50 Simplified Steel Compression Member Design

You might also like

Download as pdf
Download as pdf
You are on page 1of 16
SIMPLIFIED STEEL COMPRESSION MEMBER DESIGN William J. Rasdorf* and Stacey L. High’ Key words: Beam columns; Bending; Biaxial loads; Buckling; Columns; Combined loading; Compression members; Design; Structural engincoring; Steel State University, Box fate Professor of Civil Enginesting and Computer Science, Nerth Carclin 7908, Raleigh, NC 27805 2Research Assistant, Department of Civil Engineering, North Carolina State University, Bex 7908, Raleigh, NC 2709 Contents 1 2 o Introduction Column Design Procedure 21 Column Loading 22 Input Parameters 2.3 Basic Interaction Equations 2.4 Modified Interaction Equations Se 2.5. Derivation of the Equations . Tables 3.1 Tabular Approach to Design 3.2 Scope and Accuracy 3.3 Use of the Tables 3.4 Comparison of the AISC Method and the Modified Method . Summary Adlmowledgments References 10 10 a List of Tables 1 W 10 X 100 Data Table 2 W28X58 Data Table 2 W2&X 67 Data Table List of Figures 1 Design Example “4 16 SIMPLIFIED STEEL COMPRESSION MEMBER DESIGN William J. Rasdorf! and Stacey L. High? Abstract ‘The American Institute of Steel Construction (AIS) “Specification for the Design, Fabrication, and Erection of Structural Steel for Buildings” [10] has made manual steel colurnn design exceedingly time consuming and difficult. ‘The objective of this paper is to resent a simplified method of designing steel columns subjected to axial loads and moments for use in situations where automated design methods are inappropriate. Steel column design is based on the interaction equations of the AISC Specification. ‘These equations are presented in terms of actual and allowable stresses and much time is required by a designer to manually determine the stresses and solve the equations. To simplify their solution, the interaction equations were reformulated and a set of parameters (multipliers) was introduced into them. The parameters were investigated to determine their validity, limits, and ranges of significant influence. They were then tabulated to provide quick and easy access for use, The modified interaction equations and the tabulated parameters constitute the results of this study. They are the physical tools that enable a designer to rapidly select initial steel column sections to satisfy design requirements and specification constraints. The analysis confirms that these tools can realistically and accurately be determined. The equations ‘were algebraically derived and the tables were generated as a function of the properties of the sections. Thus, a new design method, combining the use of tabulated parameters with algebraically modified interaction equations, has been developed. This method greatly simplifies and speeds up the column section selection process 1 Introduction ‘The AISC Spe n presents a comprehensive column design procedure. ‘The sections of the Specification which govern column design are those which determine allowable stresses for members subjected to combined compressive and bending stresses. In these sections the interaction equations are specified. The difficulty is that a manual solution of the column design interaction equations is tedious as well as time consuming. The goal of the paper to present a method of simplifying column design by offering the designer the most approach to the solution of steel column design problems. "Associate Profesor of Civil Engineering and Computer Science, North Carolina State University, Box 7008, Raleigh, NC 27605, Research Accitant, Department of Civil Bagincsring, North Caroline State University, Box 7008, Raleigh, NC 27695. ‘The development of the method entailed modification of the interaction equations given in the AISC Specification and an evaluation of the factors affecting the equations. The method itself utilizes both algebraic formulas and tables. The formulas ere modified forms of the interaction equations and the tables specify values for the carefully defined multipliers of the modified equations. By substituting the appropriate multiplier values from the tables, into the modified interaction equations, the designer can quickly and accurately design steel columns in accordance with the Specification. Because the multipliers are presented in a straightforward tabular form, and fewer calculations are required when using modified interaction equations, this method is more straightforward and direct than the solution of the original equations presented in the Specification, In the development of the tabulated results presented in this paper all necessary column section properties and applicable provisions of the specification have been incorporated. In addition, the individual factors and multipliers of the interaction equations have been fully evaluated. All of the information necessary for designing steel columns subjected to either axial, Hexural, or combined loads is contained in the various tables, thus eliminating the need for referring to the AISC Manual [1]. 2 Column Design Procedure 2.1 Column Loading ‘The two besic types of loading that influence column design are axial load (P) and moment. (M). These loading types may occur separately or in varying proportions in combination with each other. Axial loads as well as other imposed loads that induce axial compressive stresses in the column are designated by P. Moment is designated by M. Moment induces bending stresses in the column and may be caused by lateral loading, applied end mo- ments, or eccentric application of an axial load. In this paper axial loade and moments are considered to be two of the given items in a column design problem statement. When axial load and moment occur simultaneously, the condition defined as combined loading exists. Combined loading is the most commonly encountered form of column loading, in steel buildings. ‘The AISC Specification presents an involved method for the design of columns subjecied to combined loading and it is this condition of loading with which this paper deals. ‘The Specification presents formulas (1.6-1a), (1.6+1b), and (1.6-2) as the requirements which members subjected to axial compression and bending (combined loading) must satisfy 19). These equations are called the interaction equations. They account for any combi- nation of loading conditions in the evaluation of allowable stresses for columns. It is these equations which were modified to provide the designer with simplified means for selv- ng the optimal section for the prescribed axial loads and ing them as well as for identify moments 2.2 Input Parameters Input parameters are values which cannot be predetermined or presumed, but which must be specified as a part of the column design problem statement. As noted above, axial loads and moments are considered to be input parameters. In addition, values for K, L, C, and Cm must also be given, These values cannot be preprogrammed into a design method or procedure as constant values since it is not possible to determine thom before @ problem is defined. When e column design problem is defined it clearly must include P, M, and KL. Eque- tions (1) and (2) show that C, and Cy, depend directly and only on M. = 75 + 1.05(My/Me) + 0.3(M;/M2) (2) On = 0.6 ~ 0.4(My/Ma) Since Ch and Cy are derived [rom an input parameter they too are considered to be problem statement input parameters, Although it is not possible to predetermine KL, Os, and Cp, it is possible to determine their limits and obtain an understanding of their influence and affect on the interaction equations, Such an analysis and evaluation has beon completed and has been incorporated into the results presented herein 2.2 Basie Interaction Equations Equations (1.6-1a), (1.6-1b), and (1.6-2) of the Specification are shown below as equations (3), (4), and (5) respectively. (3) fo/ Fa + (Cine foe)/((L— (fa/ Fez)) Foe) + (Cry Soy)! (1 - (Sai Fay)) < 1.0 @ fo/0.6Fy + fra/ Prat foy/ Foy < 1-0 ® fo[ Fat foz/Pro* foy/ Foy $1.0 If Ja/Fa < 0.15, formula (5, 1.6-2) may be substituted for formulas (8, 1.0-1a) and (4, 1.6-1b). Inspection of the interaction equations reveals that they are broken into three distinct terms. In each of the equations the first term pertains to axial compressive stresses; the second term pertains to bending stresses about the x axis; and the third term pertains to bending stresses about the y axis. Axial loads and moments are thus accounted for in terms, of axial and bending stresses Equation (2) generally governs column design between points of support, such as at locations of beam to column connections [16]. Between supports is where deffection can be at its maximum because of a lack of support or bracing. Induced axial loads tend to cause buckling stresses in the columan and what may be termed a buckling induced deflection. Applied moments, axial load eccentricity, initial crookedness, and residual stresses all induce secondary moments into the column and cause what may be termed @ bending induced deflection. These deflections are also maxirmum, relative to the ends of the column, between the support. ‘To prevent failure of a column section due to buckling induced dofiection, tho first term of equation (3) contains the factor F,, whose definition in the Specification contains appropriate factors of safety to guard against buckling failure. ‘To prevent failure of a column section due to bending induced deffection, the factors Cm and 1/(1~ fa/F2) have been introduced into the bending terms of equation (3). These factors are known as the reduction factor and the amplification factor respectively [3] Equation (4) generally governs column design at points of support. At these points there is no deflection relative to the ends of the column since the column is attached to a beam: or girder. Therefore, there can be no moment induced deflection and there is no longer @ need for the use of reduction and amplification factors as in equation (3). Buckling induced defiection is restricted at support points as well. The specification accordingly permits the use of the more liberal 0.6F, in place of F, in the axial stress term of equation (4). Thus, equation (4) is the most liberal of the interaction equations in terms of allowable stresses; however its use is restricted to the more conservative conditions of column design. Equation (5) may be used in place of both equations (3) and (4) if the axial load is relatively small. Equation (5) epplies both at supports and between supports. Its first term thus maintains the more conservative factor F. to account for buckling induced deflections ‘The amplification and reduction factors of the bending terms, however, have been omitted. ‘This omission is justified by the fact that small axial loads do not induce secondary moments and their accompanying deflections are not significant enough in magnitude to alter the results of the design process [3,8]. As a result, the safety factors in the bending stress terms of equation (5) are not needed. 2.4 Modified Interaction Equations ‘The stated objective of this paper is to simplify the manual steel column design process. Simplifying the complexity of the interaction equations and attempting to reduce the time required to obtain a solution was the means of achieving thie objective. The interaction equations were modified by algebraic substitutions resulting in equations which were simpler and offered a more iznmediate approach to the appropriate solution than those in the AISC Manual. The resulting modified equations are (6) P+ CmneZzBzMz + Cry ZyByMy < Pa (10) Za In the modified equations shown here, Om exists as it did in the original equations. P and M allow for direct substitution of loads and moments as opposed to the substitution of stresses as in the original equations. P,, B, Z, T, R, and C, are the multipliers which were algebraically derived, calculated, and tabulated. Section 2.5 presents the derivations of the multipliers. For each column section of the AISG Manual, all of the multipliers were generated and tabulated. They were generated from the equations presented herein using the known properties of the given column section specified in the manual. The tables, cou- pled with the modified interaction equations, provide the necessary tools for the simplified method of steel column design proposed here (13) 2.5 Derivation of the Equations Equations (6), (7), and (8) were derived from the AISC interaction equations (3), (4), and (5) respectively [7,14]. Equation (8) was derived as follows: ® fo/ Fa Soz/Fox + fiy/ Fry £10 Multiplying the entire equation by AF, yields: (1) Afo+ (AF fia) / Phe + (AFafiy)/Foy S AFa Knowing that AF, = P, and that Af, =P equation (11) can be modified by substitution to yield (2) P+ (Pafte)/ Fos + (Pafsy)/ Fou < P- ‘The actual bending stress must be multiplied by a factor of 12 in order to convert it to the units of foot-kips. Substituting fy = 12 M/S into equation (12) modifies its form to: (13) P+ ((12Pa)/(FoeSe)) Mz + ((12 (FivSs)) My = Pr Let B = 12P,/ FS. By substituting this newly defined multiplier into equation (15) the final form of the modified interaction equation is determined to be: @) P+ BsMs + ByMy S Pa: Equation (7) was derived from interaction equation (4) by the same steps as illustrated for the derivation of equation (8) with the exception of the axial compressive stress term. ® $/0.6Fy + fre/ Foe + fay/ Fey $1.0. Multiplying equation (4) by AF, the first term of the equation becomes: (aa) (FeAfa)/(O.6Fy) Knowing that AJq— P and that P, = P,/A the term can be further modified by substitu- tion to yield: (15) (PaP)/(A06F}). Let Cy = P,/0.6F,A. The final form of the modified interaction equation is thus determined to be: @ CaP + BeMz + ByMy < Pa. Equation (6) is also derived from interaction equation (9) in a similar manner to the derivation of equation (8). The difference between equations (3) and (5) is the introduction of the reduction and amplification factors. The reduction factor Cm remains unchanged in the modified equations. The amplification factor Z is defined as (18) Z=M(1-(fo/ Fe) Just as the interaction equations are modified to simplify them, Z also is simplified. Sub- stituting f, = P/A into equation (16) yields: (a7) Z=1)(1-(P/AF!)) Multiply by AF! to determine that: (18) Z= (AP) /((AF) ~ P). Let R= AF! by definition and equation (18) becomes: (as) Z= RRP). Equation (3) may then be expressed in its final modified form as: (6) P+ Crs nBaMe + Omy ZyByMy & Pa Equations (6), (7), and (8) thus represent the modified forms of the AISC interaction equations. These modified equations are used and referred to throughout the remainder of this paper, 3 Tables 8.1 Tabular Approach to Design ‘The authors have derived a methed of solution of the interaction equations that is obtained by both tabular and algebraic means. In this method the multipliers of the modified inter- 6 action equations are obtained from tables and are substituted into the modified equations to algebraically determine the solution The format of the tables is shown in Table I. For a given column section the table lists all of the multiplier values needed to use the modified equations for various values of KL. Values of KL range from 6 to the point at which KL = 200r- the maximum allowable slenderness ratio for compression members. The line of asterisks in the table represents the point at which the slenderness ratio (KL/r) is equal to 120. Values above this line (lower KLs) are applicable to main members and values below the line (higher KLs) are applicable to bracing and secondary members. There are two alternatives to tabularly presenting the multipliers as is proposed here. The first alternative is to simply calculate them for specific design problems as needed. Thie alternative, however, is contrary to the paper’s objective of simplifying the design process The second alternative is a graphical presentation of the multipliers. All of the information presented in the tables could be graphed in some manner, providing a means of visually presenting the multipliers for a particular section. Both approaches are reasonable; we have simply chosen to present the tabular approach here. 3.2 Scope and Accuracy A complete set of design tables has been generated for steel having yield strengths of both 36 ksi and 50 ksi. Although this paper deals primarily with wide flange columns, tables have also been generated for rectangular, square, end pipe shapes and the design procedure outlined here is applicable to these chapes as well. The various sections included in the tables are those found in the AISC Manual. The section properties used to generate the tables are those listed in the Manual. For wide flange sections the properties were obtained from a data disk supplied by AISC. The properties necessary to detive the rectangular, square and pipe tables were typed in directly by hand. The accuracy of the column design method presented here is excellent. The use of the ta- bles, multipliers, and modified equations will always result in a column section that satisfies the applicable provisions of the AISC Specification. Continued validation of the method is underway to furthor determine its optimality. In most cases the method provides not only a result, but provides the optimal result, ie., the lightest column section that satisfies the Specification. Our continued investigation is expected to yield a precise determination of the range of K values for which the method is optimal and the range for which it is conservative. Further information about this ongoing investigation, the availability of the complete set of design tables, and additional design examples can be obtained from the first, suthor. 3.3. Use of the Tables ‘The tables presented herein have been developed to enable the designer to make an accurate selection of the proper column size with the greatest accuracy and the least effort. Before using the tables the designer must determine the loads and moments to which the column is subjected to assure that the proper KL, Cs, and Cm values are calculated. Once the problem has been stated and the input parameters have been determined the designer is then ready to use the tables. The following design examples will help to illustrate the use of the design method. The mathematical and tabular solutions are provided to give the reader an ides of the accuracy which can be obtained. Figure 1 illustrates the problem and provides the input parameter values (13). For the first problem assume that a W10X100 section has been chosen as a trial section. The following values are obtained from the Table 1: Since P/P, = (135/237.9) = .57 > 0.15 both modified equations (6) and (7) [AISC equations (8) and (4)] need to be checked. The following values are obtained from equations (2) and (19) contained herein: ‘Simply substituting these values into the modified equations yields the result. Thus, © P+ CmzZeBeMz + Cry ZyByMy < Ps yields 135 + 46.5 + 36.7 = 218.2 < 237.9 and @ CaP + BeMz+ ByMy < Px yields 50.6 + 47.2 4 26.4 = 124.3 < 22709. The solution indicates that the assumed W10X100 column section satisfies the interaction ‘equations. Furthermore, because its equivalent axial load is close to the allowable P,, it is an efficient column section. The second problem illustrates the inherent capacity of the design method to allow the designer to quickly converge to the proper column section. This is a more common design 8 situation, where an experienced designer can choose an initial trial section that is close to the final answer, but is still a few weights away. ‘To complete this example, multipliers are needed for additional column sections. These are presented in Table 2 and Table 3. Using the example in Figure 1 but assuming that architectural considerations limit the column to a maximum length of fifteen feet, resolve the column design probl First, compare the actual load P (135 kips) with the allowable P,’s listed in Table 2 Table 2 shows that a W8X58, which has a P, = 251.5 at a KL = 15, could be tried for the initial trial column section. Substituting the multipliers for the W8X58 into modified interaction equations (6) and (7) results in (6) P +Cn2ZeBeMe + Cry ZyByMy < Pa yields 135 + 98.7 + 59.9= 293.6 < 251.5 and @ C.P+ B.M, + B,M, < P, yields 91.9 + 107.5 + 61.1 = 260.5 < 251.5. ‘The results chow that the W8X58 will not work and consequently, @ heavier section is necessary. Choosing a W8X67 as the second trial section, (3) P+ Cmz2_BzMz + Cry ZyBy My < Py yields 135 4.96.1 + 54.3 = 285.4< 2016 and @ CaP + BrMz+ ByMy < Py yields 92.5 + 107.8 + 60.6 = 26044 < 291.8 ‘The WSX87 would therefore prove to be a satisfactory section and is in fact the optimal solution. The key observation here is that the table lookups to determine the multiplier values and the subsequent equation solutions become very simple, enabling column sections to be chosen and checked very quickly. 3.4 Comparison of the AISC Method and the Modified Method Longhand solutions of the AISC interaction equations confirm that the sections obtained in Section 8.8 satisfy the Specification requirements and that the solutions obtained using the tables are correct. The accuracy of the tabular method is partially evident by noting that the relationship between the results of the modified method and the results of the AISC method are in the same relative proportion to each other. For example, in the first example problem, the modified method yielded results of 218.2 < 237.9 and 124.3 < 237.9. If this same problem were solved by directly using the AISC interaction equations the results would 9 have been 0.92 < 1.0 and 0.5 < 1.0, which are both in the same relative proportion to the results obtained using the modified method. In addition, the examples illustrate the ease with which trial column sections can be determined and with which a designer can converge to the most weight-ellicient section. The tabular method thus provides a fast, accurate, and efficient solution to column design problems and it does so in terms of P and M rather than in terms of f, and f,. The designer can ths work directly from loads and moments rather than converting to stresses 4 Summary The interaction equations of the AISC Specification are the basis for modern steel column. design. This paper hes been devoted to simplifying the column design process by modifying the interaction equations. ‘The accomplishment of this objective enables the designer to quickly and accurately make a proper column section selection The first step in simplifying the column design procedure was to simplify the interaction equations. ‘This was done in such a way as to create useful multipliers which were dependent on known column properties and which were also dependent on comparable values given in the problem statement, ‘The multipliers were then evaluated to determine their effect on the interaction equations, ‘The second step in simplifying the column design procedure consisted of tabularly pre- senting the multipliers. It was found that a tabular approach to determining the multipliers provided a fast efficient way of designing as demonstrated by the example column design problem solutions provided ‘The design method presented here is clearly a manual method and is therefore applicable in design situations where a manual approach is most applicable. One such situation is in the design of individual members when few columns need to be designed or redesigned. Another is in structures where automated methods do not provide a clear design time advantage ‘This will often be the case in single story structures such as schools and shopping centers. ‘The objective of a designer must be to use the most accurate, efficient, and applicable solution method for the design problem at hand. This paper has presented one such method. 5 Acknowledgments ‘The authors wish to acknowledge the significant cooperation of FDE Ltd. of Alexandria, VA, Dr. Louis Geschwindner of the Pennsylvania State University, and Ms. Leila Osteen of North Caroline State University, in completing the work. Portions of this work were supported by the National Science Foundation under Grant MSM - 8451485, a Presidential Young Investigator Award 10 10. cr 12. 13, 4. 16. References American Institute of Steel Construction, Manual of Steel Construction, Bighth Bdi- tion, AISC, New York, New York, 1980. Burgett, ..B., “Selection of a Trial Column Section,” Engineering Journal, AISC, Vol. 10, No. 2, April, 1973, pp. 56-58. “Commentary on the Specification for the Design, Fabrication and Erection of Struc- tural Steel for Buildings,” AISC, New York, New York, 1980. Disque, R.O., “Inelastic K-facior for Column Design,” Engineering Journal, AISC, Vol. 10, No. 2, April, 1978, pp. 33-35. Ginsburg, 8. and G. Kaez, “Automated Search for Optimal Standard Sections,” En- rel, 22, No. 1, January, 1085, pp. 19-21 gineering Journal, AIS' Hygienes, T.R., “Effective Column Length-Ticr Buildings,” Engineering Journal, AISC, Vol. 1, No. 1, January, 1984, p. 12 Hooper, [., and R.E. Rapp, “Simplifications in the Solution of Column Interaction Problems,” Engineering Journal, AISC, Vol. 1, No.1, Januaty, 1964, pp. 20-23. Johnson B.G,, editor, SRC Guide to Stability Design Criteria for Metal Members, John Wiley and Sons, Inc., New York, New York, 1976. Johnston, B.G., “Discussion of the Effective Length of Columns in. Unbraced Frames, Enginzering Journal, AISC, Vol. 9, No. 1, January, 1972, p. 46, Liu, W-Y., “Steel Column Bending Amplification Factor,” Engineering Journal, AISC, Vol. 2, No. 2, April, 1965, pp. 50-51 Maitra, N., “Allowable Stress for Bending Members,” Engineering Journal, AISC, Vol. 19, No. 4, October, 1982, pp. 206-208. Moore, W.B., “A General Solution for the Governing Bending Equation,” Engineering Journal, AISG, Vol. 22, No. 1, January, 1985, pp. 22-27. Phange, M., and E. Celenza, “Direct Design of Columns Subjected to Combined Loading,” Engineering Journal, AISC, Vol. 9, No.4, Octoher, 1972, pp. 142-153. Rapp, RE., “New Manual Makes Steel Column Design Rasy,” Engineering News Record, September 5, 1983, pp. 34-36 Rapp, R.E., Practical Steel Column Design, MT. Hughs Corp., New York, New York, 1965, Rice, P-F., and E.S. Hoffman, Structural Design Guide to the AISC S Buildings, Van Nostrand Reinhold Co., New York, New York, 1976. ecification for ul 17. Rubinsky, M.A., “Rapid Selection of Beam-Columns,” #agineering Journal, AISC. Vol. 5, No. 3, July, 1968, pp. 100-122. 18. Sandhu, BS., ‘Effective Length of Columns with Intermediate Axial Load,” Engi- neering Journal, AISG, Vol. 9, No. 4, October, 1972, pp. 154-156. 19. “Specification for the Design, Fabrication and Erection of Structural Steel for Build- ings,” AISC, New York, New York, 1980. 20. Yura, J.A., “The Effective Length of Columns in Unbraced Frames,” Engineering Journal, AISC, Vol. 8, No. 2, April, 1975, pp. 27-42. 21. Zweig, A., “Discussion of the Effective Length of Columns in Unbraced Frames,” Engineering Journal, Vol. 9, No. 1, January, 1972, p. 47. Nomenclature A. Area of cross section Cz Multiplier used to account for the higher allowable stresses permitted in the axial stress term of the interaction equations Cy Moment gradient coefficient for lateral-torsional buckling of beams Cm Coefficient of bending term in interaction equations dependent on column curvature caused by applied moments Fz Allowable axial compressive stress permitted in the absence of bending. Fy Allowable bending stress permitted in the absence of axial force FE Baler column stress divided by a factor of safety Fy. Specified minimum yield stress K. Bifective length coefficient E Length (feet) M_ Moment (hip-feet) ‘My Smaller end moment of a column ‘Mz Larger end moment of @ column P Axial load (kips) P, Maximum allowable axial load permitted on a column at a given KL FR. Coefficient used in the determination of the bending amplification factor T Coefficient used in the determination of the x axis moment multiplier Z Bending amplifiestion factor fa Caleulated axial compressive stress fs Calculated bending stress 1 Length (inches) 2 subscript relating symbol to strong axis bending y subscript relating aymbol to woak axie bending B Multiplier that converte moments into equivalent axial loads ee 20rk klx = kly = Soft = 135% [ SS a Figure 1: Design Example 13

You might also like