Font Size: a A A

Buckling Behavior Of High-strength And Large-sectional Angle Steel Columns

Posted on:2017-06-12Degree:DoctorType:Dissertation
Country:ChinaCandidate:K CaoFull Text:PDF
GTID:1312330485962007Subject:Structural engineering
Abstract/Summary:PDF Full Text Request
High-strength and large-sectional angle steel (HLS) columns, are angle columns with nominal yielding strength larger than or equal to 420 MPa, limb width larger than or equal to 220mm, limb thickness larger than or equal to 16mm. Compared with combined normal angles, HLS columns have the advantages of convenient assembly, etc. So using HLS columns as compressed members was a major developing direction in structural engineering, especially in steel transmission towers. The employments of HLS columns in practice have shown that, buckling strengths of these HLS columns were obviously higher than the corresponding design values in current design codes. However, researches on the method how to calculate buckling strength of these HLS columns were seldom reported, and more word need to be done on researching buckling behavior of these HLS columns.On the support of National Nature Science Foundation of China (No.51378401), investigation on buckling behavior of HLS columns were conducted, and the efforts and main findings were listed below:1) The effective length factor ? of axially compressed column ended by spring hinges (column ends which have rotational stiffness) was studied. It is proved that, ? can be determined by top and bottom ends-to-column stiffness ratios ra and rb, and the equation which ?., ra and rb, meet were derived out. Based on the equation, and with the help of secant iterative method, ? values of ra and rb, ranging from 0 to 100 were calculated. Then a table of ? values was made out for practical engineering.Furthermore, based on the date of ? values, a practical formula for calculation of ? values was proposed by fitting, which was proved to be succinct and accurate.2) Investigation on errors of column's buckling factor ?, when ignoring column ends' real rotation stiffness and considering the end was ideal-pinned or ideal-fixed, was conducted. It is shown that, errors of ? values will be non-ignorable in this situation.It is also shown that, when real rotation stiffness was close to ideal values, the errors of ? can be limited under a small level. Based on this situation, condition equations of column ends rotation stiffness were put forward, and when meeting the condition equations, errors of ? can be effectively limited below 5 percent.3) Axially compressed loading tests of HLS columns involving 6 kinds of cross section (limb width of 220mm and 250mm, thickness ranging from 20 mm to 30mm),5 kinds of slenderness (35?55), totally 90 Q420 specimens were conducted.Material tests of these Q420 steels were carried out, and stress?strain (???) curves were obtained. The real geometric data and initial bending value of each HLS specimen were measured. Rotation stiffness of the ends of the loading equipment were measured, and effective length factors ? of every specimen were calculated. Then real slenderness ? and real non-dimensional slenderness ?n of every specimen were calculated.Buckling modes and ultimate bearing strengths of every specimen were obtained through axially compressed loading tests. The non-dimensional buckling strengths, namely buckling factors ?t were then calculated. The test ?t??n data were further compared with column curves in different design codes (including Chinese standard GB50017-2003, American standard ANSI/AISC 360-10, ASCE 10-97, and European standard Eurocode 3). The compare results have shown that, current design column curves were conservative when predicating buckling strength of HLS columns, especially column b in GB50017-2003 and column b in Eurocode 3. The column curve in ASCE 10-97 was the most close to experimental results, relatively.4) ANSYS program was employed to develop finite-element (FE) models of HLS specimens. The FE models were adjusted by experimental results, and finally proved to be accurate enough when predicating buckling behavior of HLS columns.Two kinds of FE models employing ideal-elastic-plastic constitutive relation (bilinear) and constitutive relation taken hardening progress into consideration (multi-linear) were founded respectively. FE analysis results ?B of bilinear constitutive relation models and ?M of multi-linear constitutive relation models were calculated and compared with column curves in current design codes. It is shown that,?B were very close to column curve b in GB50017-2003 and column curve b in Eurocode 3. On the contray, ?M were obviously different from every current design column curves. However, ?M were close to experimental results ?t. The difference between ?B and ?M was great when ?n was small, and when ?n was large enough, ?B and ?M were close to each other.The comparison between ?t,?B and ?M shown that, FE models using multi-linear constitutive relation can reflect buckling behavior of HLS columns more accurately.5) On the condition of unifying initial imperfections, totally 60 FE models of slenderness ? ranging from 30 to 150 (?n ranging from 0.426 to 2.268) were calculated, and the buckling factor ?FE were obtained. It should be noted that, the initial bending deformations were taken as 1/1000 column length.?FE were compared with column curves in current design codes. The comparison shown that, current column curves cannot predicate buckling strengths of HLS columns accurately. This is because the influences of hardening progress in stress-strain curves were ignored. The differences between ?FE and current column curves were deeply related to non-dimensional slenderness ?n. It is suggested in this paper that, LHS columns can be divided into 3 kinds according to ?n, namely stub columns (?n?0.472), middle columns (0.472<?n?1.0), and slender columns (?n>1.0).?FE of stub columns were larger than 1.0;?FE of middle columns were small than 1.0, but still different from current column curves;?FE of slender columns were close to current column curves.6) Design method for buckling strength of HLS columns was investigated. It is proposed that, for the time being, column curve b in GB50017-2003 and column curve b in Eurocode 3 can be employed in conservative designing, and column a in Eurocode 3 can be employed in more economical designing.Furthermore, on the base of 60 FE results, a new column curve for HLS columns were drawn out, and formula of which were put forward by fitting. The proposed design formula for buckling strength ? of HLS columns can be described as:when ?n?0.472, ?=1.0; when ?n>0.472, the formula was similar to the Perry Equation in GB50017-2003, with ?2= 0.849 and ?3=0.320.The proposed design formula has the characteristic of economical, for the reason that the formula can give full play to the high bearing strength of HLS columns, and reduce the cost of steel constructions.
Keywords/Search Tags:Effective length factor, High-strength and large-sectional angle steel, buckling strength, Finite element, Column curve
PDF Full Text Request
Related items