6. Loads#

◇ EXCIT = _F (
      ◆ CHARGE = char_cine_meca/char_meca,
      ◇ FONC_MULT = function/formula/table cloth,
      ◇ TYPE_CHARGE =/"DIDI ",
                      /"FIXE_CSTE" (by default),
                      /"FIXE_PILO ",
                      /"SUIV ",
                      /"SUIV_PILO ",
  ),

6.1. General principles#

The keyword factor EXCIT makes it possible to describe a load (solicitations and boundary conditions), and possibly a multiplying coefficient and/or a type of load.

The nonlinear calculation is solved by an incremental method, so it is necessary to setup the calculation. So we apply the loads incrementally also, for example, by applying the force progressively (a ramp). The basic principle is to separate the spatial part of the load and the part (nickname) -temporal.

\[\underline{L}_{ext} \left ( {\underline{u}, t} \right ) = g \left ( t \right ) \underline{L}_{ext} \left ( \underline{u} \right )\]

The spatial part of the load \(\underline{L}_{ext} \left ( \underline{u} \right )\) is given by the keyword CHARGE (possibly including the evolution of a temperature field).

Notes:
  • In a thermo-mechanical calculation, if the initial temperature is different from the temperature reference (given in [AFFE_MATERIAU]), the deformation field associated with the initial moment may be incompatible and therefore lead to a state of constraints and associated non-zero internal variables. If one uses an incremental behavior relationship What if we do not explicitly define an initial state of constraints and internal variables (associated with an initial temperature field different from the reference temperature), the field of constraints and internal variables calculated at the first increment will not take into account only from the only variation in temperature between the initial moment and the first moment, and no any compatibility constraints associated with the initial temperature. To take Once this initial state is taken into account, it must be given explicitly, for example thanks to the keywords in ETAT_INIT. To avoid such situations that may lead to calculation errors, it is better to start a calculation by considering that you have to start from a pristine condition.

  • If an axisymmetric calculation is performed and nodal forces are imposed, these forces must be divided by \(2\pi\) (we are working on a sector of one radian) compared to to the actual loads. Likewise, if we want to calculate the resultant of the efforts, the The result must be multiplied by \(2\pi\) to get the total resultant for the structure complete. Likewise in plane stresses or in plane deformation, we work on a unit thickness: the forces (on the thickness) applied must be divided by the thickness, the real forces are obtained by multiplying the calculation efforts by the thickness.

The time part of the load \(g \left ( t \right )\) is given by the keyword FONC_MULT. If the multiplier function is not specified, a step function is applied.

It is possible to define directly \(\underline{L}_{ext} \left ( {\underline{u}, t} \right )\) in the CHARGE keyword if you use the [AFFE_CHAR_MECA_F] command (time function). In this case, you have to be careful: Indeed, if we use FONC_MULT at the same time, we multiply the load again by a function of time.

By default, TYPE_CHARGE is set to “FIXE_CSTE”: this corresponds to a loading applied to the initial and uncontrolled geometry. However, it can be a function, and, in particular, depend on the weather. A fixed load is only reevaluated at each new moment, and only if it depends on the time (defined in [AFFE_CHAR_MECA_F] and set by the moment or assigned by FONC_MULT).

6.2. Follower loads#

If TYPE_CHARGE =” SUIV “, the load is said to be « follower », i.e. it depends on the value unknowns: for example, pressure, being a loading that applies in the direction normal to a structure, depends on the updated geometry of the structure, and therefore on the movements. A follower load is reevaluated at each iteration of the resolution algorithm.

Currently, the loads that can be described as a follower are the loading of gravity for the element of CABLE_POULIE, the pressure for the models 3D, 3D_SI, D_PLAN, D_PLAN_SI, AXIS, AXIS_SI,,, C_PLAN, C_PLAN_SI, COQUE_3D and for all THM models and centrifugal force in large displacements (keyword ROTATION in [AFFE_CHAR_MECA]). It is also possible to impose that a Dirichlet charge be a follower in the case of rigidification of part of the structure by the use of LIAISON_SOLIDE (

see [AFFE_CHAR_MECA]) in large transformations.

Notes:
  • Pressure can be defined by a function that depends on geometry. In the following case, we can choose whether this dependence is on the geometry initial with the X, Y, and Z parameters of [DEFI_FONCTION] or in relation to the geometry updated with the parameters XF, YF and ZF.

  • LIAISON_SOLIDE follower loading is strictly incompatible with piloting and linear search. It can only be used with 2D isoparametric finite elements and 3D (not with structural elements such as beams, plates and shells).

  • In the case of COQUE_3D, the pressure load can be created by the keywords PRES_REP/PRES or FORCE_COQUE/PRES from [AFFE_CHAR_MECA]. The other operands of FORCE_COQUE (FX, FY…, F1, F2,…) is a step not compatible with follower loadings.

6.3. Piloted loads#

If TYPE_CHARGE is set to “FIXE_PILO”, the load is still fixed (independent of geometry) but will be managed using the keyword factor PILOTAGE. The controllable loads must come from the [AFFE_CHAR_MECA] or [AFFE_CHAR_MECA_F] operator (if not not a time-dependent function) and not be assigned the FONC_MULT keyword.

You can’t control gravity loads, centrifugal force, Laplace forces, thermal loads or initial or anelastic deformations, and bond conditions.

To have controlled loads dependent on geometry (so-called « follower » loads), we use TYPE_CHARGE =” SUIV_PILO “.

6.4. Incremental loads#

If TYPE_CHARGE =” DIDI “then Dirichlet’s conditions (imposed trips, (linear conditions) will apply to the displacement increment from the given moment under ETAT_INIT/NUME_DIDI (by default the moment when the calculation is resumed) and not on the move total.

For example for an imposed displacement (keyword DDL_IMPO from [AFFE_CHAR_MECA]) the condition will be of the form \(u-{u}_{0}=d\) where \({u}_{0}\) is the defined displacement by NUME_DIDI and not \(u=d\).