DIG­I­TAL TWIN – OVER­LAP­PING REAL WORLD WITH VIR­TUAL

Stitch World - - CONTENTS -

First fea­tured in Gart­ner Inc.’s Top 10 Strate­gic Tech­nol­ogy Trends for 2017, ‘dig­i­tal twin’ has be­come the new catch­phrase of the in­dus­tries across the globe. With dig­i­tal trans­for­ma­tion hap­pen­ing in busi­nesses, dig­i­tal twin opens a new door to sim­u­late prod­ucts and pro­cesses and con­trol their op­er­a­tions and work­ing. The con­cept of dig­i­tal twin can be de­fined as the con­ver­gence of the vir­tual world with the phys­i­cal world, that is, cre­at­ing a near real-time dig­i­tal im­age of a phys­i­cal ob­ject or process. Th­ese vir­tual rep­re­sen­ta­tions help or­ga­ni­za­tions pre­dict prob­lems in ad­vance and helps them strate­gise on its work­ing. Team StitchWorld digs deep into this con­cept and high­lights the needs and ar­eas of its im­ple­men­ta­tion in the ap­parel in­dus­try.

In­cor­po­rat­ing con­cepts like Big Data, Ar­ti­fi­cial In­tel­li­gence (AI), Ma­chine Learn­ing (ML) and In­ter­net of Things (IoT), dig­i­tal twins are now ex­plo­rative of In­dus­try 4.0 or the dig­i­tal world. Right from prod­uct devel­op­ment, th­ese vir­tual copies give con­trol over the prod­uct or process from the de­sign phase to the de­ploy­ment phase. The vir­tual copy of the phys­i­cal ob­ject is cre­ated us­ing sen­sors. Troves of data (phys­i­cal data, man­u­fac­tur­ing data, op­er­a­tional data and in­sights from an­a­lyt­ics soft­ware) are col­lected and syn­the­sized. In­for­ma­tion col­lected is fur­ther in­te­grated with AI al­go­rithms into a physics- based vir­tual model and by ap­ply­ing An­a­lyt­ics into th­ese mod­els, we get the rel­e­vant in­sights re­gard­ing the phys­i­cal ob­ject. The con­sis­tent flow of data helps in get­ting the best pos­si­ble anal­y­sis and in­sights on the ob­ject. The prime ob­jec­tive of in­te­grat­ing the vir­tual and the re­al­ity world is the in­te­gra­tion of data. The data from the phys­i­cal world is trans­mit­ted to dig­i­tal twins through the sen­sors and vice versa. This data trans­fer­ring helps to re­spond to the changes, im­prove the op­er­a­tion, and in­crease the value. Only on the ba­sis of the con­verged data

Cre­at­ing a dig­i­tal twin of a phys­i­cal prod­uct can re­flect both the de­sign­ers’ ac­tual ver­sion and the prac­ti­cal con­straints in the phys­i­cal world. The repet­i­tive iter­a­tions of prod­uct devel­op­ment takes a lot of time, and dig­i­tal twin will ac­tu­ally cut down on the time.

en­vi­ron­ment, cross anal­y­sis is pos­si­ble. A dig­i­tal twin dif­fers from the tra­di­tional CAD or merely another sen­sor- en­abled In­ter­net of Things (IoT) so­lu­tion. CAD is com­pletely en­cap­su­lated in a com­put­er­sim­u­lated en­vi­ron­ment that has demon­strated mod­er­ate suc­cess in mod­el­ling com­plex en­vi­ron­ments. So is dig­i­tal twin an en­hanced ver­sion of CAD? May be yes. On the other hand, IoT sys­tems can mea­sure things such as po­si­tion and di­ag­nos­tics for an en­tire com­po­nent but they are not able to in­ter­act be­tween com­po­nents and the full life cy­cle pro­cesses. Gart­ner Inc. be­lieves that within three to five years, hun­dreds of mil­lions of things will be rep­re­sented by dig­i­tal twins. Or­ga­ni­za­tions will use th­ese live mod­els of phys­i­cal equip­ment to proac­tively re­pair and plan for equip­ment ser­vice, to plan man­u­fac­tur­ing pro­cesses, to op­er­ate fac­to­ries, to pre­dict equip­ment fail­ure or to in­crease op­er­a­tional ef­fi­ciency, and to per­form en­hanced prod­uct devel­op­ment.

How it works?

Dig­i­tal twin can be used for phys­i­cal prod­ucts in real space,

vir­tual prod­ucts in vir­tual space, and for the con­nec­tions of data and in­for­ma­tion that ties the vir­tual and real prod­ucts to­gether. Th­ese three main func­tions of dig­i­tal twin are ex­plained in de­tail as fol­lows: 1. Con­cep­tu­alise: Cre­at­ing a dig­i­tal twin will help ob­serve the sit­u­a­tion and elim­i­nate the pos­si­bil­i­ties of breakdown.

2. Com­pare: Both the phys­i­cal prod­uct in­for­ma­tion and the vir­tual prod­uct in­for­ma­tion can be seen si­mul­ta­ne­ously with a com­mon per­spec­tive. 3. Col­lab­o­rate: A per­son can in­ter­vene and make changes in the dig­i­tal twin which can be im­me­di­ately loaded in the real sys­tem. Be­lieved to be the next driver of dis­rup­tion, global dig­i­tal twin mar­ket is pro­jected to reach US $ 13.9 bil­lion by 2022, ex­hibit­ing a CAGR of more than 36 per cent dur­ing 2017-2022, on ac­count of surg­ing adop­tion of In­dus­trial In­ter­net of Things (IIoT) and grow­ing pen­e­tra­tion of smart tech­nolo­gies and de­vices such as smart­phones, smart TVs and smart grids.

Where it can be used? Prod­uct de­sign and devel­op­ment

Cre­at­ing a dig­i­tal twin of a phys­i­cal prod­uct can re­flect both the de­sign­ers’ ac­tual ver­sion and the prac­ti­cal con­straints in the phys­i­cal world. The repet­i­tive iter­a­tions of prod­uct devel­op­ment takes a lot of time, and dig­i­tal twin will ac­tu­ally cut down on the time. It will guide the de­sign­ers to make changes in their de­sign ex­pec­ta­tions and im­prove the de­sign mod­els, achiev­ing per­son­alised prod­uct de­sign. Also, it can ac­cu­rately find the de­fect in de­sign in the vir­tual world and make rapid changes, which can lead to the im­prove­ment of the de­sign, avoid­ing te­dious ver­i­fi­ca­tion and test­ing.

Smart man­u­fac­tur­ing in dig­i­tal twin fac­tory

The whole man­u­fac­tur­ing process, from the in­put of raw ma­te­rial to the out­put of fin­ished prod­ucts, is man­aged and op­ti­mised through the dig­i­tal twin.

The vir­tual work­shop or fac­tory in­cludes the ge­o­met­ri­cal and phys­i­cal mod­els of op­er­a­tors, ma­te­rial, equip­ment, tools, en­vi­ron­ment, etc., as well as the work­ing. This can sim­u­late and eval­u­ate the dif­fer­ent man­u­fac­tur­ing strate­gies and plan­ning un­til a sat­is­fac­tory plan­ning is con­firmed. In the ac­tual man­u­fac­tur­ing ex­e­cu­tion stage, the real-time mon­i­tor­ing and ad­just­ment of man­u­fac­tur­ing process are re­alised through vir­tu­al­phys­i­cal in­ter­ac­tion and it­er­a­tion. And the prob­lems are then rapidly found out and rec­ti­fied.

Us­age mon­i­tor­ing in prod­uct

A fin­ished prod­uct dig­i­tal twin can be mon­i­tored in real time, as the prod­uct dig­i­tal twin con­tin­u­ally records the prod­uct us­age sta­tus data, uses en­vi­ron­ment data, op­er­at­ing pa­ram­e­ters, etc. Sec­ondly, the vir­tual model can con­firm about the prob­a­ble ef­fects of the dif­fer­ent en­vi­ron­men­tal pa­ram­e­ters so that they can be con­trolled. Fur­ther­more, based on the real-time data from phys­i­cal prod­uct and his­tor­i­cal data, the prod­uct dig­i­tal twin is able to ac­cu­rately pre­dict the prod­uct’s re­main­ing life, breakdown chances, etc.

Com­pa­nies at the fore­front of dig­i­tal twin­ning…

US-based GE has built dig­i­tal twins of crit­i­cal jet en­gine com­po­nents that pre­dict the busi­ness out­comes as­so­ci­ated with the re­main­ing life of the com­po­nents, like in case of gas tur­bines to de­liver the de­sired elec­tri­cal power out­put at the low­est pos­si­ble fuel con­sump­tion, and wind tur­bines col­lec­tively op­ti­mis­ing the pro­duc­tion of elec­tric­ity from wind farms. In its Si­matic pro­gram­mable con­trollers pro­duc­tion fa­cil­ity in Am­berg, Ger­many, Siemens is us­ing a com­pre­hen­sive doc­u­men­ta­tion and eval­u­a­tion sys­tem and has achieved an ex­tremely low level of er­ror in pro­duc­tion. Sch­nei­der has part­nered with Mi­crosoft to cre­ate en­ter­prise- grade ap­pli­ca­tions for Prod­uct Dig­i­tal Twins in the ar­eas of train­ing and prod­uct man­age­ment. They are us­ing the ca­pa­bil­i­ties of HoloLens to­day to lead on-site tech­ni­cians through the re­pair process from start to fin­ish. At the fac­tory level, Sch­nei­der Elec­tric’s so­lu­tion

would en­able man­u­fac­tur­ers to be­gin train­ing of the team mem­bers be­fore a fac­tory is built and pro­vide them with re­mote guid­ance. Other key play­ers of the dig­i­tal twin tech­nol­ogy are Tibco Soft­ware Inc., SAP SE, Mi­crosoft Cor­po­ra­tion, Forbesin­dia.com, and Sch­nit­ger Cor­po­ra­tion.

Ben­e­fits of dig­i­tal twin for com­pa­nies

Re­duce prod­uct de­fects: Vir­tual rep­re­sen­ta­tion of the prod­uct can un­dergo a num­ber of con­di­tions and based on the re­sults, nec­es­sary ad­just­ments can be done that can be mapped to the phys­i­cal prod­uct.

Re­duce pro­duc­tion costs: Map­ping dig­i­tal shopfloor for ca­pac­ity plan­ning and pos­si­ble down­time can cut down on the pro­duc­tion costs.

Real time mon­i­tor­ing: A user can mon­i­tor the work­ing of the prod­uct or process in real time sit­ting any­where in the world and can even change the rea­son of the de­fect.

Shorten time to mar­ket: A well-mapped prod­uct and pro­duc­tion pro­cesses will per­form ef­fi­ciently in the real world, thus sav­ing time and money in sim­u­la­tion, test­ing and anal­y­sis.

Ex­tend the life of equip­ment and as­sets: Dig­i­tal twin al­lows com­pa­nies to have a com­plete dig­i­tal foot­print of their prod­ucts, thus elim­i­nat­ing any chances of wear and tear, and ex­tend­ing their life.

Dig­i­tal twin in ap­parel in­dus­try

The his­tory of the ap­parel in­dus­try, dat­ing back to the 1960s, nar­rates the story start­ing from a man­ual form of op­er­a­tion with a seam­stress mov­ing grad­u­ally

Ac­cord­ing to IDC, by the year 2018, com­pa­nies in­vest­ing in the dig­i­tal twin tech­nol­ogy will wit­ness al­most 30 per cent im­prove­ment in cy­cle times of their crit­i­cal pro­cesses.

towards sewing ma­chines and then to ver­ti­cally in­te­grated (both back­ward and for­ward) op­er­a­tion to CMP man­u­fac­tur­ing set-up, de­scrib­ing how a lot has changed. While the re­sults were un­de­bat­able, they brought about im­mense pres­sure on the de­sign­ers and the pro­duc­tion to bring in new prod­ucts to the mar­ket and that too rather quickly. Con­nec­tion that was lost ages ago be­tween de­sign and pro­duc­tion teams led to var­i­ous nui­sances such as loss of un­der­stand­ing of man­u­fac­tur­ing among de­sign pro­fes­sion­als, con­tract man­u­fac­tur­ing mak­ing sup­ply chain less trans­par­ent, length­ier man­u­fac­tur­ing time, and in­creased num­ber of sam­ple devel­op­ment. Over the pe­riod of time, a sin­gle man­u­fac­tur­ing unit frag­mented into a num­ber of di­vi­sions or en­tirely sep­a­rate units, which made it more dif­fi­cult to map and con­trol the ap­parel in­dus­try. Now, the rise of new and in­no­va­tive tech­nolo­gies (or con­cepts) are def­i­nitely ben­e­fit­ing the in­dus­tries. While some in­dus­tries like au­to­mo­tive, aero­space, min­ing are way ahead in re­al­is­ing the need of a par­tic­u­lar tech­nol­ogy, ap­parel in­dus­try will def­i­nitely fol­low but may be a lit­tle later. In fact, frag­mented man­u­fac­tur­ing and its pro­lif­er­a­tion at a faster rate, gave birth to the need of con­nect­ing each and every process right from spin­ning to lo­gis­tics. Al­though phys­i­cally it is im­pos­si­ble to do so, dig­i­tally it is very much pos­si­ble and fea­si­ble too, thus lead­ing to a new ap­proach in ap­parel man­u­fac­tur­ing – Me­taVer­ti­cal ap­proach. The Me­taVer­ti­cal ap­proach re­fers to dig­i­tally re­con­nect­ing a frag­mented sup­ply chain us­ing Dig­i­tal Twin tech­nol­ogy, thus cre­at­ing a dig­i­tal ver­ti­cal in­te­grated fac­tory (can be called as a smart fac­tory, may be!!!).

No loose ends – Smart ap­parel fac­tory

Cre­at­ing a Me­taVer­ti­cal ap­parel fac­tory re­quires two dig­i­tal twins to be cre­ated: Prod­uct Dig­i­tal Twin and Pro­duc­tion Dig­i­tal Twin. Sam­ple ap­proval nowa­days takes around two-three weeks, thus in­creas­ing the lead time. The sit­u­a­tion wors­ens when a sam­ple is re­jected (go­ing through the sam­ples’ it­er­a­tion again). The first im­ple­men­ta­tion area of dig­i­tal twin in the ap­parel in­dus­try is to cre­ate a prod­uct (gar­ment) dig­i­tal twin. Sim­u­lat­ing a gar­ment does not mean scan­ning, none­the­less it is the the­o­ret­i­cal sim­u­la­tion, based on the math­e­mat­i­cal mod­els of colour, fab­ric, and con­structed gar­ment com­po­nents draped over an avatar. 3D sim­u­lated gar­ment al­lows de­sign­ers and buy­ers to con­form to the spec­i­fi­ca­tions of the sam­ple with­out hav­ing a phys­i­cal sam­ple. This will not only re­duce the num­ber of sam­ples de­vel­oped (less wastage) but will also re­duce the ap­proval time. Not only this, the dig­i­tal gar­ment can also show its per­for­mance based on its end-use. Data re­lated to per­for­mance (ten­sile strength, tear strength, etc.) can be added to the 3D sim­u­la­tion soft­ware. This in­te­gra­tion of data helps to de­cide fab­rics and con­struc­tion meth­ods to meet spe­cific end-use re­quire­ments. In a con­ven­tional sce­nario, a pro­to­type sam­ple is de­vel­oped first and then sent to the third-part test lab for test­ing. For ex­am­ple, a foot­ball jersey must re­sist tear­ing un­der stress, be light­weight, and ab­sorb sweat – three per­for­mance re­quire­ments for a prod­uct to with­stand the rigours of a game. For a pro­posed de­sign to be ac­cepted, the right fab­ric, fin­ish, and con­struc­tion meth­ods must be built in. Once a sam­ple is ap­proved, the next step is cost­ing and then man­u­fac­tur­ing. Won­ders of dig­i­tal twin do not end here. Data de­rived from ma­te­rial costs (from pre­vi­ous data) and

Gart­ner Inc. be­lieves that within three to five years, hun­dreds of mil­lions of things will be rep­re­sented by dig­i­tal twins. Or­ga­ni­za­tions will use th­ese live mod­els of phys­i­cal equip­ment to proac­tively re­pair and plan for equip­ment ser­vice, to plan man­u­fac­tur­ing pro­cesses, to op­er­ate fac­to­ries, to pre­dict equip­ment fail­ure or to in­crease op­er­a­tional ef­fi­ciency, and to per­form en­hanced prod­uct devel­op­ment.

Prod­uct Dig­i­tal Twin goes for dig­i­tal pro­duc­tion

Dif­fer­ent com­bi­na­tions of ma­chines, op­er­a­tors, and assem­bly lines can be an­a­lysed to com­pare through­put as well as the er­gonomic risk be­tween man­ual and au­to­mated op­er­a­tions. From a long time, con­stant pres­sure from buy­ers to cut down on costs (even by cents) has added to the woes of man­u­fac­tur­ers. Im­ple­ment­ing the dig­i­tal twin tech­nol­ogy will op­ti­mise pro­duc­tion as will elim­i­nate ar­eas adding to cost. The ad­van­tages of dig­i­tal twin reach out to brands as well. Ad­vanced soft­ware al­lows brands to de­fine prod­ucts prop­erly be­fore plac­ing or­ders. A com­plete knowl­edge of prod­ucts to be man­u­fac­tured also paves way for sub­con­tract­ing with­out any need of su­per­vi­sion and loss of qual­ity. Prod­uct dig­i­tal twin is fur­ther mov­ing ahead for man­u­fac­tur­ing, thus devel­op­ing the need to cre­ate Pro­duc­tion Dig­i­tal Twin. Cre­at­ing a dig­i­tal twin of the shopfloor means cre­at­ing a vir­tual copy of all the man­u­fac­tur­ing equip­ment, op­er­a­tors and even op­er­a­tions. In­te­grat­ing the data of prod­uct dig­i­tal twin, it is then pos­si­ble to sim­u­late the en­tire fac­tory and the chances of er­ror can be elim­i­nated even be­fore a gar­ment is ac­tu­ally pro­duced. ma­chin­ery can pre­dict its man­u­fac­tur­ing fea­si­bil­ity and cost­ing in­for­ma­tion. With real-time ac­cess to that in­for­ma­tion, the de­signer knows up­front if the pro­posed de­sign will meet pric­ing and mar­gin tar­gets.

Ac­cord­ing to BITKOM, the Ger­man As­so­ci­a­tion for In­for­ma­tion Tech­nol­ogy, Telecom­mu­ni­ca­tions and New Me­dia, every dig­i­tal twin in the man­u­fac­tur­ing in­dus­try will have an eco­nomic po­ten­tial of more than € 78 bil­lion by 2025.

Il­lus­tra­tion of a phys­i­cal and a dig­i­tal copy of an ap­parel man­u­fac­tur­ing fac­tory

Dig­i­tal twin can be used to pre­dict equip­ment fail­ure and in­crease op­er­a­tional ef­fi­ciency

Rep­re­sen­ta­tion of a vir­tual copy of a man­u­fac­tur­ing fa­cil­ity

Vir­tual twin of prod­ucts in re­tail store will en­hance shop­ping ex­pe­ri­ence in fu­ture

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.