The concepts used in problem definition with Ansys Polydata require some explanation. First, the problem that you are solving (for example, analysis of flow through a die) is referred to as a task. Within a given task, the computations that Ansys Polyflow will perform are divided into sub-tasks. There is always at least one task and one sub-task, but there may be more than one sub-task if different physics apply and need to be solved simultaneously.
A task consists of the following parameters:
Geometrical dimension of the problem (2D, 2D 1/2, 2D axisymmetric, 3D, and so on). Note that for 2D axisymmetric calculations with Ansys Polyflow, the symmetry axis coincides with the vertical (y) direction.
Numerical parameters for the calculation (maximum number of solver iterations, initial solution, convergence criteria, and so on).
For evolution calculations, the parameters controlling the continuation scheme (but not the dependence of each computed variable upon the evolution variable, which is defined for each sub-task).
Specification of a steady-state, evolution, or time-dependent calculation.
For time-dependent calculations, the parameters controlling the time-marching scheme (but not the time dependence of each variable or boundary condition, which is defined for each sub-task); for example, the initial and final time values are task attributes, but the time dependence of the flow rate along a particular boundary is a sub-task attribute.
Coupling of sub-tasks. Complex problems can involve several sub-tasks, coupled or uncoupled. By default, all sub-tasks of a given task are assembled into the same system of algebraic equations and solved simultaneously. If coupling is not required, you can request that the sub-tasks be solved sequentially.
For most cases, you will define only one task, but there are cases in which more than one task is required. For example, in an evolution calculation where the initial solution is different from the default initial solution generated by Ansys Polydata, you will need to generate two tasks: a steady-state task for generating the initial solution, and an evolution task that starts from the initial solution computed by the steady-state task.
A sub-task consists of the following parameters:
models or equations being solved (fluid flow, heat transfer, and so on)
domain of definition (that is, the region where the sub-task applies)
model parameters (relaxation times, and so on)
boundary conditions (inlet flow rate, wall temperature, and so on)
material properties (viscosity, thermal conductivity, and so on)
finite element interpolation
for evolution calculations, the dependence of each computed variable on the evolution variable
for time-dependent calculations, the time dependence of each variable or boundary condition
For simple cases (for example, flow of a homogeneous fluid through a contraction), define only a single sub-task. For others, you may need to define multiple sub-tasks. For example, if your problem involves more than one material, you will have to define a sub-task for each material, since only one set of material properties can be defined for a single sub-task.
As an example of a problem with multiple sub-tasks, consider the nonisothermal flow of a viscous fluid through an axisymmetric steel die (shown in Figure 1.2: Flow through an Axisymmetric Die).
For the fluid region, you want to compute heat transfer and motion of the fluid. For the solid (steel) region, you want to compute heat transfer only. Each of these requires a different sub-task, because each will solve different equations and use different material properties. The two sub-tasks must be solved simultaneously, because the global solution depends on the values of the variables (for example, temperature) at the interface between the two regions. Figure 1.3: Sub-Tasks shows some other examples requiring different numbers of sub-tasks.