scieee Science in your language
[en] (orig)

Application of Semi-probabilistic Methods to Verification of Series System

Abstract

Non-linear finite element analysis (NLFEA) has become a widely used tool in the reliability verification of reinforced concrete (RC) structures. Reliability assessment of RC structures is a challenging task and simplified approaches to reliability verifications are needed to allow for routine applications. Simplified semi-probabilistic methods such as the partial factor method (PFM) or the Method of Estimation of Coefficient of Variation (ECoV) may yield over- or under-conservative approximations. This contribution investigates such errors related to the applications of these two methods to series systems; the design resistances obtained by the probabilistic approach are considered as a reference level. It appears that PFM provides good approximations as the method is focused on critical members in the cases under consideration. ECoV may overestimate up to 20% in situations where NLFEA based on mean and/ or characteristic values fails to identify a dominating failure mode. The maximum observed error is attributed mainly to the failure in identifying the type of distribution of the system resistance. The presented limited analysis indicates several directions for further research (analysis of other types of structural systems, cases with non-lognormal resistances, effects of a number of system components, and their mutual correlations, and performance of advanced methods for reliability verification). Recommendations on how to identify situations when a significant error can be expected should be provided for practical applications of simplified semi-probabilistic methods.

Read accessible full text

Application of Semi-probabilistic Methods to Verification of Series System

Author: Sýkora, Miroslav
Publisher: Vysoká škola báňská - Technická univerzita Ostrava
Year: 2021
DOI: 10.35181/tces-2021-0018
Source: https://dspace.vsb.cz/bitstreams/4c88fa94-d8b7-413b-b916-0a200dda5676/download
SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 21 | NUMBER: 2 | 2021 | DECEMBER
© 2021 TRANSACTIONS OF VSB - TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 80
APPLICATION OF SEMI-PROBABILISTIC METHODS TO
VERIFICATION OF SERIES SYSTEM
Mi osla SÝKORA, Jana MARKOVÁ, Vi ali NADOLSKI
Depa men o S uc u al Reliabili y, Klokne Ins i u e, Czech Technical Uni e si y in P ague,
Šolíno a 7, P ague, Czech Republic
mi osla .syko [email p o ec ed]; jana.ma ko a@c u .cz; i ali.nadolski@c u .cz
DOI: 10.35181/ ces-2021-0018
Abs ac . Non-linea ini e elemen analysis (NLFEA) has
become a widely used ool in he eliabili y e i ica ion o
ein o ced conc e e (RC) s uc u es. Reliabili y assessmen
o RC s uc u es is a challenging ask and simpli ied
app oaches o eliabili y e i ica ions a e needed o allow
o ou ine applica ions. Simpli ied semi-p obabilis ic
me hods such as he pa ial ac o me hod (PFM) o he
Me hod o Es ima ion o Coe icien o Va ia ion (ECoV)
may yield o e - o unde -conse a i e app oxima ions.
This con ibu ion in es iga es such e o s ela ed o he
applica ions o hese wo me hods o se ies sys ems; he
design esis ances ob ained by he p obabilis ic app oach
a e conside ed as a e e ence le el. I appea s ha PFM
p o ides good app oxima ions as he me hod is ocused on
c i ical membe s in he cases unde conside a ion. ECoV
may o e es ima e up o 20% in si ua ions whe e NLFEA
based on mean and/ o cha ac e is ic alues ails o
iden i y a domina ing ailu e mode. The maximum
obse ed e o is a ibu ed mainly o he ailu e in
iden i ying he ype o dis ibu ion o he sys em esis ance.
The p esen ed limi ed analysis indica es se e al di ec ions
o u he esea ch (analysis o o he ypes o s uc u al
sys ems, cases wi h non-logno mal esis ances, e ec s o a
numbe o sys em componen s, and hei mu ual
co ela ions, and pe o mance o ad anced me hods o
eliabili y e i ica ion). Recommenda ions on how o
iden i y si ua ions when a signi ican e o can be expec ed
should be p o ided o p ac ical applica ions o simpli ied
semi-p obabilis ic me hods.
Keywo ds
NLFEA, ECoV, pa ial ac o me hod, ein o ced
conc e e, eliabili y assessmen , se ies sys ems.
1. In oduc ion
Non-linea ini e elemen analysis (NLFEA) has become a
widely used ool in eliabili y e i ica ion o ein o ced
conc e e (RC) s uc u es as he ad ancing enginee ing
knowledge makes i possible o design and build mo e
complex s uc u al sys ems. Howe e , he eliabili y
assessmen o RC s uc u es is a challenging ask due o
he g owing complexi y o nume ical s uc u al models,
he la ge numbe o andom a iables, and small ailu e
p obabili ies ela ed o Ul ima e Limi S a e e i ica ion.
The use o pa ial ac o me hod can be a p agma ic and
su icien app oach o many s uc u al sys ems. Howe e ,
he global sa e y concep o he ully p obabilis ic
app oach seem o be mo e app op ia e in he cases wi h
s ong nonlinea beha iou , he domina ing ole o ensile
and ac u e mechanical p ope ies, o wi h mul iple ailu e
modes o simila impo ance [1], [2], [3]. The ully
p obabilis ic app oach makes i possible o ealis ically
conside he andomness o inpu pa ame e s such as
ma e ial, geome ical, and load cha ac e is ics. Howe e ,
such an app oach is gene ally ime-consuming and semi-
p obabilis ic app oaches a e commonly applied in
enginee ing p ac ice.
To acili a e he ou ine applica ions o semi-
p obabilis ic me hods (bu also o a ully p obabilis ic
app oach), ope a ional ules a e being inco po a ed in o
he p esen codes o p ac ice; o ins ance:
• D a p EN 1990:2021 o he basis o design
should include gene al p o isions o nonlinea
analysis and a speci ic sec ion on he design
assis ed by nume ical simula ions (such as
ou comes o NLFEA).
• Besides he gene al conc e e-speci ic ules o
NLFEA, p EN 1992:2021 o he design and
assessmen o conc e e s uc u es should include
he in o ma i e Annex F “Nonlinea analyses
p ocedu es”. This annex p o ides ope a ional
guidance o applica ions o NLFEA along wi h he
pa ial ac o me hod (PFM), he global ac o
me hod (e.g., using he Me hod o Es ima ion o
Coe icien o Va ia ion – ECoV) o he
p obabilis ic app oach. P o isions on how o
conside NLFEA- ela ed model unce ain ies will
SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 21 | NUMBER: 2 | 2021 | DECEMBER
© 2021 TRANSACTIONS OF VSB - TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 81
be p o ided.
• d a ib Model Code 2020, p o iding a
backg ound o he conc e e s uc u es- ela ed
p o isions in Eu ocodes, is aimed o gi e de ailed
guidance ega ding he alida ion, quan i ica ion o
modelling unce ain y, and applica ion o a ious
eliabili y e i ica ion me hods including he PFM,
ECoV, and p obabilis ic app oach.
Simpli ied semi-p obabilis ic me hods such as PFM o
ECoV a e commonly de ised o yield adequa e es ima es
in mos p ac ically ele an applica ions (e.g. o design
esis ance o a s uc u al sys em in his s udy). Ine i ably,
o e - o unde -conse a i e app oxima ions migh be
ob ained in some cases whe e he me hods a e oo
simpli ied.
A de ailed s udy o such disc epancies esul s in a wide
ange o design si ua ions o p ac ical ele ance is he
subjec o he esea ch p ojec suppo ed by he Czech
Science Founda ion unde G an 20-01781S; selec ed
esul s a e p esen ed in his con ibu ion.
Fi s insigh s ega ding he pe o mance o PFM and
ECoV ha e been ob ained by:
• Ce enka J. e al. [4] who in es iga ed shea
esis ance o a es ed beam whe e he minimum o
he s i ups and conc e e con ibu ions ( hus a se ies
sys em o wo componen s) we e conside ed. The
s udy indica ed ha possibly signi ican e o s
migh be expec ed o bo h me hods in pa icula
si ua ions. Howe e , all obse a ions we e a he
uzzy due o he domina ing e ec o model
unce ain y.
• Syko a e al. [5] ook he basis om [4] bu adop ed
a mo e ealis ic model o shea esis ance whe e
he s i ups and conc e e con ibu ions (~pa allel
sys em) a e summed up. They obse ed
insigni ican e o s, wi h bo h me hods p o iding
sligh ly conse a i e es ima es.
I seems ha he e i ica ion o se ies sys ems p esen s a
challenge conce ning applica ions o PFM and ECoV. A
se ies sys em is hus analysed he e. In con as o [4], he
ollowing modi ica ions a e adop ed he e:
• Model unce ain y—commonly ea ed sepa a ely,
beyond applica ions o ECoV and possibly also o
PFM—is no conside ed he e o ob ain clea e
insigh s in o he pe o mance o he wo semi-
p obabilis ic me hods.
• The p obabilis ic models o a iables a e adop ed
om he backg ound documen s o EN 1992-1-
1:2004 and p EN 1992-1-1:2021 so as hey well
co espond o he pa ial ac o s o ma e ials.
Consequen ly, o si ua ions wi h a single ailu e
mode domina ing, he PFM and ECoV design
esis ances well ma ch hose based on he
p obabilis ic app oach, and he de iciencies o he
semi-p obabilis ic me hods in he si ua ions wi h
uly sys em beha iou can be well in es iga ed.
The con ibu ion concludes wi h a discussion abou he
limi a ions o he p esen ed analysis and he need o
u he in es iga ions.
2. Semi-p obabilis ic Me hods
Unde Conside a ion
In NLFEA, he eliabili y condi ion is commonly
o mula ed as:
Ed ≤ Rd, (1)
whe e Ed ep esen s he design alue o load e ec and Rd
is he design alue o esis ance. This con ibu ion is
ocused on es ima ing Rd by PFM and ECoV.
Acco ding o PFM, design esis ance may be
calcula ed by NLFEA using he design alues o ma e ial
pa ame e s. In his s udy, he ollowing ela ionship is
applied:
Rd,PFM = Rmod( ck / γC; yk / γS), (2)
whe e Rmod = model esis ance; ck = cha ac e is ic alue o
conc e e comp essi e s eng h; γC = pa ial ac o o
conc e e ( u he in o ma ion is in Sec ion 3); yk =
cha ac e is ic alue o yield s eng h o s eel
ein o cemen ; and γS = 1.15 – pa ial ac o o s eel.
Based on he global ac o me hod, Če enka V. [3,4]
p oposed an app oach sui able o ou ine NLFEA
applica ions – ECoV. The me hod assumes ha he
dis ibu ion o esis ance, R, can be desc ibed by a wo-
pa ame e logno mal dis ibu ion [6], which is desc ibed
by mean Rm and coe icien o a ia ion VR. The
unde lying assump ion o logno mal esis ance is
easonable o esis ances o many s uc u al sec ions,
membe s, and pe haps also o sys ems. Acco ding o
ECoV, he wo pa ame e s o a logno mal dis ibu ion can
be es ima ed as:
Rm ≈ Rmod( cm; ym), (3)
VR,ECoV = ln(Rm / Rk) / 1.65, (4)
whe e cm and ym = mean alues o conc e e comp essi e
s eng h and yield s eng h o s eel ein o cemen ,
espec i ely; and Rk ≈ R
mod( ck; yk) is he es ima e o a
cha ac e is ic alue o esis ance (es ima e o a 5% ac ile
o R). The design esis ance is hen es ima ed as:
Rd,ECoV = Rm exp(-αR β VR,ECoV), (5)
whe e αR = 0.8 is he sensi i i y ac o o a domina ing
esis ance pa ame e ; and β = 3.8 is he a ge eliabili y
index acco ding o EN 1990:2002 o Ul ima e Limi
S a es (bo h o a e e ence pe iod o 50 yea s). In Eq. (4)
and Eq. (5), he e ec o model unce ain y is in en ionally
neglec ed as discussed abo e; see also [4]. No e ha he
adop ed alues o αR and β imply ha he design alue o
sys em esis ance is de e mined as a 1.18‰ ac ile o i s
dis ibu ion.
SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 21 | NUMBER: 2 | 2021 | DECEMBER
© 2021 TRANSACTIONS OF VSB - TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 82
ECoV sepa a es alea o y and epis emic unce ain ies
and e lec s he dis inc na u e o di e en ailu e modes
(yielding o ein o cemen , ailu e o conc e e in
comp ession o ension, geome ical ins abili y) wi h only
wo NLFEA uns o es ima e Rm and Rk. As a widely
accep ed—simple and mos ly su icien ly accu a e—
me hod, ECoV has been in oduced in ib MC 2010, d a
ib MC 2020, and p EN 1992-1-1:2021.
3. Cases unde Conside a ion
Two undamen al cases o wo-componen se ies sys ems
a e conside ed in he ollowing analysis:
1. Case 1 wi h wo ailu e modes ac ing as a se ies
sys em wi h logno mally dis ibu ed componen
esis ances,
2. Case 2 is based on Case 1, bu assuming
no mally dis ibu ed componen esis ances.
Design alues ob ained by PFM and ECoV a e no malised
o hose ob ained by he p obabilis ic app oach, Rd,p ob. Fo
ins ance, when Rd,semi-p ob / Rd,p obab > 1, a semi-p obabilis ic
me hod o e es ima es design esis ance, hus being on he
unsa e side.
As discussed abo e, model unce ain y is no
conside ed in he nume ical analysis as i is ypically
ea ed sepa a ely, beyond he applica ion o ECoV. No e
ha he jus i ica ion o he alues o γC and γS acco ding
o [7] indica es ha he model unce ain y ac o s ela ed
o he ecommended alues in EN 1992-1-1:2004 a e e y
close o uni y and hus he alues o he pa ial ac o s a e
adop ed wi hou any adjus men wi h espec o model
unce ain y (in en ionally igno ed in he ollowing
analysis).
4. Se ies Sys em wi h Logno mal
Componen Resis ances (Case 1)
I is o en a gued ha simpli ied semi-p obabilis ic
me hods may ail in cases wi h se e al local ex ema ha
a e ypically caused by mul iple ailu e modes. To e i y
his, Case 1 is ocused on a simple se ies sys em ha can
well ep esen o ins ance shea esis ance o a conc e e
membe ha is de e mined as he minimum o he conc e e
and s i ups con ibu ions, Vc and Vs espec i ely:
R = min(Vc, ρ Vs), (6)
whe e ρ is a de e minis ic s udy pa ame e a bi a ily
a ied dis ega ding p ac ical cons ain s such as bounds on
ein o cemen a ios o de ailing ules. In enginee ing
applica ions, i can ep esen a ein o cemen a io, ρ =
As / Ac (wi h deno ing an a ea o s eel ein o cemen o o
sec ion). Rela ionship (6) can hen be e-w i en as:
R / Ac = min( c’, ρ s’). (7)
He ea e , sys em esis ance based on (7) is e e ed o as R
wi hou indica ing he no malisa ion wi h espec o Ac o
simpli y he no a ion.
Unce ain ies in he wo con ibu ions a e desc ibed by
coe icien s o a ia ion o c’and s’ ha accoun o
a iabili y o he espec i e s eng hs and geome ical
a iables. Mu ually s a is ically independen con ibu ions
a e desc ibed by logno mal dis ibu ions wi h he
ollowing cha ac e is ics:
• cm’ = 29.1 1MPa, V c’ = 21.3%, and c0.05’ = ck’ =
20.5 MPa,
• ym’ = 489 MPa, V y’ = 10.1%, and y0.05’ = yk’ =
414 MPa.
The coe icien s o a ia ion a e de e mined in such a way
ha he design alue o he conc e e and ein o cemen
s eng hs co espond o he alue de e mined by PFM:
cd’ = cm’ exp(-αR β V c’ ) = ck’ / 1.35, (8)
yd’ = ym’ exp(-αR β V y’ ) = yk’ / 1.15. (9)
The alue o γC is educed as he common alue o 1.5
co e s addi ional ac o o 1.15 o accoun o unce ain y
a ising om conc e e being es ed on pu pose-made
specimens in a lab, a he han in he inished s uc u e [7].
I mus be emphasised ha he adop ed alues o CoVs
a e cha ac e is ic a he o esis ances han o s eng hs;
his is why he symbol “ ’ ” is used in ela ionships om
(7) o (9).
Fig. 1 displays a iabili y o Rd,PFM / Rd,p obab and
Rd,ECoV / Rd,p obab wi h ein o cemen a io. Fo low ρ-
alues, he sys em esis ance is go e ned by he
ein o cemen con ibu ion while he conc e e con ibu ion
becomes mo e impo an wi h inc easing ρ.
I appea s ha PFM p o ides a good app oxima ion in
he case o a se ies sys em. The me hod is ocused on
c i ical membe s and hus is well sui ed o he analysis o
se ies sys ems. No e ha his conclusion is alid only o
well-calib a ed alues o he pa ial ac o s o esis ances
( ailu e modes) unde conside a ion.
Fig. 1: Va iabili y o Rd,PFM / Rd,p obab and Rd,ECoV / Rd,p obab wi h ρ
(Case 1).
R
d,semi-p ob
/ R
d,p obab
ρ
0.95
1
1.05
1.1
3% 4% 5% 6% 7% 8%
1.15
1.2
R
d,PFM
/R
d,p obab
ρ
m
=
cm
/
ym
ρ
k
=
ck
/
yk
R
d,ECoV
/R
d,p obab
ρ
d
=
(
ck
/γ
C
)/ (
yk
/γ
S
)
SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 21 | NUMBER: 2 | 2021 | DECEMBER
© 2021 TRANSACTIONS OF VSB - TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 83
Fig. 2: Va iabili y o s a is ical cha ac e is ics es ima ed by ECoV and
p obabilis ic app oach wi h ρ ( om op o bo om – mean alues,
coe icien s o a ia ion, and coe icien s o skewness).
In con as , ECoV may o e es ima e up o 18% o ρ-
alues anging om ρd o ρm. App oxima ely o ρ > ρd, he
conc e e con ibu ion becomes decisi e o sys em
esis ance while ECoV iden i ies a domina ing s eel
con ibu ion by:
• bo h analyses based on he mean and cha ac e is ic
alues o ρd < ρ < ρk,
• by analysis based on he mean alues o ρk < ρ <
ρm.
To u he in es iga e hese de iciencies, Fig. 2 shows
a iabili y o mean alues, coe icien s o a ia ion, and
coe icien s o skewness o sys em esis ance es ima ed by
ECoV and by he p obabilis ic app oach.
The unsa e e o o 18% obse ed o Rd,ECoV(ρk) is
a ibu ed o he ollowing:
1. Key is ailu e in iden i ying a ype o dis ibu ion o
sys em esis ance – igno ing he bimodal cha ac e o he
dis ibu ion and wi hou ega d o an ac ual coe icien o
skewness (Fig. 2 – posi i e ωECoV leads o a highe design
esis ance in compa ison o nega i e ωp obab). De ailed
analysis indica es ha his aspec leads o an e o o
abou 13%.
2. O he wo less impo an e ec s con ibu ing o he
unsa e e o a e a small o e es ima ion o he mean alue
(3%) and unde es ima ion o coe icien o a ia ion (2%).
In e es ing o no e is ha he e o nea ly anishes o
Rd,ECoV(ρm). Failu e o iden i y a ype o dis ibu ion and
o e es ima ed mean would lead o a signi ican
o e es ima ion, 16% + 10%, bu hey a e ou weighed by a
ma kedly o e es ima ed coe icien o a ia ion
(VR,ECoV(ρm) = 21.3% while VR,p obab(ρm) = 13.4%). Wi h
inc easing ein o cemen a io, he e o s dec ease and
anish as no sys em beha iou occu s wi h he conc e e
con ibu ion en i ely domina ing.
5. No mal Componen Resis ances
(Case 2)
Case 2 in es iga es he same se ies sys em as in Case 1,
bu he conc e e and yield con ibu ions a e now desc ibed
by no mal dis ibu ion wi h he ollowing adjus ed
cha ac e is ics:
• V c’ = 14.4%, and c0.05’ = 22.2 MPa,
• V y’ = 8.1%, and y0.05’ = 394 MPa.
while he mean alues, cm’ and ym’ emain unchanged.
Simila ly as in Equa ions (8) and (9), he coe icien s o
a ia ion a e de e mined o co espond o he pa ial
ac o s:
cd’ = cm’ (1-αR β V c’ ) = ck’ / 1.35, (10)
yd’ = ym’ (1-αR β V y’ ) = yk’ / 1.15. (11)
Fig. 3 displays a iabili y o Rd,PFM / Rd,p obab and
Rd,ECoV / Rd,p obab wi h ein o cemen a io. Va ia ion o he
a ios is simila o ha obse ed in Case 1. The ECoV e o
eaches up o 20% o Rd,ECoV(ρk) and con e ging o 8%
e en o high ein o cemen a ios. Despi e he la e is he
case wi h one domina ing componen (conc e e), ECoV
assumes a logno mal dis ibu ion which is inadequa e o
μ
R
(MPa)
ρ
15
18
21
24
3% 4% 5% 6% 7% 8%
27
30
ρ
d
ρ
m
ρ
k
ECoV
p obabilis ic
V
R
0%
5%
10%
15%
3% 4% 5% 6% 7% 8%
20%
25%
ρ
ρ
d
ρ
m
ρ
k
ECoV
p obabilis ic
ω
R
-0.5
-0.25
0
0.25
3% 4% 5% 6% 7% 8%
0.5
0.75
ρ
ECoV
p obabilis ic
ρ
d
ρ
m
ρ
k
SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 21 | NUMBER: 2 | 2021 | DECEMBER
© 2021 TRANSACTIONS OF VSB - TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 84
no mally dis ibu ed conc e e s eng h. Fo low a ios, his
e o is also p esen , bu i is o small magni ude due o a
low coe icien o a ia ion o esis ance ela ed o s eel
yielding. The unsa e app oxima ion by ECoV is now
p ima ily a ibu ed o ailu e in iden i ying a ype o
dis ibu ion o sys em esis ance. No ak L. and
No ak D. [8] adjus ed ECoV o he si ua ion when a
no mal dis ibu ion is assumed o esis ances. Fig. 3
shows ha his adjus men co ec s he ECoV es ima es o
low and high ρ- alues whe e esis ances o indi idual
componen s a e domina ing. Howe e as in Sec ion 4, he
unsa e e o o abou 20% is ound o a ios close o ρk.
Fig. 3: Va iabili y o Rd,PFM / Rd,p obab and Rd,ECoV / Rd,p obab wi h ρ
(Case 2).
6. Discussion
The p esen ed limi ed analysis o he se ies sys ems wi h
wo ailu e modes indica es some di ec ions o u he
esea ch:
1. Mos conc e e s uc u al sys ems a e deemed o ha e
p ope ies close o pa allel sys ems as hey a e o en
inde e mina e, p o iding o mul iple load pa hs.
P elimina y esul s o pa allel sys ems, pa ly p esen ed
in [5], indica e ha ECoV pe o ms e y well while PFM
ends o be a he conse a i e o pa allel sys ems when
αR = 0.8 is conside ed o bo h conc e e and s eel.
2. Posi i e co ela ions be ween ailu e modes a e
expec ed o educe he ECoV e o o bo h ypes o
sys ems. In con as , he e o s may ampli y wi h an
inc easing numbe o ailu e modes o simila impo ance.
These coun e ac ing e ec s should be nume ically
in es iga ed in he u u e.
3. The e o in cases wi h non-logno mal sys em
esis ance may be signi ican . In his con ex , si ua ions
wi h impo an a iabili y o geome ical p ope ies
(pa icula ly o small-size membe s) o wi h non-
logno mal ma e ial p ope ies (possibly o UHPC) should
be analyzed.
4. La in Hype cube Sampling (LHS) wi h a low numbe
o simula ions—also included in he d a ib MC 2020—
is o en applied o e i y he pe o mance o simpli ied
semi-p obabilis ic me hods. S a is ical unce ain y in LHS
es ima es o cases wi h mul iple ailu e modes emains o
be in es iga ed and quan i ied.
5. Fig. 2 clea ly shows ha ECoV migh conside ably
o e es ima e coe icien o a ia ion o sys em esis ance.
The ad anced a iance educ ion app oaches such as
Taylo se ies expansion o Eigen ECoV [8] may p o ide
signi ican imp o emen s, bu equi e mo e complex
in o ma ion abou s ochas ic models and ailu e egions.
Bene i s and cos s ela ed o he applica ions o hese
ad anced me hods should be u he explo ed.
7. Concluding Rema ks
P e ious pilo in es iga ions e ealed ha he Me hod o
Es ima ion o Coe icien o Va ia ion (ECoV) migh
o e es ima e he design esis ance o some se ies sys ems.
A de ailed analysis o wo s udy cases using ECoV and he
pa ial ac o me hod (PFM) p esen ed he e demons a es
ha o he se ies sys ems unde conside a ion:
• PFM p o ides good app oxima ions. The me hod is
ocused on c i ical membe s and hus is well sui ed
o he analysis o se ies sys ems. Howe e , his
conclusion is alid only o well-calib a ed alues
o he pa ial ac o s o esis ances ( ailu e modes)
unde conside a ion.
• ECoV may o e es ima e up o abou 20% in
si ua ions whe e NLFEA based on mean and/ o
cha ac e is ic alues ail o iden i y a domina ing
ailu e mode.
• The maximum obse ed e o is a ibu ed mainly
o ailu e in iden i ying a ype o dis ibu ion o
sys em esis ance; he o he wo, less impo an a e
a small o e es ima ion o he mean alue and
unde es ima ion o coe icien o a ia ion.
The p esen ed limi ed analysis indica es a numbe o
di ec ions o u he esea ch, including he analysis o :
• pa allel and mixed se ies-pa allel sys ems,
• cases wi h s ongly non-logno mal esis ance,
• e ec s o many componen s o he sys em and
co ela ions be ween componen esis ances,
• pe o mance o ad anced me hods o eliabili y
e i ica ion o RC s uc u es.
In gene al, ecommenda ions on how o iden i y si ua ions
when a signi ican e o can be expec ed should be
p o ided o p ac ical applica ions o simpli ied semi-
p obabilis ic me hods.
R
d,semi-p ob
/ R
d,p obab
ρ
0.95
1
1.05
1.1
3% 4% 5% 6% 7% 8%
1.15
1.2
R
d,PFM
/R
d,p obab
R
d,ECoV
/R
d,p obab
R
d,ECoV
/R
d,p obab
adjus ed o
no mal dis .
ρ
m
ρ
k
ρ
d

SECTION BUILDING STRUCTURES & STRUCTURAL MECHANICS VOLUME: 21 | NUMBER: 2 | 2021 | DECEMBER
© 2021 TRANSACTIONS OF VSB - TECHNICAL UNIVERSITY OF OSTRAVA CIVIL ENGINEERING SERIES 85
Acknowledgemen s
This s udy has been suppo ed by he Czech Science Founda ion
unde G an 20-01781S.
Re e ences
[1] ALLAIX, D.L., V.I. CARBONE and G. MANCINI.
Global sa e y o ma o non-linea analysis o
ein o ced conc e e s uc u es. S uc . Conc . 2013,
Vol. 14, N . 1, pp. 29-42. ISSN 1751-7648. DOI:
10.1002/suco.201200017.
[2] CERVENKA, V. Global Sa e y Fo ma o Nonlinea
Calcula ion o Rein o ced Conc e e. Be on- und
S ahlbe onbau. 2008, Vol. 103, N . 2008, pp. 37-42.
DOI: 10.1002/bes .200810117.
[3] CERVENKA, V. Reliabili y-based non-linea
analysis acco ding o ib Model Code 2010. S uc .
Conc . 2013, Vol. 14, N . 1, pp. 19-28. ISSN 1751-
7648. DOI: 10.1002/suco.201200022.
[4] CERVENKA, J., V. CERVENKA, M. SYKORA and
J. MLCOCH. E alua ion o Sa e y Fo ma s o
S uc u al Assessmen Based on Nonlinea Analysis.
In P oc. EURO-C 2018. London: CRC P ess, 2018,
pp. 669-678. ISBN 9781351726764.
[5] SYKORA, M., J. CERVENKA, V. CERVENKA, J.
MLCOCH, D. NOVAK and L. NOVAK. Pilo
Compa ison o Sa e y Fo ma s o Reliabili y
Assessmen o RC S uc u es. In P oc. ib Symp.
2019. Lausanne: In e na ional Fede a ion o
S uc u al Conc e e, 2019, ISBN 9782940643004.
[6] HOLICKY, M. In oduc ion o P obabili y and
S a is ics o Enginee s. Be lin: Sp inge -Ve lag,
2013. 181 pp. ISBN 978-3-642-38299-4. DOI:
10.1007/978-3-642-38300-7.
[7] EUROPEAN CONCRETE PLATFORM.
Eu ocode 2 Commen a y. 2008.
[8] NOVAK, L. and D. NOVAK. Es ima ion o
coe icien o a ia ion o s uc u al analysis: The
co ela ion in e al app oach. S uc .Sa . 2021, Vol.
92, pp. 102101. ISSN 0167-4730. DOI:
h ps://doi.o g/10.1016/j.s usa e.2021.102101.
Abou Au ho s
Mi osla SÝKORA was bo n in České Budějo ice,
Czech Republic. He was appoin ed Associa e P o esso a
Facul y o Ci il Enginee ing, CTU in P ague in 2015. His
esea ch in e es s include he basis o s uc u al design and
o assessmen o exis ing s uc u es, s uc u al eliabili y,
p obabilis ic op imisa ion, load modelling, isk assessmen
o echnical sys ems, and applica ions o p obabilis ic
me hods in s uc u al design.
Jana MARKOVÁ was bo n in P ague, Czech Republic.
She was appoin ed Associa e P o esso a Facul y o Ci il
Enginee ing, CTU in P ague in 2007. He esea ch
in e es s include basis o s uc u al design and assessmen
o exis ing s uc u es, s uc u al eliabili y, load modelling
wi h special ocus on clima ic ac ion e ec s, in es iga ions
o clima e change e ec s on s uc u al eliabili y, and
applica ions o p obabilis ic me hods in s uc u al design.
Vi ali NADOLSKI was bo n in Minsk, Bela us. He
ecei ed his Ph.D. deg ee om he B es S a e Technical
Uni e si y in 2015. His esea ch in e es s include
eliabili y analyses, FEA-based design, and esis ance
models o s eel s uc u es.