A General Solution for Eccentric Loads on Weld Groups

G. DONALD BRANDT

Eccentric loads on weld groups traditionally have been analyzed by elastic methods in which the individual effects of an axial load (applied through the centroid of the weld group) and a pure moment (in the plane of the weld group) were combined.1 In the 7th Edition Manual of Steel Construction2 of the American Institute of Steel Construction, as in earlier editions, such a procedure was used to create tables which could be used to determine allowable eccentric loads on selected weld groups. These tables were criticized as having non-uniform factors of safety when compared to the actual ultimate loads which the welds could support. An "ultimate strength method'' was proposed3 in which the resulting force per unit of length of each weld element is calculated from

5. The weld element which will reach its ultimate defoi mation first is the one for which the ratio of A m a x di vided by the radius to the instantaneous center is th smallest. It is assumed that deformations vary linearl with distance from the instantaneous center. 6. Consistent deformations (A) at all other weld elemen are then found from

7. The following parameters are then calculated for eac weld element:
R\i\i,i

R = Ruh(\ -

e-^Y

10 + B; 0.92 + 0.06030,M z = 75^0.01 m

-

and the total resisting forces and moments are predicted by summing the elemental forces. In the 8th Edition Manual, 4 such a procedure was used to create tables similar to those in earlier editions, but with different numerical coefficients. Butler, Pal, and Kulak 3 describe the ultimate strength method in some detail. For the sake of completeness in this report, this procedure is summarized as follows: 1. Choose an instantaneous center of rotation. (Fig. 1) 2. Assume that the resisting force on any weld element acts perpendicularly to a radius connecting that element to the instantaneous center. 3. Calculate the angle, 0, between the elemental force and the axis of the weld element. (Angle 6 is expressed in degrees.) 4. Determine the ultimate deformation which can occur on each weld element from A m a x = 0.225(0 + 5)--<7
G. Donald Brandt is Professor and Chairman, Dept. of Civil Engineering, The City College, City University of New York.

X,. = o.4e- 014 <^ Ri = *uit,,-(l e-^)^

8. By statics, calculate the corresponding applied force ai moment which hold the weld forces in equilibrium.

I.C.

Figure 1

150
ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

It should be noted that the formulas given above correspond to a V4-in.weld made using E60 electrodes. The R{ values are pounds per linear inch of weld. The final forces and moments are ultimate values. Tide 5 described the way in which these values were converted to allowable load coefficients for the AISC tables. In summary, this requires: 1. Converting from V4-in. weld to Vi6-in. weld by dividing by 4. 2. Converting from E60 electrode to E70 electrode by multiplying by 70/60. 3. Introducing a safety factor conforming to AISC Specification Sect. 1.5.3 by multiplying by 0.3. 4. Checking the shear stress on the most highly stressed weld element and, if it exceeds 21.0 ksi, reducing the load by the ratio of 21.0 divided by that shear stress. The papers by Tide 5 and by Butler et al 3 indicate that solutions have been obtained using computer programs. Butler states that the method is general, but shows details for only a C-shaped weld subjected to loading parallel to a principal axis. Tide shows results for a pair of parallel line welds and a C-shaped weld with the loads again parallel to principal axes. The AISC tables include a pair of parallel line welds, rectangular box welds, C-shaped welds, and L-shaped welds; loads are parallel to a principal axis in the first three types and parallel to a leg of the L in the last. In a previous paper, 6 the author showed how rapid solutions could be obtained for any eccentrically loaded bolt groups. The same method can be extended to weld groups. In essence, the method involves: 1. Directly finding the instantaneous center corresponding to elastic behavior of any weld arrangement for any eccentric load. 2. Directly determining the elastic solution for the maximum permissible load. 3. Directly determining an approximate value for the ultimate load. 4. Iterating to improve the approximate value. 5. Using the same procedure described by Tide to convert the ultimate load to an allowable load consistent with the AISC tables. Inasmuch as welds are continuous, it is necessary to discretize into a finite number of weld elements. A moderately large number of discrete elements is required if reasonable accuracy is to be achieved, so the rather long procedure for each weld element described above probably makes the procedure too laborious for manual calculations. Accordingly, the procedure is given here as a FORTRAN computer program. (See Appendix.)

Engineers familiar with FORTRAN should have no particular difficulty in understanding the program after reading the following discussion.
DISCUSSION OF THE COMPUTER PROGRAM

Computations for the Elastic SolutionThe length of each weld element, W, is calculated, and using these discrete elements, the centroid is located, the polar moment of inertia, / , is calculated, and the moment, M 0 , of the applied unit force about the centroid is calculated. A mapping function called FACTOR is used to transform forces to distances and it is equal to / divided by the product of the total length of weld and moment M0. The instantaneous center is located by adding to the X and Y coordinates of the centroid the quantities Py X FACTOR and +PX X FACTOR, respectively. The radius vector, D to the center of each weld element is calculated, and the largest one noted. The allowable moment about the instantaneous center of all elemental forces is XWD2/Dm3LX times the allowable force per inch of weld. The allowable (elastic) load is that moment divided by the moment of the applied unit force about the instantaneous center. This load and the identification number of the critical weld element are printed. Computations for the First Approximate Ultimate LoadThe instantaneous center located above is used as the first trial center. Calculations follow the procedure attributed to Butler3 described earlier in this paper. (Angle 0 is determined as follows: Form the dot product of the radius vector and the weld element vector. Divide this by the product of the magnitude of the two vectors; this quotient is the cosine of the angle between the vectors. Its complement is angle 6.) The (first approximate) ultimate load is the moment about the instantaneous center of all the elemental forces (Hi times Wt times D{) divided by the moment about the instantaneous center of the applied unit load. This ultimate load, and the identification number of the critical weld (having the largest force per inch) are printed, along with the coordinates of the current instantaneous center. Iterating to Improve the Approximate Ultimate LoadWhen the instantaneous center is correctly located, not only will the sum of the moments about that center be zero, but the vector sum of the forces will be zero as well. For the ultimate strength case, the X component of the force on each weld element due to the unit load can be found from A Puit and similarly for the Y component. The sum of all the Rx values plus Px will not be zero unless the instantaneous center is correct. The unbalanced

151
THIRD QUARTER / 1982

force is Fxx. Similarly for the Y direction. The vector sum of Fxx and Fyy is F> the unbalanced force. Fxx , Fyy and F are printed when the approximate ultimate load is printed. It has been found that components of a desirable shift from the previous trial center can usually be predicted from the same formulas used to locate the elastic instantaneous center. Thus, if x\ and y\ represent the coordinates of the instantaneous center for which the unbalanced force components are Fx and Fyy , the coordinates of the next center should be: x2 = xx - (Fyy X FACTOR) y2 = yx + (FxxX FACTOR) New values for the radius vectors to all weld elements are now calculated and the ultimate strength solution is repeated. In most cases, the unbalanced force decreases rapidly, and the ultimate load stabilizes after a few iterations. In the program, when the unbalanced force is less than 1% of the applied unit force, the solution is declared found. The ultimate load is converted to permissible load on a Vi6-in. weld made using E70 electrodes in the manner attributed to Tide 5 described earlier. This value is printed. It should be observed in using the program that, as in the case of AISC tables, the permissible load on the weld group is obtained by multiplying the permissible load given by the program by the number of sixteenths in the weld size. Furthermore, if the electrodes used are other than E70 (having Fv = 21 ksi), the permissible load should be multiplied by the allowable shear divided by 21 ksi. Unlike the AISC tables, no multiplication by any length should made. InputIn creating the program, provision was initially made for describing the weld configuration in two different ways. (Both of these have been retained.) 1. If the first columns on a data card contain the word LINE, the description of a weld consists of the coordinates of its starting point and ending point, and the number of equal segments into which the weld is to be subdivided. 2. If the first columns on the data card contain the word ELEMENT, the description of a weld element consists of the coordinates of its center and the length of its X and Y projections on the coordinate axes. If the first columns on the data card contain the word LOAD, the description of a normalized load (magnitude = 1) consists of its X and Y projections on the coordinate axes, and the coordinates of a point on its line of action. Observe that there may be as many data cards of LINE and/or ELEMENT type as required to describe the entire weld. Such cards may be in any order or any mixture. The last data card must be a LOAD card, since this is also a

30"

+
i"

Figure 2 signal that all data has been given. (Format for all three types of data cards is A2, 8X, 5F10.3. A fourth form of data is discussed later.) The program converts weld descriptions of either type into weld elements. These are numbered consecutively and printed, giving their identification numbers, center coordinates, and X and Y projected lengths. The LOAD data is also printed. Using the ProgramIn all the examples, the maximum permitted elastic load, Pe, the ultimate load, Pu, and the maximum permitted load, Pp, derived from the ultimate load, are listed. Example 1 (see Fig. 2): Data cards were prepared using the LINE and LOAD descriptions. Actual computer output is shown in Fig. 3 for a case where each line is subdivided into 10 equal elements. A reasonable question arises in every problem regarding the number of subdivisions required to get good accuracy. For the weld geometry and load of Example 1, computer runs were made dividing each line into 2, 5,10, 20, and 40 segments. The results are shown in Table 1. Table 1
Subdivisions Pe Pu Pp 2 1.539 26.304 1.565 5 1.253 24.783 1.472 10 1.151 24.741 1.469 20 1.100 24.593 1.460 40 1.075 24.499 1.455

Values obtained from the 7th Edition AISC Manual (for the elastic solution) and from the 8th Edition AISC Manual (for the ultimate strength solution) are, respectively, Pe = 1.05 and P p = 1.45.

152
ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

DATA FOR DESCRIPTION OF WELDS AND LOAD

LINE LINE LOAD C,.COOO 1. . 0 0 0 0 0,. 0 0 0 0 C..COOO C,. 0 0 0 0 - 1 .. 0 0 0 0 = 0 . 0000 1 . COOO 30. 5000 20 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 POLX 1.0000 l.COOO 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 I.0000 1.0000 1.0000 I.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 l.CCOO 30.5000,POLY = 10.0000 10,0000 0.0000 10.0 10.0

NUMBER OF WELD ELEMENTSi COORDINATES AND PROJECTICNS 0.0000 0.5000 1 0.0000 2 1.5000 0.0000 2.5000 3 4 0.0000 3.5CC0 5 0,0000 4.5000 oJoooo 5.5000 6 0.0000 7 6.5000 8 0. cooo 7.5000 9 o.coco 8.5000 9. 5000 10 0.0000 l.COOO 11 0.5000 12 1.COOO 1.5000 1.0000 13 2.5000 14 1 . cooo 3.5000 15 1 . cooo 4.5000 1.000 0 16 5 . 5C00 17 1.0000 6.5000 18 1.0000 7.50G0 19 1 . cooo 8.5000 1 . cooo 20 9*5000 PX= 0.CGOO,PY=

-1.0000,

0.0000

ELASTIC VALUE FOR MAXIMUM LOAD IS 1.151 MULTIPLY T H I S BY NUMBER OF SIXTEENTHS AND BY ALLOWABLE KSI / 2 l . , 0 C R I T I C A L ELEMENT I S NUMBERI 11

AT T R I A L CENTER NO. 1 X= 0.2139 FX =

X F = 0.3313 APPROXIMATE ULTIMATE LOAD = CRITICAL ELEMENT IS NUMBER I ON WHICH ULTIMATE FORCE PER INCH IS

Y= 5.0000 FY = -0.0000,

0.3313

24.748 15.6361

Figure 3. Computer printout for Example 1

153
THIRD QUARTER / 1982

AT TRIAL CENTER X=

Y* 5.0000 -0.0000 FY -0.1052 F = 0.1052 APPROXIMATE ULTIMATE LOAO * .24.743 C R I T I C A L ELEMENT I S NUMBER I ON WHICH ULTIMATE FORCE PER INCH I S 15.6238

NO. 2 0.30 8 7 FX =

AT TRIAL CENTER X=

Y= 5.0000 -0.0000 FY = 0.0344 F = 0.0344 APPROXIMATE ULTIMATE LOAD * 24.740 C R I T I C A L ELEMENT I S NUMBER 1 ON WHICH ULTIMATE FORCE PER INCH I S 15.6277

NO. 3 0.2786 FX =

AT TRIAL CENTER NO. 4 X= 0.2884 FX -

Y= 5.0000 -0.0000 FY = -0.0112 F = 0.0112 APPROXIMATE ULTIMATE LOAO = 24.741 C R I T I C A L ELEMENT I S NUMBER 10 ON WHICH ULTIMATE FORCE PER INCH I S 15.6264

AT T R I A L CENTER X=

Y= 5 . 0 000 -0.0000 FY * 0.0036 F = 0.003 6 APPROXIMATE ULTIMATE LCAO = 24.741 C R I T I C A L ELEMENT I S NUMBER 1 ON WHICH ULTIMATE FORCE PER INCH IS 15.6269

NO. 5 0 . 2 8 52 FX *

INSTANTANEOUS CENTER HAS BEEN LOCATED MAXIMUM PERMISSIBLE LOAD I S 1.469 MULTIPLY T H I S BY NUMBER OF SIXTEENTHS AND BY ALLOWABLE K S I / 2 1 . 0 Fig. 3 (cont'd)

154 ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

Example 2: Use the same weld configuration and load as in Example 1. Intuition indicates that more subdivisions should give better results, but there are two effects. One is simply the closer approximation to continuity achieved by using smaller elements. The other is the fact that the maximum stress occurs on the outside edge of a weld element rather than at its center. Obviously, more subdivisions of uniform size means that the .edge of an element and its center will be closer together. It is, however, possible to divorce the two effects by using non-uniform weld elements. For example, the top and bottom 0.001 in. of each of the welds in Example 2 are entered as individual elements, and the remainder subdivided as in Example 1. The new results are shown in Table 2. Table 2
Subdivisions Pe Pu Pp 2 0.797 23.867 1.417 5 1.009 23.977 1.423 10 1.039 24.307 1.443 20 1.047 24.374 1.447 40 1.049 24.388 1.448 Subdivisions Pe Pu Pp 2 0.800 23.919 1.420

1 L 3"
I

4"

10"

CO

v b

4 3 X

<N

x
3"
]

4"

Figure 4 AUTOMATIC in place of LINE. When this was done for both line welds, the results were as shown in Table 3. Table 3
5 1.010 23.997 1.424 10 1.040 24.317 1.444 20 1.047 24.378 1.447 40 1.049 24.391 1.448

It can be observed that the solutions stabilize with fewer subdivisions, and that the approach to the true solution is from the safe side rather than the unconservative side. The same behavior has been observed in a dozen other test cases for welds of different configurations. Example 3: Use the same weld configuration and load as in Example 1. It should obviously make very little difference if the small element welds are introduced in addition to (or even overlapping) the original welds. Therefore, if the critical weld element was unknown before the first analysis, only two additional cards need be prepared and added to the previous data for Example 1 to achieve the precision of Example 2. In Example 3, the data is:
ELEMENT ELEMENT LINE LINE LOAD 0. 1. 0. 1. 0. 10. 10. 0. 0. -1. 0. 0. 0. 1. 30.5 0.001 0.001 10. 10. 0.

Example 5 {see Fig. 4): Welds need not be vertical or horizontal, or even symmetrical. The applied load can be applied in any direction. Consider the bracket shown. Making use of the AUTOMATIC feature, and selecting the number of subdivisions for each line so that "real" elements were about 1-in. long, the data looked like this:
AUTOMATIC AUTOMATIC AUTOMATIC AUTOMATIC AUTOMATIC LOAD 7.0 3.0 0.0 0.0 3.0 0.6 11.0 11.0 8.0 2.0 0.0 -0.8 3.0 0.0 0.0 3.0 7.0 17.0 11.0 8.0 2.0 0.0 0.0 9.0 4.0 4.0 6.0 4.0 4.0

10 10

The results are Pe = 1.040, Pu = 24.312, Pp = 1.444. Example 4: Use the same weld configuration and load as in Example 1. In order to avoid the extra effort involved in specifying data for additional elements, an automatic insertion of an extra "fake" element 1/1000th as long as the subdivision size can be made at each end of any line by using the word

From the program, the value of Pe is 3.869; since the weld size is 5/i6, the allowable elastic load is 5 X 3.869 = 19.345 kips. The value of Pp is 5.452; the permissible load is 5 X 5.452 = 27.260 kips.
REFERENCES

1. Salmon, C. G. and J. E. Johnson Steel StructuresDesign and Behavior Second Edition, Harper and Row, 1980, Chap. 5. 2. Manual of Steel Construction Seventh Edition, American Institute of Steel Construction, 1970.

155
THIRD QUARTER / 1982

3. Butler, L. J., S. Pal, and G L. Kulak. Eccentrically Loaded Welded Connections Journal of the Structural Division, ASCE, Vol. 98, ST5, May 1972, pp. 989-1005. 4. Manual of Steel Construction Eighth Edition, American Institute of Steel Construction, 1980. 5. Tide, R. H. R. Eccentrically Loaded Weld GroupsAISC

Design Tables Engineering Journal, American Institute of Steel Construction, Vol. 17, No. 4, 4th Quarter 1980. 6. Brandt, G. Donald Rapid Determination of Ultimate Strength of Eccentrically Loaded Bolt Groups Engineering Journal, American Institute of Steel Construction, Vol. 19, No. 2, 2nd Quarter 1982.

APPENDIX
$ JOB $IBFT C C L I S T , K P = 2 9 , T IME=6C0,PAGES=90

1 2 3 4 5
6 7

8 9 10 li 12 13 1* 15 16 17 18 19 20 21
22 23 24 25 26 27 28 29

iTANTANEQUS CENTERS OF WELD GROUPS PROGRAM TO LOCATE I N S ' <HQ ULTIMATE RESISTANCES AND OBTAIN ELASTIC Af REAL J , M, M P , MPCG, I X , I Y , MU, LAMBDA INTEGER EL EM, TYPE INTEGER AUTO DIMENSION TEMP{5) DIMENSION X C 1 0 0 ) , Y{ 1 0 0 ) , D X ( I O O ) , D Y < 1 0 0 ) r D ( 1 0 0 > , >) 1 T H E T A ( I O O ) , R ( 1 0 0 1 , W ( 1 0 0 ) , W X U O O ) * WY( 1 0 0 ) FALLOT = 0 * 9 2 8 1 200 CONTINUE GENERALIZED INPUT WRITE i 6 , 6 0 0 0 ) 6 0 0 0 FORMAT i 1 H 1 , f DATA FOR DESCRIPTION OF WELDS AND LOAD , / I = 0 300 CONTINUE Et ( TEMPCJJ),JJ= 1,5 ) READ ( 5 , 5CC ) TYPI 500 FORMAT { A 2 , 8X, 5F 1 0 . 3 ) ELEM / EL / , LOAD / f L C / DATA LINE / L I 1 / CATA AUTO / ' A U ' / I F ( TYPE . E C . LINE GO TO 3 0 1 I F C TYPE . E Q . AUTO GO TO 3 0 1 I F ( TYPE - E Q . ELEM GO TO 302 I F ( TYPE EQ. LOAD GO TO 303 WRITE ( 6 , 6 004 ) YPE, ( T E M P { J J > , J J - 1 , 5 ) 6004 FORMAT M H O , 'UNABLE 0 DECODE THIS CARD ', /, 10X, A2 8 X , 5 F 1 0 . 3 , / , TYPE NOT STANDARD ' , / , / ,

30 31 32 33 34
35 36 37 38 39 40 41 42 43 44 45 46

GO TO 300 301 CCNTINUE i TEMPIJJ),JJ =U 5 ) WRITEI6. 6001 ) 6001 FORMAT { 1H , f LINE, 6 X , 4 F 1 0 . 4 , F 8 . 1 I. GO TO 306 IF ( TYPE .EG. LINE ) WRITE ( 6, 6005 ) 6005 FORMAT i 1H , f SHORT WE LDS ADDED AUTOMATICALLY TO LINE 306 CONTINUE CCNVERT TEMP INTO L I N E DATA SX = T E M P ( l ) SY * TEMPC2) EX = T E M P I 3 ) EY T E M P I 4 ) NSEG = TEMPJ5) XP1, YP1 = PROJECTED LENGTH OF 1 ELEMENT XP1 = ( EX - SX ) / N SEG YP1 * i EY - SY ) / N SEG XBACK, YBACK -= COORD I NATE OF P O I N T H A L F S T E P BEYOND XBACK = SX - X P 1 / 2 . YBACK = SY - Y P 1 / 2 . GO T O 3 0 4 IF (TYPE .EQ. LINEI FAKEX XP1 / 1000. FAKEY YP1 / 1000. I * I 1
WX( I ) FAKEX WYU) FAKEY X ( I ) = SX Y U i = SY

ABOVE1,/,/)

STAfU

156
ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

47 48 49 50 51 52 53

304

CCNTINUE 00 1001 . K * 1 = 1 + 1

I t

NSEG

54 55
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

W Y d ) = YP1 X ( I ) = XBACK K * X P I Y d ) YBACK K * YP1 1001 CCNTiNUE I F ( TYPE . E Q . LINE I 1 * 1 * 1 W X ( I ) = FAKEX W Y I U = FAKEY

wx( n = X P I

GO TU 305

71 72 73
74

Y d ) = EY 305 CONTINUE GO TO 3 0 0 302 CCNTINUE ( T E M P ((J J ) ,, J J = T E M P J J ) JJ WRITE ( 6 t 6002 ) , 6002 ) ELEMENT* t 3 X , 4 F 1 0 4 ) 6 0 0 2 FORMAT ( iH i C CONVERT TEMP TO ELEMENT DATA 1 * 1 * 1 X U ) = TEMPdi Y d ) = TEMPI2) WX( I ) = T E M P O ) W Y ( I ) = TEMP(4) GO TO 3 0 0 303 CONTINUE ( TEMP(JJ) , J J = WRITE ( 6 , 6003 ) 6 0 0 3 FCRMAT ( 1H , * L0A0 f , 6X , 4F10.4 ) I CCNVERT TEMP TO LOAD DATA

x m * EX

1,4 I

75
76 77 78 79 80 81

PX *

TEMP(1)

602

PY TEMP(2) PGLX = TEMP(3) POLY = T E M P U ) N = I WRITE ( 6 , 6 0 2 ) N , d , X { I ) , YC I ) , W X d ) , W I ! ) , Y I=IiN ) FCRMAT ( 1 H 1 , NUMBER OF WELO ELEMENTS = , 1 6 , / , f 1 COORDINATES AND PROJECTIONS , /,

2
82 83 C 603

(IX,

14, 4F10*4)

WRITE ( 6 , 6 0 3 ) PX, P Y , PGLX, POLY FORMAT I 1 H 0 , # P X = f , F l 0 . 4 , , PY= , F 1 0 . 4 , f 1 , POLX = ' , F 1 0 , 4 , V,PGLY = f , F 1 0 . 4 ) CALCULATE LENGTH OF EACH WELD ELEMENT

84 85
86 C 87 88 89 90 91 92 93 94 95 96 C 97 C 98 99 101

DO 101 W(U =

I * It N SORT < VtXi I)

* 2 + WYU) ** 2 )

102

CCNTINUE LOCATE CENTER OF GRAVITY OF WELD GROUP SUMX , 0 . SUMY = 0 . SUMW = 0 . DC 1 0 2 1 = 1,N SUMX = SUMX + X ( I ) * W d I SUMY * SUMY Y ( I ) * W ( I ) SUMW - SUMW> W( I ) CONTINUE XCG = SUMX / SUMW YCG = SUMY / SUMW CALCULATE MOMENT OF P ABOUT CG MPCG = PY * (POLX - XCG) - PX * (POLY CALCULATE* J J = 0. DO 1 0 3 I = 1,N

YCG)

100
101

DX(I)
DY( I )

* X( I) - XCG
* Yd) YCG

157
THIRD QUARTER / 1982

102 103 104

105 106 107 108 109 110 111 103

IX = W U ) / 12. * W Y H ) ** 2 W U J * 9Y < I ) ** 2 IY = W( i) / 12* * W X U ) ** 2 * W(I) * 0X(I) ** 2 J = J + IX + IY

CONTINUE FACTOR = XIC = - P Y Y I C = *PX ITER = 0 FPREV * CONTINUE J / C SUMW * MPCG ) * FACTOR + XCG * FACTOR YCG i.E+12

201

112 113 114 C 115

116 C 117 118 C 119 120 104

RECALCULATE DX, DY, D ANO DELTA FROM TRIAL CENTER DO 104 1=1,N DX<I) = X(I) - XIC DY<Ii YCI) - YIC CALCULATE LENGTH OF RADIUS VECTOR D(I) = SQRT { DX(I)**2 * D Y U J * * 2 I
CCNTINUE CALCULATE MCMENT OF P MP = PY * ( P G L X - X I C ) - P X * ( P O L Y - Y I C ) I F ( ITER . G E . 1 I GO TO 2 0 4 GET THE ELASTIC SCLJTION BIGD = 0 * SUMWD2 = 0 .

121
122 123

DC 107 I * 1 . N
lfiQ(l) BIGD = 203 .LE. Oil) BIGD ) GO TO 2 0 3

124
125 126 127 128 129 130

IBIGD I
CONTINUE SUMWD2 * SUMWD2 f M ( I ) * O U ) * * 2 107 CCNTINUE ELPMAX^ABS i SUMWD2 * FALLOW / ( MP * BIGD ) I WRITE ( 6 , 6 0 4 1 ELPMAX, I B I G D 604 FORMAT i 1 H 0 , ELASTIC VALUE FOR MAXIMUM LOAD I S % F 1 0 . 3 t 1 10X MULTIPLY T H I S BY NUMBER OF SIXTEENTHS , / , 2 1 0 X , f AND BY ALLOWABLE KSI / 2 1 . 0 ' , / 3 1H , ' C R I T I C A L ELEMENT I S NUMBER f , 14, / ) 204 CONTINUE DLDMIN ^ l . E + 1 2 DO 1 0 5 I = 1,N C CALCULATE ANGLE BETWEEN RADIUS VECTOR AND WELD AXIS PHl = i A R C G S U B S U W X U ) * D X C I ) W Y U ) * D Y U ) ) / 1 4W(I) * D M ) ) ) ) ) * 180. / 3.1415927 C CALCULATE ANGLE BETWEEN FORCE AND WELC A X I S T H E T A ( I ) = 9 0 . - PHI CALCULATE PRELIMINARY DELTA VALUES DELTA = 0 . 2 2 5 * ( THETACI) 5 . 0 ) * * ( - 0 . 4 7 ) DELD DELTA / D i l i C FIND SMALLEST DELTA/D I F { DELD .GE. DLDMIN ) GO TO 105 DLDMIN = OELD DMAX D ( I ) BIGDEL = DELTA 105 CONTINUE

131 132 133 134

135 136 137 138 139 140 141 142

143
144 C 145 146 147 148

BIGR -

0.
R, M

SUMM = 0 . CALCULATE REVISED D E L T A , MU, LAMBDA, RULT, DO 1 0 6 I * 1, N DELTA = BIGDEL * D ( I ) / CMAX MU = 7 5 . * EXP ( 0 . 0 1 1 4 * T H E T A ( I ) ) LAMBDA 0 . 4 * EXP C 0 . 0 1 4 6 * T H E T A U ) )

158 ENGINEERING JOURNAL / AMERICAN INSTITUTE OF STEEL CONSTRUCTION

149 150 C 151 i52 153 154 155 156 157 C 158 159 Z 160 161 162 163 164 165 166 167 168 169 170 171 172 173

2 02

106

108

RILT * I 1 0 . + THETAC I I ) / I 0 . 9 2 * 0 , 0 6 0 3 * T H E T A ( I ) ) Ril) * RULT * I 1 . - EXP i -MU * DELTA ) ) * * LAMBDA CALCULATE LARGEST R IF ( R i l l . L E . BIGR J GO TO 2 0 2 BIGR = Rill JBIG = I CONTINUE M * R ( I ) * D ( I ) * W<It SUMM=SUMM + M CCNTINUE CALCULATE RUNIT DUE TC U N I T LOADt CALCULATE PULT RUNIT * - M P / SUMM PULT = ABS ( SUMM / MP ) CALCULATE RX/P AND R Y / P SUMRX * 0 , SUMRY C. DO 1 0 8 I*lfN RX = - D Y I I J / D I D * R ( I ) * RUNIT * M( I I SUMRX SUMRX + RX RY = 0X< I-> / 0 ( 1 ) * R U I * PUNIT * U ( I ) SUMRY = SUMRY RY CCNTINUE FX = PX + SUMRX FY = PY + SUPRY

F = SQRTi ?X**2 601

FY**2

174 175 C 176 177 178 179 180 181 182 183 C 184 185 186 187 188 189 190 191 205

206

207
60 5

ITER = ITER * 1 WRITE ( 6 , 6 0 1 1 I T E R , X I C , Y I C , F X , F Y , F , PULT, J B I G , BIGR FORMAT I 1 H 0 , AT TRIAL CENTER NO. , 14, / , 1 i O X t X* f , F 1 0 . 4 , Y= , F10.4, /, 2 2 0 X , *FX * f , F 1 0 . 4 , f FY - . * , F 1 3 . 4 , / , 3 3 OX, F * f , F10.4, / 4 2 U X , 'APPROXIMATE ULTIMATE LOAD = , F 1 0 . 3 , /, 5 1 0 X , ' C R I T I C A L ELEMENT I S NUMBER , 14, / , 6 1 0 X , CN WHICH ULTIMATE FORCE PER INCH IS F 1 0 . 4 , / ! I F C ABSiFI *LE. 0 . 0 1 } GO TO 2 06 I F { ITER . G E . 30 ) GO TC 2C8 TRY NEW CENTER I F i F . L T . FPREV ) GO TO 2 0 5 FACTOR = FACTOR / 2 . CONTINUE FPREV = F XIC X I C - FY * FACTOR Y I C Y I C * FX * FACTOR GC TO 2 0 * CONTINUE CCNVERT PULT TO 1 / 1 6 WELD OF E70 ELECTRODE PMAX = PULT / 4 . * 7 0 . / 6 0 . INTRODUCE SAFETY FACTOR PMAX a o . 3 * PMAX R L I M = FALLOW / 0 . 3 * 4 . 0 * 6 0 . / 70 I F i BIGR . L E . RLIM ) GO TO 207 PMAX = PMAX * RLIM / BIGR CONTINUE WRITE ( 6 , 6 0 5 ) PMAX FCRMAT ( 1 H 0 , INSTANTANEOUS CENTER HAS BEEN LOCATED , /, 1 1 0 X , MAXIMUM PERMISSIBLE LOAD* IS , F 1 0 . 3 , / , 2 1 0 X , MULTIPLY T H I S BY NUMBER OF SIXTEENTHS f , / , 3 1 0 X , f AND BY ALLOWABLE KSI / 2 1 . 0 , / )

192
193 194 195 208 606

GO T 2 0 0 O
WRITE (6606> ITER ITERATIONS = # 15,f...TRIALS TERMINATED I FCRMAT { 1 H 0 , GO T3 2 0 0

196
$DATA

END

159
THIRD QUARTER / 1982