You should be able to find some [friction coeffients] in an engineering reference book, such as Marks' Standard Handbook, or perhaps on the Internet. The best way to obtain friction coefficients is through experimentation. Perhaps you should pick a reasonable set of coefficients and do a sensitivity analysis. It's possible the results are rather insensitive, and the value of the coefficients is unimportant. Be aware that it is sometimes important to limit the friction force with VC on Card 2 of *CONTACT. I can tell you this about exponential decay that governs the transition from sticking to sliding, and vice versa ... VREL is the relative velocity of the contacting bodies at the contact point. It is computed by LS-DYNA and has units of velocity. In order for the exponent to be dimensionless, the parameter DC also has units - the inverse of velocity. Contact interaction can involve sticking (when VREL=0) or sliding (when VREL is non-zero). When this interaction is treated numerically with a Coulomb-like friction law we must be able to smoothly transition between the static and dynamic friction coefficients - the exponential involving DC and VREL accommodates this. Please look at the relation for MU and note that it yields FS when VREL=0 - this correctly corresponds to the sticking case. Now, when DC*abs(VREL) reaches a value of about 4.0 the exponential drives the second term down so that MU is within just a couple percent of FD - this corresponds to sliding. When DC*abs(VREL) is between these values we are in the transition state and MU is a combination of FS and FD. There is no "correct" value for DC; its value is the responsibility of the analyst, who must ask himself "above what relative velocity do I consider something to be sliding?". Using this critical value of VREL the analyst then selects a value of DC, using rule of thumb stated above, such that, DC*abs(VREL) = 4.0. In this way, when VREL attains the critical value we have MU nearly equal to FD. ts Ticket#2017011610000231 __________________________________________________________ Frictional energy: LS-DYNA has been modified so that now frictional energy from *CONTACT_FORCE_TRANSDUCER is recorded in the SLEOUT file. Please note, however, that this is available only for MPP, and only for two sided force transducers. Because of the way they are implemented, it is not possible to compute frictional energy for one-sided force transducers, in either SMP or MPP. This capability was added to our Development branch (Rev 86863 for Bug 5970/8555) of the code. --------------------------------------- The "sliding interface energy" in glstat include energy due to contact penetration AND energy due to friction. Thus frictional energy is not a separate line item in glstat. Frictional energy is a line item in sleout. The sum of slave and master contact energies in sleout is equal the sliding energy in glstat. Both are equal to the sum of normal contact and friction energies. From this, you can deduce that the normal contact energy is sum(slave+master) - friction energy. ----------------------------------------- You can ask LS-DYNA to write frictional energy to the binary intfor database. The checklist for doing so is: 1. Set FRCENG to 1 on card 4 of *control_contact. 2. Include the command *database_binary_intfor (assigns an output interval). 3. Set the contact print flags to 1 (SPR and/or MPR) for contact surfaces to be included in the intfor database. 4. Include "s=" on the execution line to assign a name to the intfor database. You can postprocess the intfor database using LS-PrePost. Just read it in as you would a d3plot database. LS-PrePost labels frictional energy density as "Surface Energy Density". The frictional energy of a contact segment is the Surface Energy Density of the segment times the segment area. The summation over all segments of (segment surface energy density * segment area) gives the frictional energy component of the sliding energy in sleout. Examples: http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/energy-friction.k (includes comments) http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/slidingblock.k ______________________________________________________________ Frictional heating: To invoke frictional heating, set FRCENG=1 on *CONTROL_CONTACT. The analysis must be flagged as a coupled thermal/structural analysis using the *control_solution command. Temperature in the output is net temperature (includes heating from friction, heat generation, plastic strain, etc.). When surfaces of unequal temperature are in contact, *contact_..._thermal or *contact_surface_to_surface_thermal_friction will allow for transfer of heat across the interface due to conductance. That contact conductance can be dependent on temperature and/or contact pressure when the _thermal_friction option is invoked. (*Load_surface_stress, formerly required if the conductance is a function of pressure, is no longer required except in the special case of *mat_037. See Manual.) If there is also friction at the contact interface, the friction coefficients can be scaled based on temperature via the _thermal_friction option. ------------------------------------------ FTOSLV in *CONTACT_..._THERMAL which allows the user to specify the fraction of sliding friction energy partitioned to the slave surface was introduced ~ July 2015. NOTE this is for 3D only. The assumption had alway been a 50% - 50% sliding friction thermal energy partition to slave and master segments. The parameter (CHLM) on the CONTACT_(option)_THERMAL keyword for 3D problems is obsolete and was last used in ls-970. It was decided to use it for the fraction of sliding friction thermal energy, f, partitioned to the slave side of the contact. Then, (1-f) goes to the master side. The parameter f is a function of the material thermal properties on each side of the interface. The thermal properties are pretty much the same for a metal tool-blank-die stack up so the 50-50 partition has given acceptable answers. However, a 50-50 split is not correct for a car disk brake because we have a metal-ceramic or metal-synthetic combination between rotor and pad. I will write a tutorial on how to hand calculate f. (paraphrased from as) 7/13/15 Ticket # 2015061810000203 ------------------------------------------ Frictional heating does not work for eroding contacts. as, 8/29/2006 It is not implemented for the soft=2 option. Partitioning of the frictional heating between the slave and master side of the contact is hardwired to 50/50. as, 12/17/10 __________________________________________________________________________________ RE: More on THERMAL_FRICTION the suffix THERMAL_FRICTION allows 1) thermal contact resistance as a function of (Pressure, Temperature, Gap). Use DEFINE_FUNCTION. Velocity is not a parameter but it can be easily added. 2) mechanical friction coefficients as a function of (velocity, temperature). Ticket#2015011910000087 as, 1/19/15 ---------------------------------- *contact_AUTOMATIC_surface_to_surface_THERMAL_FRICTION contact works only for version dev and ONLY if ncpu > 0. It does NOT work in R7.0.0 irrespective of the algebraic sign of ncpu. Test case: http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/thermal_friction.k By including the word "case" on the execution line, 2 runs will be made: 100 deg and 200 deg. Compare x-force output from case100.rcforc and case200.rcforc. Aside: The non-automatic counterpart (*contact_surface_to_surface_THERMAL_FRICTION) with SHLTHK=THKOPT=1) DOES work in R7.0.0 but ONLY if ncpu>0. jd, 12/13/13 ________________________________________________________________ *PART_CONTACT treats the issue of varying friction coefficients in an ad hoc way by allowing the user to define a different friction coefficient for each part. This option requires that FS be set to -1 in *CONTACT. Any part included in the contact withOUT a *part_contact command will have friction parameters taken from card 3 of *control_contact. In reality, a friction coefficient depends on the 2 materials in contact. We can assign friction coefficients in this manner via *DEFINE_FRICTION. This option requires that FS be set to -2 in *CONTACT. _____________________________________________________ The decay coefficient is a numerical device for transitioning from static to dynamic friction and has units of 1/velocity. If no value is given for the decay coefficient, there will be no transition, and the static coefficient will be used. You can use a spreadsheet to plot net friction coefficient vs. relative velocity using the equation given in the Manual under the description of DC. As points of reference, if ... DC*|Vrel| = 0.7, then the friction coefficient is halfway between FS and FD DC*|Vrel| = 4, then the friction coefficient is essentially equal to FD. ___________Friction coef as function of pressure _______________ When FS is set to 2 in a *contact_surface_to_surface or *contact_one_way_surface_to_surface and thickness offsets are invoked (see SHLTHK in *control_contact or *contact), FD then identifies a table ID. The table lists the pressures (p1, p2, p3, ..., pn) as the abscicca values. Ordinate values for the table are not required. Immediately following the *define_table, you need to put a *define_curve for each of the pressures defined in the table. Each curve defines friction coefficient vs. relative velocity for the corresponding pressure value in the table. As an example, let's say table ID 100 is used by setting FS=2 and FD=100 in the contact definition. We want to define friction coefficient vs. relative velocity for three values of pressure. Your input might look like this: *define_table 100 $ pressure values are 5, 100, and 800 5 100 800 *define_curve 101 $ friction coef vs. rel vel corresponding to p=5 0, .3 10,.2 100,.18 *define_curve 102 $ friction coef vs. rel vel corresponding to p=100 0, .4 10,.3 100,.28 *define_curve 103 $ friction coef vs. rel vel corresponding to p=800 0, .5 10,.4 100,.38 Note that the first curve ID = table ID + 1. Each subsequent curve ID is incremented by 1. Be aware that I just fabricated the friction coefficient values above for purposes of illustrating syntax. ____________________________________________ Re: orthotropic friction There are TWO different methods for defining orthotropic friction. 1. *CONTACT_..._ORTHO_FRICTION - works for SMP/MPP - only for auto_s2s and auto_1way_s2s; must use segment sets - angle given for each segment to establish 1 and 2 directions - optional 2D table of friction coef vs. relative vel and interface pres. 2. *DEFINE_FRICTION_ORIENTATION - implemented for metal forming applications - works for SMP only.