# Convergence Issues After Introducing Interfaces, Modifying Boundary Conditions, and Converting Mohr-Coulomb to Finn/Byrne **URL:** \ **Category:** FLAC\ **Tags:** suggestions, flac3d\ **Created:** [October 8, 2026, 2:17pm UTC](https://forum.itascainternational.com/t/convergence-issues-after-introducing-interfaces-modifying-boundary-conditions-and-converting-mohr-coulomb-to-finn-byrne/2418 "2026-10-08T14:17:36Z")\ **Posts on this page:** 1\ **Page:** 1
**Author:** ![Sudipta](https://avatars.discourse-cdn.com/v4/letter/s/f475e1/32.png) [@Sudipta](https://forum.itascainternational.com/u/Sudipta)\ **Post date:** [October 8, 2026, 2:17pm UTC](https://forum.itascainternational.com/t/convergence-issues-after-introducing-interfaces-modifying-boundary-conditions-and-converting-mohr-coulomb-to-finn-byrne/2418/1 "2026-10-08T14:17:36Z")
**Title: FLAC3D – Convergence Issues After Introducing Interfaces, Modifying Boundary Conditions, and Converting Mohr-Coulomb to Finn/Byrne** Hello everyone [@itascan](https://forum.itascainternational.com/groups/itascan) I am working on a 3D FLAC3D model consisting of multiple soil and structural components, constructed in several stages. I am experiencing convergence difficulties after introducing additional interfaces and modifying the connection and boundary conditions, particularly during the fluid-flow and subsequent mechanical equilibrium stages. The previous version of my model converged successfully. However, after modifying the contact conditions and removing certain restraints, I have been unable to achieve satisfactory convergence. However, after introducing the interfaces, I encountered an issue when trying to initialize certain gridpoint properties for specific zone groups. For example, when executing a command similar to: ```auto zone gridpoint initialize fluid-modulus 0.0 ... range group 'GroupA' or 'GroupB' ``` FLAC3D returns the following error: ```auto +++ No valid gridpoints in range. ``` The zone groups were previously defined, and the issue appeared after modifying the model to include interfaces. I would appreciate guidance on whether the problem is related to interface behavior, boundary-condition changes, pore-pressure initialization, the fluid–mechanical solution sequence, or the subsequent constitutive-model conversion. ### 1. Model Description and Recent Modifications The model includes: - Mohr-Coulomb soil zones. - Elastic structural components. - Geogrid structural elements connected to adjacent zones. - Contact interfaces between different structural and soil components. - Staged construction followed by fluid-flow and mechanical equilibrium calculations. - Subsequent conversion of saturated soil zones to the Finn model using the Byrne formulation. **Previous model:** - The bottom surfaces of certain components were connected to the underlying soil using `zone attach by-face`. - Some bottom surfaces were fixed in all displacement directions. - The model successfully converged with these conditions. **Updated model:** 1. Removed the corresponding `zone attach by-face` commands. 2. Introduced interfaces between the upper components and underlying soil using `zone interface ... create by-face separate`. 3. Removed the previously prescribed bottom displacement constraints. 4. Released temporary lateral restraints after staged construction. 5. Retained the necessary external foundation boundary conditions, structural support restraints, and continuity/symmetry conditions. The objective of these changes is to allow contact-controlled interaction and relative movement instead of enforcing direct attachment or fixed displacement at the contact surfaces. ### 2. Interface Properties The newly introduced bottom interfaces have the following properties. **Soil–soil interface:** - Normal stiffness: `9.9e8` - Shear stiffness: `9.9e8` - Friction angle: `30°` - Cohesion: `0` - Tensile strength: `0` **Structural component–soil interface:** - Normal stiffness: `5.0e7` - Shear stiffness: `5.0e7` - Friction angle: `20°` - Cohesion: `0` - Tensile strength: `0` Other interfaces are also present between adjacent structural and soil components. The bottom interfaces are created during the initial construction stage, while additional interfaces are introduced during subsequent stages. ### 3. Fluid-Flow Calculation After staged construction and releasing the temporary constraints, I configure the model for fluid flow. Only the permeable soil zones participate in the flow calculation. The structural components are assigned a null fluid constitutive model. The relevant commands are: ```auto model configure fluid zone fluid cmodel assign isotropic range group 'SoilGroup1' zone fluid cmodel assign isotropic range group 'SoilGroup2' zone fluid cmodel assign null range group 'StructuralGroups' model fluid active on model mechanical active off zone gridpoint initialize fluid-modulus 1000.0 zone initialize fluid-density 1000.0 zone gridpoint initialize fluid-tension 0.0 ``` The hydraulic properties are approximately: - Soil permeability: `1.0e-9` - Porosity: `0.33–0.37` - Fluid density: `1000 kg/m³` The fluid-flow calculation is performed using: ```auto model solve fluid ratio-flow 1.0e-5 ``` ### 4. Pore-Pressure Initialization and Boundary Conditions A hydrostatic pore-pressure distribution is applied based on a specified phreatic level. Initially, the saturated and unsaturated zones are identified, and saturation is initialized accordingly. For example, the code uses: ```auto zone gridpoint initialize saturation 1.0 ... range group 'SoilGroup1' ... zone gridpoint initialize saturation 1.0 ... range group 'SoilGroup2' ... ``` The hydrostatic pore-pressure boundary condition is applied using a position-based range: ```auto zone face apply pore-pressure 8.10e4 ... gradient (0,0,-1.0e4) ... origin (0,0,-6.60) ... range position-z -6.60 1.50 ``` A separate zero-pore-pressure boundary condition is specified above the phreatic level. **One concern is that the pore-pressure boundary conditions are selected using `range position-z` without explicitly restricting the selection to specific soil boundary-face groups.** Since the updated model contains several interfaces created using `by-face separate`, I am uncertain whether this selection procedure could unintentionally include additional faces or influence the intended hydraulic boundary conditions. I would appreciate clarification on whether pore-pressure boundary conditions should be applied only to explicitly grouped exterior soil faces, particularly in models containing multiple internal interfaces. ### 5. Mechanical Adjustment After the Fluid-Flow Calculation After establishing the phreatic field, I deactivate fluid flow and reactivate the mechanical calculation while retaining the pore-pressure field. The relevant commands are: ```auto model fluid active off model mechanical active on zone gridpoint initialize fluid-modulus 0.0 ... range group 'SoilGroup1' zone gridpoint initialize fluid-modulus 0.0 ... range group 'SoilGroup2' zone gridpoint initialize displacement (0,0,0) zone gridpoint initialize velocity (0,0,0) model solve ratio-local 1.0e-4 ``` During staged construction, however, the equilibrium calculations use: ```auto model solve ratio 1.0e-4 ``` Therefore, the construction stages and the subsequent mechanical adjustment use different mechanical convergence measures. I am unsure whether this difference affects how convergence should be evaluated in the updated model, particularly with the additional contact interfaces. ### 6. Conversion from Mohr-Coulomb to Finn/Byrne After the phreatic calculation and the corresponding mechanical adjustment, I convert the saturated soil zones from Mohr-Coulomb to the Finn constitutive model using the Byrne formulation. The conversion is performed for the saturated portion of one soil group and the entire saturated underlying soil domain. The relevant commands are: ```auto model configure dynamic model large-strain on model fluid active off model mechanical active on model dynamic active off zone cmodel assign finn ... range group 'SoilGroup1' ... zone cmodel assign finn ... range group 'SoilGroup2' ... ``` After assigning the Finn model, the relevant mechanical properties and Byrne parameters are assigned: ```auto zone property constant-1 0.14736 ... constant-2 1.35391 ... flag-switch 1 ... number-latency 1000000 ... range group 'SoilGroup1' zone property constant-1 0.14736 ... constant-2 1.35391 ... flag-switch 1 ... number-latency 1000000 ... range group 'SoilGroup2' ``` The velocities and displacement reference are then reset, and another mechanical equilibrium calculation is performed: ```auto zone gridpoint initialize velocity (0,0,0) zone gridpoint initialize displacement (0,0,0) model solve ratio-local 1.0e-4 ``` After this step, the `number-latency` parameter is changed from `1000000` to `50` in preparation for a subsequent dynamic calculation. I would like to confirm whether this approach is appropriate for initializing the Finn/Byrne constitutive model after establishing hydrostatic pore pressures and introducing multiple contact interfaces. ### 7. Summary of Convergence Criteria The model uses the following solution sequence: | Calculation stage | Convergence criterion | | --- | --- | | Staged construction | `model solve ratio 1.0e-4` | | Equilibrium after releasing temporary restraints | `model solve ratio 1.0e-4` | | Phreatic fluid-flow calculation | `model solve fluid ratio-flow 1.0e-5` | | Mechanical adjustment after fluid flow | `model solve ratio-local 1.0e-4` | | Mechanical equilibrium after Finn/Byrne conversion | `model solve ratio-local 1.0e-4` | ### 8. Questions I would greatly appreciate any guidance on the following: 1. **Interfaces and convergence:** Could replacing `zone attach by-face` with contact interfaces, while removing the corresponding displacement restraints, cause convergence difficulties during the fluid-flow or subsequent mechanical stages? 2. **Interface initialization:** After creating interfaces using `by-face separate`, are additional steps necessary to establish proper initial contact, stress transfer, and mechanical equilibrium? 3. **Boundary-condition changes:** Is there a recommended procedure for releasing previously fixed contact surfaces, particularly when several interfaces are introduced during staged construction? 4. **Pore-pressure boundary selection:** Could applying hydrostatic pore-pressure conditions using only `range position-z`, without explicitly selecting exterior soil faces, cause unintended boundary-condition assignments after interfaces are introduced? 5. **Hydraulic behavior across interfaces:** Do mechanically separated contact surfaces require any additional consideration to maintain the intended hydraulic connectivity between adjacent permeable soil regions? 6. **Fluid–mechanical solution sequence:** Is it appropriate to establish the pore-pressure field with the mechanical calculation inactive, deactivate fluid flow, set the soil fluid modulus to zero, and then perform mechanical equilibration? 7. **Mechanical convergence criteria:** During the mechanical adjustment, is `model solve ratio-local 1.0e-4` an appropriate criterion? Could the difference between `ratio` and `ratio-local` explain why the new model behaves differently from the previous version? 8. **Finn/Byrne conversion:** Is converting saturated soil zones from Mohr-Coulomb to Finn/Byrne after the phreatic and mechanical equilibration stages an appropriate procedure? 9. **Equilibrium after Finn assignment:** Could the constitutive-model conversion itself cause mechanical nonconvergence? Are additional initialization steps recommended before solving again? 10. **Large-strain setting:** Could changing from `model large-strain off` during construction and fluid equilibration to `model large-strain on` before the Finn equilibrium calculation affect convergence when multiple interfaces are present? 11. **Byrne parameters:** Is temporarily assigning a very large `number-latency` value during mechanical equilibration and then reducing it before dynamic calculations an appropriate procedure? 12. **Troubleshooting strategy:** What would be the most effective way to determine whether the convergence problem originates from contact stiffness, interface separation or sliding, boundary conditions, hydraulic connectivity, pore-pressure initialization, mechanical equilibrium, or the constitutive-model conversion? I would particularly appreciate recommendations on the correct sequence for interface creation, boundary-condition release, phreatic initialization, fluid-flow equilibrium, mechanical adjustment, and Finn/Byrne initialization. Thank you very much for your time and assistance.