Reference Architecture for Gearbox Thermal-Structural Co-Simulation: Multirate Coupling with FMI 3.0, SSP 1.0 and ISO 6336

Gear mesh dynamics run at kilohertz mesh frequencies, while housing and lubricant temperatures change over minutes. The main architectural decision is therefore whether to couple the friction and thermal domains monolithically, through FMI 3.0 Model Exchange, or through multirate FMI 3.0 Co-Simulation with explicit or iterative exchange. Exchange interfaces should be packaged as SSP 1.0 system descriptions, gear-rating checks should trace back to ISO 6336 (2019 parts) and ISO/TS 6336-20:2022 for flash temperature, and OpenUSD should be used only for geometry and scene interchange, not for solver coupling.

Coupling selection criteria

Use explicit (Jacobi or Gauss-Seidel) FMI 3.0 Co-Simulation when the thermal time constant exceeds the mesh period by at least three orders of magnitude and friction depends only weakly on temperature. Switch to iterative coupling with fmi3EnterStepMode and state rollback, or to monolithic Model Exchange, when the lubricant's viscosity-temperature sensitivity makes the coefficient of friction and the mesh power loss feed back within a single macro-step.

Pre-implementation checklist

Before you write the SSP, collect the gear micro-geometry and mesh stiffness per ISO 6336-1, the lubricant viscosity at 40 and 100 °C with its pressure-viscosity coefficient, the bulk and flash temperature limits per ISO/TS 6336-20, the convection boundary conditions of the housing, and the target macro-step. Also confirm that every FMU declares canGetAndSetFMUState if you plan to use rollback.

Illustrative analytical scenario (not a measurement)

A 23-tooth pinion at 3,000 rpm gives a mesh frequency of about 1,150 Hz, so a 20 µs mesh solver step against a 50 ms thermal macro-step is a 2,500:1 rate ratio. If the energy residual at the interface exceeds 1% of the transmitted power, halve the macro-step or enable iteration. If an FMU fails, restore the last saved FMU state and fall back to a quasi-static thermal model, and log the event in the SSP run record.

How do I verify that MechSim's multirate coupling is not introducing artificial energy into the gearbox model?

Monitor the energy balance at each interface: mechanical input power minus output power should equal the mesh and bearing losses plus the change in stored kinetic energy, and that sum should equal the heat flux the thermal FMU receives. Assume the losses are piecewise constant within each macro-step. If the cumulative residual grows monotonically, reduce the macro-step, use extrapolation of the exchanged variables, or move to iterative coupling with FMU state rollback (FMI 3.0). Validate against a monolithic Model Exchange reference run on a reduced case before you scale up throughput.

Back to MechSim · Domains

← Back to MechSim
Engineering & Architecture Brief

Reference Architecture for Gearbox Thermal-Structural Co-Simulation: Multirate Coupling with FMI 3.0, SSP 1.0 and ISO 6336

Gear mesh dynamics run at kilohertz mesh frequencies, while housing and lubricant temperatures change over minutes. The main architectural decision is therefore whether to couple the friction and thermal domains monolithically, through FMI 3.0 Model Exchange, or through multirate FMI 3.0 Co-Simulation with explicit or iterative exchange. Exchange interfaces should be packaged as SSP 1.0 system descriptions, gear-rating checks should trace back to ISO 6336 (2019 parts) and ISO/TS 6336-20:2022 for flash temperature, and OpenUSD should be used only for geometry and scene interchange, not for solver coupling.

Coupling selection criteria

Use explicit (Jacobi or Gauss-Seidel) FMI 3.0 Co-Simulation when the thermal time constant exceeds the mesh period by at least three orders of magnitude and friction depends only weakly on temperature. Switch to iterative coupling with fmi3EnterStepMode and state rollback, or to monolithic Model Exchange, when the lubricant's viscosity-temperature sensitivity makes the coefficient of friction and the mesh power loss feed back within a single macro-step.

Pre-implementation checklist

Before you write the SSP, collect the gear micro-geometry and mesh stiffness per ISO 6336-1, the lubricant viscosity at 40 and 100 °C with its pressure-viscosity coefficient, the bulk and flash temperature limits per ISO/TS 6336-20, the convection boundary conditions of the housing, and the target macro-step. Also confirm that every FMU declares canGetAndSetFMUState if you plan to use rollback.

Illustrative analytical scenario (not a measurement)

A 23-tooth pinion at 3,000 rpm gives a mesh frequency of about 1,150 Hz, so a 20 µs mesh solver step against a 50 ms thermal macro-step is a 2,500:1 rate ratio. If the energy residual at the interface exceeds 1% of the transmitted power, halve the macro-step or enable iteration. If an FMU fails, restore the last saved FMU state and fall back to a quasi-static thermal model, and log the event in the SSP run record.

Technical FAQ: How do I verify that MechSim's multirate coupling is not introducing artificial energy into the gearbox model?

Monitor the energy balance at each interface: mechanical input power minus output power should equal the mesh and bearing losses plus the change in stored kinetic energy, and that sum should equal the heat flux the thermal FMU receives. Assume the losses are piecewise constant within each macro-step. If the cumulative residual grows monotonically, reduce the macro-step, use extrapolation of the exchanged variables, or move to iterative coupling with FMU state rollback (FMI 3.0). Validate against a monolithic Model Exchange reference run on a reduced case before you scale up throughput.

Related Architecture Specifications (DIGITALTWINS): drivingtwins.com · automationsim.com · prototypesim.com

More Technical Articles on mechsim.com

Synthetic Bearing Spall Growth and Acoustic Emission Vibration Benchmarks →Thermal Choking and Churning Power Loss in High-Speed Dual-Clutch Transmissions →MechSim: Reference Architecture, Interoperability & Deterministic Telemetry →