𝑜𝑝𝑒 𝑎𝑐𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑓𝑜𝑟.
➢𝐵𝑒𝑑 𝑙𝑒𝑣𝑒𝑙 𝑐ℎ𝑎𝑛𝑔𝑒: Conservation of sediment.
➢Alfa_BN=20; Alpha_BS=5, d50=250 µm.
➢ThetaSD=1 (Factor for erosion of adjacent dry cells),
➢ SedThr=0.2 (Minimum water depth for sediment computations)
Methodology
Run Delft3D 4 and Delft3D FM for exactly same configuration with
bottom evolution (~76) years
Compare bed level evolution simulated by the models
FLOW2D3D Version 6.02.13.7545M, Sep 13 2017
(Delft3D_4_03_01)
D-Flow FM Version 1.2.0.53941M, Jul 12 2018
(Delft3D_FM_beta_versie_1.4.6)
Sensitivity simulations: only M2, ThetaSD=0.1, SedThr=0.1,
AlfaBS=1/AlfaBN=1.5 (Default values)
Model domain
𝒚
𝒙
𝐨𝐩𝐞𝐧 𝐬𝐞𝐚
𝐛𝐚𝐬𝐢𝐧𝐢𝐧𝐥𝐞𝐭
𝑧 𝑏 m
𝐒𝐨𝐮𝐭𝐡
𝐖𝐞𝐬𝐭
−𝟏/𝟐𝑳 𝑺,𝒚
𝟎
𝐍𝐨𝐫𝐭𝐡
𝟏
2
3 𝟒
𝑻𝒊𝒅𝒂𝒍 𝑭𝒐𝒓𝒄𝒊𝒏𝒈:
𝐌𝟐 + 𝐌𝟒 + 𝐌𝟔
Northward propagation
Grid design
Morphodynamics: Delft3D 4 vs Delft3D FM
animation
Delft3D 4
Delft3D FM
Snapshots
t = 0 yrDelft3D 4 Delft3D FM
Snapshots
t = 0.44 yrDelft3D 4 Delft3D FM
Snapshots
t = 0.88 yrDelft3D 4 Delft3D FM
Snapshots
t = 1.32 yrDelft3D 4 Delft3D FM
Snapshots
t = 3.07 yrDelft3D 4 Delft3D FM
Snapshots
t = 10 yrDelft3D 4 Delft3D FM
 Delft3D FM seems to be too diffusive
 Problems at seaward boundary
Evolution height bed level
With Delft3D FM, bed level changes are more
rapid than with Delft3D 4 (order months)
Alfa_BS=1, Alfa_BN=1.5 (Default values)
Snapshots
t = 0 yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
Snapshots
t = 0.44 yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
Snapshots
t = 0.88 yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
Snapshots
t = 1.32 yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
Snapshots
t = 3.07 yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
Snapshots
t = 10 yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
Alfa_BS=1, Alfa_BN=1.5 (Default values)
How about the hydrodynamics?
➢Both models are run for a couple of tidal cycles
➢Comparison water elevation, velocity and discharge
Hydrodynamics: seaward boundaries
Hydrodynamics: Tidal inlet
Hydrodynamics: tidal basin
Discharge trough tidal inlet
 Hydrodynamics in the area of the tidal inlet seem to be
approximately the same
 Large differences in velocity at the seaward boundaries
Conclusions
➢Delft3D FM seems to be too diffusive, bed level changes
appear rather quickly.
➢Problems at the boundaries.
➢Is Delft3D FM ready for application?
➢Outlook: different grid size, numerical parameters, etc….
➢Other suggestions from the audience?
Thank you
Source: EMODnet, 2014
𝐖𝐚𝐝𝐝𝐞𝐧 𝐒𝐞𝐚
𝐍𝐨𝐫𝐭𝐡 𝐒𝐞𝐚
𝐕𝐥𝐢𝐞𝐥𝐚𝐧𝐝
𝐓𝐞𝐱𝐞𝐥
𝐓𝐞𝐫𝐬𝐜𝐡𝐞𝐥𝐥𝐢𝐧𝐠
𝐀𝐦𝐞𝐥𝐚𝐧𝐝 𝐏𝐢𝐧𝐤𝐠𝐚𝐭
𝐋𝐚𝐮𝐰𝐞𝐫𝐬
𝐒𝐜𝐡𝐢𝐥𝐝
𝐓𝐞𝐫𝐬𝐜𝐡𝐞𝐥𝐥𝐢𝐧𝐠, 𝐳𝐨𝐨𝐦𝐞𝐝 𝐢𝐧
Nether-
lands
North
Sea

DSD-INT 2018 Morphodynamic evolution of tidal inlet systems using Delft3D FM: An idealized model study - Nnafie

  • 1.
    Delft3D 4 Suitevs. Delft3D FM Suite (D-Morphology) Modeling the evolution of tidal inlet systems Abdel Nnafie Luitze Perk
  • 2.
  • 3.
    Test case: Tidalinlets Source: EMODnet, 2014 𝐖𝐚𝐝𝐝𝐞𝐧 𝐒𝐞𝐚 𝐍𝐨𝐫𝐭𝐡 𝐒𝐞𝐚 𝐕𝐥𝐢𝐞𝐥𝐚𝐧𝐝 𝐓𝐞𝐱𝐞𝐥 𝐓𝐞𝐫𝐬𝐜𝐡𝐞𝐥𝐥𝐢𝐧𝐠 𝐀𝐦𝐞𝐥𝐚𝐧𝐝 𝐏𝐢𝐧𝐤𝐠𝐚𝐭 𝐋𝐚𝐮𝐰𝐞𝐫𝐬 𝐒𝐜𝐡𝐢𝐥𝐝 𝐓𝐞𝐫𝐬𝐜𝐡𝐞𝐥𝐥𝐢𝐧𝐠, 𝐳𝐨𝐨𝐦𝐞𝐝 𝐢𝐧 Nether- lands North Sea
  • 4.
    Objective Verification of Delft3DFM relative to Delft3D 4 Test case: Tidal inlet-basin system Models should be able to simulate the ebb-tidal delta+ observed channel-shoal network in tidal basin
  • 5.
    Model description Hydrodynamics: Shallowwater equations ➢ 𝜕𝐷 𝜕𝑡 + 𝛻 ∙ 𝐷 Ԧ𝑣 = 0 ➢ 𝜕𝑣 𝜕𝑡 + 𝛻 ∙ Ԧ𝑣 Ԧ𝑣 + 𝑓𝑒 𝑧 × Ԧ𝑣 = −𝑔𝛻η- 𝑔 𝑢2+𝑣2 𝐶2 𝐷 Ԧ𝑣+ 1 𝐷 (𝛻 ∙ 𝐷υ𝛻) Ԧ𝑣 • Morphodynamics: ➢Van Rijn (1993), bedload + suspended load transport ➢𝐵𝑒𝑑 𝑠𝑙𝑜𝑝𝑒 𝑎𝑐𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑓𝑜𝑟. ➢𝐵𝑒𝑑 𝑙𝑒𝑣𝑒𝑙 𝑐ℎ𝑎𝑛𝑔𝑒: Conservation of sediment. ➢Alfa_BN=20; Alpha_BS=5, d50=250 µm. ➢ThetaSD=1 (Factor for erosion of adjacent dry cells), ➢ SedThr=0.2 (Minimum water depth for sediment computations)
  • 6.
    Methodology Run Delft3D 4and Delft3D FM for exactly same configuration with bottom evolution (~76) years Compare bed level evolution simulated by the models FLOW2D3D Version 6.02.13.7545M, Sep 13 2017 (Delft3D_4_03_01) D-Flow FM Version 1.2.0.53941M, Jul 12 2018 (Delft3D_FM_beta_versie_1.4.6) Sensitivity simulations: only M2, ThetaSD=0.1, SedThr=0.1, AlfaBS=1/AlfaBN=1.5 (Default values)
  • 7.
    Model domain 𝒚 𝒙 𝐨𝐩𝐞𝐧 𝐬𝐞𝐚 𝐛𝐚𝐬𝐢𝐧𝐢𝐧𝐥𝐞𝐭 𝑧𝑏 m 𝐒𝐨𝐮𝐭𝐡 𝐖𝐞𝐬𝐭 −𝟏/𝟐𝑳 𝑺,𝒚 𝟎 𝐍𝐨𝐫𝐭𝐡 𝟏 2 3 𝟒 𝑻𝒊𝒅𝒂𝒍 𝑭𝒐𝒓𝒄𝒊𝒏𝒈: 𝐌𝟐 + 𝐌𝟒 + 𝐌𝟔 Northward propagation
  • 8.
  • 9.
    Morphodynamics: Delft3D 4vs Delft3D FM animation Delft3D 4 Delft3D FM
  • 10.
    Snapshots t = 0yrDelft3D 4 Delft3D FM
  • 11.
    Snapshots t = 0.44yrDelft3D 4 Delft3D FM
  • 12.
    Snapshots t = 0.88yrDelft3D 4 Delft3D FM
  • 13.
    Snapshots t = 1.32yrDelft3D 4 Delft3D FM
  • 14.
    Snapshots t = 3.07yrDelft3D 4 Delft3D FM
  • 15.
    Snapshots t = 10yrDelft3D 4 Delft3D FM  Delft3D FM seems to be too diffusive  Problems at seaward boundary
  • 16.
    Evolution height bedlevel With Delft3D FM, bed level changes are more rapid than with Delft3D 4 (order months)
  • 17.
  • 18.
    Snapshots t = 0yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
  • 19.
    Snapshots t = 0.44yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
  • 20.
    Snapshots t = 0.88yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
  • 21.
    Snapshots t = 1.32yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
  • 22.
    Snapshots t = 3.07yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
  • 23.
    Snapshots t = 10yr Delft3D FM, A_BS=5,A_BN=20Delft3D FM, A_BS=1,A_BN=1.5
  • 24.
  • 25.
    How about thehydrodynamics? ➢Both models are run for a couple of tidal cycles ➢Comparison water elevation, velocity and discharge
  • 26.
  • 27.
  • 28.
  • 29.
    Discharge trough tidalinlet  Hydrodynamics in the area of the tidal inlet seem to be approximately the same  Large differences in velocity at the seaward boundaries
  • 30.
    Conclusions ➢Delft3D FM seemsto be too diffusive, bed level changes appear rather quickly. ➢Problems at the boundaries. ➢Is Delft3D FM ready for application? ➢Outlook: different grid size, numerical parameters, etc…. ➢Other suggestions from the audience?
  • 31.
    Thank you Source: EMODnet,2014 𝐖𝐚𝐝𝐝𝐞𝐧 𝐒𝐞𝐚 𝐍𝐨𝐫𝐭𝐡 𝐒𝐞𝐚 𝐕𝐥𝐢𝐞𝐥𝐚𝐧𝐝 𝐓𝐞𝐱𝐞𝐥 𝐓𝐞𝐫𝐬𝐜𝐡𝐞𝐥𝐥𝐢𝐧𝐠 𝐀𝐦𝐞𝐥𝐚𝐧𝐝 𝐏𝐢𝐧𝐤𝐠𝐚𝐭 𝐋𝐚𝐮𝐰𝐞𝐫𝐬 𝐒𝐜𝐡𝐢𝐥𝐝 𝐓𝐞𝐫𝐬𝐜𝐡𝐞𝐥𝐥𝐢𝐧𝐠, 𝐳𝐨𝐨𝐦𝐞𝐝 𝐢𝐧 Nether- lands North Sea