IV .
" On the Stability of Domes .
" By E. WYND)HAM TARN , M.A. Communicated by GEORGE GODWIN , Esq. Received May 5 , 1866 .
The few writers who have attempted to treat the subject of the Equilibrium of Domes mathematically , have entirely failed to obtain results that are of any practical use to the architect .
Their failure has arisen from taking a too theoretical view of the subject , and endeavouring by mathematical reasoning to find the form which a dome ought to have in order that it might stand safely .
Such a question is of no practical utility , as domes of various sizes and forms have been erected for centuries past , and the question for the architect is , -given a dome of certain form and size , what are the conditions which must obtain between it and the wall of the building it is intended to cover , in order that the whole structure may be in a condition of stability ?
The object , therefore , of this paper is to find a solution to the following problem : Given a spherical dome , built of stone or brick , of any radius and thickness , and standing on a " drum " or walls of any height ; to find the thrust of the dome on the " drum , " and the thickness that must be given to the walls in order to insure the stability of the structure .
I take the case of the dome having a spherical section , as being the form most commonly used ; but the same method of investigation will apply to domes of any form .
In this investigation I shall consider the dome as made up of a large number of arched ribs , of which the bases resting on the top of the " drum " subtend a small angle ( 9 ) at the centre , and the vertices have no thickness ; each rib having the form of a wedge cut out of the spherical shell by two planes intersecting in a vertical line through the centre , and making the small angle 0 with each other .
I shall then consider that the two wedges thus formed on opposite sides of the dome , thrust against each other at the vertex , as in the case of an ordinary semicircular arch , and by this means keep each other in equilibrium .
Of course no arch of this form if built alone could stand for a moment , as it would give way laterally ; but in the dome this is prevented by the parts on each side of the rib under consideration .
This method is adopted by Venturoli in treating on the equilibrium of domes .
Fig. 1 represents a rib of the form above described cut out of the dome , and the corresponding part of the " drum " which sustains it .
The arch will have a tendency to fall in at the crown C D , and open outwards at the haunches E F , causing the joint CD to open at D , and that at OF to open at F. We may therefore suppose that N , the thrust of the corresponding rib on the opposite side , acts at C. We shall now find the effect of the force N upon a given joint OF .
Let P be the weight of the portion of rib above OF ; x the perpendicular distance from E of a vertical from the centre of gravity of P ; y the vertical distance of CN from E. Then the moments of these forces about E are Px and Ny ; and in order that there may be equilibrium , we must have N equal to the greatest value of P. e. We have therefore to express Px in gY terms of 0 , the angle which FE makes with the vertical , and find what value of 0 makes Pa maximum .
Let r be the internal and R the external radius of the sphere , 8 the weight of a cubic foot of the material of which it is composed .
Then the volume of any portion of the rib ( p qrs CD ) is V= !
fsinO.de.dr .
; ... ... ... ( A ) so that P= .
V=3 .
q ( 1 -cos 0 ) R 7 , taking limits from 0=0 .
3§ 1866 .
] 183 Ifgis the centre of gravity of the portion of the rib whose weight is P , and G its projection on OA ; then if OG=z , we find the value of z from =SSSr3 .
sin 0 .
cos. dr .
d. dp SSSr2 .
sin 0 .
d. dO sinl SSr3 ( 1-cos 20 ) dr. .
d --3 ' g.r2..sin 0 .
dr .
dO( And since q is small , we may consider 1 --_ , nearly ; in which case 3 R4--r40sin 20 81 3-r 1-coso Now x is the perpendicular distance of gG from E ; therefore --=r sin 0 .
R-3 3 O-I-r4 -sin 20 Therefore Px=.p ( 1 -cos 0 ) t rsin 0R4-sin 20 3X 8t_r-r l-cos0 J or PxJ 8r ( R3-r3 ) sin0 ( 1 -cos 0)-3 ( R4-r4 ) ( 0-I sin 20 ) .
And since y=R-1r cos 0 , and Np - , y N. 8r ( R3-r3 ) sin 0(1-cos 0)-3 ( R4--r ' ) ( 0-sin 20 ) 24 R-r cos 0I will take q as the circular measure of an angle of 2 ?
, or -7r ='0349066 , in which case we have N= 001454 Sr(R3-r3 ) sin 0 ( 1 -cos 0)-3 ( R4--4 ) ( 0-sin 20 ) R1-r cos 0 ='001454 x N ' , say .
We have now to find what value of 0 will make N ' a maximum , and as the ordinary rules for maxima will not apply to this expression , we must find it by calculating N ' for different values of 0 .
I have done this for the case when r2 10 , R= 11 , and find that N is greatest when 0= 70 ?
; thus 0=69 ?
... ... .
-506'034 0=70 ?
... ... ... .
N'=506-241 071 ... ... ... .
N'=506-018 .
We have now to substitute in N the values of 0 , sin 0 , &amp;c. when the angle is 70 ?
, or when 0=35 9= '22173 , sin 0=-93969 , cos 0= 34202 , sin 20 90 = '64279 .
This reduces the expression for N to N=_ 007192r ( I3--r)-0039273 ( Rl-r4 ) --4 t02 We can now transpose N and P to the point E ; and in order to find the thickness of the pier , we proceed to take their moments , together with those of the part of the rib below OF , and the pier itself , about the outer bottom edge S of the pier .
If we call yS= b , the perpendicular distance from S of the direction of N , then b=H+r.cos0 , ... . .
( 2 ) where I is the height RS of the pier or " drum .
" Let Se=a , the distance from S of a vertical dropped from E ; F=the weight of the portion of the rib below OF ; c=the perpendicular distance from S of a vertical from the centre of gravity of the lower portion of the rib whose weight is F ; Q=the weight of the portion of the " drum " on which the rib stands , and q=the distance from S of a vertical from its centre of gravity ; we have then ( in equilibrium ) the equation N.b=P.a+Q.q+F.c , ... . .
( 3 ) which I will call the Equation for Equilibrium .
From this equation we can find the thickness ( t ) of the pier necessary to produce equilibrium .
In order , however , that we may have stability in the structure , we must multiply Nb by the coeffcient of stability , which we may take as 2 .
The equation from which to find the value of t then becomes 2N .
b=P.a+Q.q+F .
c , ... ... ( 4 ) which I will call the Equation of Stability .
I now proceed to find the values of P. a , Q. q , and F. c. The value of P was previously found to be P=B .
1-cosO ) R--r ; and if we substitute for q and 0 their values , P= -007656 ( R3-r3 ) .
Also a=Se=t+r(l-sin 0)=t+'06031 r ; so that we find for the value of P. a , Pa= 007656 6 ( R3-3 ) t+ 0004618 a. ( R3-r3 ) ... .
( 5 ) The value of F is found by means of the integral ( A ) , taking the limits from 0=0 to 0=-- ; which gives us 3__ -r3 F=B..cos O3r ==00398 a ( R3r ) .
Let z1 be the distance from O on OA of the perpendicular dropped from the centre of gravity of the portion of the rib whose weight is F. Then the integral ( B ) gives us , by taking the same limits of 0 as for F , --O+ sin201 3 R4 r4 220 18 lt3-r3 cos 02 And if we substitute for 0 its value , RI4 _r4 z=-*7351 a And c=OR-z1=-t+r-- , .
Hence we find F. c= -00398 a ( R3-r3 ) t+ 00398 3 { r ( R3-r3)-7351 ( R4-r4 ) } .
( 6 ) We have now to find the moment about S of the weight of the pier .
Let Q be the weight of the pier , a , the weight of a cubic foot of its material ; then Q= ? .
1 ( ( r+ t)2-r2 ) .
H= ( 2r+t ) t. H =-017453 a ( 2r +-t ) t. H. Let figure 2 represent the plan of the pier , g its centre of gravity , ST=t its thickness .
Then by the integral calculus we find 4 ( r+ t)-3 r3 sin ?
0 sin ?
I_ O3( +)2X7 ; or putting =3 ( r $t)2 _ r2 72 _2 ( r+ t)3-r__2 ( 3'+ 3 ?
t+ t2 ) .
3 ( 2.+ t)2-.2 3 ( 2r+ t ) q=Sg-=OS-Og=r+ t-Og =3 ( rt ) ( 2r+t)-2 ( 3r2+3rt+ t2 ) 3 ( 2r-+ t ) 3rt + t2 3 ( 2r+t ) We therefore obtain for the value of Qq , Qq=-017453 1(2r+ t ) t. 3r+ tt 3 ( 2r ?
t ) =*0058177 .
t2 ( 3r+ t ) ... H. ( 7 ) We can now , by adding together the several quantities ( 5 ) , ( 6 ) , ( 7 ) , and equating to the value of N6 as found from ( 1 ) and ( 2 ) , form the equation for equilibrium ( 3 ) , which will be a cubic equation in respect of t. We shall thence find the value of t necessary to produce equilibriumm in the structure , and , by equating the same quantities to 2Nb , form the equation for stability ( 4 ) , from which we get the value of t necessary to produce stability in the structure .
The following examples will serve to show how readily these formulm can be applied , and the use which the practical architect may make of them .
Example 1:Let r= 10 ft. , R=1 ft. , -= 125 lbs. , = 150 lbs. , H= 50 ft. From ( 1 ) ... N= 920092 ; , ( 2 ) ... b= 53-4202 ; therefore Nb =4915 , and 2Nb =9830 .
From ( 5 ) ... P.a= 316-767 t +191-06975 ; , , ( 6 ) ... F. c= 164-6725 t-5054 ; , ( 7). .
Q.q= 1308-98 t2+ 43-63275 t. Hence the equation for equilibrium ( 3 ) is 4915=43-63275 t3+ 1308-98 t2 481 4395 t+ 14052975 , which reduces to t3 + 30t2-+1 103t--109'42-0 .
We can solve this by means of Horner 's process , and find t= 1-7 ft. for equilibrium .
The equation for stability ( 4 ) becomes 9830=43-63275t3+ 1308-98t2+481 4395t 4140'52975 , which reduces to 430t2+1 103t-222-069=0 , from which we obtain , by Homer 's process , t=2'45 ft. for stability .
Example 2:Let r=30 ft. , R=33 ft. , H=80 ft. , 6 and a , as before .
From(1 ) ... N= 2483-84 ; , ( 2). .
b= 90-2606 ; therefore Nb =224192-8 , and 2Nb =448385-6 .
From ( 5 ) ... P. a= 8552-71 t+415476-65 ; , , ( 6 ) ... F.c= 4446-16 t+ 4094-17 ; ( 7 ) ... Q.q= 69-8124t3+ 6283-116t2 .
Adding and reducing , the equation for equilibrium becomes t3 +90t2+ 186'2t-3048-31 =0 , whence t=4-83 ft. for equilibrium .
And the equation for stability is t3+ 90t2 + 186-2t-6159'7=0 , whence t=7-07 ft. for stability .
Example 3:I will apply the formulae to the case of the dome of Sultan Mohammed 's tomb at Beejapore , described in Mr. Fergusson 's 'Handbook of Architecture .
' One peculiarity of this structure is that the inner face of the dome is set 5 ft. 6 in .
within the inner face of the wall on which it rests , as shown in fig. 3 .
; so that in calculating the values of a , &amp;c. , we must add 5-5 ft. to the thickness t ; hence we shall get a proportionately less value for t from the resulting equations .
In this dome , r=62 ft. , R=72 ft. , H=-=116 ft. ; and we will suppose , as before , that 3=125 lbs. , -=150 lbs. From ( 1 ) ... .
N= 31132 ; , , ( 2). .
b= 137-20524 ; therefore Nb =4271473-522 , and 2Nb =8542947-044 .
From ( 5). .
P. =957 ( 134920 ) ( t +55)+3-56745 ( 134920 ) =129118'44 t+1191511-78 ; ( 6). .
F. c=-4975 ( 134920 ) ( t+5 5)--4975 ( 527854-4 ) =671227t + 106567-28 ; , , ( 7 ) ... Q.=10122798 3+186xl 01 22798t2 .
Adding and reducing , the equation for equilibrium ( 3 ) becomes t3+ 186+t2 1938-6t-29373-24=0 , whence t=8-2 ft. for equilibrium .
And the equation for stability ( 4 ) becomes t3+ 186t2+ 1938'6t-71570=0 , whence t= 14-66 ft. for stability .
In the construction of this particular dome , the values obtained above for t will only apply to four points on the walls which carry the dome ; for instead of being built on a circular " drum , " this drum is made to cover a square chamber by means of " pendentives , " which are so arranged as to throw the thrust principally on the angles of the building , where the values of a and c will become very great , and it is only at the centre of each of the four sides that the above equations can be employed to compare the thrust of the dome with the strength of the walls .
At these four points the walls are about 11 ft. thick , so as to be considerably more than sufficient to produce equilibrium ; and the coefficient of stability at these points is 1-374 instead of 2 .
Had the dome therefore been built on a circular drum of 11 feet thickness , in all probability the edifice would not have stood for any length of time .
Much of the thrust of a dome may be counteracted by means of an iron belt placed round it at the point where the thrust is greatest .
This point I have shown to be at 70 ?
from the crown , or 20 ?
from the springing .

