Page images

sions are insulated in symmetrical positions within a cylindrical conducting sheath of circular section.

these three equations by k, s, and respectively ; that is, to express



[ocr errors]
[ocr errors]


CASE I.-Two-wire Cable. In the general equations (according to the notation of the first part of this communication) we have k,=kg; w,(=,(2); and w, =w 2): and it will be convenient now to denote the values of the members of


e by k the galvanic resistance in each wire per unit of length, by c the electrostatical capacity of each per unit of length when the other is prevented from acquiring an absolute charge, and by f the proportion in which this exceeds the electrostatical capacity of each when the other has a charge equal to its own; or in other words, to assume c and f so that



92 if v, and v, be the potentials in the two wires in any part of the cable where they are charged with quantities of electricity respectively 9 and q, per unit of length. The equations of electrical conduction along the two wires then become

1 (do dt kc | dx?

(2). dv, 1 d'v,

If dt

+ kc dx?

From these we have, by addition and subtraction,

1+f deg dw
1-f dew

kc dix?'

dt kc dx2 where og and w are such that v=9+w, v.=9-w

(4). If both wires reached to an infinite distance in each direction, the conditions to be satisfied in integrating the equations of motion would be simply that the initial distribution of electricity along each must be whatever is prescribed ; that is, that 0,=0.(r), and v,=02(x)

(5), when



[ocr errors]

+ f dat






Q, and o, denoting two arbitrary functions. Hence, according to Fourier, we have, for the integrals of the equations (3),

[ocr errors][subsumed][ocr errors][ocr errors][ocr errors][merged small][ocr errors][merged small][ocr errors][ocr errors]

and the solution of the problem is expressed in terms of these integrals by (4).

If now we suppose the cable to have one end at a finite distance from the part considered, for instance at the point O from which x is reckoned, and if at this end each wire is subjected to electric action so as to make its potential vary arbitrarily with the time, there will be the additional condition

v,=4,(t), and v,=42



to be fulfilled. In the other conditions, (5), only positive values of x have now to be considered, but they must be fulfilled in such a way as not to interfere with the prescribed values of the potentials at the ends of the wires; which may be done according to the principle of images, by still supposing the wires to extend indefinitely in both directions, and in the beginning to be symmetrically electrified with contrary electricities on the two sides of 0. To express the new condition (7), a form of integral, investigated in a communication to the Royal Society (* Proceedings,' May 10, 1855, p. 385), may be used; and we thus have for the integrals of equations (3),

[ocr errors][subsumed][subsumed][subsumed]
[ocr errors]


[ocr errors]

Lastly, instead of the cable extending indefinitely on one side of

the end o, let it be actually limited at a point E. If the ends of the
two wires at E be subjected to electric action, so as to make each
vary arbitrarily with the time, the new conditions to be satisfied, in
addition to the others, (5) and (7), will be

[merged small][merged small][merged small][merged small][merged small][ocr errors]

if X, and X. denote two arbitrary functions, and a the length OE.
Or, on the other hand, if they be connected together, so that a cur-
rent may go from 0 to E along one and return along the other, the
new conditions will be

[merged small][ocr errors][merged small][merged small][merged small][merged small][ocr errors]

Either of these requirements may be fulfilled in an obvious way by the method of successive images, and we so obtain the following respective solutions :

[ocr errors][subsumed][ocr errors][merged small][ocr errors][subsumed][ocr errors][subsumed][merged small][subsumed]

where F, F, E, E denote for brevity the following functions :

[merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][ocr errors][merged small][merged small][ocr errors][ocr errors][subsumed]

Each of the functions F and E is clearly the difference between two periodical functions of (&-x) and (8 + x); and each of the functions F and E is a periodical function of x simply. The expressions for these four functions, obtained by the ordinary formulæ for the expression of periodical functions in trigonometrical series, are as follows:

[subsumed][merged small][ocr errors][subsumed][ocr errors][subsumed][merged small][subsumed][ocr errors][merged small][subsumed][merged small][subsumed][subsumed][merged small]

Either (11) or (12) may be used to obtain explicit expressions for
€ the solutions (10) and (10)', in convergent series ; but of the series

so obtained, (11) converge very rapidly and (12) very slowly when t
is small; and, on the contrary, (11) very slowly and (12) very rapidly
when t is large. It is satisfactory, that, as 4 increases, the first set
of series (11) do not cease to be, before the second set (12) become,
convergent enough to be extremely convenient for practical compu-

The solutions obtained by using (12), in (10) and (10)', are the
same as would have been found by applying Fourier's ordinary pro-
cess to derive from the elementary integral e-mt sin nx the effects of
the initial arbitrary electrification of the wires, and employing a

method given by Professor Stokes* to express the effects of the variations arbitrarily applied at the free ends of the wires.

CASE II.-Three-wire Cable. The equations of mutual influence between the wires may be clearly put under the forms

cv,=9. +f(92 +93), cv,=9z+f(93+1), cv,=9z+f(+92); and the equations of electrical motion along them are then as follows:

d 9: dʻ93

dx2 da

dt dx2 dx?

191, d'92
dt - dx2



[ocr errors]

kc dqi_ dq



[ocr errors]

ke da


t dix2

[ocr errors]

ko da


If we assume
o=91+9+93, w,=291-92-93, w,=292-93-91, ws=293-9-92)
which give

Win 922


and require that w, +w,+w,=0, we find by addition and subtraction, among the equations of conduction,

do = (1+20)



doo dx

[blocks in formation]

where for w may be substituted either W., W.,, or wg.

CASE III.-Four-wire Cable. The equations of mutual influence being

co,=4.+f(92 +9) +993, and other four symmetrical with this; and the equations of motion,


dt dx2 d.x2 dr d.x

[ocr errors]

ko da


[blocks in formation]

* See Cambridge Phil. Trans. vol. viii. p. 533, “On the Critical Values of the sums of Periodic Series."

« PreviousContinue »