Download the Full Version of textbook for Fast Typing at textbookfull.

com

Computational Modelling of Concrete Structures:

Proceedings of the Conference on Computational
Modelling of Concrete and Concrete Structures 1st
Edition Günther Meschke
https://textbookfull.com/product/computational-modelling-of-
concrete-structures-proceedings-of-the-conference-on-
computational-modelling-of-concrete-and-concrete-
structures-1st-edition-gunther-meschke/

OR CLICK BUTTON

DOWNLOAD NOW

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Durability of reinforced concrete structures 1st Edition

Paul Chess

https://textbookfull.com/product/durability-of-reinforced-concrete-
structures-1st-edition-paul-chess/

textboxfull.com

Precast Concrete Structures Second Edition Kim S. Elliott

https://textbookfull.com/product/precast-concrete-structures-second-
edition-kim-s-elliott/

textboxfull.com

Self sensing concrete in smart structures 1st Edition

Baoguo Han

https://textbookfull.com/product/self-sensing-concrete-in-smart-
structures-1st-edition-baoguo-han/

textboxfull.com

Strengthening of concrete structures using fiber

reinforced polymers frp 1st Edition Hwai-Chung Wu

https://textbookfull.com/product/strengthening-of-concrete-structures-
using-fiber-reinforced-polymers-frp-1st-edition-hwai-chung-wu/

textboxfull.com
Reinforced Concrete Structures Analysis and Design (2nd
Edition) David A Fanella

https://textbookfull.com/product/reinforced-concrete-structures-
analysis-and-design-2nd-edition-david-a-fanella/

textboxfull.com

Steel-Reinforced Concrete Structures: Assessment and

Repair of Corrosion, Second Edition Mohamed Abdallah El-
Reedy
https://textbookfull.com/product/steel-reinforced-concrete-structures-
assessment-and-repair-of-corrosion-second-edition-mohamed-abdallah-el-
reedy/
textboxfull.com

Steel Fiber Reinforced Concrete Behavior Modelling and

Design 1st Edition Harvinder Singh (Auth.)

https://textbookfull.com/product/steel-fiber-reinforced-concrete-
behavior-modelling-and-design-1st-edition-harvinder-singh-auth/

textboxfull.com

Second RILEM International Conference on Concrete and

Digital Fabrication Digital Concrete 2020 Freek P. Bos

https://textbookfull.com/product/second-rilem-international-
conference-on-concrete-and-digital-fabrication-digital-
concrete-2020-freek-p-bos/
textboxfull.com

Durability and Sustainability of Concrete Case Studies for

Concrete exposures Nausherwan Hasan

https://textbookfull.com/product/durability-and-sustainability-of-
concrete-case-studies-for-concrete-exposures-nausherwan-hasan/

textboxfull.com
COMPUTATIONAL MODELLING OF CONCRETE STRUCTURES
PROCEEDINGS OF THE CONFERENCE ON COMPUTATIONAL MODELLING OF
CONCRETE AND CONCRETE STRUCTURES (EURO-C 2018), 26 FEBRUARY – 1 MARCH 2018,
BAD HOFGASTEIN, AUSTRIA

Computational Modelling of Concrete

Structures

Editors
Günther Meschke
Ruhr University Bochum, Bochum, Germany

Bernhard Pichler
Vienna University of Technology, Vienna, Austria

Jan G. Rots
Delft University of Technology, Delft, The Netherlands
CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business

© 2018 Taylor & Francis Group, London, UK

Typeset by V Publishing Solutions Pvt Ltd., Chennai, India

Printed and bound in Great Britain by CPI Group (UK) Ltd, Croydon, CR0 4YY

All rights reserved. No part of this publication or the information contained herein may be reproduced,
stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical,
by photocopying, recording or otherwise, without written prior permission from the publisher.

Although all care is taken to ensure integrity and the quality of this publication and the information
herein, no responsibility is assumed by the publishers nor the author for any damage to the property or
persons as a result of operation or use of this publication and/or the information contained herein.

Published by: CRC Press/Balkema

Schipholweg 107C, 2316 XC Leiden, The Netherlands
e-mail: [email protected]
www.crcpress.com – www.taylorandfrancis.com

ISBN: 978-1-138-74117-1 (Hbk)

ISBN: 978-1-315-18296-4 (eBook)
 Computational Modelling of Concrete Structures – Meschke, Pichler & Rots (Eds)
© 2018 Taylor & Francis Group, London, ISBN 978-1-138-74117-1

Table of contents

Preface xiii

Plenary lectures
Molecular dynamics-based structural mechanics of buildings’ resilience 3
F.-J. Ulm, K. Keremidis, R.J.-M. Pellenq & M.J.A. Qomi
Century-long durability of concrete structures: Expansiveness of hydration
and chemo-mechanics of autogenous shrinkage and swelling 15
Z.P. Bažant & S. Rahimi-Aghdam
Regularized continuum damage formulations acting as localization limiters 25
M. Jirásek
Erection of bridges and shells without formwork—challenges for the computational modelling 43
J. Kollegger, B. Kromoser & D. Suza
Network modelling of fracture processes in fibre-reinforced quasi-brittle materials 55
P. Grassl & A. Antonelli
New damage model to simulate ballistic impact on concrete targets 63
L. Pereira, L.J. Sluys & J. Weerheijm

Multiscale cement and concrete research: Experiments and modeling

Fishnet model for failure probability of nacre-like imbricated lamellar materials
and Monte Carlo verification 73
Z.P. Bažant & W. Luo
Phase-field modeling of cement paste: Where particle physics meets continuum mechanics 79
T. Petersen & F.-J. Ulm
Towards a mesoscale model of geopolymers: Interaction potential from the molecular scale 87
F. Lolli & E. Masoero
Nanoscale simulations of cement hydrates precipitation mechanisms: Impact on macroscopic
self-desiccation and water sorption isotherms 93
E. Masoero, I. Shvab, G. Di Luzio & G. Cusatis
Atomistic modeling of early hydration of C3S 103
K.M. Salah Uddin & B. Middendorf
Modeling the evolution of C3S-C3S grain interface over hydration time 109
A. Alex & P. Ghosh
Fracture properties of cement hydrates determined from microbending tests and
multiscale modeling 113
J. Němeček, V. Šmilauer, J. Němeček, F. Kolařík & J. Maňák
Testing and modelling of micro cement paste cube under indentation splitting 121
H. Zhang, E. Schlangen & B. Šavija
An adaptive concurrent two-scale FE model to predicting crack propagation in concrete 127
O.L. Manzoli, E.A. Rodrigues, L.A.G. Bitencourt Jr., T.N. Bittencourt & M. Sánchez

v
Multi-scale modelling of the mechanics of concrete based on the cement paste properties 137
G. Lifshitz Sherzer, E. Gal, E. Schlangen & G. Ye
Sensitivity estimation of cement paste properties in the microstructural characteristics 141
J.-S. Kim & T.-S. Han
Elastic-plastic multi scale approach for localization problems—the embedded unit cell 149
M. Grigorovitch & E. Gal
A continuum damage model for the simulation of concrete under cyclic loading 155
V. Kindrachuk, T. Titscher, V. Hirthammer & J.F. Unger
From slender columns to branching structures 165
R. Krč, J. Podroužek & R. Wan-Wendner

Aging concrete: From very early ages to decades-long durability

Assessing the flow characteristics of self-compacting concrete via numerical simulations
of flow tests 173
S. Kulasegaram & B.L. Karihaloo
Mixing and extrusion of printing concrete 183
Y. Yuan & Y. Tao
Formulation, validation and application of a new shotcrete model 189
M. Neuner, M. Schreter, G. Hofstetter & T. Cordes
Aging behavior of normal and high strength concretes 197
L. Czernuschka, K. Nincevic, I. Boumakis, L. Wan-Wendner & R. Wan-Wendner
A multiscale framework for the prediction of concrete self-desiccation 203
M. Pathirage, D.P. Bentz, G. Di Luzio, E. Masoero & G. Cusatis
Effect of drying on the mechanical performances of concrete 209
F. Benboudjema, F. Soleilhet, X. Jourdain & F. Gatuingt
Temperature and humidity-driven ageing of the VeRCoRs mock-up 215
J.-P. Mathieu, L. Charpin, P. Sémété, C. Toulemonde, G. Boulant, J. Haelewyn,
F. Hamon, S. Michel-Ponnelle, J.-M. Hénault & F. Taillade
Reassessment of the very long term basic creep of concrete 225
J.-M. Torrenti
Interaction between damage and time-dependent deformation of mortar in concrete:
3D FE parametric study at meso-scale 229
J. Ožbolt & S. Gambarelli
Prediction of air permeability in large RC structures using a stochastic FE THM
modeling strategy 237
E.M. Bouhjiti, J. Baroth, F. Dufour & B. Masson
Simplified modelling strategy for the thermos-mechanical analysis of massive reinforced
structures at early age 249
J. Mazars, M. Briffaut & S. Grange
Multiscale quantification of thermal expansion of concrete and thermal stresses
of concrete structures 257
H. Wang, H.A. Mang, Y. Yuan & B. Pichler
Translating environmentally-induced eigenstresses to risk of fracture for design
of durable concrete pavements 265
A. Louhghalam, T. Petersen & F.-J. Ulm
Simulation of permeable concrete to assess interrelationship between strength and permeability 275
P. Modi, M. Dave, A. Dodhiya, Y. Dadi, M. Mungule & M. Langhi

vi
A coupled chemo-mechanical damage-healing model for cementitious materials 285
A.D. Jefferson & R.E. Davies
Flow in fibrous composite materials: Numerical simulations 289
J. Kang & J. Bolander
Modeling water absorption in cement-based composites with SAP additions 295
C. Romero Rodríguez, S. Chaves Figueiredo, E. Schlangen & D. Snoeck
Full coupling between diffusion and mechanical analysis in a discrete computational
framework 305
F. Bousikhane, W. Li, G. Di Luzio & G. Cusatis
Alternative parameter to characterize pore distribution in cement paste, and its relation
to mechanical properties 311
T.-S. Han & J.-S. Kim
Influence of air voids in multiphase modelling for service life prediction of partially
saturated concrete 317
D. Eriksson & T. Gasch
Concrete frost damage due to cyclic environmental loading—experimental and numerical study 327
M. Koniorczyk, D. Gawin, A. Wieczorek & F. Pesavento
Computational modeling of combined frost damage and alkali-silica reaction to
the durability of RC bridge decks 335
F. Gong, Y. Takahashi & K. Maekawa
Multiscale modelling of ASR induced degradation in concrete 345
T. Iskhakov, J.J. Timothy & G. Meschke
Nonlinear coupling models of alkali-silica reaction and multi-directional cracked
reinforced concrete 353
Y. Takahashi, S. Ogawa, Y. Tanaka & K. Maekawa
Coupled C-M meso-scale model for ASR expansion in concrete 363
J. Liaudat, C.M. López & I. Carol
Numerical strategies for the modelling of reinforced concrete structures affected by internal
swelling in variable thermo-hydric conditions 371
B. Nedjar & C. Rospars
A finite element implementation of delayed ettringite formation in concrete structures 377
M. Malbois, B. Nedjar, L. Divet, S. Lavaud & J.-M. Torrenti
Determination of critical anodic and cathodic areas in corrosion processes of steel
reinforcement in concrete 387
J. Ožbolt, E. Sola & G. Balabanić
Challenges in corrosion detection based on numerical and experimental studies 395
M. Kwapisz, A. Vorwagner, A. Lechner & M. Rebhan

Advances in material modeling of plane concrete

Concrete mesoscopic failure analysis with the Virtual Element Method 403
G. Etse, M. Benedetto & A. Caggiano
A cyclic triaxial concrete microplane model with gradient regularization 413
I. Zreid & M. Kaliske
Simulation of cracking susceptibility in recycled concrete aggregate systems 421
A. Jayasuriya, M.J. Bandelt & M.P. Adams
On boundary layer arising from tessellation in discrete models 429
J. Eliáš

vii
Size effect in concrete under splitting tension 437
J. Suchorzewski & J. Tejchman
Numerical prediction of deterministic size effect in concrete bars and beams 447
A. Wosatko, J. Pamin & A. Winnicki
Numerical investigations on early indicators of fracture in concrete at meso-scale 457
M. Nitka & J. Tejchman
Incremental sequentially linear analysis of a notched beam 465
C. Yu, P.C.J. Hoogenboom & J.G. Rots
Cohesive crack model in discontinuous Galerkin method 473
J. Jaśkowiec
Mesoscopic simulations of crack propagation in concrete using cohesive elements 481
W. Trawiński, J. Tejchman & J. Bobiński
Configurational forces for modelling cohesive fracture in concrete 491
C.J. Pearce, A. Amar & Ł. Kaczmarczyk
A constitutive law for concrete with smooth transition from continuous into discontinuous
cracks’ description 499
J. Bobiński & J. Tejchman
Nonlocal damage formulation with evolving internal length: The Eikonal
nonlocal approach 509
G. Rastiello, C. Giry, F. Gatuingt, F. Thierry & R. Desmorat
Response scatter control for discrete element models 517
J. Podroužek, M. Marcon, J. Vorel & R. Wan-Wendner
Enhancement of multifiber beam elements in the case of reinforced concrete structures for
taking into account the lateral confinement of concrete due to stirrup 525
N. Khoder, S. Grange & Y. Sieffert

Analysis of reinforced concrete structures

Modelling of the proof load test on viaduct De Beek 535
E.O.L. Lantsoght, A. de Boer, C. van der Veen & D.A. Hordijk
Organizing an international blind prediction contest for improving a guideline for the
nonlinear finite elements analysis of concrete structures 545
A. de Boer, M.A.N. Hendriks, C. van der Veen & B. Belletti
Ultimate limit state design of three-dimensional reinforced concrete structures:
A numerical approach 553
H. Vincent, M. Arquier, J. Bleyer & P. de Buhan
Computational buckling response of reinforcing bars embedded in Hybrid Fiber
Reinforced Concrete (HyFRC) 561
I.D. Williams, C.P. Ostertag & C.A. Arteta
Analysis of RC frame corners using CDP model 569
M. Szczecina & A. Winnicki
Probabilistic cracking model for macrocrack propagation in reinforced concrete structures 579
P. Rossi, C. Nader & J.L. Tailhan
Numerical simulation of shear strength in a short reinforced concrete corbel strengthened
with composite material compared with experimental results 585
I. Ivanova, J. Assih, V. Stankov & D. Dontchev
Computational modeling and experimental testing of shear and flexure performance
of prestressing concrete T-shaped beams—safety formats 595
A. Strauss, B. Krug, O. Slowik, D. Novák & L. Novák

viii
Characterising the shear behaviour of steel fibre-reinforced concrete 605
S. Zeranka & G.PA.G. van Zijl
Shear behavior of full-scale RC slabs without shear stirrups in nuclear buildings: Experimental
and numerical modeling 615
W.S.A. Nana, S. Limam, T.T. Bui, A. Limam & S. Abouri
Numerical lower bound analysis of plate bending problems containing requirements on
shear capacity and shear-bending interaction 625
T.W. Jensen, P.N. Poulsen & L.C. Hoang
Simulation of reinforced HPFRCC deformation capacity under flexure- and
shear-dominated stress states 633
M. Pokhrel & M.J. Bandelt
Computer-aided stress field analysis of discontinuity concrete regions 641
J. Mata-Falcón, D.T. Tran, W. Kaufmann & J. Navrátil
Investigations on strength and fracture in RC beams scaled along height or length 651
J. Suchorzewski, I. Marzec, J. Tejchman & E. Korol
Modeling of quasi-brittle materials cracking using a lattice discrete element approach 661
E. Zafati, C. Oliver-Leblond & F. Ragueneau
Evaluation of safety formats for structural assessment based on nonlinear analysis 669
J. Červenka, V. Červenka, M. Sýkora & J. Mlčoch
Modelling boundary conditions imposed by loads and supports in 3D D-regions 679
C. Meléndez & P.F. Miguel
Parameter identification as the basis for finite element simulations of ultimate limit states
of concrete hinges 689
J. Kalliauer, T. Schlappal, H.A. Mang & B. Pichler
Rotational capacity of concrete beams using a crack process band 697
J.P. Ulfkjaer
An embedded rebar model for computational analysis of reinforced concrete structures
with applications to longitudinal joints in precast tunnel lining segments 705
V.E. Gall, S.N. Butt, G.E. Neu & G. Meschke
Robust design of hybrid steel fiber reinforced concrete tunnel lining segments 715
G.E. Neu, V.E. Gall, S. Freitag & G. Meschke
Structural analysis of segmented tunnel rings: A combination of transfer relations
with an interface law 725
J.L. Zhang, H.A. Mang, X. Liu, Y. Yuan & B. Pichler

Steel-concrete interaction, fiber-reinforced concrete, and masonry

Investigation of dowel action on a steel-concrete bond model for reinforced
concrete structures 737
C. Turgut, L. Jason, L. Davenne & A. Pirard
Multi-scale tension stiffening approach for the minimum reinforcement of hybrid
concrete beams 745
A.P. Fantilli, A. Gorino & B. Chiaia
Investigation of effect of local bond model on crack propagation in RC using 3D RBSM 755
L. Eddy, K. Matsumoto, K. Nagai & Z. Wang
Modeling of bond fatigue in reinforced concrete based on cumulative measure of slip 767
A. Baktheer & R. Chudoba
Fatigue life assessment of reinforced concrete members considering bond-slip 777
S. Bhowmik & S. Ray

ix
Simulation of interface behavior between steel and concrete to study loss of bond due
to reinforcement corrosion 785
B. Sanz, J. Planas & J.M. Sancho
Modeling of localization and softening behavior of passively confined concrete using
coupled RBSM and nonlinear shell FEM 795
R. Mendoza Jr., Y. Yamamoto, H. Nakamura & T. Miura
Simulation of reinforced ductile cement-based composite beams under cyclic loads 805
S.L. Billington, Y. Shao, T.E. Frank, M.J. Bandelt & D.M. Moreno
The use of trilinear softening functions for modelling the cracking of fibre reinforced concrete 813
A. Enfedaque, M.G. Alberti & J.C. Gálvez
Model comparisons for a shell structure made of textile reinforced concrete 819
T. Senckpiel & U. Häussler-Combe
Finite element research of reinforced concrete structures strengthened with Textile
Reinforced Concrete (TRC) 829
A. Schumann, R. Zobel & M. Curbach
Probabilistic multiple cracking model of elastic-brittle matrix composite reflecting
randomness in matrix, reinforcement and bond 839
M. Vořechovský, R. Chudoba, Y. Li & R. Rypl
Numerical modelling of textile reinforced concrete sandwich panels 849
I.G. Colombo, M. Colombo & M. di Prisco
A two-phased and multi-scale finite element analysis of the tensile creep behavior
of polypropylene fiber reinforced concrete 857
R. Vrijdaghs, E. Verstrynge, L. Vandewalle & M. di Prisco
Numerical simulation of the fracture behaviour of glass fibre reinforced cement:
A cohesive crack approach 867
A. Enfedaque, M.G. Alberti & J.C. Gálvez
Multiscale modeling of steel fiber reinforced concrete based on the use of coupling finite
elements and mesh fragmentation technique 877
L.A.G. Bitencourt Jr., Y.T. Trindade, T.N. Bittencourt, O.L. Manzoli & E.A. Rodrigues
Nonlinear finite element analysis of steel fiber-reinforced concrete coupling beams 889
S.-W. Kim, H.-D. Yun, S.-J. Jang, W.-S. Park, Y.-I. Jang & C.-S. Choi
Micromechanics based modelling of fibre reinforced cementitious composites 895
I.C. Mihai, A.D. Jefferson & P. Lyons
Tensile viscous response of Strain Hardening UHPFRC under high restraint
and isothermal conditions 903
M.A. Hafiz & E. Denarié
A beam-particle model to identify constitutive laws for quasi-brittle materials under complex
loading: From concrete to masonry 913
C. Oliver-Leblond, C. Giry, C. Limoge, M. Vassaux, E. Anglade & F. Ragueneau
The adobe delta damage model 921
T. Li Piani, J. Weerheijm, L. Koene & L.J. Sluys
Non-proportional loading for 3-D stress situations in sequentially linear analysis 931
M. Pari, J.G. Rots & M.A.N. Hendriks
Shear-sliding behavior of masonry: Numerical micro-modeling of triplet tests 941
F. Ferretti, C. Mazzotti, R. Esposito & J.G. Rots

Dynamic behavior: From seismic retrofit to impact simulation

Mechanical model of adhesive post-installed anchor subjected to combined force 955
Y. Takase, M. Mizoguchi & T. Wada

x
FEM analysis of post-installed adhesive anchors under combined stress 963
Y. Ishida, H. Sakata, Y. Takase, Y. Maida, Y. Shirai & T. Sato
Mechanical model of shear stress transfer of roughened concrete surface for R/C
existing member 973
T. Isozaki, U. Musya, Y. Takase, T. Abe, K. Sakamoto, T. Hiwatashi & K. Katori
A practical design method to retrofit existing RC buildings with viscous dampers 983
R. Gobirahavan & A.C. Wijeyewickrema
An improved RSA procedure to compute shear force in tall RC shear wall buildings 993
K. Khy, C. Chintanapakdee & A.C. Wijeyewickrema
Introduction of shear warping in fibre beam elements for the computation of
the nonlinear behaviour of concrete beams 1003
S. Capdevielle, S. Grange, F. Dufour & C. Desprez
Characterization of cracks of damaged concrete structures in dynamics 1009
R. Akiki, C. Giry & F. Gatuingt
High-dynamic strengthening of cementitious materials subjected to uniaxial compression 1017
E. Binder, H.A. Mang, Y. Yuan & B. Pichler
Modelling of reinforced concrete members for high impulsive loading 1023
Y. Lu & J. Xu

Author index 1031

xi
 Computational Modelling of Concrete Structures – Meschke, Pichler & Rots (Eds)
© 2018 Taylor & Francis Group, London, ISBN 978-1-138-74117-1

Preface

EURO-C 2018 represents the continuation of a series of conferences on computational methods and
numerical models for the analysis of concrete and concrete structures. EURO-C 2018 is dedicated
to Nenad Bićanić († 2016), who was one of the founding fathers of this conference series and whose
contributions significantly enriched the field of computational concrete modelling and whose influences
are still felt today. Nenad was not only an inspirational scientist, but also a warm-hearted colleague and
a dear friend. Ever since the first conference in 1984 in Split, Croatia, he was always curious about new
trends in computational modelling. He enjoyed bringing together the concrete mechanics community by
organizing unforgettable sessions, such as the “Betonstüberl” discussions, and the traditional EURO-C
skiing races. Nenad’s passing has been a great loss to his family and to all of us.

We are trying to keep his spirit alive! EURO-C 2018 will take place in Austria, in Bad Hofgastein,
from February 26 to March 1, 2018, maintaining the tradition of previous successful conferences in the
EURO-C conference series (Innsbruck 1994, Bad Gastein 1998, St. Johann im Pongau 2003, Mayrhofen
2006, Schladming 2010, St. Anton am Arlberg 2014). This series emerged as a joint activity, following
early developments in nonlinear finite element analysis and softening models for concrete, generated at
the time of the ICC 1984 conference in Split, the SCI-C conference in Zell am See and the two IABSE
Concrete Mechanics Colloquia in Delft, 1981 and 1987.
The Proceedings of EURO-C 2018 comprise 6 papers of Plenary Lecturers as well as 112 contributed
papers, grouped into 6 sections: (1) Multi-scale cement and concrete research: experiments and modelling,
(2) Aging concrete: from very early ages to decades-long durability, (3) Advances in material modelling
of plane concrete, (4) Analysis of reinforced concrete structures, (5) Steel-concrete interaction, fibre-
reinforced concrete, and masonry, as well as (6) Dynamic behavior: from seismic retrofit to impact simula-
tion. As compared to previous conferences, there are still many contributions on robustness and precision
of numerical models at the structural scale. However, trends towards the materials scale with new fibres
and 3D printable concretes, multi-scale and multi-physics frameworks, and life-cycle oriented models for
ageing and durability of existing and new concrete infrastructure are clearly visible.
We are grateful to the members of the Scientific Advisory Committee (Zdenek Bažant, Sarah Billington,
Gianluca Cusatis, Guillermo Etse, Dariusz Gawin, Christian Hellmich, Günter Hofstetter, Tony Jefferson,
Milan Jirásek, Koichi Maekawa, Jacky Mazars, Javier Oliver, Chris Pearce, Gilles Pijaudier-Cabot,

xiii
Marco Di Prisco, Ekkehard Ramm, Victor Saouma, Bert Sluys, Jean-Michel Torrenti, Franz-Josef Ulm,
Kaspar Willam, Yong Yuan) for their support and substantial efforts in the reviewing process of over
150 abstracts.
In addition to a special memorial session for Prof. Nenad Bićanić, the role of Prof. Zdenek Bažant as
the founding father of concrete fracture mechanics and a key contributor to all previous EURO-C events
will be recognized in the context of his 80th birthday.
We sincerely hope that the EURO-C 2018 Proceedings will serve as a major reference, stimulating new
research directions in the field of computational modelling of concrete and its application to the analysis
of concrete structures.

Günther Meschke, Bernhard Pichler & Jan Rots, Conference Chairmen

René de Borst & Herbert Mang, Honorary Chairmen
Bochum/Vienna/Delft/Sheffield/Vienna,
February 2018

xiv
Plenary lectures
 Computational Modelling of Concrete Structures – Meschke, Pichler & Rots (Eds)
© 2018 Taylor & Francis Group, London, ISBN 978-1-138-74117-1

Molecular dynamics-based structural mechanics of buildings’ resilience

F.-J. Ulm, K. Keremidis & R.J.-M. Pellenq

Massachusetts Institute of Technology, Cambridge MA, USA

M.J.A. Qomi
University of California, Irvine, Irvine CA, USA

ABSTRACT: Making use of the tool-box of statistical physics, a Molecular Dynamics (MD) -based
Structural Mechanics approach is proposed that is suitable for the assessment of resilience of buildings.
The paper summarizes recent developments of the approach; namely Potential of Mean Force (PMF)
expressions suitable for structural members for both two-body (stretch) and three-body (bending) interac-
tions; an original potential parameter calibration procedure based on frequency spectra of structures; an
illustration of the approach for a large-scale structure for the identification of the progressive structural
collapse and fragility curves. It is shown that the proposed MD approach through its focus on interaction
potentials is able not only to reproduce classical structural dynamics results. Most importantly, by redefin-
ing structural mechanics within the context of statistical physics, the approach provides a powerful means
of determining a damage (bond-loss) inventory of buildings required for the assessment of resilience of
buildings.

1 INTRODUCTION non-structural members (incl. windows, roofs,

etc.), and of the impact of this loss of function-
With more than 75% of catastrophic losses in the ality on residual robustness, rate of recovery, and
United States in the period of 1993–2012 caused the level of restoration of said functionality. This
by windstorms (III 2016), there is an urgent need motivates the development of an approach that
to enhance the resilience capacity of our built envi- can address both the complexity of buildings and
ronment vis-à-vis windstorms. In particular, the 18 the loss of integrity at both the component and the
coastal states on the Atlantic Ocean and Mexican system level.
Gulf, and the majority of states east of the Rocky At the core of our approach is the realization
Mountains face wind vulnerability; 45% of the that the complexity of buildings is akin to molec-
value of the US’s built environment is in the 18 ular complexity, in which many strong and weak
coastal states (Papadopoulos 2016). This economic interacting atoms define the very physical proper-
impact calls for revisiting engineering approaches ties of materials ranging from thermal properties
to structural and functional integrity of build- to mechanical and structural properties. Thus,
ings. In fact, classical systems approaches, such rather than considering a building as an ensem-
as the Federal Emergency Management Agency’s ble of elements (beams, plates, walls), we propose
HAZUS-MH (FEMA 2016), or structural engi- to consider a building as an ensemble of discrete
neering approaches at building scale (see e.g., Bar- mass points (‘atoms’) that interact via effective
bato et al. 2013, Walker 2011), condense potential bond-forces and moments, that capture all possible
wind-damage into structural fragility curves for energy states of a building from the equilibrium
specific building types and components to evalu- state to failure when subject to load. With its focus
ate the risk of damage and loss of functionality on the interaction potentials, the approach aims at
that occur due to, e.g., a windstorm, earthquake, moving beyond the limitations of continuum-ele-
flooding and so on. On a building element scale, ment-based structural mechanics approaches for
structural mechanics approaches evaluate poten- the application to buildings composed of “strong”
tial structural damage in relation to a code design interacting members (structural elements), and
limit state of structural members (see e.g. Yeo “weak” interacting members (non-structural ele-
2011). Thus, neither the system approach nor the ments). The Molecular Dynamics-based structural
case study approach provide a quantitative means mechanics approach is first presented, and a cali-
to assess the actual partial loss of functional- bration scheme for the potential energy parameters
ity of buildings due to failure of structural and are provided. The approach is then employed for

3
some sample structures showing both the com- where −ε i0 defines the well depth of the poten-
monality and the added value of the MD-based tial in the reference configuration, while ε ijn is
approach w.r.t. classical approaches for the assess- the energy parameter activated by two-pair  inter-  ij
ment of fragility curves of buildings subject to actions due to the stretch
 ij  λ ij = ( / lij )( rij − rij ) en
0

wind loads. in the link direction en rij / rij . The interaction

force at point i is readily obtained from the r.h.s.
of Eq. (3):
2 MD-BASED STRUCTURAL MECHANICS
APPROACH  N ∂U N ε ij 
ijj n
Fi = − ∑ j s = ∑ λiij enij (4)
Consider thus a structure discretized in mass ∂ri lij
points i = 1, N. The interaction energy between
these mass points is written as a series expansion We thus readily recognize that for a beam-type
of functional terms of one, two, three,... mass structure close to its equilibrium state ε ijn / liij is
points; so that the total (internal) energy of the nothing but the axial stiffness energy, EA (with
system can be written as: E = Young’s modulus and A = cross-section area).
But the concept of potential of mean force (PMF)
  
U ∑
N
i
U i (ri ) ∑
N
i j
(
U ij rij = rj − ri ) herein employed is far reaching, as it applies to
   any interactions that can be captured by energy
+ ∑i
N
j ,k
U ijk
( )
ri rj , rk + … (1) expressions. For instance, the Taylor expansion of
most non-linear expressions of potentials generate

Herein, U i (ri ) stands for the one-body term (the around the equilibrium state to harmonic expres-
ground-state energy) of the energy of mass points i = 1, sions; for instance the Morse-Potential:
   
N defined by position vector ri ; U ij (rij rj − ri )
describes two-body interactions between mass
  
points i and j; U ijk (ri rj rk ) three-body interac-
U i U sij = ε i (( − (− iij ij )) − 1)
2
(5)
tions between i, j and k; and so on. The interac-
(
where β ij ε ijn / ( ε i ) )
1/ 2
tion potentials are assumed to carry all the relevant is such that a Taylor
physics information to define the force vector at expansion of Eq. (5) around the equilibrium state
each mass point i satisfying the linear momentum (| λij | ) yields the harmonic expression (3); for
balance equation: details see e.g. Laubie et al. (2017a,b). This is illus-
  trated in Figure 1. In turn, the force that derives

∀i = 1, N ; mi
d 2 ri
dt 2
= Fi = −∇
∇ riU ri =1, N ( ) (2)
from the Morse-potential is given by:


where mi is the mass of mass point i. The key input
∂U ij
Fi =− s = 4F
∂ri
j
Fi ,max e p ( β ij λij )( ( ij ij )) e
ij
n

to the approach are thus the two-body and three

body interaction potentials, Ui, Uij and Uijk (6)

2.1 Potential-of-Mean-Force (PMF) approach

2.1.1 Two-body interactions
With a focus on making a handshake of the
molecular approach with structural mechanics
theory, it is useful to concentrate all interactions
between mass points into two body-interactions
between mass point i and j, with two-body inter-
actions defining the stretch potential between
these mass points. The simplest form of interac-
tion potentials is a harmonic potential that defines
energy states  around the equilibrium position,
that is  rij || / ij (1), defined by the rest-length
lij || rij0 || . It consists of a quadratic expansion of Figure 1. Normalized two-pair interaction energy,
the displacement arguments of the form: (U i U sij )/ ε i0 (bottom), and normalized force, Fi j / Fi ,jmax
(top), vs. normalized stretch, λij β ij / ln 2, showing that
around the equilibrium position the Morse potential
1
( )
2
U i U sij = −ε i0 + ε iijn λiij (3) (thick line) degenerates into a harmonic potential, asso-
2 ciated with a linear force-stretch relation (thin lines).

4
where Fi , max β ijj ε i0 lij ) is the maximum force 
 ϑ jjk
occurring at λij β ij , which provides a link M i jk m i ;
with strength and/or fracture behavior. ϑ i jjk

)((1 − exp (−β ))

 
2.1.2 Three-body interactions m 4m (
mmax exp − β ijk ϑ i jjkk ijk ϑ i jjk
A similar approach can be employed for three-
body interactions representing bending and tor- (11)
sional energy contribution of a link i, j and k. Such where mmax = β ε / 2 is the extremum of m at
0
three-body interactions are nothing but (linear  ijk i
ϑi k β ijk , much akin to Eq. (6) and dis-
or non-linear) rotational springs defined along played in Figure 1.
3 orthogonal axis; so that the energy reads in the
harmonic case as (see Appendix II):
2.2 Calibration
1 
U ijk = ϑ i jkj ⋅ C ijk ⋅ ϑ i jjk (7) The potential-of-mean-force approach thus
2 requires calibration of the energy parameters for
each link; that is, the bond energy, ε i0 , the stretch
where C ijk is the second order tensor of rotational
energy parameter, ε ijn , and the bending and tor-
stiffness defined in the local orthogonal coordinate
 sion energy parameters ( ij(t b) , κ ij ), or the rota-
system of the three-body link system ( n , ebij , t )
tional stiffness values C ijk . While for simple beam
and ( nk , ebik , t k ) :
or truss-types structure, these energy parameters
can be easily reconstructed from linear elastic free
⎡ 1  1   1  
−1
C ijk = 6 ⎢ b t ⊗ etij + t bij ⊗ ebij + b tikk ⊗ etik energy expressions, a more general approach is
⎣ ε ij ε ij ε ik needed to generalize the approach for any struc-
1  1  ij  ij 1  ikk  ik ⎤ tural discretization. In return, since ε ijn and C ijk
+ t b ⊗ ebik + n ⊗ en + n ⊗ en ⎥ are defined close to the (undeformed) equilibrium
ε ik 2κ ij 2κ ik ⎦ position, it is recognized that these parameters can
(8) be obtained from an analysis of the linear elastic
response of the system. In return, the bond energy
with ( ij , ε , ij ) and ( ik, ε , ik ) the bending and
b
ij
b
ik ε i0 requires consideration of the rupture of the
torsional link energies of link ij and ik, respec- bonds, associated with fracture mechanisms.
tively. For instance, for a straight beam element,
ε ijb 12 EI bb / lij , ε tij EItt / liij , κ ij = GJ / liij , with E = 2.2.1 Elastic calibration from eigenfrequencies
Young’s modulus, G = shear modulus, Ibb, Itt, J the Consider thus a structure composed of N mass
second-order (bending and torsional) moments of points. The system has thus 3 N degrees of freedom,
geometric inertia of the beam in the local coordinate associated with 3 N eigenfrequencies ωk (k = 1,3 N)
system. Analogously to the pair-interaction, non- that can be determined from solving the eigenvalue
harmonic expressions can be employed for three- problem considering harmonic potentials:
point interactions as well; e.g. a Morse potential:
det ([ k ] ω k2 [ m ]) = 0 (12)
⎛
( (

)) ⎞
2
U ijk
= εi − − β iijk ϑ i jjk ⎟⎠ (9)
⎝ where [m] and [k] stand for respectively the
   (3 N × 3 N) mass matrix and stiffness matrix that
where ϑ i jk ( (ϑ i jk ⋅ ϑ i jk )1/ 2 , with the Morse are obtained from the linear momentum balance
parameter β ijk ijk ε i0 )1/ 2 chosen such that equation (2), with [k] the Hessian matrix of the
a Taylor expansion around the equilibrium state overall energy of the system evaluated around the
entails the harmonic expression (7). The contri- equilibrium position:
bution of the three-point interaction to the right-
hand-side of Eq. (2) is readily found to be: ∂ 2U
[k ] = |
2 ri 0
(13)
 
 δ {u}
N ∂U
ijk
∂ϑ i jjk  jk ∂U ijk
Fi = − ∑ j k  = ∑ j k M i ⋅  ; M i = −  jjk
N
jk
∂ri ∂ri ∂ϑ i with {u} the vector assembling all displacement
(10) degrees of freedom of the mass points. Defined
 around the equilibrium position (for which non-
jk
with M i the moment vector, which is entirely harmonic potentials degenerate to harmonic
defined by the PMF expression; for instance, in expressions), [k] is thus a unique function of the
case of the Morse potential (9): local stretch and rotational energies; ε ijn and C ijk .

5
 Consider thus a ‘measured’ discrete frequency
Gcδ Γ ij ε0 1
spectrum of a structure, ω lexxp (l M ). This meas- λij ≤ 2 = 2 i = (16)
urement can be obtained either from the Fourier ε ij
n
ε ij β ij
analysis of actual dynamic measurements on the
structure, or from a frequency analysis obtained by Similar stretch-based rupture relations can be
other continuum-element based linear-elastic anal- derived for non-harmonic potentials; for instance,
ysis of a structure, such as Finite Element Method. for the Morse potential, Eq. (5),
Provided that the measured frequency spectrum
has a sufficient number of frequencies, a minimi- 1 Gcδ Γ ij
zation of the error between ‘measured’ and ‘model’ λij ≤ − ln(1 − α i ) α i = (17)
frequencies thus provides a means to calibrate 3 N β ij ε i0
local stretch and bending energies, ε ijn and C ijk:
Otherwise said, knowledge of the critical stretch
∑ (ω ) and the (material) fracture energy provides a means
3N 2
min ω exp
x (14)
ε ijn ,C ijk m =1 m n
to calibrate the ground-state energy ε i0 (noting that
the two-body stretch energy ε ijn is known from the
where ω nexxp is the ‘experimental’ frequency asso- frequency calibration; see Section 2.2.1).
ciated with the n-th experimental eigenmode, for A similar approach can be employed for rupture
 exp
x of the three-body interactions leading to restrict-
which the experimental eigenvector,  vn pairs
with the m-th model eigenvector vm ; that is, mak- ing the angle norm to a critical value; e.g. for a har-
ing use of the orthogonality of eigenvectors: monic potential (7):

n argma
g xl ,M (vnexpx vm ) (15)  ε0
ϑ i jk ≤ 2 i =
1
(18)
C ijk : 1 β ijk
Since frequency values are integrated values that
account for a large range of eigenmodes ranging Similarly for the Morse potential (9):
from collective to local eigenmodes, the discre-
tized structure thus captures the same frequency  1 Gcδ Γ ij
content as the “real” structure within the limits ϑ i jk ≤ − ln(1 − α i ) α i = (19)
imposed by the level of mass point discretization. β ijk ε i0
This approach for potential parameter calibra-
tion has much in common with what is known, in
solid state physics, as the vibrational densities of 2.3 Examples
state (or phonon density of state) at low tempera-
The relevance of the PMF approach for struc-
tures (see e.g. Qomi et al. 2015). These collective
tural elements is illustrated here for two types of
harmonic vibrations, the so-called phonons, are
structural elements, namely for bending moment
closely related via the Hessian (i.e. Eq. 13) to the
bearing fastener connections employed in timber
free energy of a molecular system, and thus to the
and steel structures; and steel-reinforced concrete
energy parameters of the system. Which is why the
sections.
frequency approach (Eqn. 14 and 15), employed
here—by analogy—for structural systems, pro-
2.3.1 Timber-moment connections
vides a convenient way to calibrate the energy
In timber structures, connections are critical struc-
parameters of the structural system that define the
tural elements which often limit the load bearing
(close-to-) equilibrium state.
capacity of the members they connect. Such con-
nections are achieved with different fastener tech-
2.2.2 Fracture property calibration niques ranging from riveted nail connections (see
When a link ruptures, the energy that is stored Fig. 2), to dowels and bolted connections. Given
into the link between the equilibrium position the importance of these connections, a wealth
and rupture is dissipated into the creation of frac- of experimental research has been dedicated to
ture surface δΓij, with GcδΓij the dissipated energy assessing the moment-angle relations ( )
per link rupture, entailing a release of the energy of such connections (Hattar et al. 1995; Gehloff
U i U sij → 0. That is, if the stretch potential et al. 2010; Wang et al. 2015). In such tests, two
expression is specified, the dissipated energy pro- timber members connected with a specific fastener
vides a means to define a link-fracture criterion. are subjected to a bending moment and corre-
For purpose of illustration consider the harmonic sponding angles are recorded. A typical example
expression (3). For U i U sij → 0, we readily obtain is shown in Figure 2, together with the Morse-
a stretch criterion for rupture; that is: fit obtained from fitting the experimental energy

6
Figure 2. Illustration of Morse-fit of Moment-angle
relationships for wood connections. Shown is a riveted
connection (composed of a steel plate and nails, shown
in the bottom right, lengths in mm) of two beams made
of glued laminated Alberta spruce, tested by Hattar and
Cheng (1995). The Morse parameters obtained by fit-
ting the energy are βijk = 8.47; ε i0 = 5.18 kNm; αi = 0.52.
The reported maximum experimental moment was max
maxMexp = 22.0 kNm [1 kNm = 1 kJ].

U ijk ∫M expx
dϑ to the Morse-potential expression
(9). The fit provides values for the Morse-curvature
parameter, βijkj , from which the ground-state energy, Figure 3. Moment-Curvature Relationship for a
εi β ijkj , and the dimensionless frac- 0.4 × 0.4 m2 concrete section of concrete compressive
ture parameter, α i ( )
β ijkϑ cexxp , [with ϑ cexxp strength 25 MPa, steel reinforced with ρ = 3.5%,
according to Charif et al. (2015). The Morse energy
the angle at failure], are obtained.
fit is obtained for β ijk ij ε i0 liij0 31.3 kN, and
αi = 0.64. The maximum moment max Mexp = 384 kNm.
2.3.2 Reinforced concrete sections [1 kNm = 1 kJ].
It is common practice in reinforced concrete struc-
tural engineering to represent the moment-load
bearing capacity in terms of non-linear moment-
curvature relations. In contrast to moment-angle moment capacity. The calibrated values, ε i0 / li0j
relationships which define the interaction at a and β ijk liij0 , are then rescaled, in structural simula-
specific structural connection, moment-curvature tions, for a specific mass-point discretization lij0 .
relations are section properties that encapsulate The quality of such an energy calibration using the
the non-linear stress-strain responses of the con- Morse potential is shown in Figure 3 as well.
stituent materials; i.e. concrete and reinforcing These examples thus illustrate that molecular
steel in the case of reinforced concrete members. potentials devised to capture the potential energy
A typical example is shown in Figure 3, com- of interatomic bonds can be effectively employed
posed of an initial linear branch, followed by a for structural elements. Shown here for the har-
monotonic increasing segment up to a maximum monic case and the Morse potential, there is a
moment-section capacity and a descending branch wealth of other potential expressions to capture a
before the section fails. For such section properties, large range of relevant structural behaviors.
it is appropriate
 to calibrate the Morse-potential
by letting || ϑ i jk || κ liij0 in Eqn. (9) and (19), with
κ the curvature. The Morse parameters which are 3 APPLICATION
thus obtained from fitting the three-body interac-
tion energy U ijk lij0 ∫M dκ are the ground state With a closed-system of relations for the calibra-
energy per unit length, ε i ij β ijk liij0 ), tion of the potential parameters thus established,
and the curvature parameter, β ijk liij0 , where max this section aims at applying the MD-inspired
Mexp is the maximum (experimental) section structural mechanics approach to a resilience

7
Figure 4. DOE-inspired office building (top) and dif- Figure 5. Frequency calibration of potential parame-
h lij0 = 5 and
ferent discretization levels (bottom), N h/ ters of DOE-building: Model vs. ‘measured’ fundamental
N = 10 (with h the storey height). periods for different discretization levels, N h/ lij0 .
‘Measured’ frequencies were obtained from finite element
simulations of the structure using beam elements for col-
evaluation of a typical office building, inspired umns and plate elements (with both dispalcement and
from the Department of Energy Reference Build- rotational degrees of freedom, and a FE-discretization
ing Library for Medium Offices (Fig. 4): a 3-storey of N = 5). The average quadratic error vs. discretization
building of slab size 24 × 24 m2, 0.1 m slab thick- level of the calibration is displayed in the inlet, showing
ness, and storey height h = 3.35 m, resting on 16 the convergence of the discrete approach with increasing
columns of cross-sectional area 0.4 × 0.4 m2. Dif- mass discretization.
ferent levels of discretization are herein employed
to check the sensitivity and convergence of the dis- showing the high efficiency of the MD-inspired
crete approach. approach compared to classical continuum-based
approaches around the equilibrium state.
3.1 Calibration
3.2 Progressive collapse due to pressure loading
The building is discretized considering 4 con-
stant discretization lengths (columns and slabs): Figure 6 displays a sample output of one MD-sim-
N h / lij0 = , 5,10 and 15. For the elastic cali- ulations, when the structure is subjected to a pres-
bration, we consider, as ‘experimental’ frequency sure load (modeled as concentrated forces on mass
spectrum, the fundamental frequencies (eigenval- points with a load ramp fast enough to avoid a
ues and eigenvectors) obtained by Finite Element dynamic rebounce). The figure shows the evolution
Analysis of the same structure discretized by beam of the free energy Uijk of all three-body interactions
and plate elements exhibiting both displacement and of the external work achieved by prescribed
and rotational degrees of freedom (in contrast to pressure load along mass point displacements, as
the discrete approach which has only displacement well as the deflection of the top floor in function
degrees of freedom defined by the position vector of the potential energy. Following a quasi-linear
of each mass point). load-displacement response which entails (1) a
Figure 5 displays the results of the frequency quadratic evolution of the potential energy in
calibration procedure, Eqn. (14–15), together function of the prescribed forces respectively dis-
with the average quadratic error in function of placement, the structure reaches (2) its maximum
the discretization level (inlet of Fig. 5). The error capacity to store externally supplied work (by
in frequency reduces with the discretization level pressure forces) into recoverable free energy. As
close to ∼N−2, showing the high convergence rate a consequence, bending bonds break leading to a
of the discrete approach with increase in discreti- successive drop of the bending interaction energy
zation levels. That is, the discrete approach with (curve labeled ‘free energy’), which continues until
∼3N degrees-of-freedom accurately captures the elastic storage capacity is exhausted. This results
frequency spectrum of a structure defined by ∼6N in (3) the external work to approach the potential
displacement and rotational degrees of freedom, energy (1:1 – line).

8
 Figure 7. Progressive collapse analysis of structure:
Potential energy and bond energy release rate vs. number
of (unbroken) bonds. For μ > 0, the bond-rupture
process can be considered as (1) stable as long as the
Figure 6. Sample output of MD-calculations in terms
slope ∂μ/∂Nb ≥ 0; and for (2) ∂μ/∂Nb < 0, is it unstable;
of free (bending) energy, external work by pressure forces
which can be associated with a progressive collapse. (here
applied to mass points, and deflection all in function of
for a potential energy Epot ≈ 1.4 MJ, thus well beyond the
the overall potential energy of the structure.
maximum free energy peak shown in Figure 6).

An important engineering question is what it is mechanics, to which one can refer as bond energy
the load that defines the onset of progressive struc- release rate: it is the thermodynamic driving force
tural failure. In a first approach, it is of interest of the bond breakage dNb ≤ 0, and derives from the
to analyze the evolution of the potential energy systems potential energy Epot = F – WF (with F ≈ U
in function of the number of bonds, Nb. This is the Helmholtz free energy and WF the work by
shown in Figure 7. To fully appreciate the dis- prescribed forces along mass point displacements,
played result, it is of importance to realize that the ri ri 0 . Herein, we assume that the average kinetic
simulation are carried out in the so-called micro- energy of all mass points, Ek, is negligible as far as
canonical or NVE-ensemble, where N stands for the evaluation of the free energy is concerned; that
the number of mass points, V for the volume of is, F = U – TS ≈ U, with T ∼ Ek the temperature
the simulation box and E for the overall energy, i.e. and S the entropy.) With a focus on the bending
potential energy and kinetic energy; all assumed interaction energies, this is precisely what Figure 7
constant in the chosen NVE ensemble. Implicitly, displays: the potential energy of the structure vs.
this ensemble is the “engineering ensemble” of the number of bonds, and the derived bond-energy
choice for structural dynamics simulations. In fact, release rate according to Eq. (20). It is then intrigu-
a constant number of mass points simply means ing to push the fracture mechanics analogy, at the
that no mass can leave the simulation box (closed structural scale, one step further and define the
system), even if all bonds to its neighbors are bro- rules of structural bond breakage by analogy with
ken; a constant volume simply specifies the size fracture mechanics in the form:
of the box in which the structure resides; whereas
a constant energy means that the sole source of d b ≤ 0; ( μ − μ ) dNb = 0
μ − μc ≤ 0 dN (21)
internal energy dissipation is due to irreversible
damage, fracture or plastic evolutions; and that the where μc is the critical bond energy release rate
dissipated energy by bond breakage is transferred (analogue of the fracture energy) at which bond
into kinetic energy. With this “engineering ensem- breakage occurs. It is a structural property defining
ble” definition in mind, it is of interest to consider a structural energy dissipation capacity, such that
two successive states, and calculate the internal dis- −μcdNb is the energy dissipated in the creation of
sipation due to bond breakage; that is—by analogy broken bonds. Consider then a situation in which
with classical fracture mechanics: μ(Nb) = μc. For that any additional bond breakage,
Nb + dNb < Nb occurs in a stable manner, it must
∂E pot be μ(Nb + dNb) ≤ μc; thus, considering a Taylor
dD = − μdN
dNb ≥ 0 μ = (20)
∂N b expansion:

where μ (of dimension of energy) is recognized as ∂μ

μ ( Nb ddNb ) μ (Nb ) + N b ≤ μc
dN (22)
the equivalent of the energy release rate in fracture ∂N b

9
 That is, much akin to the stability analysis of whereas σ is the standard deviation. For purpose of
fracture processes (while noting that dNb < 0): example, we consider a coefficient of variation of
CV = 50% for wood structure with, and CV = 30%
∂μ ∂ 2 E pot for RC structures. The results are displayed in Fig-
= ≥0 (23) ure 8, in terms of the cumulative distribution func-
∂N b ∂N b 2
tion (CDF) of the failure load, expressed here in
corresponding wind-speeds V p.
p An interesting
As the bond-energy release rate thus increases in
observation is that a CV of 30% for RC structures
the process of bond rupture (dNb < 0), the collapse
translates into a CV of the failure load of only 7%
can be considered to be progressive. This is shown
(wind speed CV = 7%); whereas a CV of 50% for
in Figure 7.
wood structures entails a CV of the failure load
of 13% (CV = 3%). This significant reduction in
3.3 Fragility curves uncertainty from the bond-level to the structural
scale highlights the capacity of the structural sys-
With a computationally efficient tool thus in place,
tem to redistribute externally supplied energy (load)
it becomes possible to develop fragility curves by
throughout the system, thus showing the value of
considering—for a typical class of structures—a
heterogeneity on limit loads related to bond fracture.
variability of the bond energy input parameters.
Similar results have been found for highly heteroge-
This is illustrated here for, respectively, a RC—and
neous porous materials (Laubie et al., 2017b,c).
a wood structure, designed—deterministically-
for similar limit loads. To obtain fragility curves,
400 simulations are carried out for each structure
with randomly assigned, normally distributed limit 4 CONCLUDING REMARKS
angles for wood structures and limit curvatures for
RC-structures; that is, The need for accurate predictions of resilience of
buildings subject to extreme events requires a new

ϑ i jk ( )
κ liij0 ≤ ε (ϑ c σ ) (24) generation of engineering tools that permit a damage
assessment below the limit loads for which structures
are classically designed. This damage of structures
where ϑ c is the mean value defined by Eq. (19) of both non-structural and structural members can
calibrated from experimental data (Figs. 2 and 3), be viewed as a sequence of bond breakage. The MD-
inspired method here proposed addresses this bond
breakage explicitly in a discrete way: rather than
extending the classical realm of continuum based
structural members into the non-linear regime, the
MD-approach focusses on the elastic and in-elastic
interactions between mass points that discretize the
bulk of the structure. Provided a sufficient level of
discretization, the MD-approach provides an effi-
cient way to capture the energy transformations
witnessed by a structure when subject to load. The
following points deserves attention:
1. There are several advantages of the MD-
method compared to e.g. Finite-Element-based
continuum approaches; such as the absence of
instabilities in MD, related in FEM to the loss
of positive definiteness of the stiffness matrix
and associated mesh-dependency issues; effi-
ciency (O(N) for MD vs. O(N2) ÷ O(N log (N) of
Figure 8. Example of fragility curves for a a reinforced FEM; ease of bond removal etc., which makes
concrete structure and a wood structure with timber-mo- the MD-based method highly efficient for large-
ment connections. The curves represents the cumulative scale recursive simulations required for e.g. fra-
distribution function (CDF) of the velocity correspond- gility curve determination.
ing to wind limit loads of the structures with randomly
assigned normally distributed rupture angles (for wood)
2. With a focus on internal interaction energy
and rupture curvature (for RC-structures). [For reference, description, the MD-approach permits a
reported wind speeds of the September 19, 2017 Hurri- straightforward link with other energy-based
can Maria that devastated Puerto Rico were 90–125 mph, mechanics approaches, such as elasticity and
with wind gusts of 175 mph]. linear and non-linear fracture mechanics. This

10
 common denominator provides a clear path for support was provided by ICoME2 Labex (ANR-
calibration of the interaction potentials, from 11-LABX-0053) and the A*MIDEX projects
measured frequency spectra for energy parame- (ANR-11-IDEX-0001-02) cofunded by the French
ters defining the equilibrium configuration, and program “Investissements d’Avenir,” which is man-
from almost classical test data (such as moment- aged by the ANR, the French National Research
curvature curves) for bond energy parameters. Agency. All simulations were carried out with the
Existing interatomic potential expressions are open source code LAMMPS (Large-scale Atomic/
readily adapted to permit a representation of Molecular Massively Parallel Simulator), dis-
bond-forces and moments. Similarly, based tributed by Sandia National Laboratories, a US
upon thermodynamics derivations, the occur- Department of Energy laboratory.
rence of progressive collapse mechanisms can be
detected and used for structural optimization.
3. The simulation results presented in this paper REFERENCES
were obtained within the context of NVE-
ensemble simulations. This means that any Barbato M, Petrini F, Unnikrishnan VU, Ciampoli M.
form of energy dissipation is restricted to bond- 2013. Performance-based hurricane engineering
(PBHE) framework. Struct Saf 2013;45:24–35.
breakage, and does not consider other sources Charif A., Shannag M.J., Dghaither S. 2015. Ductility of
of dissipation, such as viscous phenomena. The reinforced lightweight concrete beams and columns.
application of the proposed technique within Lat. Am. J. Solids & Structures 2015, 11, 1251–1274.
the context of other ensembles is a natural exten- Cimellaro GP, Reinhorn AM, Bruneau M. 2010. Frame-
sion. For instance, considering an NVT ensem- work for analytical quantification of disaster resil-
ble (where T stands for the temperature of an ience. Eng Struct 2010;32:3639–49.
outside bath with which the system is in contact) Federal Emergency Management Agency (FEMA), 2016.
would naturally entail consideration of a linear Hazus | FEMA.gov 2016. http://www.fema.gov/hazus
velocity-dependent damping in the equation of Gehloff M., Closen M., Lam F. 2010. Reduced edge
distances in bolted timber moment connections with
motion, which at “low enough prescribed tem- perpendicular to grain reinforcements. World Confer-
perature” would lead to the static response of the ence on Timber Engineering 2010, Paper 123.
system (see Appendix I for temperature consid- Hattar C.P., Cheng, J.J.R. 1995. Development of moment
eration). Here future work is needed in order to connections in glued-laminated Alberta spruce and
identify the link between engineering approach pine timber. Report A5020–121. Canadian Forest
and physics approach enabled through a com- Service and Land and Forest Services, Edmonton,
mon language rooted in energy and dissipation Alberta, Canada.
representations; much akin to recent results Insurance Information Institute 2016. Catastrophes:
obtained in the field of mesoscale poroelasticity Insurance Issues | III 2016. http://www.iii.org/
issue-update/catastrophes-insurance-issues.
of heterogeneous media (Monfared et al. 2017). Laubie, H., Monfared, S., Radjai, F., Pellenq R., Ulm,
4. The fragility curves thus determined can be used F.J. 2017a, Effective potentials and elastic properties in
in convolution expressions with probability of the Lattice-Element Method: Isotropy and Transverse
load occurrence to provide a rational criterion Isotropy, J. Nanomech. Micromech. 7(3): 04017007.
for the probability of failure of structures in spe- Laubie, H., Monfared, S., Radjai, F., Pellenq R., Ulm,
cific environments (Noshadravan et al. 2017). F.J. 2017b, Disorder-induced stiffness degradation
These probabilities have become key to evalu- of highly disordered porous materials, J. Mech. Phys.
ate the resilience of structures and communities, Solids 106207–228.
and become now available for a much larger class Laubie, H., Radjai, F., Pellenq R., Ulm, F.J. 2017c, Stress
transmission and failure in disordered porous materi-
of structures. Ultimately, we envision that such als, Phys. Rev. Lett. 119, 075501.
MD-based structural simulations can be used Laubie, H., Radjai, F., Pellenq R., Ulm, F.J. 2017d,
together with CFD simulations to evaluate the A potential-of-mean-force approach for fracture
resilience of a city’s building stock—an urgent mechanics of heterogeneous materials using the lat-
task in view of the perils of global warming. tice element method. Journal of the Mechanics and
Physics of Solids 105, 116–130.
Monfared, S., Laubie, H., Radjai, F., Pellenq R., Ulm,
ACKNOWLEDGEMENT F.-J. 2017, Mesoscale poroelasticity of heterogeneous
media, J. Nanomech. Micromech. 7(4): 04017016.
Multihazard Mitigation Council 2005. NATURAL
This research was carried out by the Concrete HAZARD MITIGATION SAVES: An Independent
Sustainability Hub (CSHub@MIT), with funding Study to Assess the Future Savings from Mitigation
provided by the Portland Cement Association Activities Volume 1 -- Findings, Conclusions, and
(PCA) and the Ready Mixed Concrete Research & Recommendations. 2005.
Education Foundation (RMC E&F). The CSHub@ Noshadravan, A., Miller T.R., Gregory J.G. 2017, A Life-
MIT is solely responsible for content. Additional cycle Cost Analysis of Residential Buildings Including

11
 Natural Hazard Risk, Journal of Construction Engi- APPENDIX II: FROM 2-BODY
neering and Management 143 (7), 04017017. BENDING INTERACTIONS TO 3-BODY
Papadopoulos A. 2016. Resilience -- The Ultimate INTERACTIONS—LINK WITH BEAM
Sustainability -- Lessons from Failing to Develop a THEORY
Stronger and Safer Built Environment. Resilience
Action Fund.
Qomi MJA., Ulm F.J., Pellenq R. 2015. Physical ori- There are different ways to account for three-body
gins of thermal properties of cement paste. Physical interactions related to bending, namely one that
Review Applied 2015; 3, 064010. starts at the bond level ij, the other by considering
Shinozuka M, Feng MQ, Lee J, Naganuma T. (2000) jumps in angular variations at all mass points. The
Statistical analysis of fragility curves. J Eng Mech first approach forms much of the basis of classical
2000;126:1224–31. beam theory; whereas the second is at the core of
Walker GR. 2011. Modelling the vulnerability of build- MD-type approaches, expressed e.g. by Eq. (7). This
ings to wind—a review. Can J Civ Eng 2011;38:1031–9. appendix shows the formal link between the two
Wang M., Song X., Gu X., Zhang Y., Luo L. 2015. Rota-
tional behavior of bolted beam-to-column connec-
approaches in the context of harmonic potentials.
tions with locally cross-laminated glulam. J. Struct. More specifically, the bond-level approach con-
Eng. 2015, 141(4), 04014121. siders, in addition to the stretch potential, U sij , a
Yeo DH, Simiu E. 2011. High-Rise Reinforced Concrete bending potential, U bij , constructed in a way to
Structures. Database-Assisted Design for Wind. J. satisfy per (mass-less) link ij the force and moment
Struct. Engrg 2011; 137(11): 1340–1349. balance equations (Laubie et al. 2017a):

   ∂(U sij + U bij )

APPENDIX I: ON MASS-POINT Fi j F ji = 0; Fi j = −  (A-4)
∂δ i
DISCRETIZATION WITHIN THE LIMITS
OF “CLASSICAL” MECHANICS      ∂U ij
Mij M ij + rij F ji = 0; M i j = − b (A-5)
∂ϑ i
The employment of a molecular simulation
approach within the framework of classical      
mechanics – as opposed to Quantum Mechanics – with rij rj − ri ; δ i ri − ri 0 stands for  the
requires that the discretization respects the rules displacement vector of mass point i, and ϑ i for
of classical mechanics (i.e. Newton’s law (2)) as the vector of angular rotations. In order to satisfy
expressed by the de Broglie wavelength: (A-4) and (A-5), this requires to develop bending
interactions in function of both  lateral
 displace-
p
h ments θ ijb ij ij
b ( ij ij )  b (δ j
ij
δ i  ij ϑ i ) and
lij0  λ = (A-1) θ ijt ij ( ij ij )  t  (δ j δ i ij ϑ i )andangle
2π m
mkkBT t
 
variations, ϑ ijb (ϑ j − ϑ i ) ⋅ ebij and ϑ ij (ϑ j −ϑ i )⋅ etij
ij
where lij0 is the distance between mass points i and in the link’s local orthonormal basis ( n b , t );
j, h is the Planck constant, m is the particle mass that is, in the case of harmonic expressions while
(typically m ρ A Allij0 with ρ the mass density of neglecting torsion effects (Laubie et al. 2017a):
the material and A the cross-sectional area of the
1 ⎡ 2⎤
( ) 1
( )
2
structural member, kB is the Boltzmann constant, U bij = ε ijb ⎢ θ iijb − ϑ ijtθ iijb + ϑ iijt ⎥
and T the absolute temperature. It thus follows: 2 ⎣ 3 ⎦
1 ⎡ 2⎤
( ) 1
( )
2
+ ε ijt θ iijt − ϑ ijbθ ijt + ϑ ijb ⎥ (A-6)
h2 / 3 2 ⎣ 3 ⎦
lij0  (A-2)
(2πρ A
AT )
1/ 3

Or, alternatively, in terms of the moment vectors:

Considering h = 6.626070040(81) × 10 −34
Js; kB =
1.38064852(79) × 10−23 J/K; hence: 2 ⎡
b ( ( )
M i j t ) − M i j M ij t + M ij ,t ⎤⎥
2 2
U bij
ε iij ⎣ ⎦
1′ / 3
⎛ ρA AT ⎞ 2 ⎡
(
+ t ⎢(M i j b ) − M i j b M ij b + M ij ,b ⎤⎥ )
2 2
lij0 ⎜ 1.7169 × 10 −15 (A-3) ⎣ ⎦
(A-7)
⎝ kg m 2 K ⎟⎠ ε iij

Eqn. (A-1) to (A-3) thus exclude simulations at where ε ijb and ε ijt are bending  inter- 
(or close to) zero (Kelvin) temperature. A similar action energyy parameters,
 and M i j l M i j ⋅ elij
restriction applies to the Lattice Element Method and M j il
M j ⋅ el (l = t,bb) . If we let ε ij =
i ij b

(see Laubie et al. 2017d). 12EII bb / lij , ij 12 EItt / lij , we readily recognize

12
   
that expressions (A-6) and (A-7) are nothing but [of the form Uτij  12 ij (ϑ j ϑ i ) enij = 2κ1iij (M i j n )2
the (free and complementary) bending energy of [with M i j M i j ⋅ enij M ij ⋅ enij the link’s tor-
a linear elastic beam finite element with displace- sion moments satisfying momentum balance], one
ment and rotation degrees of freedom. obtains expression (8) of the rotational stiffness
The previous 2-body bending interaction expres- tensor C ijk .
sions can be recast in terms of a 3-body interaction Since the link energy expressions U bij and Uτij
energy of a mass point i connected to two other satisfy moment equilibrium (A-5), the global angu-
mass points j and k, when equating the energies of lar momentum is automatically satisfied; for which
the two links: reason the MD-approach requires only satisfying
the linear momentum (2) and thus the determina-
1  
U bijk (1 ij
2
) ( )
U b + U bik = ϑ i jkj ⋅ C ijk
2
b
⋅ ϑ i jjk (A-7)
tion of the forces from Eqn. (6) and (10):

 N ∂U
ijj
N ∂U
ijk
   Fi = − ∑ j s + ∑ j 
where ϑ i jk ϑ i j − ϑ ik is the jump in angle at point ∂ri ∂ri
i, whereas C ijk b
is the second order tensor of rota- 
 N  ∂ϑ jjk
= ∑ j Fi j enij + ∑ j M i jk ⋅ i
N
tional stiffness that relates to the link energies (A-9)
( ijb , ε ijt , ε ikb , ε ikt ) by: ∂ri
 
⎡1  1   where M i jk = C ijk ⋅ ϑ i jjk ; and, for small rotations
( )
−1
b
C ijk = 6 ⎢ b etij ⊗ etij + t ebij ⊗ ebij (around the equilibrium position, for which
ε
⎣ ij ε iij
harmonic expressions hold; see Fig. 1):
1  ik  k 1  ik  ik ⎤
+ b et ⊗ t + t eb ⊗ eb ⎥ (A-8)    
ε iik ε iik ⎦  rij × riik − rij0 × riik0
ϑ i jk ≅   (A-10)
rij0 riik0
If, in addition, harmonic torsion terms are
considered in the link interaction potential

13
Exploring the Variety of Random
Documents with Different Content
 The principal gullies beginning from the North were Asmak Dere,
Aghyl Dere, Chailak Dere and Sazli Beit Dere. The last of these ran
down to what, on the 6th of August, was the Northern extremity of
the Anzac position. Between it and Chailak Dere, a spur left the main
ridge of Sari Bair and ran down towards the sea: after it came into
Christian hands, this spur was christened Rhododendron Ridge and
played an important part in the August fighting. The portion of the
Sari Bair range, which was joined by Rhododendron Ridge, was
known as the Chunuk Bair and here the battle was to rage most
fiercely. It culminated to the northward in a summit called Hill Q.,
and thence the range trended eastward to Koja Chemen Tepe, the
culminating height of the position and the objective of the Suvla
force. Half-way down the slope of the Chunuk Bair facing the Gulf of
Saros, was a patch of cultivation known as The Farm. The whole of
the seaward face of the Chunuk Bair was covered with prickly scrub
about four feet high and cut by narrow ravines running down to the
Aghyl Dere which starts just below The Farm.
On the night of August 6th General Godley had launched his
attack northward from Anzac. By 1.30 a.m. on the 7th the mouths of
the Chailak Dere and Aghyl Dere had been seized and a strong
lodgment made on Damakjelik Bair, a detached hill between the
Asmak and Aghyl Deres. This lodgment protected the left flank of
the assault on the Chunuk Bair which was then launched.
By dawn the left assaulting column had forced its way up the
Aghyl Dere, and the Indian Infantry Brigade had occupied The Farm,
while on the extreme left the 4th Australian Brigade had reached the
Asmak Dere, and were advancing towards Koja Chemen Tepe. The
advance of the New Zealanders up the Chailak Dere had been
slower, but soon after 6 a.m. they had stormed the Turkish trenches
on Rhododendron Ridge, and established themselves at the point
where that ridge joins the Chunuk Bair. At the same time they got
into touch with the Indian Brigade on their left. Preparations were
made for an assault on the main Chunuk Ridge, but the troops were
terribly exhausted by their night marches in an impossible country,
and the arrival of Turkish reinforcements made further advance by
daylight impossible. It was decided to allow the troops to rest, and
attack again just before dawn on the 8th.
For this attack the New Zealanders, Australians and Indians who
had taken part in the first day’s fighting were reinforced by six
battalions of the 13th Division. On the right the assault from
Rhododendron Ridge on the Chunuk Bair was successful, and a firm
footing on the crest was gained; but the centre attack was unable to
advance much further than The Farm, and the attempt on Koja
Chemen Tepe was unsuccessful. The General resolved to attack
again under cover of darkness, and called up the two battalions of
the 29th Brigade, which had not already been allotted any duty, to
take part in it.
The Hampshires and Royal Irish Rifles had moved at 1 a.m. on the
8th from their bivouacs in Shrapnel Gully, to Rest Gully. This gully
was situated near the southern end of the great sap which ran
northward from Anzac Beach towards what was known as No. 2
Post. The cove of Anzac itself, between the headlands of Hell Spit
and Ari Burnu, though often swept by Turkish fire, was concealed
from the enemy’s view by overhanging cliffs. To the northward,
however, the beach was commanded throughout its length by the
heights of the Chunuk Bair, and men moved on it by daylight at their
peril. In order to facilitate movement by day, Australians and New
Zealanders working by night had dug a sap wide and deep enough
to hold a mule, which ran northward parallel with the sea for nearly
a mile. This had acquired the name of “ The Anzac Sap.”
About 10 a.m. on the 8th, the Hampshires and Rifles fell in, and
followed Brigade Headquarters along this sap in single file, until they
reached its northern end at No. 2 Post. At this point General Godley
had established his headquarters, and here the two battalions
collected and waited for the greater part of the day. Late in the
afternoon they again moved northwards, and entered the area which
had just been won from the enemy. Here they came under fire from
hostile snipers, but worse was to come. They had been ordered to
move up the Chailak Dere, but the Turks were well aware that this
was one of the few paths by which reinforcements could approach
the Chunuk Bair, and were shelling its entrance persistently.
In small parties the men dashed through the barrage, and in most
cases got off without heavy losses. Lieutenant Graham Martyr’s
platoon of the Irish Rifles, however, was unlucky, and was almost
annihilated. Having passed this dangerous spot, the whole long
procession moved on in Indian file up the deep bed of the Dere.
Progress was slow, since the gully was half choked already with
supplies and reinforcements going up to the hills, as well as with the
wounded coming down. As dusk fell the two battalions bivouacked
on the slopes leading down to the Gully. They did not however have
much time for rest, since at 9.15 p.m. they were aroused to take
part in the assault on the Chunuk Bair. For this, three columns were
being organized, the Rifles and Hampshires being allotted to the
centre column, which was under the command of Brigadier-General
A. H. Baldwin, who had previously commanded the 38th Brigade.
Besides the two 10th Division battalions, General Baldwin had also
the 6th East Lancashires and 5th Wiltshires, which belonged to the
13th Division. The column which was to move on the right of the
centre column was commanded by Major-General F. E. Johnston,
and consisted for the most part of New Zealanders. It was intended
to operate from and extend the territory already gained on the
Chunuk Bair. To the left a column under Major-General H. V. Cox,
consisting of the 4th Australian Brigade, the Indian Brigade, and four
battalions of the 13th Division, was to attack Hill Q. at the northern
end of the Chunuk Bair.
General Baldwin’s column was entrusted with the task of moving
up the Chailak Dere and attacking Hill Q. from the south-west, with
its flanks protected by the columns on the right and left. The
intention of the Commander-in-Chief had been that this centre
column should start from the Chailak Dere and deploy behind the
line already occupied by the New Zealanders, moving thence at
dawn along the crest of the Chunuk Bair to assault Hill Q.
Unfortunately, however, this complicated manœuvre miscarried, as
the guides allotted to the column missed their way, with the result
that the troops, after alternately marching and halting all through
the night, found themselves at dawn on the 9th in the Aghyl Dere at
the foot of the Chunuk. The column on the left had been more
fortunate, and its head succeeded in reaching its objective,
occupying the col which connects Hill Q. with the Chunuk Bair.
Hardly however had the Ghurkas and South Lancashires gazed on
the town of Maidos and the Dardanelles crowded with transports
bringing up reinforcements for the enemy, when they were shelled
off the position, which was promptly re-occupied by the Turks.
Meanwhile General Baldwin’s column was closing up and getting
into formation for the attack. The men went forward with splendid
spirit, but the task they were called on to perform was beyond
human power. Not only did the enemy’s shrapnel fire redouble its
force, but the whole of the left flank was enfiladed by hostile
machine-guns, which almost wiped out the East Lancashires. In this
advance many of the officers of the Rifles were wounded. To climb
the Chunuk in broad daylight in the face of an enemy well supplied
with machine-guns and possessing observation posts from which he
could direct the fire of his still unsubdued artillery, was a harder feat
than the storming of the breach of a hostile fortress in the
Napoleonic wars, since the distance to be covered was so long and
so rugged, that it was impossible to maintain the impetus of the
charge. An attempt to find easier ground to the left failed, and so
the Rifles and Hampshires took up their position behind the crest of
a small under-feature which jutted out some three hundred yards
from The Farm.
General Baldwin was accompanied to this position by General
Cooper and the staff of the 29th Brigade, who, since the whole
Brigade had been allotted piecemeal to different Commanders, came
up to assist in passing orders. At 9 a.m. a company and a half of the
Hampshires under Major Pilleau were ordered to move up the slope
to the right and try to get in touch with the New Zealanders of
General Johnston’s column. While doing so they came under heavy
shrapnel fire, but succeeded in working their way up to that part of
the ridge which was in the hands of the New Zealanders.
 BRIGADIER-GENERAL R. J. COOPER, C.V.O., C. B.
COMMANDING 29TH BRIGADE

The position thus gained was maintained throughout the 9th, the
Hampshires holding a line down the seaward slopes of the Chunuk
Bair, and then turning almost at right angles towards the north-east
along the crest of the under-feature above The Farm. The Rifles
prolonged this line on the left to a point where it was taken over by
the two battalions of the 38th Brigade. This left flank was somewhat
in the air, as the flank-guard on the Damakjelik Bair was more than a
mile in rear of the line. The only protection to this flank was that
afforded by the Left Column under General Cox, which had
succeeded in occupying Hill Q. at dawn and had been driven off it.
These had now withdrawn to the line of the Asmak Dere, but they
were terribly exhausted. The Australians and Indians had been
marching and fighting in a tropical climate for forty-eight hours
without relief, while the New Army battalions had lost heavily,
especially in officers.
Throughout that day Baldwin’s column lay out on the face of the
Chunuk Bair. Pinned to their positions by the Turkish shrapnel which
hailed on them without respite, they suffered terribly from the
scorching rays of the sun. Shade there was none, for the scrub was
so prickly that it was impossible to crawl underneath it, while
nothing short of direct cover afforded any protection from the sun
vertically overhead. Water was terribly scarce; although wells had
been discovered in the bed of the Aghyl Dere, it was a task of great
difficulty to convey the water up to the troops, since part of the
Aghyl Dere was swept by the enemy’s fire. The torments of thirst
were increased by the fact that the only food available for the men
was salt bully beef and hard dry biscuit. It was an effort to swallow
more than a few mouthfuls, and to the weakness caused by enteritis
was added the weakness of inanition.
The casualties did not appear heavy, but they steadily mounted
up, and in the course of the day each of the 29th Brigade battalions
lost about a fifth of its strength. Night brought relief from the sun,
but no rest, for the battalions were ordered to entrench themselves
where they stood. The exhausted men were incapable of heavy
labour, but a narrow shallow trench was gradually excavated. Night
too gave an opportunity to send the wounded away, for after hasty
dressing had been applied by battalion medical officers they had, of
necessity, been obliged to await a convenient occasion for their
removal. The nearest hospital was four miles away on the shore at
Anzac, and a terrible burden thus fell on the stretcher-bearers, who
had to carry their comrades all this distance. Every man who could
limp or hobble down to the beach, walked, but the serious cases
were numerous, and the battalion establishment of stretcher-bearers
(which had not been fixed with such an abnormal campaign in view)
found itself severely taxed. During the night the New Zealand
Brigade on the right of the Hampshires, was withdrawn and relieved
by part of the Wiltshires and Loyal North Lancashires, and also by
the 6th Leinsters.
Dawn came, and with it the Turkish counter-attack. Throughout
the night their artillery had thundered unceasingly, but before
daybreak it redoubled in violence. As the light grew, an enormous
mass of the enemy threw itself against the battalions holding the
lodgment effected by the New Zealanders on the crest of the
Chunuk Bair, while further hordes moving down from the north and
Hill Q. attacked Baldwin’s column at The Farm. The two battalions on
the crest were almost annihilated, and the ground they held was
lost. Fortunately, however, as was described in the last chapter, the
momentum of the attack was checked by our artillery.
The Turks moving down the crest of the Chunuk were in full view
of the fleet, and the fire brought to bear on them was so terrific that
their reinforcements were unable to penetrate the barrage. They
pressed on against Rhododendron Ridge, but were stopped by the
concentrated fire of ten New Zealand machine-guns which were
placed in position by a famous Hythe musketry expert. But although
for the time the danger was lessened and the Turkish losses were
enormous, yet the fact that the two battalions holding the Ridge of
the Chunuk had been driven back, left the right flank of the
Hampshires dangerously exposed. Although its losses were very
heavy, this company and a half which had been sent out to maintain
connection with the ridge succeeded in holding its ground.
The remainder of the Hampshires were now up in the firing line on
the right of The Farm position, but were losing very heavily. Colonel
Bewsher who commanded them had been seriously wounded in the
head about 6 a.m., and was resting before making his way down to
the beach when a wounded sergeant-major informed him that there
appeared to be no officers left unhurt. He, therefore, wounded as he
was, returned to the firing line, and discovered that although there
were still two captains with the detached company and a half, the
remainder of the battalion had not only lost all its officers but all its
company sergeant-majors and quartermaster-sergeants as well. One
machine-gun had been put out of action by a shell, but the men
were holding their ground manfully.
Meanwhile, on the left, the hostile attack developed with even
greater force. Orders had been received to send the 5th Wiltshires to
relieve the New Zealanders on the crest of the Chunuk, but one
company had been retained as its withdrawal would have left part of
the line completely unmanned. A company of the 9th Warwicks had
come up to relieve the Wiltshires, but were found to be very weak.
There were also on the left in addition to the Royal Irish Rifles,
about 50 men, all that remained of the East Lancashires, and a few
Ghurkas and Maoris belonging to the left column who had retired
down the hill and joined General Baldwin.
Against these few exhausted men, less than a thousand in all, the
Turks were free to throw the whole of their reserves, since by this
time (dawn, Tuesday) it was clear that the advance from Suvla was
not likely to get much further. They came on again and again,
covered by a very heavy shrapnel fire, and again and again they
were driven back. Our losses, however, were terribly heavy and they
could afford to lose ten men to our one, for our last reserves (except
for one battalion five miles away) were already up in the firing line.
Worst of all were the casualties in officers. The dawn was misty and
just as it began to grow light General Baldwin was killed. Almost at
the same instant General Cooper fell, severely wounded in the lungs.
Colonel Bradford of the Rifles was then the senior officer with the
column, but just as he was informed that the command devolved
upon him, he, too, fell seriously wounded.
In quick succession, Major Morphy, the second-in-command of the
Rifles, received a bad wound in the thigh, and Major Eastwood, their
Adjutant, was killed. Very shortly afterwards Captain McCleverty, the
Brigade-Major, was hit by a bullet which passed through both cheeks
and broke his jaw, while Major Wilford of the Rifles, on whom the
command of his battalion had devolved and who had exhibited great
courage and resource, sustained a severe wound in the head.
Colonel Bewsher of the Hampshires, who had been wounded twice
but was able to stand, then took over the command of all that was
left of General Baldwin’s force. The oft-repeated attacks continued,
nearly all the junior officers were down, and though our thin line
was never actually pierced yet in many places the enemy came so
near that they fought with our men at close quarters. In an effort to
repulse a rush of this kind on the left about 9 a.m. Captain Gerald
Nugent, Staff Captain of the 29th Brigade, fell, revolver in hand,
leading his men forward. His death was a sorrow to the whole
brigade, for he was a man in a thousand. The surliest cynic who
cultivated a grievance against all Staff Officers found himself quite
unable to resist Nugent’s kindness of heart and wonderful charm of
manner. The manner of his death was suited to his bright and
unselfish life.
About this time Colonel Bewsher came to the conclusion that the
position was untenable. On the right the enemy had re-occupied the
crest of the Chunuk Bair and were pressing the Hampshires hard,
while on the left General Cox’s column had retired to the Damakjelik
Bair in rear, leaving the Chunuk completely exposed on that flank.
There appeared nothing to prevent the Turks from establishing
themselves in the Aghyl Dere and so cutting the only line of
communication. The casualties, too, had been terrible. Every staff
officer on the hill was either killed or wounded. The Hampshires and
Rifles had only four officers left between them and the English
companies were in just as bad case. The fight had been raging for
over four hours, the men were utterly exhausted, and there was no
sign of reinforcements. Colonel Bewsher, therefore, ordered a
retirement which was carried out in a regular and orderly manner.
This little mixed force, drawn from seven different units, comprising
in its number men from Winchester and Salisbury, Birmingham,
Burnley and Otago, Belfast and Khatmandu, had held a weak
position against enormous odds, with little food and less water, for
over 24 hours, and when they retired had hit the enemy so hard that
they were not pursued.
Even then they were not disposed of, for at the bottom of the hill
a staff officer (Captain Street) who was arranging to send up water
and ammunition, called to them to come on again and they
responded. The Hampshires on the right under their last officer, the
Rifles in the centre, and the Wilts and Warwicks on the left, turned
their faces again to the Hill of Death and advanced once more. The
effort was futile for by this time the Turkish line was strengthened by
machine-guns, but it was heroic, a vindication of the power of the
spirit of man to soar above hunger and thirst and the imminent fear
of death, and place itself on a level with that of the heroes.
Both battalions had suffered terribly. The Hampshires, who had
gone into action on the morning of the 9th, with a strength of
approximately twenty officers and over 700 men, had at noon on the
10th one combatant officer (Captain Hellyer) and not more than 200
men fit for duty. A few more who had lost their direction in the
retirement rejoined in the course of the following day. The Rifles
were in nearly as bad a condition. They were commanded by their
junior captain, who had only been promoted to that rank at Mudros,
and two subalterns were all the combatant officers that he had
under him. The men, too, had been driven back in small parties and
had been scattered, and it was clear that neither of the battalions
was in a position to fight again for some days. Fortunately for their
personal well-being, both of their quartermasters had survived the
fight. Lieutenant Dowling of the Rifles had toiled unceasingly in
drawing and attempting to send up rations, water, and above all,
ammunition. The Rifles, too, had obtained devoted service from their
doctor, Lieutenant Adam, R.A.M.C., who had worked like a hero in
dealing with the hundreds of cases that had passed through his
hands.
The Hampshires had found their quartermaster a tower of
strength. Not only had Lieutenant Saunders worked magnificently
throughout the fight, but in the difficult days of reorganization, he
turned his hand to anything and acted as Adjutant and Company
Commander and in any other capacity in which he could be of use.
In spite of the misfortunes of his battalion he remained cheerful and
imperturbable, and it was refreshing to look at his beaming, bearded
face. In recognition for the good work he had done he was awarded
the Military Cross. A quartermaster is described as a non-combatant
officer, and his services are not always fully recognized, but in
Gallipoli he was exposed to fully as much danger as anyone else,
while the load of responsibility on his shoulders was far greater. Any
negligence on his part meant that his battalion would go hungry and
thirsty and lack ammunition at a pinch. Soldiers will agree that no
man does more important work and better deserves recognition than
a good quartermaster.
Meanwhile, the last battalion of the brigade was hurrying towards
the scene of action. At 7 a.m. on the morning of the 10th the
Connaught Rangers received orders to prepare to move at once. The
detached company, which had been doing fatigue work at Brown’s
Dip all night, was hastily recalled, and in less than an hour the
battalion moved off. It was necessary for them to take a circuitous
route to the beach for fear that the Turkish observers on Gaba Tepe
should notice that the right of the Anzac position was being
weakened. At 9 a.m. Anzac Cove was reached, and the battalion
hurried on northwards. As it entered the long sap leading to No. 2
Post, it began to realise the severity of the fighting for the first time,
for the sap was full of wounded.
Most of these wounded, too, belonged to the Leinsters,
Hampshires, and Irish Rifles, and their number made it clear that the
brigade had suffered heavy losses. It was only, however, when
checks in the march allowed an opportunity of speaking to the less
seriously injured that the full extent of the casualties became clear.
The officers of the Rangers heard with growing sorrow that the
whole Brigade Staff were either killed or seriously wounded, and that
the Rifles and Hampshires had practically ceased to exist. They saw
carried past them, with drawn set faces, half masked by dry and
clotted blood, men who had worked and played with them at the
Curragh and Basingstoke, whose wives and children were their
friends. Even in the pale, unwashed, unshaven faces and strained
and suffering eyes of the less seriously wounded who paused to
speak to them, they read the realization of the ordeal that lay before
them. Behind all was the thought of the friends lying up on the slope
of the Chunuk Bair, whose families would never look on them again.
It was an unnerving ordeal for a young regiment, but, fortunately,
there was little time for reflection, and the Rangers hurried on. At
No. 2 Post there was a short halt, while Colonel Jourdain interviewed
General Birdwood and General Godley, who informed him that the
Turks had broken through a section of the line, and that his battalion
was placed under the command of General Cox to help him to
retrieve matters. He was exhorted to move forward as quickly as
possible, as the need for reinforcements was urgent. Accordingly,
before the rear of the battalion had extricated itself from the sap,
the head was in motion again. It must be borne in mind that except
for the brief information which the Colonel had received from
General Birdwood, officers and men alike were completely ignorant
of the previous operations. They knew nothing of the extension of
the Anzac position northward on the night of the 6th, nor of the
repeated attacks on the Chunuk Bair; above all, they were unaware
that a landing had taken place at Suvla. It was, however, clear to
them that they were in new country, for up to No. 2 Post they had
moved by well-trodden paths protected at any point of danger by
saps and sandbags. Now they were in open country, with the sea on
their left, and on the right a range of low foothills, which in places
sank sufficiently to enable them to see the ridge of the Chunuk high
above them.
Here and there accoutrements hurriedly cut off a wounded man
showed that Turkish shrapnel and snipers had to be reckoned with,
but there appeared to be a momentary lull in the fighting. Past the
mouth of the Chailak Dere the Rangers hurried in single file sweating
under the pitiless sun past Bauchop’s Hill, and over a low nek into
the Aghyl Dere. Here, again, their progress grew slower, for the gully
was narrow and filled with wounded and mules and resting Ghurkas.
It was stiflingly hot, and the smell of the mules and the dust, shut in
tightly between the high scrub-fringed banks of the gully, were
almost unendurable. The Rangers moved forward for a hundred
yards at a time, until at 11.15 a.m. General Cox’s headquarters were
reached.
The halt there was a brief one for the Rangers were at once
directed to place themselves under the orders of Brigadier-General
W. de S. Cayley commanding the 39th Brigade, for the purpose of
reinforcing his line. Below General Cox’s headquarters, the Aghyl
Dere forked into two branches, one coming from the Damakjelik Bair,
the other, the southern branch, from the foot of the Chunuk. Along
this southern branch the Rangers went in single file for about four
hundred yards, passing an extemporized dressing-station crowded
with Ghurkas in slouch hats, and broad, baggy shorts, until they
reached a point where a spur ran down from the Damakjelik Bair
and gave a certain amount of protection against rifle fire from the
Chunuk. Here, General Cayley had established his headquarters in
the narrow protected area; in rear of it were crowded all that
remained of three or four English battalions. Above, the crest was
lined by Sikhs. Into this zone of safety the Rangers hurried, and
after forming up, lay down to rest while their Colonel went to
General Cayley for orders. The General was established in an
observatory of boughs, which gave some shelter from the view of
snipers on the Chunuk, and after giving Colonel Jourdain and the
officers who accompanied him a very welcome cup of tea, he
proceeded to explain the situation.
Although General Baldwin’s column had been driven from The
Farm position, yet, apparently, it had not yet been occupied by the
Turks. It was believed that they were greatly exhausted and had
been much discouraged by the heavy losses inflicted on them by our
artillery, and it was considered that it might be possible to re-occupy
The Farm position. Accordingly “A” and “B” Companies of the
Connaught Rangers were ordered to advance up the Aghyl Dere,
climb the slopes of the Chunuk Bair as far as The Farm, and occupy
the position which was reported to have been partly entrenched.
The men were much exhausted, since they had marched about
seven miles in the noonday heat without regular halts.
They were allowed an hour’s rest, and endeavours were made to
fill their water-bottles, but very little water was obtainable, as the
allowance at Anzac had been reduced to a pint a day per man. Extra
ammunition was given out, and sandbags and entrenching tools
were carried by the men. About two in the afternoon, “B” Company,
who were to keep The Farm on their right hand, led off into the
scrub on the left of the gully. “A” Company followed them, and for
about two hundred yards were able to work along the bed of the
Dere itself, crouching under the high bank to avoid the bullets which
whistled overhead.
Although the main body of the enemy had retired behind the main
crest of the Chunuk Bair, yet they had pushed forward snipers and
machine-guns in sufficient numbers to render the advance of the
two companies a decidedly unpleasant proceeding. A sudden turn in
the direction of the gully brought the commander of “A” Company,
who was at the head of his column, face to face with a long bare
stretch of sand running for three hundred yards straight in the
direction of the Chunuk Bair, which was filled with corpses and with
the equipment that showed where a wounded man had fallen.
Instinctively, he ran forward as the bullets began to throw up the
sand all round him, and was followed by his signallers and observers
and the men of the leading section. For about fifty yards they ran on
until they reached a spot where a cross gully, running down from
Rhododendron Ridge, afforded some protection from the pitiless
machine-gun fire, but in that fifty yards half of the dozen men had
fallen. Accordingly, the subaltern of the leading platoon was sent
back to warn the remainder of the company, not to attempt to use
the Dere, but to work their way through the scrub on its right. He
ran the gauntlet successfully and the advance continued slowly.
Unfortunately, it had been impossible to give the men any definite
objective, as from below The Farm was invisible, and many of them
lost their way in the thick undergrowth, but about a platoon and a
half of each company found its way through the bushes fringing the
Aghyl Dere and commenced the ascent of the Chunuk Bair. Once
they began to climb they were comparatively free from the
attentions of the snipers and machine-guns, since the lower slopes
of the hill were dead ground, but the climb itself was almost
intolerable. The ascent was extremely steep, and covered in scrub,
in which lurked enormous boulders. The sun was still tropical, and
the men, most of whom carried picks or shovels, as well as their
weapons, were heavily laden. Often a man was obliged to lay down
his rifle to haul himself up a rock, and found it an almost intolerable
burden to have to take it up again. It was only by halting and resting
every ten minutes that it was possible to make any progress. The
officers, who did not know that they might not find the whole
position in the hands of the Turks, did their utmost to retain in the
men a sufficient reserve of energy to enable them to charge if it
proved necessary. As The Farm came in sight three hundred yards
ahead, an irregular extension was formed on the hillside, and the
two companies got into touch again. “B” on the left, “A” on the right,
pressed forward to reach their objective. It was unoccupied.
Unoccupied by the Turks, indeed, yet there were many relics of
the struggle that had been waged there at dawn. A narrow ditch
hardly a foot deep showed where an attempt had been made to
entrench the position, while scattered round it were sandbags and
entrenching tools, rifles and bandoliers of ammunition in a confusion
so unnatural that it seemed horrible. Normally, such things are
carefully stored and arranged, and even more carefully accounted
for, and to see them thrown broadcast about a bare hillside was
desolate indeed. Among them lay the men who had used them;
some groaning for water, while others, under the influence of the
scorching sunshine, had already begun to give forth the unspeakably
foul sweet odour of corruption that in those August days tainted half
the hills and valleys of Gallipoli. The sight was depressing enough,
but at least the enemy were not there, and the men would be able
to rest before they had to fight.
 As the senior officer on the position was congratulating himself on
this, a concealed machine-gun opened on the right about two
hundred yards away. The right flank of “A” Company was in full view
of it, and both Captain Massy, who commanded there, and a
subaltern with him were wounded. Captain Massy, however,
remained calm, and after binding up his comrade’s wounds as neatly
as a man with a bullet-hole through his right arm was able to, he
withdrew his men to join the remainder of the company on the left.
These were screened from the direct view of the hostile machine-
gunners by bushes, but the gun was firing at every sound, which
made movement, and still more digging, impossible. Gradually,
however, sandbags were filled, and a traverse made of them, which
protected the men as long as they lay still. A few picked shots were
detailed to fire at intervals into the bushes where the invisible
machine-gun appeared to be, and the knowledge that they were
retaliating encouraged them greatly. Further comfort was given by
the capture of a Turkish sniper, who had been found lurking in the
bushes behind us. None of the men had ever seen a Turk before,
and the general curiosity as to his appearance served to distract the
men’s minds from their immediate prospects.
These, as they presented themselves to the officer who found
himself temporarily in command, were by no means cheerful. The
trench which the men were supposed to hold would require at least
six hours’ work before it would give decent protection from shrapnel.
It was also badly sited and only gave a field of fire of a few yards.
The men available for work on it were few in number and very
weary. There was sufficient food and plenty of ammunition, but
water was very scarce, for those who possessed sufficient self-
control to refrain from drinking during the weary climb, had been
unable to resist the entreaties of the wounded, and had allowed
them to empty their water-bottles.
The only road by which supplies of any kind could be obtained
was the Aghyl Dere, which was swept by the enemy’s fire. In
addition, it was also known that very few reinforcements were
obtainable. Finally, both flanks of the position were “in the air,” the
right being already dominated by a hostile machine-gun, which was
placed so as to enfilade the line. It was clear that if, after dark, the
Turks were to attack, the detachment would be in a hopeless
position, and were bound to be either captured or destroyed.
However, orders had been given that the line was to be held, and
there was nothing to be done but obey them. The men were,
therefore, instructed to rest until darkness made it possible for them
to improve their position, and all ranks lay down and awaited the
enemy’s attack.
Before it developed, however, General Cayley sent orders that the
detachment was to withdraw at dusk, bringing with it all the
wounded who were lying on the face of the hill. Major Money, of “B”
Company, who had now taken over the command, at once detailed a
party under Lieutenant Blake to cover the withdrawal, and as it was
within an hour of sunset, began to collect the wounded at once.
These for the most part belonged to the East Lancashire and
Wiltshire Regiments, with a few of the Royal Irish Rifles. They had
lain out from dawn to dusk under the burning rays of the
Mediterranean sun without food, water, or attention, and suffering
agonies.
By the time they had been collected, the sun was setting, and the
pilgrimage of pain began. There were no stretchers, nor were even
waterproof sheets available, so that each wounded man had to be
carried by his shoulders and legs. The mountain was pathless, and in
the growing darkness the bearers made many a false step, which
must have caused torture to the sufferers. Some shrieked with pain,
others showered blessings on the heads of the men who were
saving them from an agonizing death by thirst, and in the growing
dusk, the load of misery was slowly carried to the foot of the hill. To
the credit of the Turkish machine-gunners it must be said that they
made no attempt to fire as soon as they perceived that wounded
were being removed.
On this, as on other occasions in Gallipoli, we were glad to be able
to respect the chivalry of our foes. An attempt was made to bring
down some of the rifles and equipment that were scattered over the
face of the Chunuk Bair, but there were hardly enough men to carry
them, and some had to be abandoned. It was after 7 p.m. before
the covering party withdrew, being the last British troops to occupy
the Chunuk Bair. Among them was Captain Massy, who, ignoring his
wound, had insisted on remaining till all the wounded had been
removed. For his gallantry on this occasion he was awarded the
Military Cross.
It was dark before the Aghyl Dere was reached, and the Rangers
were glad to find that the two remaining companies of their battalion
had been employed in entrenching a line on each side of the gully
and making sandbag traverses on each side of it. All the wounded
who had fallen in the earlier fighting had been dressed and
removed. This was a feat requiring extraordinary courage and
endurance on the part of the battalion stretcher-bearers. They had
been obliged to go into the exposed section of the Agyhl Dere under
a storm of bullets, in order to bring out the wounded, and yet they
not only did so, but often dressed the man’s wounds under fire
before they removed him. Then after the Medical Officer had treated
him they had to bear their heavy burden all the way to the beach,
returning only to plunge into the fire-swept zone again and rescue
another comrade.
There were no men in the force who did their duty more
strenuously and fearlessly than the stretcher-bearers of the 5th
Connaught Rangers on the 10th of August, 1915, and officers who
had grumbled at having to allot some of their best and strongest
men for non-combatant duties realized how well it was that they had
done so. Nor must the part played by the medical officer be
forgotten. Lieutenant J. I. O’Sullivan, Royal Army Medical Corps,
found himself confronted by the débris of two brigades, but he rose
to the occasion magnificently. Unpacking his paniers under a bush
just behind the line, he not only worked on till long after dark
without a rest, but remained cheerful and encouraging through it all.
Only those who passed through his hands know what they owe to
him.
 So at sunset on August 10th ended the Battle of Sari Bair, which
had begun on the night of the 6th. It had been hard fighting, and
Mr. Ashmead Bartlett, the newspaper correspondent, has described it
as the hardest battle in which British soldiers have been engaged
since Inkerman. Those who took part in it, however, prefer to think
of General Godley’s restrained but deeply significant testimony:—

“I do not believe that any troops in the world could have

accomplished more. All ranks vied with one another in the
performance of gallant deeds, and more than worthily upheld the
best traditions of the British Army.”

Note.—Since this chapter was written, Brigadier-General Cooper has been awarded
a C.B., and Colonel Bewsher of the Hampshires, and Major Wilford, Indian
Army (attached Royal Irish Rifles) have received the D.S.O. for their services
in this action.
 CHAPTER V

S U V L A B A Y A N D C H O C O L AT E H I L L

“Death is nothing; but to live vanquished and without

glory is to die every day.”—Napoleon.

If you sail up the western coast of the Gallipoli Peninsula, soon after
passing Anzac Cove, you will notice that the hills which have fringed
the shore all the way from Cape Helles begin to run further inland,
and that a gradually widening strip of level ground becomes visible
between the cliffs and the sea. The coast line, too, which has
hitherto pointed north and south, turns in a north-westerly direction,
and thus increases the extent of plain until it culminates at the end
of four miles in a cape known as Nibrunesi Point. Two miles north of
Nibrunesi is another promontory called Suvla Point, and these are
the two extremities of a semi-circular bay, which had no name on
the original maps issued to the army, but which was soon to be well
known as Suvla Bay. It is a name which has brought sorrow to many
homes, and which will be perpetually associated with failure, but
there are many glorious memories associated with it.
There are old and historic regiments that think more proudly of
Maiwand and Chillian-wallah than of victories gained with less stern
fighting; and it may well be, that in the future the four Fusilier
regiments from Ireland and the Royal Irish Regiment will be glad to
remember that their service battalions fought at Suvla. A year later,
at Salonica, when the gates of the Supply Depôt were christened
after great battles of the war, the name of Suvla was thought not
unworthy to be associated with those of Ypres and Verdun. Greater
glory no man could ask for, and none of the few survivors of the
10th Division could pass that gate without a throb of pride.
Suvla was well suited to a landing, since the beach shelved gently
and offered a long slope of sand on which lighters could run ashore.
West of Nibrunesi Point an isolated hill, known as Lala Baba, rose to
a height of a hundred and fifty feet close to the shore, while behind
this was the curious feature known as the Salt Lake. In August, this
was dry and presented a surface of white sticky mud nearly a mile
across gleaming brightly in the sun. North-east of the Salt Lake the
ground rose gently till it culminated in Tekke Tepe, nine hundred feet
high and four miles inland. South of Tekke Tepe and about three
miles east of the Salt Lake, was the village of Anafarta Saghir in a
cultivated valley. South of this again was a lower ridge known as
Scimitar Hill, and then another valley containing the village of Biyuk
Anafarta. South of Biyuk Anafarta the ground rose steeply to form
the main chain of the Sari Bair. Between the two Anafartas and the
Salt Lake was a cultivated plain, studded with little cornfields and
isolated olive trees, but from this plain, nearly two miles inland, rose
two isolated hills, about two hundred feet high, known as Yilghin
Burnu (or Green Hill) and Chocolate Hill.
 The landscape was finally framed by a high crest running inland in
a north-easterly direction from Suvla Point, falling steeply in cliffs to
the Gulf of Saros on the north, but presenting a gentler slope to the
southern plain. This ridge reached a height of 400 feet near the sea
and was there called the Karakol Dagh, while further inland, where it
maintained an average height of 600 feet, it was known as the
Kiretch Tepe Sirt. From its crest could be seen the whole of the plain
enclosed by Tekke Tepe, Sari Bair and Damakjelik Bair, on which the
battle was destined to be fought.
 SUVLA, SHOWING LALA BABA AND THE SALT LAKE
(From a water colour by Captain Drummond Fish, Royal Irish Rifles)

The Commander-in-Chief had planned that the transports

conveying the 11th Division from Imbros were to leave as soon as
night fell on the 6th, and effect their landing under cover of
darkness. The 10th Division, having a longer voyage (Mudros being
60 and Mitylene 120 miles away) were intended to reinforce them on
the following day. It was believed that the Turks would be taken by
surprise, and that little or no resistance was to be anticipated. Three
landing places had been arranged for; one known as Beach A in
Suvla Bay itself, the others, Beach B and Beach C, on the shore
south of Nibrunesi Point. The three Brigades of the 11th Division
landed simultaneously, and met with slight resistance from a Turkish
picket entrenched on Lala Baba. The hill was, however, taken with
the bayonet, and the whole of the beaches made good, while the
11th Manchester Regiment drove the enemy’s outposts on the
Karakol Dagh back on to the Kiretch Tepe Sirt. By the time this much
had been gained, day dawned and the first portion of the 10th
Division began to appear on the scene.
This consisted of the 31st Brigade and the two battalions of Royal
Dublin Fusiliers, which had been waiting at Mitylene, the whole force
being under the command of Brigadier-General F. F. Hill. Early in the
afternoon of the 6th, the battalions had left the transports, on which
they had spent nearly a month, and transferred themselves to
trawlers and channel steamers. At sunset they weighed anchor and
steamed northward, all, except a few on board, being completely
ignorant of their destination. The lights on the shore told them that
they had passed Achi Baba, and as they steamed by Anzac, the
noise of battle at Lone Pine and on Sari Bair reached them from afar.
Just as the pale morning light began to make it possible to
distinguish the difference between sea and land, the ships anchored
off Nibrunesi Point.
In the original plan of operations it was designed that the 11th
Division should form the right wing and the 10th the left of the
advance, and with this scheme in view it had been arranged to land
the 10th on Beach A, inside Suvla Bay. The landing at Beach A
during the night had, however, been considerably delayed owing to
the fact that many of the lighters had run aground in the shallow
waters of the bay, and the Naval Authorities had, therefore, decided
to land General Hill’s force on Beach C below Nibrunesi Point. At the
same time, General Hill was directed to reinforce the 11th Division,
placing himself under the orders of Major-General Hammersley, who
commanded that unit.
The process of disembarkation began about 5.30 a.m., the first
two lighters taking to the shore a company of the 6th Inniskillings
and a company of the 5th Royal Irish Fusiliers, as well as General Hill
and his staff.
It was at once clear to all that the Turks had not been completely
taken by surprise. The scrub which covered the slopes of all the
surrounding hills, combined with the scattered olive groves to make
it impossible to detect the numbers of the enemy, but it was obvious
that they were well supplied with artillery. Their shrapnel was
bursting fiercely over the men of the 11th Division as they moved
forward, and as soon as the lighters reached the beaches, an
effective barrage was at once established there. Even the troops
awaiting disembarkation were under fire, and suffered the painful
experience of having to lie down, closely packed together, and
unable to retaliate. The lighters were obvious and easy targets, and
in one boat alone the 7th Dublins lost an officer and seventeen men.
On the whole, however, the force was lucky, and the casualties on
landing were not heavy. Little could be done to keep down the
hostile artillery fire, since the enemy’s guns were well concealed,
and but few of our batteries had landed. Two mountain-guns on Lala
Baba kept up a constant fire, and the warships co-operated, though
lack of facilities for observation rendered their fire comparatively
ineffective.
General Hill reached the landing place two hundred yards south of
Nibrunesi Point about 6.30 a.m. Leaving orders for units as they
landed to rendezvous on the seaward side of Lala Baba, he went in
search of General Hammersley in order to ascertain his wishes. At
this time the Turkish detachments, which had been watching the
beaches, were retiring slowly across the wooded plain which
stretches between the Salt Lake and Anafarta Saghir, pursued by the
11th Division. This pursuit, however, was considerably impeded by
the fact that two small eminences, each about a hundred-and-sixty
feet high, about half-a-mile from the south-eastern corner of the Salt
Lake, were still in the enemy’s hands. These positions were
afterwards known as Chocolate Hill and Green Hill respectively, the
Turkish name for the range being Yilghin Burnu. As long as the Turks
held these knolls, they were in a position to bring enfilade artillery
fire to bear on the advance across the Anafarta plain; and
accordingly General Hill was directed to co-operate with two
battalions of the 11th Division in their capture. This order had
unfortunately the result of making any future junction with the
portion of the Division under Sir Bryan Mahon’s command
impossible, since that was directed to guard the left flank of the
advance, while General Hill’s force was to move to the extreme right.
Owing to this detachment of a Brigade and a half, the work of the
Staff tended to become more difficult.
By the time that General Hill rejoined his force with these orders,
he found that the 6th Inniskillings and 5th Royal Irish Fusiliers had
reached the rendezvous under Lala Baba. Two companies of the 7th
Dublins under Major Lonsdale, the second-in-command, had also
arrived there, and the remainder of the battalion, followed by the
6th Dublins and 6th Royal Irish Fusiliers, were coming up. The latter
unit had been put ashore some way down the beach, and had had
to march a considerable distance in order to reach Lala Baba.
The process of disembarkation and assembly had naturally taken a
considerable time, and it was not till close on noon that the advance
began. In order to reach the northern shores of the Salt Lake, and
get in touch with the 11th Division, the units of General Hill’s force
had to pass over a narrow neck of land between the Salt Lake and
the sea, on which the hostile artillery had carefully registered. Every
minute it was swept by bursts of shrapnel, and the only way in
which it crossed was by a section at a time rushing over it and
trusting to luck. It was a trying ordeal for young troops engaged in
their first action, but they faced it cheerfully. The 7th Dublins in
particular were much encouraged by the example of their Colonel.
As an old soldier, he knew that there were times when an officer
must be prepared to run what would otherwise appear unnecessary
risks; so while everyone else was dashing swiftly across the neck, or
keeping close under cover, it is recorded that Colonel Downing—a
man of unusual height and girth—stood in the centre of the bullet-
swept zone, quietly twirling his stick. The sight of his fearlessness
must have been an inspiration to his men.
As soon as each battalion had crossed the neck, it formed up on
the low ground north of the Salt Lake, under the slight amount of
cover afforded by a low eminence known as Hill 10. When all had
got across, the advance eastward began. The crossing of the neck
had occupied a good deal of time, and it was close on 3 p.m. For
more than four hours the sun had been directly overhead, a blinding
glare was reflected from the shining surface of the Salt Lake, and
the heat was almost overpowering. Few of the men had slept during
the night, since excitement and the discomfort caused by their
closely-packed quarters on board the fleet sweepers had combined
to keep them awake. Except for a cup of tea about 3 a.m., and a
mouthful hastily swallowed before moving off, they were fasting, and
already many of the more improvident had emptied their water-
bottles. In addition, these young soldiers who had never seen war
before, had been since four in the morning exposed to shrapnel fire,
with but little chance either of taking cover or of retaliating. They
had seen their comrades fall stricken at their sides without the
consolation of knowing that the enemy was suffering to an equal
extent. However, the prospect of action was encouraging, and it was
with confident faces that they turned towards the foe. Their one
desire was to come to close quarters with the enemy on their
immediate front, but he was invisible.
From the low ground across which they were moving little could
be seen but the masses of scrub backed by the semi-circle of hills,
and only broken by the minarets of Anafarta. The three leading
battalions (6th Inniskilling Fusiliers, 5th Royal Irish Fusiliers and 7th
Dublins) crossed the dry bed of the Asmak Dere, and began to turn
southward towards Chocolate Hill. Up to this point the left flank of
the movement had been protected by the troops of the 11th
Division, who were advancing in the direction of Anafarta, but every
yard gained to the southward tended to throw this flank more and
more into the air. Though invisible, the enemy was making his
presence felt. Round white balls of shrapnel were continually forming
overhead, and out of the dense bushes rifle bullets came whizzing
past the men’s heads. Now and then a Turkish sniper was caught,
sometimes festooned in boughs to enable him to escape notice; but
the casualties caused by snipers were not so serious on the first day
as they became later. The heaviest losses were caused by the
artillery, for near the sea the scrub was thinner, and the long lines of
men slowly advancing were plainly visible to the enemy’s observers
on the surrounding hills. Occasionally too, a Taube buzzed overhead,
making its observations with comparative impunity, since except on
the ships, there were no anti-aircraft guns.
Still the men pressed on, driving the Turks through the scrub
before them. It was unpleasant work, particularly for officers, since
little or nothing was known, either of the country or of the strength
of dispositions of the enemy, and at any moment a platoon might
have found itself confronted by a heavy counter-attack launched
from the depths of the scrub, or enfiladed by hidden machine guns.
Also, it proved a good deal harder to keep in touch with other units
than it had in training days at the Curragh or in the Phœnix Park.
The danger of pushing on too fast and finding oneself isolated was
no imaginary one, but was alarmingly illustrated by the disaster
which befell the 1st/5th Norfolks four days later. Nor did the tropical
heat, which wore out and exhausted the men, help to quicken the
movement. All these considerations combined with the pressure
exercised by the enemy on the left flank of the Royal Irish Fusiliers
tended to make the advance slow.
The dispositions of the force for the attack were as follows:—
On the right “A” and “B” Companies of the 6th Royal Inniskilling
Fusiliers were in the firing line, supported by “C” and “D” Companies
of the same unit; and by the 6th Royal Irish Fusiliers who had been
brought up from the reserve. The 5th Royal Irish Fusiliers were on
the left, having “A” and “B” Companies in the firing line and “C” and
“D” in support. Owing to the fact that the left flank was exposed,
this battalion was gradually being compelled to face in a south-
easterly direction, with the result that a gap began to appear
between it and the 6th Inniskillings. This gap was filled by “A”
Company of the 7th Royal Dublin Fusiliers, closely supported by “D”
Company (“The Pals”) of the same unit. The 6th Royal Dublin
Fusiliers, who had been the last to come ashore, were still in
reserve, and the 5th Inniskillings had not yet landed.
 BRIGADIER-GENERAL F. F. HILL, C.B., C.M.G., D.S.O.
COMMANDING 31ST BRIGADE

Steadfastly the Fusiliers went forward, moving on a line parallel to

that which they had taken in the morning, but in the opposite
direction. As they passed the Salt Lake, the Inniskillings, who were
on exposed ground, suffered severely, as many of the men stuck in
the swamp. Land-mines, too, which exploded on contact, were
encountered and caused losses, while the shrapnel burst overhead
unceasingly. Nothing, however, could have been more encouraging
to the men than the demeanour of their leader. Wherever the danger
was greatest General Hill was to be found, calm and collected, trying
to save the men as much as possible. His fearlessness, his complete
disregard of personal danger, set an inspiring example, and officers
and men alike went forward more cheerfully, thanks to the lead
given them by their General.
As the advance continued high explosive shells were mingled with
the shrapnel, and though they did not claim so many victims, they
were infinitely more trying to the strained nerves of the weary men
in the ranks. By 5 p.m. they had come within 300 yards of the hill,
and were under a heavy rifle fire. By this time the men were very
weary. They had had a long voyage of 120 miles under most
uncomfortable conditions, they had been under unceasing artillery
fire for more than twelve hours, they had marched more than five
miles burdened by rifle and ammunition through the noon of a
tropical day, and it was no wonder that they were exhausted.
Chocolate Hill, too, was a formidable proposition: though only a
hundred and sixty feet high, it rose steeply from the plain, and it
was now obvious that it had been carefully prepared as a defensive
position, for its sides were seamed by trenches. Though it was
impossible to ascertain how strongly those trenches were held, yet it
was clearly imperative that the men should have a rest before
making the assault.
While the fleet and the batteries that had now been landed
bombarded the position, the men of General Hill’s force lay down in
their ranks on the sun-baked ground, firing a shot from time to time,
but with abundant leisure to look about them. On their right they
could see the white houses and tiled roofs of Anafarta Saghir, while
to the left they gazed across the shining white surface of the Salt
Lake, past Lala Baba, to the bay crowded with warships and
transports and hurrying launches, and to the calm and splendid peak
of Samothrace. Many of “D” Company (“The Pals”) of the 7th
Dublins were men who had taken degrees at Trinity or the National
University, and they may well have recalled past studies and thrilled
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com