Houdini 20.0 Nodes Dynamics nodes

FEM Solver dynamics node

Sets and configures a Finite Element solver.

On this page
Since 13.0

The FEM Solver DOP sets objects to use the FEM solver.

If an object has this DOP as its Solver subdata, it will evolve itself as a cloth object.

This solver recognizes the following special subdata, if present:

It also recognizes most standard subdata:

Before each solve, the finite element solver reads the simulation state of each object from the attributes on the simulation geometry Geometry to get the previous state. After the solve, the new state is written to attributes the Geometry. In addition, the solver may maintain an EmbeddedGeometry. This would typically be a more higher-resolution geometry that moves and fractures along with Geometry. The embedded geometry can consists of polygons or tetrahedrons (or a mix of these two).

The finite element solver approximates the physics of continuous materials by splitting them up into a finite number of elements. In the case of the Solid object, the elements are determined by the tetrahedrons. In the case of the Cloth Object, the elements are determined by triangles and quadrangles. The resolution of the tetrahedrons and the orientations of the individual tetrahedrons have little influence over the overall movement; as long as the overall solid shape is the same the behavior is roughly the same (except for very coarse meshes). The finite element method (FEM) treats the elements an approximation of a continuous material. This property makes the results very predictable when you simulate the same shape using a lower-res and a higher-res mesh.

Parameters

Substepping

Substeps

This is the number of substeps per frame. The higher this value, the better the quality and the accuracy of your simulation will be. However, the solve time per frame may increase. When you're having trouble with a simulation’s quality or stability, the first thing to try is increasing the Substeps. One particular case where this may be needed is when you have fast moving objects that collide. When the relative velocity is large compared with the feature size, you may need to manually increase the Substeps to get better looking results.

Capabilities

Enable Collisions

When this is disabled, no collisions will happen in the simulation, regardless of any other collision settings on the solver or on any of the objects that are solved. This is a convenient toggle that allows you to quickly see how an object behaves without collisions.

Enable Fracturing

When this is disabled, no fracturing or tearing will occur on any of the Solid Objects or Cloth Objects attached to this solver.

Accuracy

Absolute Tolerance Implicit Solve

The absolute tolerance that is used to decide when the finite element solver has found a good enough approximate solution for the implicit integration step. This roughly indicates the amount of error allowed in the acceleration that is computed in each integration step. It is not recommended to increase this tolerance too much from its default.

Relative Tolerance Implicit Solve

The relative tolerance that is used to decide when the finite element solver has found a good enough approximate solution for implicit integration step. In contrast with the absolute tolerance, this parameter has no units. It is not recommended to increase this tolerance from its default, because this may produce results that are unstable or poor quality. In some types of simulations, it may be necessary to reduce the Relative Tolerance Implicit Solve to a lower value, say 0.0001 or even smaller. For example, the tolerance may need to be lowered in fracturing simulations with very stiff objects.

Max Collision Passes

This is the maximum number of times, within a single substep, that the solver is allowed to detect and resolve new collisions. The collision resolution step in each pass may introduce new, secondary collisions that must be resolved in later passes.

Attributes

Create Quality Attributes

This creates float-valued primitive attribute quality on the simulation geometry. This attribute allows you to find primitives that have a bad effect on the simulation. The closer the value is to 1, the better. Bad tetrahedrons have a value close to 0 and should be removed or improved.

Create Energy Attributes

This creates float-valued vertex attributes on the simulation geometry. These attributes potentialdensity, dissipationdensity, and kineticdensity show the potential energy density, the density of energy dissipation, and the kinetic energy, respectively.

Create Fracture Attributes

This allows the fracturecount point attribute to be created. Any point that was fractured at any time in the simulation will have a fracturecount of 1 or higher ever after.

Advanced

Float Precision

This determines the floating point precision that is used internally by the finite element solver. Float 32 bit uses less memory and is generally faster than Float 64 bit. However, the extra accuracy of 64-bit floating point numbers may be needed when you are simulating object with very high overall stiffness or when your geometry is positioned very far away from the origin.

Integrator Type

This determines the type of integration that is used. The default is Implicit, which is the most stable option for simulations that have fracturing stiff objects. The Implicit Order 2 allows more lively looking simulations of higher quality for about the same solve time. This is best used when doing finite element simulation of flesh on characters.

Max Fully Implicit Passes

For highly nonlinear finite element simulations, increasing this parameter may result in better quality without having to increase the sub-steps. This may be needed for simulations in which the deformations of the tetrahedrons are relatively large.

Attributes

The finite element solver will recognize and use attributes on the simulated geometry. In the DOP network, this simulation geometry is attached to the simulated object as a sim-data with name Geometry. When an object is created, then the geometry and all the corresponding attributes are read from the Initial Geometry. This includes the standard position and velocity point attributes P and v.

The finite element solve supports input attributes and output attributes. Some attributes, such as the simulation state, are both input and output attributes. The input attributes include multiplier attributes for material properties, fracture attributes, and attributes for controlling target positions and corresponding hard/soft constraints. The output attributes include optional attributes for tet quality, energy densities, FEM node forces, collision info attributes and fracture info attributes.

Material Property Multiplier Attributes

Each of the material properties of a simulated object can be locally modified using multiplier point attributes. As a rule, each of the material properties in the Model tab of an object can be affected by a multiplier attribute. As a rule, the name of the parameter is the name of the attribute. The name of the attribute is the name that is displayed after “Parameter:” when you hover over a parameter with your mouse cursor.

You can locally change the material properties of the object using point attributes. For example, you can make some polygons resists stretching and bending more than other polygons. These attributes work as multipliers for the parameters in the Model tab: The stiffness multiplier is a convenient way to modify the local stiffness for all object types that are recognized by the finite element solver:

Name Class Type Description
stiffness Point Float Multiplier for all types of stiffness.
dampingratio Point Float Multiplier for all damping ratios.
massdensity Point Float Multiplier for all mass densities.

For solid objects, the following multiplier point attributes can be used to modify the local behavior:

Name Class Type Description
solidstiffness Point Float Multiplier for both the shape stiffness and the volume stiffness of a Solid Object.
solidshapestiffness Point Float Multiplier for the shape stiffness of a Solid Object.
solidvolumestiffness Point Float Multiplier for the volume stiffness of a Solid Object.
solidmassdensity Point Float Multiplier for the mass density of a Solid Object.

Collision Control Attributes

The FEM solver looks at collision identifiers to decide which primitive pairs are allowed to collide. The rule is that a pair of primitives may collide if they have the same collision identifier. (This mechanism may possibly be extended in a future release, allowing the user to specify exactly which collision identifier pairs may collide.) Collisions can be suppressed altogether for certain primitives by setting the special value -1. A collision id may be specified separately for the interior and the exterior side of each polygon and tetrahedron. The exterior side of a polygon is decided using the winding order convention, just like the normal direction. If interiorcollisionid is not specified, then the default collision id of 0 is used for triangles, but interior collisions are disabled for tets. If exteriorcollisionid is not specified, then the default collision id of 0 is used for both tets and triangles.

Consider, as an example, an FEM muscle simulation, we may want the muscles to collide only with the interior side of the skin polygons, so the exteriorcollisionid for those polygons may be set to -1 (disable).

Name Class Type Description
exteriorcollisionid Primitive Integer Collision identifier for the exterior side of a polygon or tet surface
interiorcollisionid Primitive Integer Collision identifier for the interior side of a polygon or tet surface

The following attributes can be used to locally multiply the Repulsion and Friction parameters:

Name Class Type Description
repulsion Primitive Float Multiplier for the Repulsion of an FEM Object
friction Primitive Float Multiplier for the Friction of an FEM Object

Material Space Attributes

The attribute materialP can be thought of as the positions of the simulated object in the material space. materialP is the undeformed configuration relative to which the current position P determines the deformation of a simulated object. materialP must stay the same throughout the entire simulation. The finite element solver relies on assumption that the materialP attribute remains unchanged from frame to frame; it should never be modified externally (e.g., through a SOP solver) otherwise bad simulation results will be produced.

In the case where no Rest Shape is specified and nor restP attribute is provided, materialP can be thought of as a permanent rest position. If no animation of the rest position is required in a sim, only materialP should be specified (no restP). At any stage in the simulation, it is the mapping from materialP to the current P that determines the deformation of tets in simulated objects. The deformation in turn defines the energy stored inside the object.

To help determine the anisotropic behavior of solids, including fiber contraction, the solver can makes use of local UVW frames. These UVW frames may be specified directly using the vertex/point attributes materialU, materialV, and materialW. Alternatively, they may be inferred from UVW positions that may be specified by a vertex/point position attribute materialuvw. The FEM solver embeds the UVW directions within the material space that may be specified by the attribute materialP. To get the correct idea of how this works, UVW directions that are fed into the FEM solver should be visualized relative to the material position materialP.

For FEM muscle simulations, the easiest way to specify the muscle fiber direction is through vertex/point attribute materialW. It is fine if no materialU and materialV directions are specified in this case, as the solver will infer arbitrary materialU and materialV directions from materialW in that case.

The attribute materialuvw can be used to specify a UVW parametrization of the material space. The U, V and W directions that are implied by materialuvw matter if the anisotropic controls are used or when the fiber controls are used on a simulated object. For the FEM muscle simulation use case, the fiber controls are an important tool for controlling muscle contraction.

Similar to materialuvw, the materialuv attribute can be used to specify UV directions for cloth. This attribute is essential for triangle meshes, in particular, to define the warped and weft directions for cloth.

Name Class Type Description
materialP Point or Vertex Vector Material position of each point, defining the material space
materialU Point or Vertex Vector The U direction within the material space
materialV Point or Vertex Vector The V direction within the material space
materialW Point or Vertex Vector The W direction within the material space
materialuvw Point or Vertex Vector Local material uvw coordinates for each point or vertex of a tet.
materialuv Point or Vertex Vector Local material uvw coordinates for each point or vertex of a polygon or polysoup.

Material Property Multiplier Attributes

fracturepart Primitive Integer Partitions the object into unbreakable parts. Must be either -1 (no part) or a nonnegative number that indicates a part.
enablefracturing Point/Vertex Integer Locally enable/disable fracturing for points or vertices.
fracturethreshold Point/Vertex Float Multiplier for the object’s Fracture Threshold.

Fracturing Control Attributes

When you create a simulation with fracturing, it is recommended to specify chunks of tetrahedrons that you want to stay together. Otherwise, the fracturing process may create a very large amount of separate pieces, many of which may consist of single tetrahedrons. For this purpose, you can assign a nonnegative integer to each chunk using the fracturepart attribute. In areas where you don’t want to specify parts, you can set fracturepart to -1, which means that each primitive in that region will become its own part. Real-life materials tend not to be equally strong everywhere. For realistic results, it is recommended to vary the Fracture Threshold locally using the vertex attribute fracturethreshold.

fracturepart Primitive Integer Partitions the object into unbreakable parts. Must be either -1 (no part) or a nonnegative number that indicates a part.
enablefracturing Point/Vertex Integer Locally enable/disable fracturing for points or vertices.
fracturethreshold Point/Vertex Float Multiplier for the object’s Fracture Threshold.

Drag Force Control Attributes

The behavior of the drag force can be modified locally using the following attributes:

normaldrag Primitive Float Multiplier for the object’s Normal Drag.
tangentdrag Primitive Float Multiplier for the object’s Tangent Drag.

Reference Attributes

The attribute baseP can be used to specify a generic base position for all the object points. This attribute’s values must not be changed during a simulation. When the user does not specify baseP, the solver creates this point attribute based on the point positions on the creation frame. This attribute is used as a fallback; whenever the user does not specify materialP, the attribute baseP is read instead. In the same way, baseP is used as a fallback for when no restP or targetP attributes are provided. Finally, baseP is used to bind the simulated and the embedded geometry, in the embedded workflow (e.g., a T-pose). This embedded binding looks at the baseP position attribute on both the simulated geometry and the embedded geometry. If no baseP attribute is provided by the user on the embedded geometry, the solver creates the baseP attribute on the embedded geometry based on the position P at the creation frame.

If Allow Changing Rest is enabled on the FEM Solver, then the attribute restP may be used to modify rest positions. The attribute restP can be used to specify an animated rest position for all the object points. For example, at each frame restP may be modified in a SOP Solver before the finite element solver. Among other things this makes it possible to create plastic deformation kinds of effects. When the rest should stay the same during an entire simulation the attribute restP should not be used. In that case, it is sufficient to specify only an attribute materialP, which would act as a permanent, unchanged rest position. If no attribute materialP is specified, the solver falls back to the baseP attribute that gets automatically created at the creation frame. table>>

Target Attributes

Target attributes can be used to make a simulated object partially follow a target animation. The attribute targetP can be used to specify a target position for each object point. When you use the Import Target Geometry option on the simulated object, the targetP will be set automatically every frame. Alternatively, you can create and modify these attributes yourself, using a Multi Solver and a SOP Solver. The target positions and velocities allow the user to mix animation and simulation in a very stable way (assuming the Target Strength and Target Damping parameters have been set on the object). You can set the Target Strength and Target Damping parameters on the object to express how strongly the object should match the target position and velocity, respectively. This is a way to create soft constraints. You can use the pintoanimation to create hard constraints that make the simulated points follow targetP exactly.

Name Class Type Description
targetP Point or Vertex Vector Target position of each point.
targetstrength Point Float Multiplier for the object’s Target Strength. If this attribute is missing, a multiplier of 1 is used at all points.
targetdamping Point Float Multiplier for the object’s Target Damping. If this attribute is missing, a multiplier of 1 is used at all points.
pintoanimation Point Int When 1, the point is hard constrained to the target animation (e.g., targetP). When zero, the point is unconstrained.

Fiber Attributes

The fiberscale point attribute acts as a multiplier for the rest strain in the fiber direction. The fiber direction itself can be specified using the materialW vertex/point attribute. Among other things, this is useful for FEM muscle simulations. If the fiberscale is changed from 1 to 0.5, then the muscle wants to be half as long as before in the direction of the fiber. If you animate the fiberscale in a SOP Solver such that it decreases from 1 to a smaller value, you will cause a muscle contraction in the sim.

The fiberstiffness point attribute acts as a multiplier for the stiffness along the fiber direction. The fiber direction of the material is determined by the W axis of the materialuvw coordinates. fiberstiffness works as a multiplier on top of all the other material property multipliers, including the anisotropic multipliers. If the fiberstiffness changed from 1 to 10, then the stiffness along the fiber direction becomes 10 stronger than before. This can be used to control how strong and how quick the effect of muscle flexing using the fiberscale attribute takes effect.

For fiberscale/fiberstiffness to have the desired effect, it is important that UVW directions are specified. A material-space UVWs for FEM muscles can be specified using the materialuvw point/vertex attribute. By providing materialuvw as a vertex attribute, you are able to provide a local UVW space for each individual tet, which gives you to option of providing a separate UVW frame to each tet.

Name Class Type Description
fiberstiffness Point Float Multiplier for stiffness along the fiber direction, the W direction implied by materialuvw.
fiberscale Point Float Multiplier for the rest strain along the fiber direction, the W direction implied by materialuvw.

State Attributes

Below is a list of attributes that are maintained internally by the solver. Each of these attributes is written to at the end of each solve and read from at the start of the next solve. You should not modify any of these attributes yourself. When you do, the solver is likely to become unstable and you will get bad results. However, you can inspect the values in these attributes in your network for visualization or for the creation of secondary effects.

At each frame, the finite element solver computes a new physical state for each simulated object. The physical state of the object is represented by the point attributes P and v, representing the position and velocity, respectively. The solver’s integration scheme maintains additional attributes a for acceleration and j for jerk.

The point attributes P, v, a, and j store the current integration state of the object. These attributes should not be modified during the simulation because the finite element solver will become unstable and produce low-quality results.

Name Class Type Description
P Point Vector Do not modify! Current position of each object point.
v Point Vector Do not modify! Current velocity of each object point.
accel Point Vector Do not modify! Current acceleration of each object point.
jerk Point Vector Do not modify! Current jerk of each object point.

Embedded Geometry Attributes

These attributes are created on the Embedded Geometry of the Solid Object. The parent attribute is maintained by the embedding code itself, and should not be modified. The baseP point attribute can be provided on the Embedded Geometry by the user to control the binding between the simulated geometry and the embedded geometry. If no baseP is provided, it will be copied from the point positions stored in P at the creation frame. The alignment happens relative to the baseP point attribute on the simulated geometry. If the simulated geometry has a materialP vertex or point attribute, then this attribute takes precedence, allowing control per vertex, rather than per point, if necessary. When you want to ensure that embedded geometry ends up on the desired side of a fracture between simulated geometry, you can use the combination of vertex attributes baseP on the embedded geometry and restP on the simulated geometry. This allows you to line up the embedded geometry with the separate parts in the simulated geometry, for example using the Exploded View SOP. The fracturepart attribute allows you to make sure that the embedded geometry follows the right parts when it gets fractured. When both the simulated and the embedded geometry have the fracturepart attribute, the finite element solver will parent embedded geometry to simulated geometry that has the same fracture part.

Name Class Type Description
parent Primitive Float The index of a parent primitive in the simulated geometry.
baseP Point Float Base positions used for alignment with simulated mesh.
fracturepart Point or Vertex Float Optional user-specified fracture part ID.
P Point Float Positions that correspond to the deformed state.
v Point Float Velocities that correspond to the deformed state.
N Point or Vertex Float Normals that correspond to the deformed state.

Optional Output Attributes

These are attributes that are optionally generated by the solver, when the generation is enabled on the simulated object. These attributes can be useful for visualization, for example, using the Finite Element Visualization SOP. Additionally, these attributes may be used to create secondary effects, for example, particles flying off in regions where fracturing occurs. The optional output attributes are also expected by the Finite Element Visualization SOP.

The following attribute is generated when Create Quality Attributes is turned on:

Name Class Type Description
quality Primitive Float A quality metric between 0 (worst) and 1 (best)

Finite element simulation tends to be sensitive to the quality of the incoming primitives. Low quality primitives may slow down, destabilize or lock a finite element simulation. Low quality primitives are best avoided by using the Tet Embed as a tool to create your tet mesh. Although various quality metrics exist for tetrahedra, the one that’s generated by the solver in this attribute is the one that best matches Houdini’s finite element solution.

The solver generates energy-density attributes for each object that has Create Energy Attributes turned on. The material property settings in the Model tab and the corresponding multiplier attributes result in potential energy, energy dissipation and kinetic energy. For each of these three contributions, local densities are computed within the solver. These densities and quantities derived from them are used to determine the motion and behavior of the objects that are solved by the finite element solver.

Name Class Type Description
potentialdensity Point Float The local density of deformation energy
dissipationdensity Point Float The local density of the rate of energy loss
kineticdensity Point Float The local density of the kinetic energy

The potentialdensity attribute is directly affected by the stiffness parameters in the Model tab. The kineticdensity is proportional to the mass density that is specified for the object. The dissipationdensity is related to the damping settings.

If Create Fracture Attributes is enabled on the simulated object, then the fracturecount point attribute is created. The point attribute fracturecount maintains for each point, the number of times that the point has been involved in a fracture. So any point with a nonzero value of fracturecount has been involved fracturing.

Name Class Type Description
fracturecount Point Integer The number of times a point was fractured during the simulation

Legacy Attributes

In most situations where you want to influence a finite-element simulation, you will want to use soft constraints to achieve this, for example, target constraints, region constraints, or the target strength/damping settings on the object. These are first-class solver features that work in a stable way with the solver and should produce high quality results when used correctly. Purely for backwards compatibility, a force attribute is still supported. Because the force attribute lacks essential information that the solver needs, this attribute cannot be relied on when stability and quality are important. When setting up a new sim, alternatives such as soft targeting, region constraints and animated rest positions should be considered instead of the force attribute.

Name Class Type Description
fexternal Force Vector External force density
force Force Vector Another name for external force density

Inputs

First Input

This optional input can be used to control which simulation objects are modified by this node. Any objects connected through this input and which match the Group parameter field will be modified.

If this input is not connected, this node can be used in conjunction with an Apply Data node, or can be used as an input to another data node.

All Other Inputs

If this node has more input connectors, other data nodes can be attached to act as modifiers for the data created by this node.

The specific types of subdata that are meaningful vary from node to node. Click MMB an input connector to see a list of available data nodes that can be meaningfully attached.

Outputs

First Output

The operation of this output depends on what inputs are connected to this node. If an object stream is input to this node, the output is also an object stream containing the same objects as the input (but with the data from this node attached).

If no object stream is connected to this node, the output is a data output. This data output can be connected to an Apply Data DOP, or connected directly to a data input of another data node, to attach the data from this node to an object or another piece of data.

Locals

ST

The simulation time for which the node is being evaluated.

Depending on the settings of the DOP Network Offset Time and Scale Time parameters, this value may not be equal to the current Houdini time represented by the variable T.

ST is guaranteed to have a value of zero at the start of a simulation, so when testing for the first timestep of a simulation, it is best to use a test like $ST == 0, rather than $T == 0 or $FF == 1.

SF

The simulation frame (or more accurately, the simulation time step number) for which the node is being evaluated.

Depending on the settings of the DOP Network parameters, this value may not be equal to the current Houdini frame number represented by the variable F. Instead, it is equal to the simulation time (ST) divided by the simulation timestep size (TIMESTEP).

TIMESTEP

The size of a simulation timestep. This value is useful for scaling values that are expressed in units per second, but are applied on each timestep.

SFPS

The inverse of the TIMESTEP value. It is the number of timesteps per second of simulation time.

SNOBJ

The number of objects in the simulation. For nodes that create objects such as the Empty Object DOP, SNOBJ increases for each object that is evaluated.

A good way to guarantee unique object names is to use an expression like object_$SNOBJ.

NOBJ

The number of objects that are evaluated by the current node during this timestep. This value is often different from SNOBJ, as many nodes do not process all the objects in a simulation.

NOBJ may return 0 if the node does not process each object sequentially (such as the Group DOP).

OBJ

The index of the specific object being processed by the node. This value always runs from zero to NOBJ-1 in a given timestep. It does not identify the current object within the simulation like OBJID or OBJNAME; it only identifies the object’s position in the current order of processing.

This value is useful for generating a random number for each object, or simply splitting the objects into two or more groups to be processed in different ways. This value is -1 if the node does not process objects sequentially (such as the Group DOP).

OBJID

The unique identifier for the object being processed. Every object is assigned an integer value that is unique among all objects in the simulation for all time. Even if an object is deleted, its identifier is never reused. This is very useful in situations where each object needs to be treated differently, for example, to produce a unique random number for each object.

This value is also the best way to look up information on an object using the dopfield expression function.

OBJID is -1 if the node does not process objects sequentially (such as the Group DOP).

ALLOBJIDS

This string contains a space-separated list of the unique object identifiers for every object being processed by the current node.

ALLOBJNAMES

This string contains a space-separated list of the names of every object being processed by the current node.

OBJCT

The simulation time (see variable ST) at which the current object was created.

To check if an object was created on the current timestep, the expression $ST == $OBJCT should always be used.

This value is zero if the node does not process objects sequentially (such as the Group DOP).

OBJCF

The simulation frame (see variable SF) at which the current object was created. It is equivalent to using the dopsttoframe expression on the OBJCT variable.

This value is zero if the node does not process objects sequentially (such as the Group DOP).

OBJNAME

A string value containing the name of the object being processed.

Object names are not guaranteed to be unique within a simulation. However, if you name your objects carefully so that they are unique, the object name can be a much easier way to identify an object than the unique object identifier, OBJID.

The object name can also be used to treat a number of similar objects (with the same name) as a virtual group. If there are 20 objects named “myobject”, specifying strcmp($OBJNAME, "myobject") == 0 in the activation field of a DOP will cause that DOP to operate on only those 20 objects.

This value is the empty string if the node does not process objects sequentially (such as the Group DOP).

DOPNET

A string value containing the full path of the current DOP network. This value is most useful in DOP subnet digital assets where you want to know the path to the DOP network that contains the node.

Note

Most dynamics nodes have local variables with the same names as the node’s parameters. For example, in a Position DOP, you could write the expression:

$tx + 0.1

…to make the object move 0.1 units along the X axis at each timestep.

See also

Dynamics nodes