Setting thermal solver parameters

Steady state convergence criteria

Automatic specifies that the criterion of maximum temperature change of any element is defined from the previous iteration. When temperature change of all elements is below this level, the solution is converged.

Specify allows you to define the Maximum Temperature Change criterion. Typically a value of 0.001°C will give adequate results. However, temperature stability is not a true indication of model convergence. In this case, you should always check the overall heat balance.

Select the steady state Heat Imbalance check box to establish an additional convergence criterion based on global heat balance.

• Select Global Fraction to apply a convergence criterion based on the model heat imbalance normalized by the heat flow into the sinks (fixed-temperature elements).

• Select Absolute to define the heat balance criterion based on the difference between the heat applied to the model and the heat flow into the sinks.

The steady state Iteration Limit option specifies a limit on number of iterations for the steady state analysis. This sets a hard limit on solution time.

Transient convergence control

• The transient Maximum Temperature Change option specifies the maximum allowable temperature change of any element from the previous iteration. When temperature change of all elements is below this level the solution is converged. Typically a value of 0.1 Δ°C will give adequate results. However, temperature stability is not a true indication of model convergence. You should always check the overall heat balance.

• The transient Iteration Limit option specifies a limit on number of iterations the thermal solver performs at each time step of the transient analysis. This sets a hard limit on solution time.

Thermal solver

• A Relaxation Factor can increase the stability of the thermal solution. At every iteration, the solver calculates new temperatures based on the element heat imbalance. The new temperature Tnew is calculated by:Tnew = (1 - δ) × T old+ δ × T eqwhere δ is the relaxation factor and T eq is the element's equilibrium temperature, at which its heat flows are balanced. Values of less than 1 will damp the iterative solution process, increasing both stability and solution time. For models that are highly non-linear (for example, dominated by radiation or fluid effects), the iterative solution can oscillate or even diverge. Use smaller relaxation factors (to increase stabilizing effect) combined with a higher iteration limit to ensure convergence for such models. Values of 0.1 or less may be required. For linear models (conduction dominated), set the Relaxation Factor to 1 (off) to minimize solve time.

• The Convergence Trace option outputs to the solver log file a comprehensive diagnostic output of the model's convergence parameters at every iteration. It can be used to determine if a solution is oscillating, diverging, or converging.

Duct flow solver

In addition to the maximum temperature change convergence criterion, the duct flow solver also uses the pressure convergence criterion that you set in the Pressure Convergence Criterion box.

Iteration Limit sets the value for the maximum number of duct flow iterations the thermal solver performs for a given time step without converging.

As for the thermal solver, you can select the Convergence Trace check box to output to the solver log file, the model convergence parameters for the duct flow network at every iteration.

Conjugate gradient solver

The conjugate gradient solver uses a biconjugate gradient method with a Newton Raphson scheme for non-linear terms to solve the system of equations describing the thermal model. This solution algorithm is in general much faster than the Jacobi method, particularly for ill-conditioned linear problems. You can control the behavior of the conjugate gradient solver through the following parameters.

The Convergence Norm specifies the type of norm that is used to calculate the convergence residuals for temperatures.

The Convergence Criterion tells the conjugate gradient solver when converged results are achieved. For the MAX option, the Convergence Criterion is set to 1e-07 by default. For the L2 option, the Convergence Criterion is set to 1e-08 by default. The default option is MAX.

Iteration Limit sets a limit to the number of iterations the conjugate gradient solver will perform. Sometimes, for hard to converge models, the default of 100 can be too low, a value of 300 is recommended for these cases.

The Preconditioner Matrix Fill Value specifies the maximum number of additional terms in each row of the preconditioning matrix. If the conjugate gradient solver reaches its iteration limit without converging, the fill value is automatically increased by the Preconditioner Matrix Fill Value, the preconditioning matrix is recreated and the solution process is repeated until convergence is achieved.

Element discretization

Element Center Method

Creates conductances between element centers.

Element CG Method

Creates conductances between boundary elements and between the element CG and its boundary elements. When you select this method you can also specify which methods for Shell Accuracy, Solid Element Capacitance Distribution, and Element Convention for Temperature Constraint you want to use.

You can set the Element Discretization, Shell Accuracy, Solid Element Capacitance Distribution, and Element Convention for Temperature Constraint customer defaults to specify the methods that are selected by default in the Solver Parameters dialog box.

Tip:

To find a customer default, choose File→Utilities→Customer Defaults, and click Find Default .

Boundary elements

Boundary elements are created on each 3D solid element surfaces, 2D planar element edges, and 1D element ends except if the surface of the solid element or the edge of a planar element on which it is to be created is already occupied by another element. Boundary elements are only created for elements that have non-zero thermal conductivity.

Triangular element	Triangular element with its three 1D boundary elements	Conductances between the CG of the triangular element and its three 1D boundary elements

Duct thermal discretization

Both Advection and No Conduction and Advection and Conduction methods assume a constant temperature profile along the length of the duct element equal to the center of gravity temperature. These methods lump the effect of external heat and, in the case of advection and conduction method, conduction to the element inlet, where the temperature drops abruptly from the inlet temperature to the constant center of gravity temperature.

The Exponential Advection method assumes an exponential temperature profile along the length of a duct element due to the conduction to non-duct elements. The exponential profile is used to calculate the element average temperature and the element outlet temperature. The outlet temperature is then used as the inlet temperature of the next downstream element in the duct network.

You can set the Duct Thermal Discretization customer default to specify the default for the duct thermal discretization method.

Thermal coupling

You can select the Elements Part of a Thermal Coupling Do Not Convect to Environment check box to ensure that thermal coupling elements do not convect to the environment when you define a thermal coupling and a convection to environment boundary condition on the same selection.

How do I

How to choose values for Maximum Normalized Velocity Change and Maximum Normalized Pressure Change

Learn more

Adjusting solver parameters

Setting flow solver parameters

Setting coupled solver parameters

Understanding radiation parameters

Setting relaxation factors for the flow solver

Defining a time step for a flow analysis

Understanding the freeze flow field options

Quick links

Command reference

Pre/Post video examples

Bulk Entry Descriptions

Simcenter 3D tutorials

Browse Simcenter 3D help by product area

Simcenter 3D Thermal/Flow, Electronic Systems Cooling, and Space Systems Thermal boundary conditions

Setting thermal solver parameters, Simcenter 3D 2021.1 Series

© 2020 Siemens

window.mainLanguage="en_US"

window.delivId=""

window.projectId=""

MathJax.Hub.Config({ TeX: { extensions: ["autoload-all.js"] }, tex2jax: { displayMath: [ ] }, "SVG": { scale: 125 } });