Metal Forming Application

Home » Knowledge Base » Metal Forming » Metal Forming Application


The first doc­u­ment­ed pub­li­ca­tions to thrust LS-DY­NA® on­to the in­ter­na­tion­al stage were two pa­pers [1,2], pub­lished in In­ter­na­tion­al Con­fer­ence of FE-Sim­u­la­tion of 3-D Sheet Met­al Form­ing Process­es in Au­to­mo­tive In­dus­try, May, 1991, Zurich, Switzer­land:

  1. Im­prov­ing Stan­dard Shell El­e­ments Fric­tion Mod­els and Con­tact Al­go­rithms for the Ef­fi­cient So­lu­tion of Sheet Met­al Form­ing Prob­lems with LS-DY­NA3D, by J.O. Hal­lquist, D.W. Still­man, K. Schweiz­er­hof and K. Weimar.
  2. Ex­plic­it Time In­te­gra­tion and Con­tact Sim­u­la­tions for Thin Sheet Met­al­form­ing, by K. Schweiz­er­hof and J. O. Hal­lquist;

The con­tent of the two pa­pers are self-ex­plana­to­ry from the pa­per ti­tles.

The year of 1991 proved to be a very sig­nif­i­cant and pro­duc­tive year for LS-DY­NA in form­ing sim­u­la­tion, which al­so in­clud­ed three more pub­li­ca­tions in VDI BERICHTE NR. 894, 1991 [3,4,5]. The pa­pers dis­cussed de­tailed eval­u­a­tion and ap­pli­ca­tion of LS-DY­NA in three-di­men­sion­al sheet met­al form­ing in var­i­ous small lab­o­ra­to­ry parts and in full-scaled in­dus­tri­al stamp­ing con­cerns at Vol­vo [4] and Mer­cedes-Benz [5]. Con­sid­er­ing the time they were pub­lished, the ac­tu­al work must have been done at least sev­er­al years ear­li­er. There would be many years gone by be­fore oth­er spe­cial­ized stamp­ing sim­u­la­tion soft­ware be­came com­mer­cial­ly avail­able.

The year of 1991 was al­so the be­gin­ning of a par­a­digm shift for the en­tire tool and die in­dus­try world­wide. Stamp­ing die de­vel­op­ment us­ing hand-made plas­ter mod­els for as long as the in­dus­tri­al ex­ist­ed would be quick­ly re­placed by the CAD and CAE tech­nol­o­gy that use soft­ware such as LS-DY­NA. With­in 10 years, hand-de­vel­op­ments of stamp­ing dies with plas­ter mod­el would all but dis­ap­pear from the tool and die in­dus­try in North Amer­i­ca. Soft dies that used to prove var­i­ous draw die de­vel­op­ment con­cepts are no more. Tra­di­tion­al method of ‘heat & beat’ was re­placed by the cut­ting edge CAE sim­u­la­tion tech­nol­o­gy. Along with the shift, a whole new pro­fes­sion of stamp­ing sim­u­la­tion en­gi­neers was born, some of them from skilled, tra­di­tion­al die shop work­ers. There was a huge ex­plo­sion in pro­duc­tiv­i­ty and ef­fi­cien­cy in the stamp­ing man­u­fac­tur­ing en­gi­neer­ing.

From day one, the de­vel­op­ment of LS-DY­NA has been in­ten­sive­ly fo­cused on ac­cu­ra­cy and func­tion­al­i­ty. This de­vel­op­ment phi­los­o­phy has al­lowed LS-DY­NA to be­come a pre­mier stamp­ing sim­u­la­tion soft­ware. LS-DY­NA is the soft­ware of choice for the most ad­vanced and com­plex stamp­ing process­es cal­cu­la­tions. The ac­cu­ra­cy of LS-DY­NA is well doc­u­ment­ed, through in­ter­na­tion­al bench­marks in sheet met­al form­ing, such as the NU­MISHEET se­ries of con­fer­ences. Be­cause of this fo­cus, LS-DY­NA has passed many in­de­pen­dent bench­marks and eval­u­a­tions by its users against ex­per­i­men­tal mea­sure­ments and was the first com­mer­cial soft­ware to be in­ten­sive­ly ap­plied in the vir­tu­al en­vi­ron­ment to dri­ve in­dus­tri­al stamp­ing man­u­fac­tur­ing en­gi­neer­ing and pro­duc­tion. It has be­come an in­te­grat­ed crit­i­cal path in the ad­vance man­u­fac­tur­ing en­gi­neer process in ma­jor­i­ty of OEMs and their sup­pli­ers. To­day’s LS-DY­NA, with its pow­er­ful MPP tech­nol­o­gy, can take on the most chal­leng­ing in­dus­tri­al stamp­ing prob­lems that no oth­er soft­ware can do, such as com­put­ing with ease de­tailed stamp­ings with 10 mil­lion plus shell or sol­id el­e­ments both ex­plic­it­ly and im­plic­it­ly, with in­cred­i­ble speed. At LSTC, we form the im­pos­si­ble.

As much as we would like to em­pha­size how pow­er­ful LS-DY­NA is, the im­por­tance of hu­man-ware is un­doubt­ed­ly just as crit­i­cal. Ex­pe­ri­ences has proven, ex­tra­or­di­nary suc­cess is af­ford­ed to those en­gi­neers who take the time to un­der­stand the ca­pa­bil­i­ties of LS-DY­NA, to study in de­tails its re­sults, and to ap­ply in every as­pects of stamp­ing man­u­fac­tur­ing en­gi­neer­ing.

  • Ref­er­ence:
  1. J.O. Hal­lquist, D.W. Still­man, K. Schweiz­er­hof and K. Weimar: Im­prov­ing Stan­dard Shell El­e­ments Fric­tion Mod­els and Con­tact Al­go­rithms for the Ef­fi­cient So­lu­tion of Sheet Met­al Form­ing Prob­lems with LS-DY­NA3D, In­ter­na­tion­al Con­fer­ence of FE-Sim­u­la­tion of 3-D Sheet Met­al Form­ing Process­es in Au­to­mo­tive In­dus­try, May, 1991, Zurich, Switzer­land.
  2. K. Schweiz­er­hof and J. O. Hal­lquist: Ex­plic­it Time In­te­gra­tion and Con­tact Sim­u­la­tions for Thin Sheet Met­al­form­ing, In­ter­na­tion­al Con­fer­ence of FE-Sim­u­la­tion of 3-D Sheet Met­al Form­ing Process­es in Au­to­mo­tive In­dus­try, May, 1991, Zurich, Switzer­land.
  3. P.C. Gal­braith, M.J. Finn, S.R. MacEwen, A.R. Carr, K.M. Gaten­by, T.L. Lin, G.A. Clif­ford, J.O. Hal­lquist and D. Still­man: Eval­u­a­tion of an LS-DY­NA3D Mod­el for Deep Draw­ing of Alu­minum Sheet, VDI BERICHTE NR. 894, 1991.
  4. K. Mat­ti­as­son, L. Bern­sprang, A. Ho­neck­er, E. Schedin, T. Ham­mam , and A. Me­lander: On the Use of Ex­plic­it Time In­te­gra­tion in Fi­nite El­e­ment Sim­u­la­tion of In­dus­tri­al Sheet Form­ing Process­es, VDI BERICHTE NR. 894, 1991.
  5. M. Grober and K. Gru­ber: Nu­mer­i­cal Sim­u­la­tion of Sheet Met­al Form­ing of Large Car Body Com­po­nents, VDI BERICHTE NR. 894, 1991.

Application Fields

LS-DY­NA is the most ver­sa­tile soft­ware avail­able com­mer­cial­ly, ow­ing to its de­vel­op­ment strat­e­gy of one scal­able code that in­te­grates mul­ti-physics, mul­ti-stage, and mul­ti-scale ca­pa­bil­i­ties. Ap­pli­ca­tion of LS-DY­NA in stamp­ing man­u­fac­tur­ing en­gi­neer­ing is ful­ly process de­pen­dent, lim­it­ed on­ly by the imag­i­na­tions of its users, and has proven to be ap­plic­a­ble (but not lim­it­ed) in the fol­low­ing ar­eas:

  • Grav­i­ty load­ing (im­plic­it)
  • Binder clos­ing/­set­ting (im­plic­it & ex­plic­it)
  • Form­ing (thin, thick shells and solids)
  • Flang­ing (think shell & solids, im­plic­it & ex­plic­it)
  • Hem­ming (press & roller)
  • Spring­back (free stand­ing or on fix­ture nets)
  • Spring­back com­pen­sa­tion of stamp­ing die and line dies
  • As­sem­bly sim­u­la­tion (clamp­ing forces, per­ma­nent set, spring­back, etc.)
  • Ston­ing for sur­face de­fects of ex­te­ri­or pan­els
  • Dy­nam­ic pan­el trans­fer (trans­fer press line)
  • Pan­el drop­ping (on­to fix­ture) sim­u­la­tion
  • Stamp­ing op­ti­miza­tion (draw beads, ma­te­r­i­al prop­er­ties & tool geom­e­try, etc.)
  • Var­i­ous sta­t­ic & dy­nam­ic load­ing of struc­tures
  • Dent­ing and snap-through sim­u­la­tion
  • One-step stamp­ing ini­tial­iza­tion for sub­se­quent process (crash sim­u­la­tion, etc.)
  • Stamp­ing scrap shed­ding process
  • Tube-bend­ing/­Hy­dro-form­ing
  • Hot/­warm stamp­ing & su­per­plas­tic form­ing
  • Mag­net­ic met­al form­ing

Most of the ap­pli­ca­tions al­so fea­ture 2-D as well as 3-D el­e­ments.

Rapid Advancement

At LSTC, new fea­tures re­lat­ed to sheet form­ing are rapid­ly added, in­ten­si­fied and com­plete­ly dri­ven by cus­tomer de­mand. This al­lows us to meet the spe­cif­ic needs of new stamp­ing ap­pli­ca­tions, whether it is a new ma­te­r­i­al or a new process, and a grow­ing cus­tomer base.

Some of the re­cent­ly de­vel­oped fea­tures in­clude:

  • Mod­i­fied Yoshi­da kine­mat­ic non-lin­ear hard­en­ing mod­els, for Ul­tra-High Strength Steel (UHSS) & Alu­minum stamp­ing and spring­back sim­u­la­tion, with Hill’s and Bar­lat’s yield cri­te­ria
  • Stamp­ing die com­pen­sa­tion for spring­back (in pro­duc­tion use since ear­ly 2006) with new fea­tures de­vel­oped:
    • – Glob­al, lo­cal, it­er­a­tive, on draw die, trim­ming die & flang­ing die
    • – Ac­cel­er­at­ed it­er­a­tive com­pen­sa­tion
  • New fea­tures in im­plic­it/­ex­plic­it de­vel­op­ment and ap­pli­ca­tion
    • – Grav­i­ty, binder clos­ing, flang­ing sim­u­la­tion
    • – Con­tact-based scrap fall sim­u­la­tion
  • Ston­ing sim­u­la­tion for sur­face de­fects of ex­te­ri­or sur­face pan­els
  • Di­rec­tion­al and pres­sure sen­si­tive fric­tion mod­el
  • Fail­ure pre­dic­tion with Forma­bil­i­ty In­dex (F.I.) for form­ing lim­its with non-lin­ear strain path
  • One-step stamp­ing ini­tial­iza­tion for crash/­dura­bil­i­ty
  • As­sem­bly sim­u­la­tion (clamp­ing forces, per­ma­nent set, spring­back, etc.)
  • Ston­ing for sur­face de­fects of ex­te­ri­or pan­els
  • Dy­nam­ic pan­el trans­fer (trans­fer press line)
  • Pan­el drop­ping (on­to fix­ture) sim­u­la­tion
  • Stamp­ing op­ti­miza­tion (draw beads, ma­te­r­i­al prop­er­ties & tool geom­e­try, etc.)
  • Var­i­ous sta­t­ic & dy­nam­ic load­ing of struc­tures
  • Dent­ing and snap-through sim­u­la­tion
  • One-step stamp­ing ini­tial­iza­tion for sub­se­quent process (crash sim­u­la­tion, etc.)

Mutual Benefits

LS-DY­NA ben­e­fits by work­ing with the in­dus­try’s best and bright­est. The close work­ing re­la­tion­ship has al­lowed LS-DY­NA to stay in the fore­front of the form­ing tech­nol­o­gy. Our cus­tomers al­so ben­e­fit from a more ca­pa­ble soft­ware that can han­dle the most de­mand­ing, chal­leng­ing and state-of-art man­u­fac­tur­ing ap­pli­ca­tion. For the past two decades, LSTC has been the on­ly FEA soft­ware com­pa­ny se­lect­ed to par­tic­i­pate in all ma­jor in­dus­tri­al/­gov­ern­ment con­sor­tium projects in the Unit­ed States, with some of the projects in­clud­ing SPP, SCP, DFE, and A/­SP-IRG DOE NSP, par­tic­i­pat­ed by all ma­jor U.S. au­to­mo­tive OEMs and ma­te­r­i­al sup­pli­ers.

Cus­tomer feed­back and col­lab­o­ra­tions are para­mount to LS-DY­NA’s fu­ture. The ex­pe­ri­ence shared by our cus­tomers is one of the dri­ving forces for the ad­vance­ments and im­prove­ments of our soft­ware.

Metal Forming Process Basics

The process be­low shows a typ­i­cal draw de­vel­op­ment process of an au­to­mo­tive stamp­ing from fi­nal fin­ished prod­uct to the com­plete draw die faces.

There are many types of draw die process, as shown be­low:

From left to right, these dif­fer­ent draw process­es are ex­plained be­low:

  • Air Draw
    • – Sin­gle ac­tion. 3-piece die sys­tem with 1 piece up­per (cav­i­ty) and 2 piece (binder and punch) low­er; up­per cav­i­ty moves down.
  • Tog­gle Draw
    • – Dou­ble ac­tion. 3-piece die sys­tem with 2 piece up­per (binder and punch) and 1 piece (cav­i­ty) low­er; up­pers moves down.
  • Air Draw with Pres­sure Pad
    • – Sin­gle ac­tion. Pres­sure pad added on top of reg­u­lar air draw.
  • Stretch Draw (four piece)
    • – Dou­ble ac­tion. 4-piece die sys­tem with 2 piece up­per (up­per binder and cav­i­ty) and 2 piece low­er (low­er binder and cav­i­ty). Up­per binder moves down to close with low­er binder, mov­ing to­geth­er for a cer­tain dis­tance then up­per cav­i­ty comes down to close with punch.
  • Crash Form
    • – Sin­gle ac­tion with no binder. Up­per and low­er tool takes the same shape and up­per moves down.

One of the most im­por­tant die process­es in­cludes trim and flang­ing dies (line dies). What hap­pens in the trim­ming and flang­ing process of­ten in­flu­ence what needs to be done ahead in the draw de­vel­op­ment. An­tic­i­pat­ing how sheet met­al will flow in the down­stream process re­quires years of stamp­ing ex­pe­ri­ences. Sim­u­la­tion of the line die process­es of­ten helps avoid tens of thou­sands of dol­lars of draw die re­work as a re­sult of ei­ther wrin­kling or split in the flang­ing process. The pic­ture be­low shows typ­i­cal flang­ing process­es:

Die (struc­ture) de­sign be­gins as soon as prod­uct and process de­signs are com­plet­ed. Shown be­low is a draw die de­sign. De­signs of oth­er line dies, such as trim and flang­ing dies, etc. are in­clud­ed in this process.

At the same time as the draw die faces and line dies are de­signed, stamp­ing sub-as­sem­bly process de­sign is con­duct­ed. This type of process in­volves var­i­ous types of hem­ming process, as shown be­low:

The fol­low­ing links pro­vide up­dat­ed key­word man­u­al pages re­lat­ed to met­al form­ing; check back of­ten for up­dates and ad­di­tion­al new key­words.

Element Technology

The most used el­e­ments for sheet form­ing are of el­e­ment type #2, and #16.

In-Plane Integration Point

El­e­ment Type #2 (Be­lytschko-Tsay shell), as shown above, fea­tures the fol­low­ing,

  • One in-plane in­te­gra­tion point (re­duced in­te­grat­ed shell)
  • Suit­able for in-plane stretch­ing
  • Suit­able for thick­ness/­thin­ning pre­dic­tion
  • Ide­al for form­ing ap­pli­ca­tion where no fur­ther spring­back sim­u­la­tion is de­sired
  • De­fined with the vari­able ELFORM=2 in *SEC­TION_­SHELL
  • De­fault el­e­ment in LS-Pre­Post Met­al Form­ing eZ-Set­up in Lev­els 1 and 2 – for ear­ly fea­si­bil­i­ty and for forma­bil­i­ty on­ly.

El­e­ment Type #16 (Be­lytschko-Tsay shell), as shown above, fea­tures the fol­low­ing,

  • Four in-plane in­te­gra­tion points (ful­ly in­te­grat­ed shell)
  • Suit­able for non-uni­form in-plane stress dis­tri­b­u­tion
  • Suit­able for bend­ing & un­bend­ing pro­duc­ing non-uni­form in-plane stress­es
  • Ide­al for sit­u­a­tion where stress ac­cu­ra­cy is more im­por­tant
  • Ide­al for form­ing sim­u­la­tion where en­su­ing spring­back sim­u­la­tion is de­sired, for line-die and hem­ming sim­u­la­tion, ston­ing, and for all im­plic­it cal­cu­la­tion
  • Costs slight­ly more CPU time com­pared to El­e­ment #2
  • De­fined with ELFORM=16
  • De­fault in LS-Pre­Post Met­al Form­ing Ap­pli­ca­tion GUI in Lev­el 3 – for spring­back

Oth­er el­e­ments in­clude el­e­ment types #25 (Be­lytschko-Tsay shell), and #26 (ful­ly in­te­grat­ed shell) with thick­ness stretch.

Out-of-Plane Integration Point

Il­lus­trat­ed be­low, a type 16 el­e­ment with 5 out-of-plane in­te­gra­tion points is shown. The num­ber of out-of_­plane in­te­gra­tion point is de­fined through the vari­able NIP in key­word *SEC­TION_­SHELL.

The num­ber of out-of-plane in­te­gra­tion points (NIP) con­trols how ac­cu­rate­ly one can cap­ture the through the shell thick­ness stress-strain in­for­ma­tion. Typ­i­cal through the thick­ness stress pat­terns in met­al form­ing are shown be­low:

Each of the cas­es rep­re­sents the fol­low­ing,

  • Case I – a pure elas­tic bend­ing dis­tri­b­u­tion
  • Case II – elas­tic and plas­tic bend­ing
  • Case III – bend­ing with in-plane stretch­ing
  • Case IV – dis­tri­b­u­tion af­ter spring­back

The fol­low­ing il­lus­trates the con­se­quences on the se­lec­tion of NIP in cap­tur­ing the stress dis­tri­b­u­tion in the cas­es,

  • NIP=2 – caus­es no er­ror for Case I; cause er­rors for Case II, III, IV
  • NIP=3 – no er­ror for Case I; er­ror for case II, III; big er­ror for Case IV
  • NIP=5, or 7 – no er­ror for Case I; small er­ror for Case II, III, and IV.

Oth­er fre­quent­ly used el­e­ments in met­al form­ing are var­i­ous 3-D sol­id and 2-D el­e­ments. Sol­id el­e­ments are de­fined with the vari­able ELFORM in *SEC­TION_­SOL­ID,

  • ELFORM=-2 – ful­ly in­te­grat­ed S/­R sol­id, for poor as­pect ra­tio, ac­cu­rate for­mu­la­tion
  • ELFORM=2 – ful­ly in­te­grat­ed S/­R sol­id
  • ELFORM=1 – con­stant stress sol­id el­e­ment

2-D el­e­ments are de­fined with the vari­able ELFORM in *SEC­TION_­SHELL,

  • ELFORM=12 – plane stress (ex­plic­it and im­plic­it)
  • ELFORM=13 – plane strain (ex­plic­it and im­plic­it)
  • ELFORM=14 – ax­isym­met­ric sol­id area weight­ed (ex­plic­it on­ly)
  • ELFORM=15 – ax­isym­met­ric sol­id vol­ume weight­ed (ex­plic­it and im­plic­it)

Contact and Friction Modeling

Contact Types

The con­tact in­ter­faces in met­al form­ing are the FORM­ING types of con­tact:


Type ‘a’ is the most com­mon­ly used con­tact types in stamp­ing sim­u­la­tion. Type ‘b’ treats both the blank and the rigid bod­ies equal­ly, and it is slight­ly more time con­sum­ing. Type ‘b’ is need­ed in some cas­es where mesh den­si­ty be­tween blank and tools dif­fer too much, and is al­so used fre­quent­ly in im­plic­it sim­u­la­tion.

There are some fea­tures spe­cif­ic and unique to the FORM­ING type of con­tact:

  • Re­gard­less what­ev­er thick­ness is spec­i­fied in *SEC­TION_­SHELL, the rigid bod­ies (tools) have no thick­ness;
  • Neg­a­tive thick­ness off­set us­ing the vari­able MST in *CON­TACT… for the tools is fre­quent­ly and con­ve­nient used for up­per/­low­er die thick­ness off­set;
  • Dis­joint/­un­con­nect­ed el­e­ments in tool mesh­es are al­lowed; LS-Pre­Post of­fers the best au­to­mat­ic tool mesh­ing for form­ing sim­u­la­tion;
  • Sheet blank must be slave sur­face; tools must be mas­ter sur­faces.
  • Mesh adap­tiv­i­ty works the best with Form­ing type con­tact; and a host of oth­er fea­tures are de­vel­oped specif­i­cal­ly for Form­ing con­tact.

The con­tact thick­ness off­set (vir­tu­al off­set) can be se­lect­ed in the eZ Set­up in the first Set­up page of the LS-Pre­Post Met­al Form­ing Ap­pli­ca­tion GUI, as marked be­low. The amount of the tool­ing off­set by de­fault is 1.1 times blank thick­ness.

In LS-DY­NA, draw bead is al­so mod­eled through con­tact, which will be dis­cussed in the Draw Bead Mod­el­ing sec­tion.


The most com­mon type of fric­tion used for met­al form­ing is the coulomb fric­tion mod­el. The de­fault co­ef­fi­cient of fric­tion (COF) in LS-Pre­Post is 0.12. This de­fault val­ue can be changed in the eZ Set­up, un­der Con­trol/­Show.

A more ad­vanced fric­tion mod­el is ori­en­ta­tion de­pen­dent and pres­sure sen­si­tive. This is done through the use of *DE­FINE_­FRIC­TION_­ORI­EN­TA­TION, where a de­tailed de­scrip­tion is pro­vid­ed.

Material Model

LS-DY­NA has over 200 types of ma­te­r­i­al mod­els avail­able, with rough­ly the fol­low­ing mod­els are re­lat­ed to sheet met­al form­ing. Among the list be­low the first ten are the most fre­quent­ly used:

  • MAT_­037: MAT_­TRANS­VERSE­LY_­ANISOTROP­IC_­ELAS­TIC_­PLAS­TIC (1948 Hill’s anisotropy mod­el)
  • MAT_­036: MAT_­3-PA­RA­ME­TER_­BAR­LAT (1989 Bar­lat’s mod­el, alu­minum)
  • MAT_­133: MAT_­BAR­LAT_­YLD2000 (alu­minum)
  • MAT_­125: MAT_­KINE­MAT­IC_­HARD­EN­ING_­TRANS­VER­SLY_­ANISOTROP­IC (Yoshi­da non-lin­ear kine­mat­ic hard­en­ing rule with M37)
  • MAT_­226: MAT_­KINE­MAT­IC_­HARD­EN­ING_­BAR­LAT89 (Yoshi­da with M36, alu­minum)
  • MAT_­242: MAT_­KINE­MAT­IC_­HARD­EN­ING_­BAR­LAT2000 (Yoshi­da with M133, alu­minum)
  • MAT_­122: MAT_­HILL_­3R (1948 pla­nar anisotropy w/­ three R’s, Yoshi­da op­tion)
  • MAT_­020: MAT_­RIGID (For rigid tools)
  • MAT_­024: MAT_­PIECE­WISE_­LIN­EAR_­PLAS­TIC­I­TY (solids)
  • MAT_­165: MAT_­PLAS­TIC_­NON­LIN­EAR_­KINE­MAT­IC (Lemaitre & Chaboche)
  • MAT_­136: MAT_­CORUS_­VEG­TER (Veg­ter yield)
  • MAT_­113: MAT_­TRIP (For austenis­tic stain­less TRIP steel)
  • MAT_­033_­96: MAT_­BAR­LAT_­YLD96
  • MAT_­244: MAT_­UHS_­STEEL

One of the new fea­tures LS-DY­NA of­fers is the fail­ure (neck­ing) pre­dic­tion of sheet met­al un­der the non-lin­ear strain path. Cur­rent­ly this ca­pa­bil­i­ty is im­ple­ment­ed in *MAT_­036 and *MAT_­037. The de­tails are doc­u­ment­ed in the Met­al Form­ing Re­lat­ed Key­words.

Adaptive Mesh Refinement and Fusion

Adap­tive mesh re­fine­ment is a crit­i­cal fea­ture in met­al form­ing. It helps in re­duc­ing the mass scal­ing in­duced in­er­tia ef­fect, es­pe­cial­ly for large, un­sup­port­ed ex­te­ri­or pan­el stamp­ings. It con­trols the num­ber of el­e­ments by cre­at­ing in lo­ca­tions where they are need­ed most. It is an im­por­tant tool in in­creas­ing the com­pu­ta­tion­al ef­fi­cien­cy. It should be used with cau­tion, as in­creased num­ber of mesh adap­tive lev­els ac­cu­mu­lates more adap­tiv­i­ty in­duced er­ror, es­pe­cial­ly in form­ing for spring­back sim­u­la­tion.

The key­word is *CON­TROL_­ADAP­TIVE. The ‘look-for­ward’ re­fine­ment op­tion en­ables blank mesh re­fin­ing start­ing at a user spec­i­fied ap­proach­ing dis­tance to tool­ing. The lev­els of re­fine­ment, fre­quen­cy of re­fine­ment, min­i­mum el­e­ment size al­lowed, and the an­gle cri­te­ri­on can be spec­i­fied. In the eZ Set­up, these vari­ables are au­to­mat­i­cal­ly set for the three lev­els and users have the op­tion to mod­i­fy each vari­able be­fore job sub­mis­sion, as shown in the LS-Pre­Post Met­al Form­ing Ap­pli­ca­tion .

Mesh fu­sion is al­so pos­si­ble, and is done through the use of card #3 in the key­word. Shown be­low is an ex­am­ple of fu­sion on the NU­MISHEET 2005 Cross Mem­ber, where the mesh re­fines first as the sheet blank goes through the draw die ra­dius, then fus­es to­geth­er in the rel­a­tive­ly flat area of the draw wall.

Spec­i­fy­ing Mesh Fu­sion Cri­te­ria

Adap­tive Mesh Re­fine­ment and Fu­sion on NU­MISHEET’05 Cross Mem­ber

For lo­cal­ized form­ing and flang­ing, where mesh­es are de­formed in spe­cif­ic lo­cal ar­eas, the adap­tive box can re­duce num­ber of el­e­ments and avoid un­nec­es­sary re­fine­ment in ar­eas not need­ed. De­tailed in­for­ma­tion can be found in *CON­TROL_­BOX_­ADAP­TIVE. The box­es are easy to cre­ate in LS-Pre­Post, through Mod­el/­En­ti­ty menu, as shown be­low.

Cre­ation of Adap­tive Box to Con­trol Lo­cal Mesh Re­fine­ment

LS-Pre­Post Menu Ac­cess to Adap­tive Box Cre­ation

Draw Bead Modeling

There are two meth­ods to mod­el the draw beads, used to con­trol the blank flow dur­ing form­ing.

Method 1 – Analytical line bead

  • Draw bead is de­fined by:
    • – a con­sec­u­tive list of slave nodes, or,
    • – beams
  • Plas­tic strain in sheet due to bend­ing and un­bend­ing can be de­fined op­tion­al­ly
  • Equiv­a­lent re­strain­ing force is in­put and cal­cu­lat­ed based on the ten­sile strength
  • Re­cent­ly de­vel­oped fea­ture au­to­mat­i­cal­ly gen­er­ates mul­ti­ple line beads with equal amount of force di­vid­ed among them, ide­al for cas­es where the fi­nal sheet blank edge po­si­tion is re­al­ly close to the line bead
  • Ad­van­tage:
    • – less re­fined blank mesh in the bead re­gion
    • – faster com­pu­ta­tion ef­fi­cien­cy
  • For Lev­el 1 and some Lev­el 2 (ac­cu­rate) sim­u­la­tion; not for Lev­el 3 sim­u­la­tion (see LS-Pre­Post Met­al Form­ing Ap­pli­ca­tion)
  • Ac­ces­si­ble through eZ Set­up in LS-Pre­Post, un­der Draw­Bead menu.

Method 2 – Real draw bead

  • Draw bead shape is an in­te­gral part of the tools; very ac­cu­rate
  • Fine mesh is need­ed in the draw bead area
  • Small­er time step is re­quired (set DT2MS ac­cord­ing­ly) to avoid dy­nam­ic ef­fect
  • For Lev­el 2 (ac­cu­rate) and es­pe­cial­ly Lev­el 3 (for spring­back/­hard die re­lease)
  • Draw bead shape/­geom­e­try can be de­ter­mined pre­cise­ly with this lev­el of sim­u­la­tion and re­leased for NC ma­chin­ing of the hard tools
  • Press ton­nages pre­dic­tion are more ac­cu­rate

Re­al draw beads pro­vide more ac­cu­rate re­sults. The ‘pull-in’ from the blank edge, and more im­por­tant­ly, the ‘tight­en­ing’ of the sheet blank with­in the punch open­ing dur­ing the bead clos­ing can be more re­al­is­ti­cal­ly sim­u­lat­ed. In cas­es where por­tion of the fin­ished part is de­signed on the binder for ma­te­r­i­al uti­liza­tion pur­pose, the skid, or im­pact marks can be vivid­ly sim­u­lat­ed, fol­lowed by a spring­back analy­sis to see how the marks af­fect the qual­i­ty of the stamp­ings. Some of LS-DY­NA’s met­al form­ing users use on­ly re­al beads for all hard die de­ci­sions, or at least for the ma­jor clo­sure pan­els and deep-drawn parts. A com­par­i­son be­tween re­strain­ing forces by re­al draw beads and test da­ta is list­ed be­low:

Implicit Application

Im­plic­it sta­t­ic al­go­rithm is most­ly ap­plied in the grav­i­ty ini­tial­iza­tion of the blank, in binder clos­ing, and in flang­ing sim­u­la­tion. Im­plic­it tech­nol­o­gy has the unique ad­van­tage of the to­tal ab­sence from the in­er­tia ef­fect. This ad­van­tage is es­pe­cial­ly im­por­tant in binder clos­ing sim­u­la­tion of large ex­te­ri­or pan­el, where over 90% of the blank is un­sup­port­ed in the die cav­i­ty; and in free flang­ing process, where large ar­eas of the flange are just tak­ing a free ride for the most part of the process, not go­ing in­to con­tact (sup­port) with the flang­ing steel un­til the last few mil­lime­ter of the trav­el. Dy­nam­ic in­er­tia ef­fect is most promi­nent in area of the un­sup­port­ed blank in these process­es. Im­plic­it sta­t­ic tech­nol­o­gy in LS-DY­NA of­fers a much more re­al­is­tic sim­u­la­tion re­sults in these sit­u­a­tions.

With the ad­vance of com­put­er hard­ware and MPP tech­nol­o­gy with­in LS-DY­NA, full-scaled in­dus­tri­al pro­duc­tion parts can now be done in im­plic­it sta­t­ic with very rea­son­able amount of turn-around time. With suf­fi­cient mem­o­ry, SMP is ef­fi­cient for mod­els over 100,000 el­e­ments on the blank. For blank with over 100,000 el­e­ments, the name of the game is MPP.

The im­plic­it sta­t­ic ca­pa­bil­i­ty is of­fered through the use of the key­word *CON­TROL_­IM­PLIC­IT_­FORM­ING, where de­tailed us­age is pro­vid­ed.

Springback and Compensation

Spring­back and com­pen­sa­tion tech­nol­o­gy in LS-DY­NA is the sin­gle most im­por­tant tool re­spon­si­ble for mil­lions of dol­lars in cost sav­ings re­sult­ing from no or re­duced num­ber of die re­cuts. This pow­er­ful fea­ture is bat­tle hard­ened through vig­or­ous in­dus­tri­al ap­pli­ca­tion and prove-out, and is avail­able through the use of the fol­low­ing key­words, where all ca­pa­bil­i­ties and many ex­am­ples are pro­vid­ed:


The main fea­tures of the it­er­a­tive spring­back com­pen­sa­tion tech­nol­o­gy in LS-DY­NA in­clude:

  • Dis­place­ment based. User set the com­pen­sa­tion scale fac­tor, usu­al­ly be­tween 0.7~1.0, un­changed through­out the it­er­a­tive process.
  • Com­pen­sa­tion amount is based on the fac­tor and spring­back amount, in the op­po­site di­rec­tion of the spring­back.
  • In the en­gi­neer­ing stage of the dies, suc­cess­ful com­pen­sa­tion hinges on the use of ac­cu­rate spring­back pre­dic­tion from LS-DY­NA®,
    • In­put: mesh af­ter form­ing, mesh af­ter spring­back, orig­i­nal tool mesh, bridge dis­place­ment;
    • Out­put: com­pen­sat­ed tool mesh;
    • Process de­pen­dent – sim­u­late the en­tire die process.
  • In lat­er stage of hard die build, use of scan da­ta makes ac­cu­rate com­pen­sa­tion pos­si­ble.
  • “Al­ter­nat­ed dual di­rec­tion­al tar­get ap­proach­ing”: usu­al­ly need 3~4 it­er­a­tions to reach tar­get (about +-0.5mm).

Global & Iterative Springback Compensation of Draw Dies

  • This is used where an en­tire part needs to be com­pen­sat­ed. An ex­am­ple of such ap­pli­ca­tion can be found here.
  • Ad­di­tion­al­ly, two im­por­tant is­sues, “sym­met­ric bound­ary con­di­tion”, and “3-D trim curves map­ping” must be ad­dressed, as shown be­low:

Local & Iterative Springback Compensation of Draw Dies

  • This is used where one or more lo­cal­ized re­gions with­in a part needs to be com­pen­sat­ed. The re­gion(s) is de­fined by two en­closed curves. In ad­di­tion to ex­am­ples pro­vid­ed here, an­oth­er ex­am­ple on the NU­MISHEET 2005 deck­lid in­ner is shown be­low:

Compensation of Trim Dies

  • This is used where a (to be) trimmed part needs to be com­pen­sat­ed. The need aris­es where ad­di­tion­al spring­back has oc­curred be­tween end of draw­ing and end of trim­ming, caus­ing the trimmed pan­el not ‘nest­ing’ on the orig­i­nal trim post. The post is of­ten com­pen­sat­ed so when the trim pad comes down to push the drawn pan­el against the post the pan­el will not be moved out of po­si­tion, re­sult­ing in bad trim lines. An ex­am­ple of such ap­pli­ca­tion can be found here.

Compensation of Flanging Dies

  • This is used where a flang­ing area needs to be com­pen­sat­ed. An ap­pli­ca­tion ex­am­ple is shown be­low:

Application of Iterative Compensation in Later Stage of Die Construction

Your hard tool pan­els are hit, and analy­sis of the scan da­ta re­veals there is a need to com­pen­sate for di­men­sion­al de­vi­a­tion. This typ­i­cal­ly is caused by a late prod­uct change that was not cap­tured dur­ing the En­gi­neer­ing stage, or by a change in die process, tool­ing, etc. LS-DY­NA of­fers two meth­ods to han­dle the sit­u­a­tion. In this case, the spring­back pan­els used for the com­pen­sa­tion in­put is ex­act, since the scanned STL file will be used as the tool to ob­tain the spring­back mesh.

  • “Push” with Im­plic­it Sta­t­ic – An ex­am­ple can be found here.
  • “Crash” form­ing with Ex­plic­it Dy­nam­ic or Sta­t­ic Im­plic­it – The scan da­ta is used as rigid tools for the up­per and low­er dies to form the trim pan­el from base­line it­er­a­tion “ITER0”.

LS-PrePost® Metal Forming Application

LS-Pre­Post pro­vides the most up-to-date sup­port for met­al form­ing fea­tures in LS-DY­NA. This is ac­com­plished through the “eZ Set­up” GUI in the Met­al Form­ing menu ac­ces­si­ble through AP­PLI­CA­TION pull-down menu. The goal of the eZ Set­up is to take the bur­den off the users in cre­at­ing LS-DY­NA in­put decks that uti­lize the lat­est fea­tures for sheet met­al form­ing.

The eZ Set­up is ca­pa­ble of set­ting up a typ­i­cal en­tire stamp­ing sim­u­la­tion process in­clud­ing grav­i­ty, binder clos­ing, draw­ing, trim­ming and spring­back sim­u­la­tion, or any in­di­vid­ual stamp­ing process, or a com­bi­na­tion of any process­es, as shown in the fig­ure be­low:

All nec­es­sary in­for­ma­tion is in­put through the GUI dri­ven menu ac­cord­ing to the process se­lect­ed, and out­put in­to a set of files, used to sub­mit to a LS-DY­NA sim­u­la­tion. The CASE dri­ver al­lows the user to sub­mit a sin­gle sim­u­la­tion that con­sists of all the process de­fined. Fur­ther­more, the col­or cod­ed (yel­low: to be de­fined; green: al­ready de­fined) GUI sys­tem and the Next but­ton make it dif­fi­cult to miss any de­f­i­n­i­tion need­ed for the process­es se­lect­ed.

The eZ Set­up al­so in­cludes a ma­te­r­i­al li­brary cov­er­ing the most com­mon ma­te­r­i­al types used for the sim­u­la­tion. The ma­te­r­i­al files are lo­cat­ed in the LS-Pre­Post in­stal­la­tion di­rec­to­ry, un­der a di­rec­to­ry called lspp_­matlib. The ma­te­r­i­al spec­i­fi­ca­tions are read­i­ly iden­ti­fi­able from the file name.

The in­put files are ful­ly pa­ra­me­ter­ized through the use of a set of ASCII con­trol files that are spe­cif­ic to the draw types, ac­cu­ra­cy/­speed lev­els de­sired. These con­trol files are lo­cat­ed in the LS-Pre­Post in­stal­la­tion di­rec­to­ry, un­der a di­rec­to­ry called lspp_­form­ing:

There are three form­ing con­trol lev­els avail­able in the eZ Set­up, as seen in the fol­low­ing fig­ure. These lev­els reg­u­late the bal­ance be­tween ac­cu­ra­cy and speed, for dif­fer­ent sim­u­la­tion ob­jec­tives:

  • Lev­el 1 – for ear­ly fea­si­bil­i­ty
  • Lev­el 2 – for forma­bil­i­ty on­ly
  • Lev­el 3 – for spring­back

Lev­el 1 pro­vides the fastest pos­si­ble turn-around CPU time while main­tain­ing the nec­es­sary over­all ac­cu­ra­cy; Lev­el 2 pro­vides de­tailed ac­cu­ra­cy with fast CPU time; Lev­el 3 us­es all the bells and whis­tles nec­es­sary in the form­ing for the ac­cu­rate spring­back pre­dic­tion and com­pen­sa­tion.

These lev­els are de­signed specif­i­cal­ly each of the stat­ed pur­pose, and the de­fault vari­able de­f­i­n­i­tions such as tool speed, mass scal­ing, adap­tive re­fine­ment lev­els, etc., can be shown with the but­ton Show, and mod­i­fied, if need­ed, as shown be­low:

The fol­low­ing chart pro­vides a per­for­mance com­par­i­son among the three lev­els:

The eZ Set­up uti­lizes the lat­est au­to­mat­ic po­si­tion­ing fea­ture and is ca­pa­ble of the fol­low­ing:

  1. Grav­i­ty load­ing with Im­plic­it Sta­t­ic, pre-bend­ing op­tion­al;
  2. Im­plic­it sta­t­ic binder clos­ing for both air and tog­gle;
  3. Three-piece air and tog­gle draw;
  4. Trim­ming in both 2-D and 3-D;
  5. Spring­back sim­u­la­tion with im­plic­it sta­t­ic;

Examples and Verification

  1. NU­MISHEET 2001 BM4

  2. Forma­bil­i­ty In­dex (F.I.) for *MAT_­037An ex­am­ple be­low shows an area of a stamp­ing with high­ly non-lin­ear strain paths, re­sult­ing in a safe pre­dic­tion when plot­ting with a tra­di­tion­al FLD curve. Plot­ted us­ing the Forma­bil­i­ty In­dex, it ac­tu­al­ly al­ready reached the neck­ing lim­it.
  4. *MAT_­226 (*MAT_­KINE­MAT­IC_­HARD­EN­ING_­BAR­LAT89)Test­ing re­sults on a sin­gle el­e­ment in var­i­ous cyclic load­ing con­di­tions are shown be­low.

    Sin­gle El­e­ment Test­ing – Mul­ti-cy­cle Uni­ax­i­al Ten­sion/­Com­pres­sion

    Sin­gle El­e­ment Test­ing – Mul­ti-cy­cle Bi­ax­i­al Ten­sion/­Com­pres­sion

    Sin­gle El­e­ment Test­ing – Mul­ti-cy­cle Shear

    Sin­gle El­e­ment Test­ing – Mul­ti-cy­cle Plane Strain

    An in­dus­tri­al ex­am­ple shown on a hood in­ner in­di­cates un­der the same draw bead con­di­tion, *MAT_­226 re­sults in more lo­cal­ized thin­ning in plane strain con­di­tion than in *MAT_­125. Wrin­kling con­di­tions are more se­vere in *MAT_­226 than in *MAT_­125.

    *MAT_­125 on a Hood In­ner (Cour­tesy of Ford Mo­tor Com­pa­ny Re­search and En­gi­neer­ing Lab­o­ra­to­ry)

    *MAT_­226 on a Hood In­ner (Cour­tesy of Ford Mo­tor Com­pa­ny Re­search and En­gi­neer­ing Lab­o­ra­to­ry)

  5. Con­straint-free Spring­back Pre­dic­tion with In­er­tia Re­liefThis unique fea­ture al­lows for a con­straint-free spring­back sim­u­la­tion, pre­ferred in many tool and die shops. All rigid body modes are re­moved with­in the solver au­to­mat­i­cal­ly to avoid sin­gu­lar­i­ty. Typ­i­cal­ly, a two-step process is need­ed.
    1. First, in the usu­al spring­back sim­u­la­tion run, bound­ary con­di­tions are re­moved. The fol­low­ing key­word is added to ex­tract sev­en eigen­val­ues from the struc­ture:
      • $NEIG
      • 7
    2. Sec­ond, when eigen­val­ue sim­u­la­tion is com­plet­ed, d3eig file is read in­to LS-Pre­Post for view­ing of eigen­val­ues and eigen­vec­tors. A thresh­old fre­quen­cy is se­lect­ed as a val­ue be­tween the 6th and 7th eigen­val­ues, and set as the vari­able THRESH(=0.02 be­low) in the fol­low­ing key­word:
      • 1, 0.02

    Since the first six modes are typ­i­cal­ly small and the 7th mode is a few or­ders larg­er than the 6th mode, it is pos­si­ble to skip the first step all to­geth­er, and use the fixed val­ue of 0.02 for all THRESH. The fol­low­ing ex­am­ple shows the tech­nique of in­er­tia re­lief used on NU­MISHEET 2002 Fend­er Out­er. It is rec­om­mend­ed to use the DY­NAIN file gen­er­at­ed by LS-DY­NA run of a pre­vi­ous process.

  6. Lin­ear Sta­t­ic Analy­sis on Struc­turesLin­ear sta­t­ic sim­u­la­tion can be ap­plied on die struc­ture analy­sis. It is typ­i­cal­ly used to de­ter­mine if the die is un­der the fa­tigue lim­it and if it meets the de­flec­tion re­quire­ment un­der var­i­ous loads. It can al­so be used to op­ti­mize the die struc­ture for lean die man­u­fac­tur­ing. The fol­low­ing pro­vides a val­i­da­tion of a can­tilever beam mod­eled with both shells and sol­id el­e­ments:

  7. Dent­ing Analy­sisOne of the ap­pli­ca­tion area of the im­plic­it ca­pa­bil­i­ty is the sta­t­ic dent­ing sim­u­la­tion. An sphere-shaped in­den­ter is ap­plied with Force or Dis­place­ment, with sin­gle or mul­ti­ple cy­cles, to be pushed against a pan­el to be stud­ied. Dent depth on the pan­el is stud­ied af­ter the in­den­ter is with­drawn. The fol­low­ing shows dent­ing sim­u­la­tion re­sults, based on the mod­el from SAE pa­per 2002-01-0789.

AVI Library

  1. Deep draw­ing (tog­gle), trim­ming, mul­ti­ple flang­ing sim­u­la­tion of a box­side out­er pan­el – cour­tesy of Au­todie LLC.
  2. Scrap trim­ming and shed­ding sim­u­la­tion of a trim die in­volv­ing by-pass and aer­i­al cams – cour­tesy of Ford Mo­tor Com­pa­ny.
  3. Im­plic­it binder clos­ing in air – NU­MISHEET’02 fend­er out­er.
  4. Im­plic­it binder clos­ing in tog­gle – NU­MISHEET’05 deck­lid in­ner.
  5. Im­plic­it flang­ing sim­u­la­tion – NU­MISHEET’02 fend­er out­er.