Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 153 (2016) 277 282

XXV Polish Russian Slovak Seminar Theoretical Foundation of Civil Engineering

Girders of hydraulic gates optimal design

Igor S. Kholopova, Mikhail I. Balzannikova, Vadim Yu. Alpatova *, Alexey.V. Solovieva
a
Samara State University of Architecture and Civil Engineering, Molodogvardeyskaya St., 194, Samara, 443001, Russia

Abstract

Metal gates of hydraulic structures are supplied with load bearing girders. Girders, crossbeams and gates require much steel when
being manufactured. Girders of hydraulic gates rational design leads to considerable cost savings. The paper describes steel
hydraulic girders optimal design. Steel hydraulic girders optimal design is based on the strength theory introduced by M. Huber,
R. Mises, H. Hencky and on the theory of mirror functions. With the help of out-form volume we can either calculate the optimal
form for a girder with variable cross-section or the optimal height for a girder of uniform cross-section. The paper introduces
hydraulic girders calculations according to the authors' methodology.
2016
2016TheTheAuthors.
Authors. Published
Published by by Elsevier
Elsevier Ltd.Ltd.
This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the XXV Polish Russian Slovak Seminar Theoretical
Peer-review
Foundationunder responsibility
of Civil of the organizing committee of the XXV Polish Russian Slovak Seminar Theoretical Foundation
Engineering.
of Civil Engineering.
Keywords: Metal gates; Hydraulic structures; Girders; Optimal design; Mirror functions.

Nomenclature

deformation energy per unit of volume

[] deformation energy per unit of volume for direct tension
Poisson's ratio
liquid limit of the material
elastic module
normal stress in i point of the bar
bar capacity
, , internal stress in i point of the bar

* Corresponding author. Tel.:+7-902-320-37-37;

E-mail address: avu75@mail.ru

1877-7058 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the XXV Polish Russian Slovak Seminar Theoretical Foundation of Civil Engineering.
doi:10.1016/j.proeng.2016.08.115
278 Igor S. Kholopov et al. / Procedia Engineering 153 (2016) 277 282

cross-section shape coefficient

distance from the neutral axis to the outer fiber c of the cross-section
this cross-section the radius of gyration at bending
cross-section form non-dimensional coefficient in shear
cross-sectional height of the bar
cross-section area of the bar
bar length

1. Introduction

There are several well-known methods of building structures optimal design [1, 2, 5, 6, 10] that can be applied
while designing building structures cross-beams or girders. In this paper, the authors suggest applying the method of
optimal design to girders of hydraulic gates.
Steel hydraulic girders optimal design is based on the strength theory introduced by M. Huber, R. Mises, H.
Hencky [3]. According to this theory, strength in this case is provided on condition that:

(1 + ) /3 ( 1) (1)

For a plane stress problem, direct stress in i point of the bar can be determined from the formula:

= + (2)

Specific potential energy of the shape change in i point of the bar is

+ + [] (3)

As a result of transformations we can introduce a new equation from the strength condition. This equation is
mathematically determined as non-linear modular form i bar capacity [4-6]:

= , (4)

Shape coefficients i i are cross sections accessible characteristics and can be easily derived from gauge
material or calculated. With given cross-sectional height, cross-section form coefficients and are taken as
constant.
It makes possible to divide our optimal design problem into two parts: 1) the problem of optimum size defining
when h=const; 2) the problem of optimum cross-sectional height.

2. Bar optimal design with the use of out-form

For a statically indeterminate bar with uniform cross-section the volume function (determined according to the
strength condition in i point and with regard to Formulae (3) and (4)) can be presented as it follows:

Vi = li
2

0,433 ( N ip + N ij x j ) + 2 ( N ip + N ij x j )( M ip + M ij x j ) i +
(5)

i [U ] E + ( M ip + M ij x j ) 2 i2 + 3 (Qip + Qij x j ) 2 i2

In general terms, Formula 5 can be given as

 Igor S. Kholopov et al. / Procedia Engineering 153 (2016) 277 282 279

n n n
Vi = Fi li = Aijk x j xk + 2 Bij x j + Ci (6)
j =1 k =1 j =1

Formula 6 is a non-linear modular form [6, 8] as the quadric form under the radical sign is in itself a quadratic
sum and is absolutely positive. According to [9] quadric form radicals under the radical in the area V I can be only
imaginary with the minimum V i min >0, and actual if V i min =0. Figures 1,2 present a diagram of a modular form of
statically indeterminate system with either one or two indeterminates.

Fig. 1. Modular form diagram Fig. 2. Surface of modular form volume

Let us present a modular form for each point and every cross-section as well as a corresponding hypersurface.
When one-dimensional (Fig. 1, 2) they are a line of crosscut hyperboles. There is a multitude of volume points
(1, 2, 3)V ik inside every dome-shaped hypersurface which give strength in k point. For p points at X j =const4 the
strength is given by intersection of the sets 1 2 p =. An explicit equation for a volume value choice can
be put down as a mathematical unit firstly introduced by Yu. Radtsig [10, 17] and known as volume out-forms. The
term "out-form" means that the largest surface area is chosen out of the multitude (as Figure 3 shows) with the
vector certain value:

Fig. 3. Modular form out-form volume diagram

Out-form goes as follows:

V = V1 || V2 || || Vp (7)

A double-bar here means: or V 1 , or V 2 , or V 3 , or V (the greatest of them). A double-bar may go horizontally. As

far as the out-form as an envelope of convex sets intersection, it is also dome-shaped. This characteristic is very
important when we chose a proper optimization method. Figure 3 shows that if we need to choose the least volume
we have to firstly find the minimum out of maximum rating defined by the out-form, that is V opt =minmax (minimax
280 Igor S. Kholopov et al. / Procedia Engineering 153 (2016) 277 282

decision). While sorting parameter values X and calculating the correspondent value of the out-form one can find the
optimal point.
If the cross-section is variable along its length (piece-wise variable) it is possible to find an out-form for every
part of the uniform cross-section (8) in case its height is also permanent. If the bar is statically determinate then
every k modular form value is static and we need to choose maximum value V k to find an out-form as there is no
min choice among max in X j spacing.
If this statically determinate bar is variable in its cross-section along its length, we can find the optimal shape or
height for a uniform cross-section bar while using the form coefficient and . For that we need to find the
dependence and from the cross-sectional height h [8]. If we take I-beams (according to All Union State
Standard - GOST - 8239-89) and compare all the coefficients and for outer fibers, fibers on flange and wall
junction and neutral fibers we'll see that there is an approximate hyperbolic dependence between 3 and and linear
dependence between h and (as Fig. 6 shows).
Paper [7] introduces statistical research proving that the radius of cross-section gyration of a defined height h-
const in flexural and compressed flexural elements does not depend on their other characteristics if their
construction and calculation constraints are taken into account. The radius of gyration changes only if cross-
sectional height is changed. We can see almost the same dependence in respect of the width of (flange) section b.

Fig. 4. Dependency diagrams of cross-section form coefficients for I-beams

(according to All Union State Standard (GOST) 8239-89) x , y , x , y from cross-sectional height h

Cross-section form coefficients for rolled sections, fabricated sections of simple and complex types, U-sections,
etc can be approximate by functions like the following:

= a1 + a2 / h = c1 + c2 h (8)

These statistical dependencies here let us define two problems:

1) determining statically determinate bar of a variable cross-section optimal shape;
2) determining statically determinate bar of a uniform cross-section optimal height;
It should be noted that the second problem is a special case for the first one.
Let us put functions (8) into the strength condition (3). Here we get a modular form of the volume as regard to h
in k point of the bar.

Vic (h) = Fili = li / R [ N ic2 + 2 N ic M ic (a1 + a2 / h)ic + M ic2 (a1 + a2 / h) 2k + 3 Qic2 (c1 + c2 h)ic2 ] (9)

For a part of the bar with d z length the elementary capacity can be presented as an out-form for n fiber points:

dV = || V1 (h) ||V2 (h) || V3 (h) || ......Vn (h) || dz (10)

 Igor S. Kholopov et al. / Procedia Engineering 153 (2016) 277 282 281

Double-bar in the out-form here means "or / or" (that is either the figure before the line or the figure after is
bigger). The bar capacity with l length and variable cross-section presents an integral:

l
V = || V1 ||V2 || V3 || ......Vn || dz (11)
0

Equation (11) is an integral modular form of a particular type. It cannot be found through elementary functions
and presents in itself area covered by modular forms maximum ordinates. It is possible to work out this integral
form (11) using a numerical method with the help of a PC. We just need to reduce a continual problem to a discreet
problem while dividing the bar into several parts l separated by cross-sections.
To make automation of this problem simpler we suggest analyzing it in matrix form. Force matrix here can be
derived from equations of equilibrium for each i cross-section (i = 1, 2, ..., m). Force matrix-vector consists of
submatrixes containing bending moments, axial and shearing forces in i cross-section.

S = [ S1 , S 2 , S3 ,....., S m ] (i = 0, 1, 2, ..., m) , (12)

Here:

S i = [ N i , M i , Qi ] (13)

Each fiber characteristics are determined by a diagonal matrix D ik (k fiber, i bar) of the following type:

Dik = diag (1, ,1.732 ) ik (14)

Each i cross-section area is chosen according to the maximum value of out-forms determined for fiber points of
junction. Area ik in k point of i cross-section possible value can be calculated by the following matrix algorithm:

Fik = 1 / R {S Dik E1}{( S Dik E1}T (15)

In equation (17) E 1 is an identity matrix used for other matrixes multiplying:

E1 = diag (1,1,1) (16)

Elements of the possible point values of the matrix can be used to put down the area out-form corresponding to i
cross-section. It is necessary to choose the maximum possible value of he area for every cross-section. It provides
strength conditions if h j is permanent. Then within the determined minimum discreet cross-section heights h j (j = 1,
2, ..., d) we find the minimum value among those maximum and find the corresponding height h as function (17)
shows:

|| F1 (h1 ) || F2 (h1 ) || F3 (h1 ) ||...... || Fk (h1 ) || ..... || Fn (h1 ) ||= max

|| F (h ) || F (h ) || F (h ) ||...... || Fk (h2 ) || .... || Fn (h2 ) ||= max (17)
Fi = 1 2 2 2 3 2 = min
..........................................................................................
|| F1 (hd ) || F2 (hd ) || F3 (hd ) ||...... || Fk (hd ) || ... || Fn (hd ) ||= max

The volume between two cross-sections of the bar can be calculated while multiplying the length and two
minimax areas half-sum:

= ( + + 1)/2 (18)
282 Igor S. Kholopov et al. / Procedia Engineering 153 (2016) 277 282

Consequently, the volume of all bar is equal to

m
V = V
i =1
i
(19)

In statically determinate systems force matrix S i can be easily found from equilibrium conditions. In statically
indeterminate systems though it is not quite easy to calculate the height corresponding to the minimum volume as
force matrix S is in this case h function, and h is the function of rigidities correlation. Thus, this tasl is non-linear.
We believe the solution here should be iterational and should present a sequential decision of two simpler tasks:
1) volume miniaturization in spacing X i when h=const;
2) volume miniaturization of h when x j =const.

3. Conclusion

In real life, the construction also requires the account of the cross-section flanges and walls local buckling
resistance, bending limitations, limitations connected with width, depth and size unification of flanges and walls. It
definitely can lower the whole economical effect. Still if we take all these limitations into consideration, the whole
effect of steel saving is about 10-20% compared with a girder or a beam with uniform cross-section. So, the solution
for this problem is quite actual but it requires thorough design of corresponding algorithms and programs.

References

[1] Chang-Yu Cui, Bao-Shi Jiang, A morphogenesis method for shape optimization of framed structures subject to spatial constraints,
Engineering Structures, Volume 77, 15 October 2014, Pages 109-118, ISSN 0141-0296, http://dx.doi.org/10.1016/j.engstruct.2014.07.032.
[2] Lazaros Elias Mavromatidis, A review on hybrid optimization algorithms to coalesce computational morphogenesis with interactive energy
consumption forecasting, Energy and Buildings, Volume 106, 1 November 2015, Pages 192-202, ISSN 0378-7788,
http://dx.doi.org/10.1016/j.enbuild.2015.07.003.
[3] I.I. Goldenblat, V.A. Kopnov, Strength and plasticity criteria of construction materials, Moscow, Mashinostrojenie, 1968, p. 192.
[4] Ya., I.Olkov, I.S. Kholopov, Steel pre-stressed trusses optimal design, Moscow, Stroyizdat, 1985, p. 155.
[5] I.S. Kholopov, Bar systems optimization as applied to SAE (computer-aided engineering systems), Doctoral dissertation< Moscow, 1972.
[6] I.S. Kholopov, Mathematical modeling of bar systems volume by use of modular forms (textbook), Kuibyshev State University ed., 1982,
p.60.
[7] I.S. Kholopov, I.V. Loseva, Constructive and calculation constraints and their use in for discreet optimization of steel column, Higher
,QVWLWXWLRQVEXOOHWLQ$UFKLWHFWXUHDQG&LYLOHQJLQHHULQJ
[8] I.S. Kholopov, Statically determinate bars with variable cross-sections: optimization by use of modular forms of FE volume, Building
Mechanics and Structures Calculation, 1981, pp. 14-17.
[9] N.V. Efimov, E.R. Rozendorn, Linear algebra and multidimensional geometry, Moscow, 1970, p. 336.
[10] Yu.A. Radtsig, Statically indeterminate trusses of minimum weight, Kazan University ed., 1969, p. 286.
[11] I.S. Kholopov, Optimization model for statically indeterminate bar systems under several loadings, Higher Institutions bulletin, Architecture
DQG&LYLOHQJLQHHULQJ
[12] I.S. Kholopov, M.I. Balzannikov, A.O. Lukin, A.V. Solovjev, On spatial lattice metal structures use in roofing of mechanical equipment
URRPVRIK\GURHOHFWULFSRZHUVWDWLRQV9HVWQLNRI0*68SS-41.
[13] I.S. Kholopov, A.V. Solovjev, A.O. Lukin, Welted I-Girders of variable cross-VHFWLRQZLWKURXQGKROHVSS-30.
[14] I.S. Kholopov, I.V. Loseva, Solid cross sections of steel frames: optimization on the basis of discrete model, Higher education
establishments news: Architecture and civiOHQJLQHHULQJSS-58.
[15] I.S. Kholopov, M.I. Balzannikov, V.Yu. Alpatov, On spatial lattice metal structures use in roofing of mechanical equipment rooms of
hydroelectric power stations, Vestnik SGASU, Vestnik of SSUACE, Town Planning and AUFKLWHFWXUHSS-232.
>@ 0, %DO]DQQLNRY %* ,YDQRY  0LNKDVHN +\GUDXOLF HQJLQHHULQJ VWUXFWXUHV PDQDJHPHQW V\VWHP 9HVWQLN 0668 6FLHQWLIic-
7HFKQRORJLFDO-RXUQDOSS-124.
[17] V.Yu. Alpatov, A.A. Lukin, The Method of Mirror Functions in Optimal Design of Pre-stressed Spatial Structures, Procedia Engineering,
Volume 111, 2015, Pages 20-29, ISSN 1877-7058, http://dx.doi.org/10.1016/j.proeng.2015.07.030.