CC BY
76
ОБЩИЕ ПРОБЛЕМЫ СТРОИТЕЛЬНОЙ НАУКИ И ПРОИЗВОДСТВА. УНИФИКАЦИЯ И СТАНДАРТИЗАЦИЯ В СТРОИТЕЛЬСТВЕ
УДК 693
V.V. Nadol'skiy, M. Holicky*, M. Sykora*
Belarusian National Technical University
*Czech Technical University in Prague, Klokner Institute
COMPARISON OF RELIABILITY LEVELS PROVIDED BY THE EUROCODES AND STANDARDS OF THE REPUBLIC
OF BELARUS
Comparison of reliability levels of steel structures designed according to the Eurocodes and to the standards of the Republic of Belarus is provided. The main differences between the basic principles of both standards (such as load combinations, the system of partial factors) with a particular focus on design of steel structures are demonstrated. The main parameters characterizing load effects and resistances are compared on the general level. Probabilistic models of basic variables are adjusted to relevant conditions of the Republic of Belarus. In the numerical example, reliability of steel elements is analysed for different combinations of permanent and variable actions. It appears that the standards of the Republic of Belarus assure a lower reliability level than the Eurocodes (reliability indices ranging between 2.0 and 3.5). The main reason for this difference is attributed to the specification of design values of permanent and variable loads. As for both systems of standards under consideration, the reliability of structures exposed to the snow load is significantly lower than the reliability of structures exposed to other types of the load; therefore, further harmonization is required. Further studies concerning more complicated structural elements made of various steel grades are needed.
Key words: Eurocodes, partial factor, probabilistic analysis, reliability level, structural reliability, variable action.
1. Introduction
In January 2010 European standards (Eurocodes) for the design of building structures were introduced in the Republic of Belarus. Since that time both Eurocodes (EN) and national standards of the Republic of Belarus can be applied for structural design. In the following text national standards of the Republic of Belarus for the design of steel structures are referred to as SNiP standards.
Presented comparison of design rules and associated reliability levels provided in EN and SNiP standards is motivated by a foreseen implementation of Eurocodes in the Republic of Belarus. It is expected that the feedback obtained from this study and experience from practical applications of Eurocodes will provide background materials for development of National annexes, maintenance and also for a future improvement of Eurocodes.
Eurocodes recognise the responsibility of regulatory authorities in each country and guarantee their right to determine values related to regulatory safety matters at national level. The national decision concerning alternative values of reliability elements and other Nationally Determined Parameters are a matter of national safety and economic conditions. The submitted study is focussed on comparison of the reli-
ability levels provided by EN and SNiP standards. It should provide valuable background information for decisions concerning Nationally Determined Parameters in Eurocodes to be applied in Belarus. The comparison is based on a probabilistic approach consisting of the following steps:
1. Identification of main differences between basic principles of the both standards (such as load combinations, system of partial factors) with a particular focus on design of steel structures.
2. Comparison of the main parameters characterizing load effects and resistances.
3. Reliability analysis of common steel members using conventional probabilistic models.
2. Partial factor method for steel structures
Both EN and SNiP standards are based on the concept of limit states in conjunction with the partial factor method. Consequently, general principles of reliability verification of steel members given in TKP EN 1993-1-1 [1] and SNiP II-23 [2] are similar (Table 1).
Table 1 — Reliability conditions for structural members
Resistance of cross-sections Buckling resistance of members
TKP EN 1993-1-1 [1] Ed < z f /yMn d — J y ' M0 Ed < z X fy /Ym1
SNiP II-23 [2], SNiP 2.01.07 [3] Y F/z < Y R =Y R / Y ' n — ' c y ' c yn ' m Yn f/(9 z) < Yc Ry
Notation is adopted from [1, 2], the partial factor yn is taken from [3].
In Table 1 Ed and F denote the design value of the action effect, see Section 2.2; z — geometric characteristics of a cross-section of the member (area, section modulus etc.); f — the characteristic value of the yield strength; yM0 — the partial factor for resistance of a cross-section (for strength verifications); yM1 — the partial factor for resistance of the member (for stability verifications); x and 9 — the reduction factor for relevant buckling mode; Ry — the design value of the yield strength; Ryn — the normative value of the yield strength; yc — the partial factor for working condition; yn — the partial factor for importance of a structure.
2.1. Comparison of the parameters characterizing resistances
There are some differences in the definition of geometric characteristics associated with the plastic verification and local buckling. In TKP EN 1993-1-1 [1] a direct calculation of plastic section modulus is applied while in SNiP II-23 [2] plastic deformation is captured by the coefficients c, c, c Other differences comprise also approach to local buckling.
Additional differences are observed in the definition of x and 9 factors according to TKP EN 1993-1-1 [1] and SNiP II-23 [2]. In [1] x is defined on the basis of different initial imperfections depending on the type of cross-section and plane buckling. Comparison of these parameters requires a separate study and the stability verification is thus not considered here.
According to TKP EN 1990 [4] the characteristic value of the yield strength f of the steel is defined as a 5 % fractile, which corresponds to the definition of the 'normative values' of the yield strength Ryn according to SNiP II-23 [2]. According
to [2] design value of the yield strength Ry is determined by dividing the normative value Ryn by the partial factor for the material property ym. The factor ym is given in [2] for different standards for steel production; in this study ym = 1.025 is accepted in accordance with the General Specifications GOST 27772 [5].
The partial factors yM0 and yM1 apply to the resulting resistance for strength and stability verifications, respectively. These factors take into account the uncertainty of basic variables included in the model of resistance (such as unfavourable deviation of material properties from its characteristic value), the uncertainty of resistance model and possibly other effects. According to the National Annex to TKP EN 19931-1 [1] Ym0 = Ym /Yc and Ymi = Ym /YC is taken.
Working condition factor gc accounts for different condition (simplification of design models, local increase of the strength properties of steel, values of the initial curvature of the elements [6]), see Table 6 in [2]. For simplicity yc = 1 is assumed here.
In addition the following assumptions are made for the presented reliability analysis:
1. Elastic behaviour of structural members is considered only.
2. The analysed members and their cross sections are not susceptible to buckling.
2.2. Combination of actions
According to [3, 4] design value of an action for the most unfavourable combination of loads is to be determined. The load combination rules in TKP EN 1990 [4] are based on Turkstra's rule [7]; i.e. the leading variable action is described by its maximum value while the other variable actions are approximated by combination values. In load combinations given in SNiP 2.01.07—85 [3] combination factors are applied to all variable actions and identification of the leading variable action is not needed. In practical applications this can be advantageous since several load combinations may need to be considered when using TKP EN 1990 [4] and when the leading action cannot be easily determined (e.g. columns of multi-storey buildings, combinations of axial and shear forces and bending moments).
In the following analysis the combination of three actions is considered: permanent action G, leading variable action Qx (imposed or, alternatively, snow load) and accompanying variable action Q2 (wind load).
Load combination rules according to TKP EN 1990
In [4] design procedure for the fundamental load combination in permanent design situations introduces two alternative load combination rules, denoted here as A and B. Assuming linear behaviour of structural members when variables G, Qx and Q2 and their characteristic values Gk, Qk1 and Qk2 denote load effects.
Combination scheme A
Considering the formula (6.10) in [4], the design value of action effect Ed is given in terms of the partial factors yg, Yq 1 and Yq 2 for the permanent and variable actions and the combination factor as:
Ed = YGGk + Yq,1Qk,1 + Yq,2 Vq,2 Qk,2 (1)
It is recommended to consider equal factors for both variable actions, Yq1 = Yq2.
Combination scheme B
An alternative procedure is provided by twin expressions (6.10a) and (6.10b) where the combination factor y is also applied to Q in the first expression and a reduction factor is applied to the permanent action in the second expression:
Ed = Yg Gk + Ye,1 Ve,1 Qk,1 + Yq,2 Ve,2 Qk,2 (2)
Ed = 4 Yg Gk + Yq,1 Qk,1 + Yq,2 Vq,2 Qk,2 (3)
The less favourable action effect obtained from (2) and (3) should be considered.
Note that EN 1990 allows further modification of the alternative B by simplification of equation (2) where permanent loads are considered only. However, this scheme is not allowed in TKP EN 1990 [4] and thus it is not considered hereafter. Moreover, this scheme yields lower reliability levels than Combination schemes A and B [8, 9].
Load combination rules according to SNiP 2.01.07—85 — Combination scheme C
According to Clauses 1.4, 1.7 and 1.8 [3] a variable action can be divided into long-term and short-term loads. The parameters defined in [3] are denoted by the symbol «*» in the following text.
When a load combination includes a permanent and more than one variable load, design values of variable load effects shall be multiplied by the following combination factors: ^ = 0.95 for long-term loads and ^ = 0.9 for short-term loads. When a load combination includes a permanent load and one single variable load (long-term or short-term), the combination factors do not apply. The design value of action effect F is given in a similar way as in EN 1990 [4]:
7—l * * . * * * , * * /Л * / * \
F = Yg Gk + Yq,1 Vq,1 Qk,1 + Yq,2 Vq,2 Qk,2 (4)
F = Yg* Gk* + max[Yq,1* Qk,1*; Yq,2* Qk,2*] (5)
For a combination of variable actions the less favourable action effect obtained
from (4) and (5) shall be considered.
2.3. Reliability differentiation
In persistent design situations TKP EN 1990 [4] allows reliability differentiation through modification of partial factors yf. Partial factors for adverse load effects are multiplied by a factor kFI. According to SNiP 2.01.07 [3] the partial factor Yn takes into account possible economic, social and environmental consequences of failure. Numerical values of partial factor Yn and the reliability classification are provided in GOST 27751 [10]. Since definition of reliability classes in TKP EN 1990 [4] and GOST 27751 10] are similar, kFI = 1 (RC2) and Yn = 0.95 (type II) are accepted for a medium reliability class in this study.
2.4. Comparison of the parameters characterizing load effects
Permanent actions
The factors yg and 4 are assumed to be 1.35 and 0.85, respectively as recommended in [4]. According to [3] the partial factor yg* for the permanent actions depends on the component parts of the action (see table 1 [3]); for common structures an averaged value yg* = 12 is used.
The characteristic (normative) values of permanent actions are defined in both considered standards in the same way and thus Gk* = Gk [3, 11].
Snow loads
Partial factor for the snow load recommended in [4] is ys = 15. Combination factor y0S for the snow load according to the National Annex to TKP EN 1991-1-3 [12] and' TKP EN 1990 [4] is assumed to be 0.6.
According to Clause 5.7 [3] with the change No. 1 [13] the partial factor ys* for snow load is also 1.5 for any structural member except for roof members for which YS* becomes: ys* = 15 for Gk* / Sk* > 0.8, ys* = 1.6 for Gk* / Sk* < 0.8.
According to Clause 1.8 [3] the snow load described by a total normative value is considered as a short-term action and thus = 0.9.
Inconsistencies in models for snow loads on roofs in documents [3, 12] include differences in shape, exposure and thermal coefficients. For example, for monopitch roofs (for angle of pitch of roof in the range of 0-30o) the shape coefficient according to [12] is 0.8 and according to [3] is 1. In addition, there are some distinctive features in the appointment of the normative value of snow load on the ground. According to [3, 12] unity exposure and thermal coefficients can be taken for common structures.
In accordance with principles of the present suite of Eurocodes [4, 12] the characteristic value of the snow load on the ground sk is specified as a 98 % fractile of annual extremes (a 50-year return period). According to notes to Table 4 in [3] the characteristic (normative) value of snow load is a mean value of annual extremes obtained on the basis of sufficiently long observations (for a period longer than 10 years).
Given the above, we obtain Sk*=^, where ^ is the mean value of annual extremes. Assuming the Gumbel distribution of annual maxima the characteristic value according to [12] becomes:
Sk = Som = ^ [1 - (0.45 + 0.78 ln(-ln0.98)) VJ (6)
This leads to Sk* = 0.41Sk for VS = 0.55 which is a typical coefficient of variation for annual maxima of the snow loads in the Republic of Belarus [14—16].
Considering differences between snow maps in TKP EN 1991-1-3 [12] and SNiP 2.01.07—85 [3] with change No. 1 [13], the approximation Sk* / Sk ~ 0.66 can be accepted. For further analyses Sk* / Sk = 0.66 * (1 / 0.8) = 0.83 is then considered taking into account the major difference in the shape factor.
Imposed loads
Partial factor for the imposed load is Yq = 15 [4]; combination factor is assumed to be 0.7 according to the table A.1.1 [4] (Category B: office areas).
According to Clauses 1.7 and 1.8 [3] imposed load is divided into long-term and short-term loads. According to Clause 3.6 [3] partial factor Yq* for the imposed load depends on the total normative value of load:
Yq* = 1.3 for the total normative value lower than 2.0 kPa
Yq* = 1.2 for the total normative value equal or higher than 2.0 kPa.
For the total (long-term) and reduced (short-term) normative values the combination factor ^q* is 0.9 and 0.95, respectively. The ratio between reduced and total values for loads on floors of office buildings (Table 3 in Clause 2 of [3]) is
Qk.reduced / Qk °.35.
In accordance with National Annex to EN 1991-1-1 [11] the characteristic value of imposed load is adopted from table 3 of SNiP 2.01.07—85 [3] and thus Qk* = Qk.
Wind action
Partial factor for the wind action is yw = 1.5; combination factor y0W is 0.6 [4]. According to Clause 6.11 [3] the partial factor yw* for wind action is 1.4. According to Clause 1.8 [3] the wind action is considered as a short-term action therefore the load combination factor is yW* = 0.9.
Similarly as for the snow loads there are differences in the assessment of wind actions according to [3, 17]. However, analysis of the [3, 17] showed that the definition of the characteristic (normative) value of wind pressure w0 [3] is similar to the definition of basic velocity pressure qb in [17]. For the purposes of this paper only the
differences between w0 and qb are considered. Effects of pressure, orography, roughness and other coefficients require additional studies.
According to 6.4 [3] characteristic value of wind pressure can be determined by the relationship:
w0 = 0.61 v02 (7)
where v0 — wind velocity at 10 m above ground level for the location of type A (coastal areas, lakes and reservoirs, desert, steppe, forest-steppe and tundra), corresponding to the 10-minute interval averaging and being exceeded on an average every five years (probability of exceedance by annual maxima is then 0.2, see relationship (9) below).
The value of the basic velocity pressure according to expression (4.10) in [17] is defined as:
qb = 0.5 p vb2 (8)
where p — density of the air, the recommended value is 1.25 kg/m3, vb — basic wind velocity calculated from expression (4.1) [17]:
v, =c,. c кп (9)
b dir season b,0 v '
where cdir is the direction factor taking into account the wind direction (the recommended value is 1.0); c — seasonal factor (the recommended value is 1.0);
season
vb0 — the fundamental value of the basic wind velocity - the characteristic 10 minutes mean wind velocity, irrespective of wind direction and time of year, at 10 m above ground level in open country terrain with low vegetation such as grass and isolated obstacles with separations of at least 20 obstacle heights (Clause 4.2(1)P [17]).
The wind actions calculated using [17] are characteristic values (See [4], Clause 4.1.2). They are determined from the basic values of wind velocity or the velocity pressure. In accordance with Clause 4.1.2 (7)P [4] the basic values are characteristic values having annual probabilities of exceedance of 0.02, which is equivalent to a mean return period of 50 years (see relationship (10)). Based on the Gumbel distribution it follows:
SNiP: v0 = ц1{1- [0.45+0.78 ln(-ln(1 - 0.2))] V} (10)
EN: v^o = ЦД1- [0.45+0.78 ln(-ln(1 - 0.02))] V} (11)
where Vv — coefficient of variation of annual maxima of the wind velocity. For the Republic of Belarus it can be assumed Vv = 0.12 [16, 18] and then the ratio Wk* / Wk can be approximated as follows:
Wk w0 _ 0.61v0 _ 0.61
2 A ^ A , /А „г . А ПОЛ / . /1 A A 1-, \2
Wk qb 0.5pvb °.5p
1 - (0.45 + 0.78ln(-ln(l - 0.2))) • 0.12
l- (0.45 + 0.78 ln(- ln(l — 0.02))) • 0.12
0.68 (12)
Considering differences in wind maps in National Annex to [17] and Annex 5 to [3], the ratio Wk* / Wk should be reduced to about 0.65. This value is accepted in the following reliability analysis.
The differences in the definition of characteristic values and partial factors according to the EN and SNiP standards are summarised in Table 2.
Table 2 - Comparison the characteristic values and partial factors according to the EN and SNiP standards
Parameters Partial factors applied in the study for EN for SNiP
Permanent load Gk* / G = 1 Yg = 1.35 ; Ç = 0.85 YG*= 1.2
Imposed load q: / 6k=1 Ye = L5 ; Vo,e =a7 Ye* = 1.3 or 1.2; v = °.9
Snow load S * /S. = 0.83 k k YS = L5 ; Vo.s = °.6 Y* = 1.5 or 1.6; yS* = 0.9
Wind action Wk* / Wk = 0.65 Yw = 1.5 ; Vo.w = a6 Yw' = 1.4; V = 0.9
Yield strength R /f = 1 yn J y Ymo = 1.025 Y = 1; Y = 1.025 ' c 7 ' m
Reliability differentiation — к = 1 Y = 0.95 n
3. Reliability analysis
3.1. Limit state function
Reliability of generic steel members designed using the Combination schemes given above is analysed by probabilistic methods. The limit state function g(X) is written as follows:
g(X) = KR z fy - KE (G + + Q^yeJ (13)
where KR and KE are random variables characterizing the uncertainty in resistance and load effect models, respectively. These variables are used to account for inaccuracy and imprecision of the accepted theoretical models.
The reference period of 50 years is taken into account. In two alternatives imposed (Q50years) or snow (S50yeais) loads are considered as the leading action (Q150years); wind (W1year) is an accompanying load (Q21year).
Turkstra rule [7] is applied to describe combination of time-variant loads. The dominant load is described by 50-year maxima (corresponding to 50-year reference period). For simplification an accompanying wind load is approximated by annual maxima in both alternatives. Note that different models for wind action combined with either imposed or snow load would be used in a more advanced analysis; more details are provided e.g. in [19].
3.2. Probabilistic models of basic variables
Yield strength
The yield strength of steel is described by a two-parameter lognormal distribution. The coefficient of variation is considered by the realistic value 0.08 [20]. The mean of the yield strength is obtained as f = fy exp(1.65F/y) and leads to 1.14-times the characteristic value.
Model uncertainties
The model uncertainties are described by the lognormal distribution [20]. Rolled sections subject to bending about a strong axis and no stability phenomena are hereafter taken into account. The mean 1.15 and the coefficient of variation 0.05 of the model uncertainties in resistance follow from evaluation of a number of tests reported in the background document to Eurocode 3 [21]. The statistical properties of the model uncertainties in load effect are considered in accordance with [20].
2/2013
Permanent actions
Normal distribution is widely accepted for the permanent actions; this is in good agreement with theoretical assumptions and actual experimental data. According to [3, 11] the characteristic value for a permanent load corresponds to a 50 % quan-tile of the distribution, pG=Gk. The coefficient of variation is commonly VG = 0.1.
Snow load
In the National Annex to TKP EN 1991-1-3 [12] Weibull, Gumbel or Frechet distributions are recommended for annual maxima of the snow load on the ground S1year. Previous studies [14, 16] indicate that the Gumbel distribution can be a suitable theoretical model for S1year in Belarus; that is why this distribution is accepted in this study. For available meteorological data [14, 16] 50-year maxima of snow load can be described by the mean value mS=1.04 Sk and by a coefficient of variation VS=0.2. It is noted that the shape coefficient is described in the reliability analysis by its nominal value; but in general it should be described as a random variable [20, 25].
Imposed load
For imposed loads no country-specific differences are assumed and thus the probabilistic model is based on data accepted in other countries. In numerous studies [8,9,14,16] the probabilistic models of imposed loads follow from recommendations of JCSS [20]. It is noted that these models are in good agreement with results of studies conducted in 1980's [22—24].
Wind load
For available meteorological data [16] the parameters of annual wind pressure maxima can be approximated as follows: pW = 0.58 Wk and VW = 0.37.
All the probability models used for the calculations are presented in Table 3.
Table 3 — Probability models of basic variables
Variable symbol Dist. mx / X
Permanent load G Normal 1 0.1
Imposed load (50 years) Q50years Gumbel 0.6 0.35
Snow load (50 years) ^50years Gumbel 1.04 0.20
Wind action (1 year) W, lyear Gumbel 0.58 0.37
Yield strength fy Lognormal 1.14 0.08
Resistance uncertainty Lognormal 1.15 0.05
Load effect uncertainty KE Lognormal 1 0.1
3.3. Reliability measures
The failure probability pf is the basic reliability measure. It can be determined on the basis of a limit state (performance) function g(X) defined in such a way that a structure is considered to be reliable if g(X) > 0 and to fail if g(X) < 0. In Annex C of [4] an alternative measure of reliability is conventionally defined by the reliability index p, which is related to the probability of failure pf = ®(~P); here 0( ) is the cumulative distribution function of the standardised normal distribution.
4. Comparison of reliability indices
To cover a wide range of load combinations load ratios x and k are introduced. The load ratio x denotes the ratio of characteristic variable actions to the total characteristic load given as:
X=(ew+ Qk,2) / (Gk+ ew+ ek,2) (14)
Variable load ratio k = Qk 2/Qk 1 characterizes the relationship between accompanying and leading variable actions.
Two common combinations of variable actions are considered:
In alternative 1 the leading imposed load is combined with an accompanying wind action (floor beams);
In alternative 2 the leading snow load is combined with an accompanying wind action (roof beams).
The reliability analysis is conducted by the FORM method considering a 50-year reference period; the results are shown in Figures 1—3. Variation of the reliability index with the load ratio is shown for alternative 1 with a single imposed load (k = 0) in Figure 1. For Combination scheme C two values of the partial factor Y * are taken (see Section 2.4); equation (4) is not considered.
4.6 4.2 3.8 3.4 3.0 2.6 2.2 1.8
A
B ■
■ - O:
0.2
0.4
0.6
0.8 1 X
в
0
Fig. 1. Variation of the reliability index with the load ratio for alternative 1, k = 0 (single imposed load) and Combination schemes A, B and C
Figure 2 shows variation of the reliability index with the load ratio for alternative 2 with a single snow load (k = 0). For Combination scheme C the partial factor Y* changes at x » 0.6 that corresponds to Gk* / Sk* = 0.8.
Fig. 2. Variation of the reliability index with the load ratio for alternative 2, k = 0 (single snow load) and Combination schemes A, B and C
Figure 3 indicates variation of the reliability index with the load ratio for alternative 1 with к = 0.9 (combination of dominant imposed and accompanying wind loads).
It is noted that in Combination Scheme C Equation 5 is more unfavourable for the variable load ratio к > 0.2 for any combination of imposed, snow and wind loads. Equation 4 thus applies for very low values of к and its practical significance seems to be small.
Fig. 3. Variation of the reliability index with the load ratio for alternative 1, k = 0.9 (combination of imposed and wind actions) and Combination schemes A, B and C
The following observations are made from the results of the reliability analysis:
For all the Combination Schemes a significant variation of reliability level with the load ratio is observed (particularly for the leading snow load).
For design based on both EN and SNiP standards, reliability of structures exposed to the snow load is significantly lower than for structures subjected to the imposed load.
SNiP documents for the design of steel structures yield lower reliability levels in comparison with Eurocodes (difference in the reliability index by about 1).
SNiP load combination schemes lead to similar variation of the reliability index with the load ratio as expression (6.10) in TKP EN 1990 [4] (Combination scheme A).
Twin expressions (6.10a) and (6.10b) (Combination scheme B) yield the most uniform reliability levels.
For x > 0.7 (imposed load) and x > 0.3 (snow load) the reliability index is lower than a common target level of 3.8 and reliability of structural members designed according to Eurocodes seems to be insufficient [4].
Considering load combination rules according to Eurocodes, cases with a single variable action (k = 0) seem to be more critical than when combination of variable actions occurs (k > 0). It follows that combination factors may be somewhat conservative since the same reliability should be generally achieved for structures exposed to a single variable action or combination of variable actions.
For design based on SNiP standards a sudden jump in the reliability index appears for the leading snow load due to the change of the partial factor.
It is emphasized that the presented results are significantly dependent on the assumed models for basic variables including model uncertainties. Moreover, various simplifying assumptions concerning snow and wind loads are accepted in this study. Therefore, the obtained results should be considered as indicative.
Conclusions
The presented comparison of reliability levels provided by Eurocodes and by the standards of the Republic of Belarus for design of steel structures indicates that:
Standards of the Republic of Belarus lead to lower reliability levels than Eurocodes (reliability indices ranging from 2.0-3.5), this can be mainly attributed to differences in specification of permanent (partial factor) and variable loads (characteristic values).
Except for the shift in reliability levels similar variation of the reliability index with load ratios is observed for combination schemes according to Eurocodes and standards of the Republic of Belarus.
Standards of the Republic of Belarus do not provide explicit guidance for target reliability levels; however, they provide a complex system of partial factors enabling adjustments to different conditions of a structure.
For both considered systems of standards, reliability of structures exposed to snow load is significantly lower than for structures subjected to imposed load and further harmonisation of reliability levels is required.
For a more accurate comparison of reliability levels, further studies concerning more complicated structural elements made of various steel grades are needed. Further research should be focused on the optimisation of target reliability levels for conditions of the Republic of Belarus, which would provide recommendations for appropriate reliability levels. Similar studies will be conducted also for new updated versions of SNiP which were issued in the Russian Federation in 2011.
Acknowledgments
This study project has been conducted within the project CZ/11/LLP-LdV/ TOI/134005 Vocational Training in Assessment of Existing Structures, funded with support from the European Commission (programme Leonardo da Vinci). The paper reflects views only of the authors, and the Commission cannot be held responsible for any use which may be made of the information contained therein. Results of the project LG11043 have been utilised.
References
1. TKP EN 1993-1-1:2009. Evrokod 3. Proektirovanie stal'nykh konstruktsiy. Chast' 1-1. Obshchiepravila ipravila dlyazdaniy. [EN 1993-1-1 Eurocode 3: Design of Steel Structures. Part 1-1: General Rules and Rules for Buildings] Minsk, STROYTECHNORM Publ., 2009.
2. SNiPII-23-81. Stal'nye konstruktsii [Construction Norms and Rules II-23—81. Steel Structures]. Moscow, Gosstroy Publ., 1991.
3. SNiP 2.01.07-85. Nagruzki i vozdeystviya [Construction Norms and Rules 2.01.0785. Loads and Actions]. Moscow, Gosstroy Publ., 1999.
4. TKP EN 1990:2011. Evrokod. Osnovy proektirovaniya konstruktsiy [EN 1990 Eurocode: Basis of Structural Design]. Minsk, STROYTECHNORM Publ., 2011.
5. GOST 27772—88. Prokat dlya stroitel'nykh stal'nykh konstruktsiy. Obshchie tekh-nicheskie usloviya [State Standards 27772—88. Rolled Products for Steel Structures. General Specifications].
6. Posobiepoproektirovaniyu stal'nykh konstruktsiy (kSNiPII-23—81* «Stal'nye kon-struktsii») [Handbook of Design of Steel Concrete Structures (based on Construction Norms and Rules II-23—81*. Steel Structures)]. Moscow, TsNIISK im. Kucherenko Publ., 1989, 148 p.
7. Turkstra C.J. Theory of Structural Design Decisions. SM Studies Series no. 2. Ontario, Canada. Solid Mechanics Division, University of Waterloo, 1970.
8. Holicky M. and Retief J.V. Reliability Assessment of Alternative Eurocode and South African Load Combination Schemes for Structural Design. Journal of the South African Institution of Civil Engineering, vol. 47, no. 1, 2005, pp. 15—20.
9. Gulvanessian H. and Holicky M. Eurocodes: Using Reliability Analysis to Combine Action Effects. Proceedings of the Institution of Civil Engineers Structures & Buildings. August 2005, vol. 158, no. SB4, pp. 243—252.
10. GOST27751—88 Nadezhnost' stroitel'nykh konstruktsiy i osnovaniy. Osnovnye polozheniya po raschetu [State Standard 27751—88. Reliability of Structures and Foundation Soils. Principal Provisions for Analysis].
11. STB EN 1991-1-1:2007. Evrokod 1. Vozdeystviya na nesushchie konstruktsii. Chast ' 1-1. Udel'nyy ves, postoyannye i vremennye nagruzki nazdaniya. [EN 1991-1-1 Eurocode 1: Actions on Structures. Part 1-1: General Actions. Densities, Self-weight, Imposed Loads for Buildings]. Minsk, STROYTECHNORM Publ., 2007.
12. TKPEN 1991-1-3:2009. Evrokod 1. Vozdeystviya na konstruktsii. Chast' 1-3. Obsh-chie vozdeystviya. Snegovye nagruzki [EN 1991-1-3 Eurocode 1: Actions on Structures. Part 1-3: General Actions. Snow Loads]. Minsk, STROYTECHNORM Publ., 2009.
13. Izmenenie №1 SNiP 2.01.07—85 «Nagruzki i vozdeystviya» [CHANGES №1 to Construction Norms and Rules 2.01.07—85. Loads and Actions]. Minsk, Ministry of Architecture and Construction of the Republic of Belarus, 2004.
14. Tur V.V., Markovskiy D.M. Kalibrovka znacheniy koeffitsientov sochetaniy dlya vozdeystviy pri raschetakh zhelezobetonnykh konstruktsiy v postoyannykh i osobykh ra-schetnykh situatsiyakh [Calibration of Load Combination Factors Used in Design of Reinforced Concrete Structures in Persistent and Accidental Design Situations]. Stroitel'naya nauka i tekhnika [Construction Science and Machinery]. 2009, № 2(23), pp. 32—48.
15. Tur V.V. Obespechenie nadezhnosti stroitel'nykh konstruktsiy v svete trebovaniy evropeyskikh i natsional'nykh normativnykh dokumentov po proektirovaniyu [Assurance of Reliability of Building Structures in the Context of Requirements of European and National Design Standards]. Perspektivy razvitiya novykh tekhnologiy v stroitel'stve i podgotovke inzhenernykh kadrov: sbornik nauchnykh statey. [Prospects for Development of New Technologies in the Construction Industry and Training of Engineers: Collection of Research Papers]. Grodno, GrGU Publ., 2010, pp. 480—497.
16. Markovskiy D.M. Kalibrovka znacheniyparametrov bezopasnosti zhelezobetonnykh konstruktsiy s uchetom zadannykh pokazateley nadezhnosti [Calibration of Safety Parameters for Reinforced Concrete Structures based on the Target Reliability Indices]. Brest, 2009.
17. TKP EN 1991-1-4:2009. Evrokod 1. Vozdeystviya na konstruktsii. Chast' 1-4. Obsh-chie vozdeystviya. Vetrovye vozdeystviya. [EN 1991-1-4 Eurocode 1: Actions on Structures. Part 1-4: General Actions. Wind Actions/ Minsk, STROYTECHNORM Publ., 2009.
18. Archives of meteorological observations at meteorological stations in Belarus, Ukraine, Russia, Poland and the Baltic States. Available at: http://pogoda.by/zip. Date of access: 20.02.2012.
19. Sykora M., Holicky M. Comparison of Load Combination Models for Probabilistic Calibrations. In Faber M.H., Köhler J., Nishijima K. (eds.). Proceedings of 11th International Conference on Applications of Statistics and Probability in Civil Engineering. ICASP11, 1-4 August, 2011, ETH Zurich, Switzerland. Leiden, the Netherlands, Taylor & Francis/Balke-ma, 2011, pp. 977—985.
20. JCSS Probabilistic Model Code, Zurich. Joint Committee on Structural Safety, 2001. Available at: www.jcss.byg.dtu.dk.
21. Eurocode 3. Editorial Group Background Documentation to Eurocode No. 3 Design of Steel Structures Part 1. General Rules and Rules for Buildings, Background Document for Chapter 5 of Eurocode 3, Document 5.01, 1989.
22. Rayzer V.D. Metody teorii nadezhnosti v zadachakh normirovaniya raschetnykh parametrov stroitel'nykh konstruktsiy [Methods of the Reliability Theory Applicable to Problems of Standardization of Design Parameters of Building Structures]. Moscow, Stroyizdat Publ., 1986, 192 p.
23 Bulychev A.P. Vremennye nagruzki na nesushchie konstruktsii zdaniy torgovli [Temporary Loads on Bearing Structures of Retail Stores]. Stroitel'naya mekhanika i raschet sooruzheniy [Structural Mechanics and Analysis of Structures]. 1989, no. 3, pp. 57—59.
24. Gordeev V.N. Nagruzki i vozdeystviya na zdaniya i sooruzheniya [Loads and Actions on Buildings and Structures]. Gordeev VN., Lantukh-Lyashchenko A.I., Pashinskiy V A., Perel'muter A.V., Pichugin S.F., Perel'muter A.V., editor. Moscow, ASV Publ., 2007, 482 p.
25. Holicky M., Sykora M. Partial Factors for Light-Weight Roofs Exposed to Snow Load. Bris R., Guedes Soares C., Martorell S., editors. Supplement to the Proceedings of the European Safety and Reliability Conference ESREL 2009. Prague, Czech Republic, 7—10 September 2009. Ostrava: VSB Technical University of Ostrava, 2009, p. 23—30.
Поступила в редакцию в январе 2013 г.
About the authors: Nadol'skiy Vitaliy Valer'evich — master of sciences, assistant lecturer, Department of Metal and Timber Structures, Belarusian National Technical University (BNTU), 65 prospekt Nezavisimosti, Minsk, 220013, Republic of Belarus; Nadol-skivv@mail.ru; +375 259 997 991;
Holicky Milan — Doctor of Philosophy, Professor, Deputy Director, Klokner Institute, Czech Technical University in Prague (CTU), Solinova 7, 166 08, Prague 6, Czech Republic, milan.holicky@klok.cvut.cz; +420 2 2435 3842;
Sykora Miroslav — Doctor of Philosophy, researcher, Klokner Institute, Czech Technical University in Prague (CTU), Solinova 7, 166 08, Prague 6, Czech Republic, miroslav. sykora@klok.cvut.cz; +420 2 2435 3850.
For citation: Nadol'skiy V.V., Holicky M., Sykora M. Comparison of Reliability Levels Provided by the Eurocodes and Standards of the Republic of Belarus. Vestnik MGSU [Proceedings of Moscow State University of Civil Engineering]. 2013, no. 2, pp. 7—21.
В.В. Надольский, М. Голицки, М. Сыкора
СРАВНЕНИЕ УРОВНЕЙ НАДЕЖНОСТИ СТАЛЬНЫХ КОНСТРУКЦИЙ, ОБЕСПЕЧИВАЕМЫХ ЕВРОКОДАМИ И СТАНДАРТАМИ РЕСПУБЛИКИ БЕЛАРУСЬ
Выполнено сравнение уровней надежности стальных конструкций, запроектированных в соответствии с Еврокодами и стандартами Республики Беларусь. Показаны основные различия между базовыми принципами обоих систем нормативных документов (например, в сочетаниях нагрузок, системах частичных коэффициентов) с акцентом на проектирование стальных конструкций. Сопоставлены основные параметры, характеризующие сопротивление и воздействия (эффекты от воздействий). Вероятностные модели базисных переменных скорректированы с учетом фактических условий Республики Беларусь. На численных примерах выполнен анализ надежности обобщенного стального элемента для различных комбинаций постоянных и переменных воздействий. Показано, что стандарты Республики Беларусь приводят к меньшему уровню надежности, чем Еврокоды (индекс надежности в диапазоне 2,0...3,5). Основная причина этого различия связана с нормированием расчетных значений постоянных и переменных воздействий. Для рассматриваемых систем стандартов надежность конструкций, подверженных воз-
действию снеговой нагрузки, значительно ниже, чем для конструкций, подверженных действию полезной нагрузки, и поэтому требуется дальнейшее согласование расчетных положений. Необходимы дальнейшие исследования, касающиеся более сложных конструктивных элементов, изготовленных из стали различных марок.
Ключевые слова: Еврокоды, частичный коэффициент, вероятностный анализ, уровень надежности, надежность конструкции, переменные воздействия.
Библиографический список
1. ТКП EN 1993-1-1:2009. Еврокод 3. Проектирование стальных конструкций. Часть 1-1. Общие правила и правила для зданий. Минск : РУП «Стройтехнорм», 2009.
2. СНиП II-23—81*. Стальные конструкции. М. : Госстрой, 1991.
3. СНиП 2.01.07—85. Нагрузки и воздействия. М. : Госстрой, 1999.
4. ТКП EN 1990:2011. Еврокод. Основы проектирования конструкций. Минск : РУП «Стройтехнорм», 2011.
5. ГОСТ 27772—88. Прокат для строительных стальных конструкций. Общие технические условия.
6. Пособие по проектированию стальных конструкций (к СНиП II-23—81*. Стальные конструкции) / ЦНИИСК им. В.А. Кучеренко Госстроя СССР М. : ЦИТП Госстрой СССР, 1989. 148 с.
7. Turkstra C.J. Theory of Structural Design Decisions, SM Studies Series No. 2. Ontario, Canada: Solid Mechanics Division, University of Waterloo. 1970.
8. Holicky M. and Retief J.V. Reliability Assessment of Alternative Eurocode and South African Load Combination Schemes for Structural Design. Journal of the South African Institution of Civil Engineering, vol. 47, no. 1, 2005, pp. 15—20.
9. Gulvanessian H. and Holicky M. Eurocodes: Using Reliability Analysis to Combine Action Effects. Proceedings of the Institution of Civil Engineers Structures & Buildings. August 2005, vol. 158, no. SB4, pp. 243—252.
10. ГОСТ 27751—88. Надежность строительных конструкций и оснований. Основные положения по расчету.
11. СТБ EN 1991-1-1:2007. Еврокод 1. Воздействия на несущие конструкции. Часть 1-1. Удельный вес, постоянные и временные нагрузки на здания. Минск, РУП «Стройтехнорм», 2007.
12. ТКП EN 1991-1-3:2009. Еврокод 1. Воздействия на конструкции. Часть 1-3. Общие воздействия. Снеговые нагрузки. Минск : РУП «Стройтехнорм», 2009.
13. Изменение № 1 СНиП 2.01.07—85. Нагрузки и воздействия / Министерство архитектуры и строительства Республики Беларусь. Минск, 2004.
14. Тур В.В., Марковский Д.М. Калибровка значений коэффициентов сочетаний для воздействий при расчетах железобетонных конструкций в постоянных и особых расчетных ситуациях // Строительная наука и техника. 2009. № 2 (23). С. 32—48.
15. Тур В.В. Обеспечение надежности строительных конструкций в свете требований европейских и национальных нормативных документов по проектированию // Перспективы развития новых технологий в строительстве и подготовке инженерных кадров : сб. научных статей. Гродно : ГрГУ, 2010. С. 480—497.
16. Марковский Д.М. Калибровка значений параметров безопасности железобетонных конструкций с учетом заданных показателей надежности : автореф. дисс. ... канд. техн. наук. Брест, 2009.
17. ТКП EN 1991-1-4:2009. Еврокод 1. Воздействия на конструкции. Часть 1-4. Общие воздействия. Ветровые воздействия. Минск : РУП «Стройтехнорм», 2009.
18. Архивы метеорологических наблюдений по метеостанциям Беларуси, Украины, России, Польши и Прибалтики [Электронный ресурс] / Гидрометцентр Республики Беларусь. Режим доступа: http://pogoda.by/zip. Дата обращения: 20.02.2012.
19. Sykora M, Holicky M. Comparison of load combination models for probabilistic calibrations. In Faber M.H., Köhler J., Nishijima K. (eds.), Proceedings of 11th International Conference on Applications of Statistics and Probability in Civil Engineering ICASP11, 1-4 August, 2011, ETH Zurich, Switzerland. Leiden (The Netherlands): Taylor & Francis/Balkema, 2011. pp. 977—985.
20. JCSS Probabilistic Model Code, Zurich: Joint Committee on Structural Safety, 2001. www.jcss.byg.dtu.dk.
21. Eurocode 3 Editorial Group Background Documentation to Eurocode No. 3 Design of Steel Structures Part 1 - General Rules and Rules for Buildings, Background Document for Chapter 5 of Eurocode 3, Document 5.01, 1989.
22. Райзер В.Д. Методы теории надежности в задачах нормирования расчетных параметров строительных конструкций. М. : Стройиздат, 1986. 192 с.
23. Булычев А.П. Временные нагрузки на несущие конструкции зданий торговли // Строительная механика и расчет сооружений. 1989. № 3. С. 57—59.
24. Нагрузки и воздействия на здания и сооружения / В.Н. Гордеев, А.И. Лантух-Ля-щенко, В.А. Пашинский, А.В. Перельмутер, С.Ф. Пичугин ; под общ. ред. А.В. Перельму-тера. М. : Изд-во АСВ, 2007. 482 с.
25. Holicky M., Sykora M. Partial Factors for Light-Weight Roofs Exposed to Snow Load. In Bris R., Guedes Soares C., Martorell S. (eds.), Supplement to the Proceedings of the European Safety and Reliability Conference ESREL 2009, Prague, Czech Republic, 7—10 September 2009. Ostrava: VSB Technical University of Ostrava, 2009, pp. 23—30.
Об авторах: Надольский Виталий Валерьевич — магистр, ассистент кафедры металлических и деревянных конструкций, Белорусский национальный технический университет, 220013, г. Минск, проспект Независимости, д. 65, Nadolskivv@mail.ru;
Голицки Милан — доктор философии, профессор, заместитель директора Института Клокнера, Чешский технический университет в Праге, Чешская Республика, 166 08 Прага 6, ул. Солинова, д. 7, milan.holicky@klok.cvut.cz;
Сыкора Мирослав — доктор философии, научный работник Института Клокнера, Чешский технический университет в Праге, Чешская Республика, 166 08 Прага 6, ул. Солинова, д. 7, miroslav.sykora@klok.cvut.cz.
Для цитирования: Надольский В.В., Голицки М., Сыкора М. Comparison of reliability levels provided by the eurocodes and standards of the Republic of Belarus // Вестник МГСУ. 2013. № 2. С. 7—21.
