]\gt ; 1888 .
] roblems on the Cross-sections of Pipes , sharpened Hamiltonian Numbers , . . . .
and that consequently the relation\mdash ; Hi , . . . .
may be written in the form\mdash ; 3 .
4 .
5 .
6 . . . . .
( 2 .
) . .
The comparison of of with that given by ( 1 ) furnishes us with an equation which , after several reductions have been made in which special attention must be paid to the order of the quantities under consideration , ultimately leads to the determination of the values of , . . . .
in succession .
III .
' Hydraulic Problems on the -sections of Pipes and nnels.\ldquo ; By HENRY , Professor of Applied Mathematics and Mechanism in the Royal College of Science for .
Received March 14 , 1888 .
In that division of hydromechanics which is devoted to the investigation of the flow of liquids through pipes and open channels , the resistance due to the friction of the contained liquid against the sides of the pipes or channels has led to expressions for the velocity as a function of the dimensions and shape of the cross-section commonly designated as the hydraulic mean depth .
This qua1ltity is defined as the quotient of the area of the crosssection of the liquid by that part of its perimeter in contact with the pipe or channel .
In a full pipe this perimeter is identical with that of the pipe 's cross-section , and in practice this is generally a circle .
It is ako proved from ) Calculus of Variations that a circle is the closed curve which , under a given perimeter , has the largest area , and by the same processes of analysis a segment of a circle appears to be that which includes the greatest area between its arc and its chord .
If we call the hydraulic mean depth of a pipe or channel bounded by a curved outline , its definition gives the condition Prof. H. Hennessy .
Hydraulie Problems on [ Apr. 19 , the limits of the integrals are taken between the same points on the curve .
If is given , then the problem is to find the curve which makes fydx maximum for the given value of .
This is a well-known isoperimetrical problem* , for by the principles of the Calculus of Variations we have in this case\mdash ; here is arbitrary , and therefore which gives ' and the equation of a circle with radius .
This result proves that for a full pipe the circle gives the greatest hydraulic mean depth , but it does not tell what is the particular arc of a which gives the greatest quotient for the area of the segment between itself and its chord divided by itself .
This is best done by the ordinary methods of maxima and minima as follows :\mdash ; Let represent the angle subtended at re by the segment of the circle radius in , then\mdash ; his 'History of the Calculus of Variations , ' p. 69 , Todhunter has made a remark on this problem ; namely , that if the curve instead of being closed were required to pass through two given fixed points with the arc between these points of a given length , the constants of integration would not be arbitrary , and there would be two equations from the fact of the circle )assing through the given points and another arising from the given length .
solution here g1ven avoids the ne essity of such equations by employing the well-known properties of an arc of a circle and its included segment.\mdash ; March 29 , 1888 .
1888 .
] the of Pipes and Channels .
When this may be satisfied either by , or by some arc between and .
The root , substituted in the value of , makes this positive and equal to , as may be eatiily shown by expanding and .
Let now , and by successive trials we shall find that nearly satisfies the cquation With this value is and are ative and is also negative , showing that the result gives a maximm for which in this case becomes , nearly .
The hydraulic mean depth of a full pipe or of a half-full pipe of circular section is , hence that for a section less by about three- tenths of the perimeter of the circle is greater .
The area of the section of greabest hydrftulic mean depth is or of the entire .
If the pipe is horizontal the of liquid contained in it is ) to the cross-section , hence a circular pipe under snch condition has greatest hydraulic mean when it is nearly seven-eighths full ; or when the liquid has fallen from the state so as to its surface AB the of an arc of .
The versed sine of this arc is nearly , being the FIG. 1 .
S 104 Prof. H. Hennessy .
Hydraulic Problems on [ Apr. 19 , diameter , so that for a pipe of 2 feet internal diameter the greatest hydraulic mean depth would be when the surface of the liquid had fallen below the top by , or nearly inches .
As the velocity of the liquid is nearly as the square root of the hydraulic mean depth , the pipe filled to this height would carry liquid with a velocity slightly greater than when completely full .
This conclusion is only when the effective head of liquid is due solely to the inclinatio1l of the pipe .
When ttJe level of the liquid within the pipe falls the hydraulic mean depth tends towards its minimum value , and its decrease becomes rapid as the arc diminishes ; thus if is a very small angle .
But , where is the length of an arc of a semicircle ; hence if the 4th power of is negligible we have Although pipes and conduits for water supply are usually quite full , those for drainage purposes are most commonly only partly filled with liquid , and the amount of liquid is liable to fluctuations .
This has led to the adoption for drainage pipes of an oval curve for the outline of cross-section , with the longer axis of the oval vertical and terminated at bottom by an arc of greater curvature than at top .
The form of this cross-section suggests an inquiry as to how far a curve which has been often treated in isoperimetrical problems would satisfy the conditions for giving a favourable hydraulic mean depth in an open channel with fluctuating contents .
We have seen that a particular arc of a circle gives a maximum for the of the area of the segment divided by the perimeter of ths arc , and we shall find that there is a particular catenary which gives a maximum for a corresponding quotient of the area included between its perimeter and its chord .
If as usual we make the directrix the axis of the parameter , and the length of the curve , then adopting the usual notation , and but in this case , as the area whose quotient divided by the perimeter is to be a maximum is the snce between the rectangle under the coordinates and and the area included bebween the curve , its parameter , and the directrix , we have manifestOy\mdash ; and as , this may be written 1888 .
] the Cross-sections of Pipes and Channels . .
The shape of the curve depends on the rolation veen .
its pnrameter and its lellgtb , heuce we must find the value of which nlakcs a maximum in the above expression .
The problem seems therefore to amount this elementary statical question :lexibl .
and chain is attached to two supports on the same horlzontal line ; required the distance between the supports so as to make the of the surface included between the chain and the zontal l the greatest possible ; or given the perimeter of a catenary to find the chord , so that the area between itself and the culve shall be a maxlmum .
The above expression gives\mdash ; If we , and make , we have By successive tl.ials this equation may be satisfied by substituting , whence .
This value substituted in the expresslon for gives a negative result , and therefore the value of is a maximum when .
When , and nearly .
With further we ] find therefore\mdash ; nearly , Prof. H. Hennessy .
draulic Pr.oblems on [ Apr. 19 , But the depth of the curve below its chord is or In this inquiry is the perimeter of the half curve , so that the total perimeter , the chord , and depth are respectively in the ratio of the numbers 3 , 2 , and 1 , or the chord of the catenary of greatest area for a given perimeter is twice the depth , and the length of the curve is three times the depth .
The outline of this curve is readily shown by attaching a fine chain of 3 units of length to supports at a distance FIG. 2 .
The chord AB The arc ADB of the catenary of 2 units The catenary which would give a maximum hydraulic mean depth for an open channel is therefore one whose depth is the radius and chord the diameter of a semicircle .
On substituting the value of found above in the equation for , we shall find that the mean depth of the catenary under consideration is nearly or , where is the total perimeter of the curve .
In a semicircle the hydraulic mean depth is , or nearly , hence the hydraulic mean depth of the catenary of maximum area is nearly equal to that of a semicircls of equal perimeter .
But a channel formed by the outline of such a catenary would when the iquid falls , not be liable to so rapid a reduction of hydraulic mean depth as in the semicircle .
For small arcs of a circle it has been shown that this is proportional to the square of the ] subtended at the centre .
In the catenary if is the an made by the tangent with the directrix , it is also the angle made by the radius of curvature with the axis of , which in this case coincides with the axis of depth , and as a a , 1888 .
] the Cross-sections of Pipes hannels .
' when and higher powers al.e omitted ; and remembering , we may write such an arc\mdash ; .
An arc of the semicircle at its base subtending the angle when is small the value , as already pointed out .
Hence for a circular channel and for one formed by a catenary of equal perimeter and maximum area , the hydraulic mean depth for small segments subten ding equal angles would be greater for the latter .
On looking at the outline of such a catenary inscribed in a semicircle , this resuIt seems to be confirmed , and the curve approaches the oval which experience has led engineers to adopt for the section of pipcs carrying fluctuating quantities of liquid .
The general result of the .
inquiry may be summed up in the following conclusions : all pipes and conduits employed to convey liquid for consumption or for milling , the circular f ; ection is best , as the level of the liquid in the pipe is rarely , if ever , below lf the diamete ] ' .
drainage such a folun is also the best if the liquid rarely falls half the diametel , but if it is ]iable to fnll nearly to the bottom of the pipe or conduit , an oval f , such as that actually recommended , is the best .
If the pipe is likely to bc as often half full as slightly filled , it is probable that some advantage would be gained by employ .
the catenary of maximum for a given perimeter for the lowel ' part of the oval .
A pattern for this form can be always readily constructed by remembering the reIatious 1 , 2 , 3 for the depth , -he chord , and the of the c1lrve .
In designing the base of the pipe , it is only necessary , as already pointed out , to hang a fine of 3 units between supports placed at 2 units on the same horizontal line .
It is.well known that in a ular notch or triallgulal .
channel , the sidcs of whioh are at angles , the velocity of the liquid varies but little with the depth , and it is possible to conceive that a channel may have such a form as to make such a variation extremcly small .
If wo suppose the surface of the liquid in an open channel to be bounded by the of the cross-scction of the anncl , we shall have as before the hydraulic mean depth\mdash ; Problems on the Cross-sections of Pipes , .
[ Apr. 19 , and if we make constant\mdash ; ( fydx ) the limits of the rals in both cases being taken on the same points of the curve .
From this it follows that\mdash ; which on integrating gives This result indicates a catenary with its convexity turned to the chord and to the axis of , but between the limits and the value of becomes inary , the constant being the hydraulic mean , which must be very small in such a case as here supposed , if we take from to , and such a notch or channel might be approximately realised by two arcs of a catenary with ameters corresponding to the small arbitrary value of FIG. 3 .
A notch or channel with such a cross-section would have an almost coTlstant hydraulic mean depth , but it would be inapplicable to any useful purposes in the application of hydraulics .
The cross-sections of rivers and navig , ble canals are regarded chiefly with reference to permanence , the question of their anlic mean depth is less important than in the case of water 1888 .
] On the Heating ffects of supply and pipes .
In canals the trapezoidal section is that which experience has almost universally established as tho best wherever canals are carried ordinary earth , and the rectangular section is only adopted when the sides are composed of coherent such as rock or masonry .
The semicircular section for an open channel would not approximate to the shapes usually adopted in canals , but ib may be worth .
that tlJe outline of the of greatest area approaches more nearly to such shapes . .
' On the Heating Effects of Electric Currents .
No. III By .
H. PREECE , F.R. .
Received March 15 , 1888 .
I have a great deal of pains to verify the dimensions of the currents , as detailed in my paper read on December 22 , 1887 , equired to fuse different wires of such thicknesses that the law is strictly followed ; and I snbmit the following as the final values of the constant " " \ldquo ; for the different metals:\mdash ; Inches .
Centimetres .
Millimetrcs .
Copper . . . . . . . . . . . . . . . . . .
10 , 2,530 .
7,585 1,873 PIatinum . . . . . . . . . . . . . . . .
5 , 172 Gcrman silver . . . . . . . . . . . .
5:2.30 1,292 Platinoid . . . . . . . . . . . . . . . .
4 , 1,173 Iron . . . . . . . . . . . . . . . . . . . .
3,148 Tin . . . . . . . . . . . . . . . . . . . .
1,642 Alloy ( lcad and tin to 1 ) 1 , Lead . . . . . . . . . . . . . . . . . . . .
1 , : ; 79340 .
With these constants I have calculated the two ving tables , which I hope will be found of some use and value:\mdash ;

