scieee Open visual document viewer

Legion Bridge in Prague - Assessment of Stone Arches

Vokál, Marek

Abstract

The article deals with available engineering assessment methods for the stone vault bridges according to relevant standards. The article summarizes variable methodologies used for design and assessment of masonry vaults in last period. All methods were applied for assessment of the Legion Bridge over Vltava River in Prague and results were compared. The models were validated by results of static load test.

Full text

SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER Legion B idge in P ague – Assessmen o S one A ches Ma ek VOKÁL, Michal DRAHORÁD Depa men o Conc e e and Mason y S uc u es, Facul y o Ci il Enginee ing, Czech Technical Uni e si y in P ague, Thaku o a 7/2077, 166 29 P ague 6, Czech Republic. ma ek. ok[email p o ec ed], mic[email p o ec ed] DOI: 10.35181/ ces-2019-0023 Abs ac . The a icle deals wi h a ailable enginee ing assessmen me hods o he s one aul b idges acco d- ing o ele an s anda ds. The a icle summa izes a i- able me hodologies used o design and assessmen o mason y aul s in las pe iod. All me hods we e ap- plied o assessmen o he Legion B idge o e Vl a a Ri e in P ague and esul s we e compa ed. The mod- els we e alida ed by esul s o s a ic load es . Keywo ds Mason y a ch b idge, Vaul , B idge, Assess- men , Load ca ying capaci y, Th us line 1. In oduc ion When assessing exis ing b idges, he equi emen o su icien mechanical esis ance and s abili y is usually gi en by he maximum load ha he s uc u e is able o ca y sa ely - he maximum weigh o he oad ehicle which can pass he b idge unde he speci ied condi- ions, i.e. he load ca ying capaci y. I is necessa y o know he ac ual ma e ial cha ac e - is ics o he ca ying elemen s o he s uc u e and he geome y o he s uc u e as a basis o de e mining he load-ca ying capaci y. Deg ada ion o he s uc u e in he cou se o ime changes bo h he ma e ial and geo- me ic pa ame e s o he ca ying and non-ca ying el- emen s o he b idge. The e o e, inding ou he deg ee o deg ada ion is c i ical o de e mine he load-ca ying capaci y. De e mining ac ual de ec s o he b idge, he deg ee o deg ada ion and he eal b idge geome y a e objec i e o diagnos ics. The load ca ying capaci y is in luenced by he ac ha he b idge s uc u es we e designed o be loaded acco ding o he s anda ds alid a he ime o b idge cons uc ion. The equi emen s o load o be ans- e ed by he b idges a e inc easing. The Legion B idge was designed o be loaded by a he d o ca le, and a load es o a ca le he d was also ca ied ou on he b idge o Empe o F ancis, which is name o he Legion B idge a he ime o comple ion. The e o e, a numbe o his o ic s uc u es do no ha e he load ca ying ca- paci y ha he adminis a o o b idge wan s (e en in s a e wi hou de ec s and deg ada ion). De e mina ion o he load bea ing capaci y o he exis ing b idge oday is go e ned by he same p inciples as he design o new s uc u e (see applicable echnical s anda ds and egula ions EN, DIN, and MVL). In case o mason y aul s uc u es his ask is com- plica ed by he s uc u al beha iou and ypical p op- e y o he ma e ial - a negligible ensile s eng h, see [1] and [9]. Due o hese ac s, he p ocedu es o de e - mina ion o he load ca ying capaci y a e non-linea and mus in ol e a la ge numbe o pa ame e s wi h signi ican a iabili y. The e o e, i is e y di icul o se up a simple analy ical model and special p og ams de eloped di ec ly o aul s uc u es a e usually used. In case o oad b idges h ee kinds o load ca ying capaci y (acco ding o [6]) can be calcula ed : •Vn - no mal load ca ying capaci y o he b idge – ep esen s he maximum weigh o one ypi- cal lo y, which may pass he b idge wi hou any es ic ions (posi ion limi ed by sa e y ba ie s only). •V - exclusi e load ca ying capaci y – ep esen s he maximum weigh o one uck, which can pass he b idge as single ehicle in any posi ion o lane espec i ely (no o he a ic loads excep pedes- ians is pe mi ed). I can mo e anywhe e on he b idge. c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 71 SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER •Ve - excep ional load ca ying capaci y - which ep esen s he maximum weigh o special ehi- cle, which can pass he b idge unde special con- di ions (speci ied eloci y, speci ied pa h and ec- cen ici y). De e mina ion o each load ca ying capaci y p esen s sepa a e calcula ion in bo h limi s a es (SLS and ULS) conside ing all o he ele an loads ( empe - a u e, wind, lood e c.). The calcula ion o load ca ying capaci y was ca ied ou using he esul s o diagnos ics acco ding o chap- e 3. The condi ions o eliabili y used o e i y he s uc u e see in chap e 4. The me hods o de e min- ing he in e nal o ces and s esses see in chap e 5. 2. B ie his o y and desc ip ion o he b idge The Legion b idge connec s he Old own wi h he Lesse own o P ague h ough he S řelecký Island. The o egoe b idge o Legion B idge was B idge o Empe o F an išek and i was second b idge in P ague inished in 1841. The Legion B idge was cons uc ed a he posi ion o B idge o Empe o F an išek. The con- s uc ion o new b idge an om 1898 o 1901. In con- as o he o iginal b idge, he supe s uc u e is made o a massi e s one aul s uc u e. The cons uc ion o he b idge consis s o nine la aul s o di e en spans: 26.6+34.3+38.5+42.0+27.8+27.8+31.9+28.7+25.6m. Two aul s abo e he S řelecký Island a e aul s o ci - cula segmen s, he o he aul s a e ellip ical. The con- s uc ion o he b idge is made o g ani e blocks wi h a gap o 12−15 mm illed wi h cemen mo a . The s one acade o on walls o ligh sands one and ed g an- i e symbolizing na ional colo s. The b idge has wide oo ways and mo o way lanes, elec ic acks we e also buil on he b idge, ams un along an exis ing b idge om June 17, 1901. The Legion B idge is an immo able cul u al monu- men and is he e o e p o ec ed acco ding o he p o- isions o law On S a e Monumen Ca e, as amended. Due o he ac ha i is a building loca ed in he e i o y o he P ague His o ical Rese e (PPR), he p o isions o he Go e nmen Dec ee On he His o ical He i age in he Capi al Ci y o P ague. The his o - ical monumen in P ague, ep esen ing he his o ical cen e o P ague, was included in he UNESCO Wo ld He i age Lis in 1992. 3. Diagnos ics o he b idge In o de o ind he geome ic and ma e ial p ope ies o he b idge, ollowing p ocedu es we e ca ied ou . Fo de ailed esul s see [11]: 1. de ailed mapping o all isible damages, 2. geode ic su ey, 3. geo echnical bo eholes, 4. co e bo eholes om he supe s uc u e (ge ing E, b, m), 5. geo ada su ey o hickness o a ches, 6. con inuous measu emen o empe a u es, 7. di ing su ey o subs uc u e, 8. s a ic and dynamic load es – see chap e 7.5. 3.1. De ailed mapping As an example o de ailed mapping o isible dam- ages, which was done oge he wi h he acous ic es o deg ada ion o s ones o a ches, ollowing igu e is shown: Fig. 1: Mapping o isible damages in he spand el wall nea pie No. 2 Because o he c acks in he spand el walls, he wall is conside ed jus as a load in he models, no as a ca ying elemen . c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 72 SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER 3.2. Geo echnical bo eholes Pie s numbe 2, 3, 4, 6, 8 we e d illed om he b idge deck o he ounda ions. F om ha , we ge he in- o ma ion abou hickness and ma e ial p ope ies o asphal laye s, back ill, mason y o pie s and ounda- ion g ound. I was ound ou , ha mason y o pie has much lowe s eng h, han he supe s uc u e, see chap e 3.3. 3.3. Co e bo eholes Comp essi e s eng h was es ed on 34 g ani e spec- imens om supe s uc u e. The measu ed a e age s eng h was b= 132 MPa. Mo a o supe s uc- u e has a e age s eng h 25.8MPa, es ed on 114 specimens. Comp essi e s eng h was es ed on 131 g ani e and sedimen a y ock specimens om subs uc u e. The measu ed a e age s eng h was m= 86 MP a. Mo - a o supe s uc u e has a e age s eng h 25.2MPa, es ed on 169 specimens. The cha ac e is ic comp essi e s eng h o mason y k=K α b β m,(1) was calcula ed 28.9MPa o supe s uc u e and 22.8MPa o subs uc u e (which a e e y high al- ues), whe e: Kis coe icien depending on ype o mason y and mason y elemen s, αis coe icien depending on ype join s and mo a , βis coe icien depending on ype o mo a . The Young modulus o g ani e o supe s uc u e was ound 41.3GPa on 17 specimens, Young modulus o g ani e o subs uc u e was ound 31.1GPa on 29 spec- imens. 3.4. Geo ada su ey o hickness o a ches The Geo ada me hod (GPR) is based on he p inciple o ansmi ing high- equency elec omagne ic wa es in o he examined en i onmen and hen egis e ing he wa e image o e lec ed wa es. The wa e image is a ec ed by local inhomogenei ies, especially wi h di - e en conduc i i y and o he elec omagne ic p ope - ies. Inhomogenei ies can ha e bo h plana cha ac e (discon inui y su aces, s uc u al in e aces) and lo- cal cha ac e (eg, ca i ies, e c.). Inhomogenei ies a e mani es ed by ampli ica ion o he egis e ed e lec ed signal ampli ude. In Figu e 2 he e a e blue lines ep- esen ing he a ch shape aken om geode ic su ey. The yellow a ea and whi e lines a e aken om he Geo ada me hod. Fig. 2: Geo ada me hod in span No. 9 Measu ing in longi udinal di ec ion o he b idge showed good ag eemen wi h he a chi e documen a- ion. On he aul o span 7 was ound 0.15 −0.2m conc e e laye , which was no expec ed. Measu ing in ans e se di ec ion showed cons an hickness, which was expec ed. 4. Limi s a es 4.1. Ul ima e limi s a e In he ul ima e limi s a e (ULS) he beha iou o he s uc u e jus be o e he collapse is in es iga ed. Fo he bea ing capaci y de e mina ion load ac o s acco ding o app op ia e EN a e conside ed (1.35 o dead load and 1.35 o li e load). Gene ally, i is as- sumed ha plas ic hinges a e ully de eloped h ough he s uc u e. Fo de ails on mason y a ch b idges be- ha iou see in [1] and [2]. Resis ances o axial and shea o ces a he ULS can be w i en acco ding o [3] as: NRd = db(h−2eu),(2) VRd = ( k0+ 0.4σd)b(h−2eu)/γM,(3) whe e: dis he design s eng h o mason y in com- p ession, k0is he cha ac e is ic alue o ini ial shea s eng h a no mal s ess equal o 0, b, h is wid h o heigh - espec i ely, euis he eccen ici y o he esul- an axial o ce in c oss-sec ion a he ul ima e limi s a e, σdis he design comp essi e s ess in he com- p essed a ea a he ul ima e limi s a e ( he s ess is uni o mly dis ibu ed, see Figu e 3), 0.4is he coe i- cien o ic ion in he mason y join , γMis he ac o o he ma e ial. c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 73 SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER Fig. 3: S ess dis ibu ion o mason y c oss-sec ion a he ul i- ma e limi s a e. 4.2. Se iceabili y limi s a e The se iceabili y limi s a e (SLS) desc ibes he be- ha iou o he s uc u e unde o dina y ope a ing con- di ions. Ful illing he condi ions o se iceabili y limi s a e p o ides he equi ed p ope ies and beha iou o he s uc u e h oughou i s li e ime. In e ms o se - iceabili y limi s a e, c ack wid h and s uc u al s ess unde ope a ing load a e e i ied (load ac o equals o 1.0). In e ms o e i ica ion o aul s uc u es, i is necessa y o e i y he maximum axial s ess in he c oss-sec ion and he heigh o he comp essed a ea a he c oss-sec ion (see [4] and [3]). Elas ic beha iou o he s uc u e is conside ed wi h a linea dis ibu- ion o axial s ess in he comp essed a ea o he c oss sec ion. Tensioned pa o he sec ion is excluded o s ess de e mina ion (see Figu e 4 ): Fig. 4: S ess dis ibu ion o mason y c oss-sec ion a he se - iceabili y limi s a e. σn,max =NEk 3b(h−2e)≤0.45 k,(4) hc≥h 2−→ e≤h 3,(5) e=MEk NEk ,(6) whe e: MEk, is cha ac e is ic momen caused by load, NEk is cha ac e is ic no mal o ce caused by load. 5. Me hods o calcula ion - a ch 1. G aphical me hod (con ols all he equi emen s) 2. Linea calcula ion (con ols all he equi emen s) (a) Beams – 2D o 3D (b) Plane-s ess elemen s – 2D (c) 3D solid elemen s 3. Non-linea calcula ion (con ols SLS equi e- men s) (a) 2D – plane (b) 3D – solid elemen s – no used in his a icle 4. Equilib ium me hod — Limi S a e:Ring (con ols only collapse o he s uc u e) 5.1. G aphical me hods Va ious g aphical me hods had been used un il com- pu e aided design came o enginee ing p ac ice. I p o ides e y simple and quick design app oach inde- penden on a ch b idge shape. G aphical me hods a e based on he h us line de e mina ion. Th us o ce a he c oss sec ion can be ound as cen oid o he axial s ess diag am. When he h us line is known, s ess a he c oss-sec ion can be calcula ed as well. The me hod o inding he h us line uns in ollowing o de (symme ic a ch acco ding o [5]): 1. Di ide he a ch in 2 symme ic pa s, ind he weigh o one hal and om he geome y o a ch we ind o ce H (ho izon al o ce in he op o a ch) 2. Di ide one hal o a ch in o se e al pa i ions ( e - ical lines can be used o di ide) 3. D aw g aphical ep esen a ion o all pa s – Fi - size o ec o in chosen scale, ac s in i s cen oid 4. Fo ce H we loca e o example in he uppe bound o c oss sec ion co e and eac ion in he lowe bound o c oss sec ion 5. Fo ge ing he esul ing o ce R1 in i s pa i ion, we g aphically add he o ce Fi o H, o ge ing nex esul ing o ces in each pa i ion, we g aphi- cally add he o ces Fi o p e ious esul ing o ce Basic p inciple o calcula ion is shown in Fig.5 and [8]. c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 74 SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER Fig. 5: Basic p inciple o he g aphical me hod. 5.2. Linea calcula ion S uc u al analysis using linea calcula ion me hod was done in wo di e en op ions. In he i s one he a ch was ep esen ed by sequence o beams in i s cen e line. The back ill was ep esen ed by e ical beams p o ided in app op ia e longi udinal dis ance, see Fig- u e 6. While modelling using linea calcula ion, one should no o ge , ha his calcula ion me hod doesn’ ake in o accoun ma e ial non-linea i y (and geome y changes due o excluding he ensioned pa o c oss sec ion). Fo he second op ion he a ch and back ill we e modelled by he 3D solid elemen s wi h a ious mechanical p ope ies – see Figu e 11. Fig. 6: Linea 2D beam model. 5.3. Non-linea calcula ion Fo pe o ming an non-linea analysis and assessmen o he s uc u e, he ma e ial – mo a and mason y elemen s - is homogenized o p ese e i s p ope ies in ela ion o he eal beha iou o he s uc u e o i s pa . I is assumed ha he dimensions ( hick) o he mason y elemen s and join s be ween hem do no signi ican ly a ec he dis ibu ion o s ess in he ma- son y elemen . The eal s ess-s ain diag am o he mason y shows non-linea beha iou (see [1]) pa icu- la ly due o negligible ensile s eng h. In his a icle, i is conside ed ha he ma e ial ac s only in comp es- sion and when he ensile s ess occu s, c acks open up, see Figu e 7). I , subsequen ly, (e.g. in ano he load combina ion) he ensile s esses in he c oss sec ion disappea , he c acks close and he c oss-sec ion ac s again as ull. Fig. 7: Non-linea beha iou o mason y. The s uc u al model in p og am Midas was p e- pa ed using plane-s ess elemen s, modelling he join s be ween he g ani e blocks as se o elas ic links wi h he p ope y "Comp ession only", see Figu e 8. Fig. 8: Elas ic links be ween nodes in join s o mason y. 5.4. Equilib ium me hod on igid blocks Limi S a e:RING is a special analysis so wa e o checking he load bea ing capaci y o he aul in he plane o he longi udinal sec ion o he b idge s uc- u e, including he load dis ibu ion by he back ill. I uses equilib ium equa ions on he pa s o aul ac as a igid bodies, he s uc u e is di ided in o igid bod- ies depending on o ming o he plas ic hinges in loca- ions wi h he lowes heigh o comp essed a ea. Load dis ibu ion is conside ed acco ding o Bousinesq, see example in Figu e 9. Fo mo e de ails see [7]. c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 75 SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER Fig. 9: Ve ical a ic load dis ibu ion (dispe sion) o he aul conside ed o a ch modelling in Limi S a e:RING so - wa e. 6. Me hods o calcula ion - ans e se di ec ion 6.1. Linea calcula ion 1) Beam elemen s Modelling using his me hod is done by ep esen ing he a ch by beams in i s middle line and di iding he back ill in chosen in e al o ep esen i by beams as well, see Figu e 6. I he model shown in Figu e 6 is copied se e al imes in he ans e se di ec ion he analysis model ep esen ing he a ch as a body can be a anged. The s i ness o ans e se beams is chosen as a s i ness o a ch. Such 3D linea model can be used o s udy he e ec o eccen ici y o li e load on he bending momen s dis ibu ion in he ans e se di- ec ion – see Figu e 10. Fig. 10: Linea 3D beam model. 2) Solid elemen s Ano he op ion o linea analysis is o model he body o he s uc u e by he 3D solid elemen s – see Figu e 11. Bo h longi udinal and ans e se di ec ion can be modelled by his way. Fig. 11: 3D solid model. 6.2. E ec i e wid h P inciples o "modelling" o he b idge span 4 in ans- e se di ec ion using e ec i e wid h can be seen om Figu e 12. This way o modelling is used in mos codes. I conside s conse a i e idea, ha non-loaded lane o a ch doesn’ ca y any load. So he shea and bend- ing s i ness be ween loaded and non-loaded elemen s is conside ed equal ze o. Fig. 12: Calcula ion o e ec i e wid h. 7. Resul s 7.1. G aphical me hod Resul o g aphical me hod o span 4 a e shown in Figu e 13. Fig. 13: G aphical solu ion o he span 4 om he a chi e doc- umen a ion. c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 76 SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER 7.2. Linea calcula ion The bending momen s on he beams a e shown in Fig- u e 14. Fig. 14: Bending momen on he beams ep esen ing a ch in longi udinal di ec ion. Fig. 15: Resul ing s ess om he load o exclusi e load model. 7.3. Non-linea calcula ion P incipal s ess in he a ch o span 4 can be seen in Figu e 16. The legend o cu es in he Figu e 17 is ollowing: •sw means sel weigh , •Nmeans non-linea combina ion, •oh means hea -up, •och means cooling, •4V nT!means load by li e load. Fig. 17: No mal s ess dis ibu ion in he middle o span 4 e - sus he c oss sec ion heigh . 7.4. Shea be ween he blocks As a demons a ions o ailu e mode, pic u e o span 4 om p og am Limi S e:RING is a ached: Fig. 18: Failu e mode o span 4 The esul ing shea o ce ac ing on he c oss sec- ion depends on he geome y o he a ch. In beam model, i depends on chosen geome y o axis o he beams. In solid model, i depends on choice o ma e- ial model – linea o non-linea . In Fig. 20 can be seen, ha changes in geome y acco ding o ma e ial non-linea i y a ec s esul ing shea o ce highly. Fig- u e shows esul ing shea o ce om he linea beam model wi h changed axis nea he bo om o a ch. 7.5. S a ic load es Six ehicles o weigh 31.6,31.2,31.7,31.5,31.7,31.5 we e used o he s a ic load es . Tes was ca ied ou acco ding o [10] in spans 3, 4, 5, 6. In his a icle, only esul s o span 4 a e p esen ed. The measu ed displace- men s a e qui e small in compa ison o p eciseness o measu emen , which is equal 0.25 mm. Fou unc ions o ans e se e ical displacemen a e compa ed in he Figu e 21: 1. linea beam model, 2. linea solid model, 3. "non-linea beam model", 4. "non-linea solid model". c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 77 SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER Fig. 16: P incipal s ess om he non-linea model. Fig. 19: Changes o geome y o beam axis depending on ec- cen ici y o load Resul s om he e ec i e wid h calcula ion is no p esen ed in he plo , because i is e iden , ha i is e y conse a i e – esul ing displacemen on he le and igh edges would be ze o, which is no ue. The unc ion "non-linea beam model" and "non- linea solid model" was c ea ed by mul iplying he lin- ea esul by knon, which was calcula ed as knon = wnon−linea /wlinea , whe e wnon−linea is displacemen om he 2D non-linea model loaded by sel weigh and wlinea is displacemen om he linea solid model loaded by sel weigh . The esul ing "non-linea " calcula ed displacemen s a e in he ange o measu ed displacemen s +- p ecise- ness (which is equal o 0.25 mm). The esul s app o e he non-linea beha iou o he b idge. Fig. 20: Resul ing shea o ce depending on eccen ici y o load Fig. 21: Displacemen compa ison 8. Compa ison o me hods and discussion 8.1. Sel weigh – he main load In he Figu e 22 we can see he eccen ici y om he linea , non-linea model and om g aphical solu ion om he a chi e documen a ion. I is e iden , ha c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 78 SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 19 |NUMBER: 2 |2019 |DECEMBER eccen ici y om he non-linea model is highe , as ex- pec ed, because ension, which is allowed in he linea model pushes he esul an h us line o he cen oid o he c oss sec ion. The non-linea analysis is mo e ime consuming (on he e o o he enginee as well as he e o o he compu e ), bu he eal beha iou o he s uc u e is be e desc ibed by he non-linea model (see he beha iou o mason y in Figu e 7 and 21). The esul s om he non-linea model a e mo e dange ous and a e close o he limi s a es. Fig. 22: Compa ison o linea , non-linea and g aphical me hod o span 4. 8.2. T a ic load and i s dis ibu ion in ans e se di ec ion Load dis ibu ion in he ans e se di ec ion was com- pa ed on h ee models – 3D solid, 3D beam model and e ec i e wid h model. Resul can be ound in Figu es 14, 15 and 12. 3D linea solid model assumes linea be- ha iou in all di ec ions, he e o e gi es he mos non- conse a i e esul s. The mos conse a i e model is e ec i e wid h model, because i en i ely excludes pa o c oss sec ion. The eal beha iou is somewhe e be- ween. The beam model gi es esul s be ween he wo men ioned me hod, in opinion au ho is he e o e he mos eal. The beam model is much simple and he s i ness in ans e se di ec ion can be easily changed. Real 3D solid model (which conside s non-linea be- ha iou ) is complica ed wi h he ac , ha we don’ know many c ucial cha ac e is ics o mason y – such as bending and shea s i ness o mo a be ween he blocks – nei he in longi udinal no he ans e se di- ec ion. Tha is he eason he 3D solid model is ec- ommended jus o special s uc u es and i we know he pa ame e s. 8.3. Final esul s o load ca ying capaci y – no mal s ess The inal esul s o load ca ying capaci y e lec s bo h he esul s o modelling he a ch i sel and modelling o ans e se di ec ion. Tab. 1: Resul ing load ca ying capaci y – no mal s ess. (m. means model) A ch m. T ans e se m. Vn V Ve Linea 3D beam 41 122 230 Non-linea e ec i e wid h 32 83 185 RING e ec i e wid h 46 105 182 I is in gene al known, ha Limi S a e:RING gi es non-conse a i e esul s. Non-linea model gi es con- se a i e esul s, i s because o he me hod o assess- ing he load dis ibu ion in ans e se di ec ion, sec- ond because o modelling o he a ch i sel . The lin- ea model esul s a e be ween wo men ioned me hod. Fo Ve and V , he beam model gi es he mos non- conse a i e esul s. The load is concen a ed o small s ip o he a ch in e ec i e wid h model, bu in 3D beam model all he beams ca y pa o he load. 8.4. Final esul s modelling – shea s ess When calcula ing he esis ance o mason y blocks o he shea , c ucial is ic ion coe icien . In Limi - S a e:RING, de aul alue 0.6 is conside ed (acco ding o labo a o y es s, see [12]). Fo compa ison, alues om 0.597 o 0.705 we e ob ained acco ding o [13]. Acco ding o [3], coe icien 0.4 should be used. Fo he compa ison, he coe icien 0.4 is used o all mod- els. Tab. 2: Resul ing load ca ying capaci y – shea s ess. (m. means model) A ch m. T ans e se m. Vn V Ve Linea 3D beam 0 0 0 Non-linea e ec i e wid h 42 111 240 RING e ec i e wid h 35 79 161 As can be seen in Figu e 20 and 19, geome y changes a e c ucial and model, ha doesn’ conside ha , gi es un eal shea o ces. This is he case o lin- ea beam models o aul s o low sagi a. 9. Conclusion Se e al models o he Legion b idge we e ca ied ou . Resul s we e alida ed by he s a ic load es . Mod- elling o a ches showed, ha he mos eal beha iou o a ches desc ibes he non-linea model, because i conside s he non-linea i y, which impac s he calcu- la ion he mos – negligible s eng h in ension. The esul s om non-linea modelling a e non-conse a i e in compa ison o o he me hods. Modelling o load dis- ibu ion in ans e se di ec ion showed, ha 3D solid c 2019 TRANSACTIONS OF THE VSB-TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 79