Chapter VII: Mains and Service-Pipes--Subsidiary Apparatus (1)

The process by which acetylene is produced, and the methods employed for purifying it and rendering it fit for consumption in dwelling-rooms, having been dealt with in the preceding pages, the present chapter will be devoted to a brief account of those items in the plant which lie between the purifier outlet and the actual burner, including the meter, governor, and pressure gauge; the proper sizes of pipe for acetylene; methods of laying it, joint-making, quality of fittings, &c.; while finally a few words will be said about the precautions necessary when bringing a new system of pipes into use for the first time.

THE METER.--A meter is required either to control the working of a complete acetylene installation or to measure the volume of gas passing through one particular pipe, as when a number of consumers are supplied through separate services under agreement from a central supply plant. The control which may be afforded by the inclusion of a meter in the equipment of a domestic acetylene generating plant is valuable, but in practice will seldom be exercised. The meter records check the yield of gas from the carbide consumed in a simple and trustworthy manner, and also serve to indicate when the material in the purifier is likely to be approaching exhaustion. The meter may also be used experimentally to check the soundness of the service-pipes or the consumption of a particular burner or group of burners. Altogether it may be regarded as a useful adjunct to a domestic lighting plant, provided full advantage is taken of it. If, however, there is no intention to pay systematic attention to the records of the meter, it is best to omit it from such an installation, and so save its initial cost and the slight loss of pressure which its use involves on the gas passing through it. A domestic acetylene lighting plant can be managed quite satisfactorily without a meter, and as a multiplication of parts is undesirable in an apparatus which will usually be tended by someone not versed in technical operations, it is on the whole better to omit the meter in such an installation. Where the plant is supervised by a technical man, a meter may advisedly be included in the equipment. Its proper position in the train of apparatus is immediately after the purifier. A meter must not be used for unpurified or imperfectly purified acetylene, because the impurities attack the internal metallic parts and ultimately destroy them. The supply of acetylene to various consumers from a central generating station entails the fixing of a meter on each consumer's service-pipe, so that the quantity consumed by each may be charged for accordingly, just as in the case of public coal-gas supplies.

There are two types of gas-meter in common use, either of which may, without essential alteration, be employed for measuring the volume of acetylene passing through a pipe. It is unnecessary to refer here at length to their internal mechanism, because their manufacture by other than firms of professed meter-makers is out of the question, and the user will be justified in accepting the mechanism as trustworthy and durable. Meters can always be had stamped with the seal of a local authority or other body having duly appointed inspectors under the Sales of Gas Act, and the presence of such a stamp on a meter implies that it has been officially examined and found to register quantities accurately, or not varying beyond 2 per cent. in favour of the seller, or 3 per cent, in favour of the consumer. [Footnote: It may be remarked that when a meter-- wet or dry--begins to register incorrectly by reason of old age or want of adjustment, its error is very often in the direction that benefits the customer, _i.e._, more gas passes through it than the dials record.] Hence a "stamped" meter may be regarded for practical purposes as affording a correct register of the quantities of gas passing through it.

Except that the use of unalloyed copper in any part of the meter where it may come in contact with the gas must be wholly avoided, for the reason that copper is inadmissible in acetylene apparatus (_see_ Chapter VI.), the meters ordinarily employed for coal-gas serve quite well for acetylene. Obviously, however, since so very much less acetylene than coal-gas is consumed per burner, comparatively small meters only will be required even for large installations of acetylene lighting. This fact is now recognised by meter-makers, and meters of all suitable sizes can be obtained. It is desirable, if an ordinary coal-gas meter is being bought for use with acetylene, to have it subjected to a somewhat more rigorous test for soundness than is customary before "stamping" but the makers would readily be able to carry out this additional test.

The two types of gas-meter are known as "wet" and "dry." The case of the wet meter is about hall-filled with water or other liquid, the level of which has to be maintained nearly constant. Several ingenious devices are in use for securing this constancy of level over a more or less extended period, but the necessity for occasional inspection and adjustment of the water-level, coupled with the stoppage of the passage of gas in the event of the water becoming frozen, are serious objections to the employment of the wet meter in many situations. The trouble of freezing may be avoided by substituting for the simple water an aqueous solution of glycerin, or mixture of glycerin with water, suitable strengths for which may be deduced from the table relating to the use of glycerin in holder seals given at the close of Chapter III. The dry meter, on the other hand, is very convenient, because it is not obstructed by the effects of frost, and because it acts for years without requiring attention. It is not susceptible of adjustment for measuring with so high a degree of accuracy as a good wet meter, but its indications are sufficiently correct to fall well within the legalised deviations already mentioned. Such errors, perhaps, are somewhat large for so costly and powerful a gas as acetylene, and they would be better reduced; but it is not so very often that a dry meter reaches its limit of inaccuracy. Whether wet or dry, the meter should be fixed in a place where the temperature is tolerably uniform, otherwise the volumes registered at different times will not bear the same ratio to the mass of gas (or volume at normal temperature), and the registrations will be misleading unless troublesome corrections to compensate for changes of temperature are applied.

THE GOVERNOR, which can be dispensed with in most ordinary domestic acetylene lighting installations provided with a good gasholder of the rising-bell type, is designed to deliver the acetylene to a service-pipe at a uniform pressure, identical with that under which the burners develop their maximum illuminating efficiency. It must therefore both cheek the pressure anterior to it whenever that is above the determined limit to which it is set, and deliver to the efferent service-pipe acetylene at a constant pressure whether all or any number of the burners down to one only are in use. Moreover, when the pressure anterior to the governor falls to or below the determined limit, the governor should offer no resistance--entailing a loss of pressure to the passage of the acetylene. These conditions, which a perfect governor should fulfil, are not absolutely met by any simple apparatus at present in use, but so far as practical utility is concerned service governors which are readily obtainable are sufficiently good. They are broadly of two types, viz., those having a bell floating in a mercury seal, and those having a diaphragm of gas-tight leather or similar material, either the bell or the diaphragm being raised by the pressure of the gas. The action is essentially the same in both cases: the bell or the diaphragm is so weighted that when the pressure of the gas exceeds the predetermined limit the diaphragm or bell is lifted, and, through an attached rod and valve, brings about a partial closure of the orifice by which the gas flows into the bell or the diaphragm chamber. The valve of the governor, therefore, automatically throttles the gas-way more or less according to the difference in pressure before and after the apparatus, until at any moment the gas-way is just sufficient in area to pass the quantity of gas which any indefinite number of burners require at their fixed working pressure; passing it always at that fixed working pressure irrespective of the number of burners, and maintaining it constant irrespective of the amount of pressure anterior to the governor, or of any variations in that anterior pressure. In most patterns of service governor weights may be added when it is desired to increase the pressure of the effluent gas. It is necessary, in ordering a governor for an acetylene-supply, to state the maximum number of cubic feet per hour it will be required to pass, and approximately the pressure at which it will be required to deliver the gas to the service-pipe. This will usually be between 3 and 5 inches (instead of about 1 inch in the case of coal-gas), and if the anterior pressure is likely to exceed 10 inches, this fact should be stated also. The mercury-seal governors are usually the more trustworthy and durable, but they are more costly than those with leather diaphragms. The seal should have twice or thrice the depth it usually has for coal-gas. The governor should be placed where it is readily accessible to the man in charge of the installation, but where it will not be interfered with by irresponsible persons. In large installations, where a number of separate buildings receive service-pipes from one long main, each service-pipe should be provided with a governor.

GASHOLDER PRESSURE.--In drawing up the specification or scheme of an acetylene installation, it is frequently necessary either to estimate the pressure which a bell gasholder of given diameter and weight will throw, or to determine what should be the weight of the bell of a gasholder of given diameter when the gas is required to be delivered from it at a particular pressure. The gasholder of an acetylene installation serves not only to store the gas, but also to give the necessary pressure for driving it through the posterior apparatus and distributing mains and service-pipes. In coal-gas works this office is generally given over wholly or in part to a special machine, known as the exhauster, but this machine could not be advantageously employed for pumping acetylene unless the installation were of very great magnitude. Since, therefore, acetylene is in practice always forced through mains and service-pipes in virtue of the pressure imparted to it by the gasholder and since, for reasons already given, only the rising-bell type of gasholder can be regarded as satisfactory, it becomes important to know the relations which subsist between the dimensions and weight of a gasholder bell and the pressure which it "throws" or imparts to the contained gas.

The bell must obviously be a vessel of considerable weight if it is to withstand reasonable wear and tear, and this weight will give a certain hydrostatic pressure to the contained gas. If the weight of the bell is known, the pressure which it will give can be calculated according to the general law of hydrostatics, that the weight of the water displaced must be equal to the weight of the floating body. Supposing for the moment that there are no other elements which will have to enter into the calculation, then if _d_ is the diameter in inches of the (cylindrical) bell, the surface of the water displaced will have an area of _d^2_ x 0.7854. If the level of the water is depressed _p_ inches, then the water displaced amounts to _p_(_d^2_ x 0.7854) cubic inches, and its weight will be (at 62° F.):

(0.7854_pd^2_ x 0.03604) = 0.028302_pd^2_ lb.

Consequently a bell which is _d_ inches in diameter, and gives a pressure of _p_ inches of water, will weigh 0.028302_pd^2_ lb. Or, if W = the weight of the bell in lb., the pressure thrown by it will be W/0.028302_d^2_ or 35.333W/_d^2_. This is the fundamental formula, which is sometimes given as _p_ = 550W/_d^2_, in which W = the weight of the bell in tons, and _d_ the diameter in feet. This value of _p_, however, is actually higher than the holder would give in practice. Reductions have to be made for two influences, viz., the lifting power of the contained gas, which is lighter than air, and the diminution in the effective weight of so much of the bell as is immersed in water. The effect of these influences was studied by Pole, who in 1839 drew up some rules for calculating the pressure thrown by a gasholder of given dimensions and weight. These rules form the basis of the formula which is commonly used in the coal-gas industry, and they may be applied, _mutatis mutandis_, to acetylene holders. The corrections for both the influences mentioned vary with the height at which the top of the gasholder bell stands above the level of the water in the tank. Dealing first with the correction for the lifting power of the gas, this, according to Pole, is a deduction of _h_(1 - _d_)/828 where _d_ is the specific gravity of the gas and _h_ the height (in inches) of the top of the gasholder above the water level. This strictly applies only to a flat-topped bell, and hence if the bell has a crown with a rise equal to about 1/20 of the diameter of the bell, the value of _h_ here must be taken as equal to the height of the top of the sides above the water-level (= _h'_), plus the height of a cylinder having the same capacity as the crown, and the same diameter as the bell, that is to say, _h_=_h'_ + _d_/40 where _d_ = the diameter of the bell. The specific gravity of commercially made acetylene being constantly very nearly 0.91, the deduction for the lifting power of the gas becomes, for acetylene gasholders, 0.0001086_h_ + 0.0000027_d_, where _h_ is the height in inches of the top of the sides of the bell above the water- level, and _d_ is the diameter of the bell. Obviously this is a negligible quantity, and hence this correction may be disregarded for all acetylene gasholders, whereas it is of some importance with coal-gas and other gases of lower specific gravity. It is therefore wrong to apply to acetylene gasholders formulæ in which a correction for the lifting power of the gas has been included when such correction is based on the average specific gravity of coal-gas, as is the case with many abbreviated gasholder pressure formulæ.

The correction for the immersion of the sides of the bell is of greater magnitude, and has an important practical significance. Let H be the total height in inches of the side of the gasholder, _h_ the height in inches of the top of the sides of the gasholder above the water-level, and _w_ = the weight of the sides of the gasholder in lb.; then, for any position of the bell, the proportion of the total height of the sides immersed (H - _h_)/H, and the buoyancy is (H - _h_)/H x _w_/S + pi/4_d^2_, in which S = the specific gravity of the material of which the bell is made. Assuming the material to be mild steel or wrought iron, having a specific gravity of 7.78, the buoyancy is (4_w_(H - _h_)) / (7.78Hpi_d^2_) lb. per square inch (_d_ being inches and _w_ lb.), which is equivalent to (4_w_(H - _h_)) / (0.03604 x 7.78Hpi_d^2_) = (4.54_w_(H - _h_)) / (H_d^2_) inches of water. Hence the complete formula for acetylene gasholders is:

_p_ = 35.333W / _d^2_ - 4.54_w_(H - _h_) / H_d^2_

It follows that _p_ varies with the position of the bell, that is to say, with the extent to which it is filled with gas. It will be well to consider how great this variation is in the case of a typical acetylene holder, as, if the variation should be considerable, provision must be made, by the employment of a governor on the outlet main or otherwise, to prevent its effects being felt at the burners.

Now, according to the rules of the "Acetylen-Verein" (_cf._ Chapter IV.), the bells of holders above 53 cubic feet in capacity should have sides 1.5 mm. thick, and crowns 0.5 mm. thicker. Hence for a holder from 150 to 160 cubic feet capacity, supposing it to be 4 feet in diameter and about 12 feet high, the weight of the sides (say of steel No. 16 S.W.G. = 2.66 lb. per square foot) will be not less than 12 x 4pi x 2.66 = 401 lb. The weight of the crown (say of steel No. 14 S.W.G. = 3.33 lb. per square foot) will be not less than about 12.7 x 3.33 = about 42 lb. Hence the total weight of holder = 401 + 42 = 443 lb. Then if the holder is full, _h_ is very nearly equal to H, and _p_ = (35.333 x 443) / 48^2 = 6.79 inches. If the holder stands only 1 foot above the water-level, then _p_ = 6.79 - (4.54 x 401 (144 - 12)) / (144 x 48^2) = 6.79 - 0.72 = 6.07 inches. The same result can be arrived at without the direct use of the second member of the formula:

For instance, the weight of the sides immersed is 11 x 4pi x 2.66 = 368 lb., and taking the specific gravity of mild steel at 7.78, the weight of water displaced is 368 / 7.78 = 47.3 lb. Hence the total effective weight of the bell is 443 - 47.3 = 395.7 lb., and _p_ = (35.333 x 395.7) / 48^2 = 6.07 inches. [Footnote: If the sealing liquid in the gasholder tank is other than simple water, the correction for the immersion of the sides of the bell requires modification, because the weight of liquid displaced will be _s'_ times as great as when the liquid is water, if _s'_ is the specific gravity of the sealing liquid. For instance, in the example given, if the sealing liquid were a 16 per cent. solution of calcium chloride, specific gravity 1.14 (_vide_ p. 93) instead of water, the weight of liquid displaced would be 1.14 (368 / 7.78) = 53.9 lb., and the total effective weight of the bell = 443 - 53.9 = 389.1 lb. Therefore _p_ becomes = (35.333 x 389.1) / 48^2 = 5.97 inches, instead of 6.07 inches.]

The value of _p_ for any position of the bell can thus be arrived at, and if the difference between its values for the highest and for the lowest positions of the bell exceeds 0.25 inch, [Footnote: This figure is given as an example merely. The maximum variation in pressure must be less than one capable of sensibly affecting the silence, steadiness, and economy of the burners and stoves, &c., connected with the installation.] a governor should be inserted in the main leading from the holder to the burners, or one of the more or less complicated devices for equalising the pressure thrown by a holder as it rises and falls should be added to the holder. Several such devices were at one time used in connexion with coal-gas holders, and it is unnecessary to describe them in this work, especially as the governor is practically the better means of securing uniform pressure at the burners.

It is frequently necessary to add weight to the bell of a small gasholder in order to obtain a sufficiently high pressure for the distribution of acetylene. It is best, having regard to the steadiness of the bell, that any necessary weighting of it should be done near its bottom rim, which moreover is usually stiffened by riveting to it a flange or curb of heavier gauge metal. This flange may obviously be made sufficiently stout to give the requisite additional weighting. As the flange is constantly immersed, its weight must not be added to that of the sides in computing the value of _w_ for making the correction of pressure in respect of the immersion of the bell. Its effective weight in giving pressure to the contained gas is its actual weight less its actual weight divided by its specific gravity (say 7.2 for cast iron, 7.78 for wrought iron or mild steel, or 11.4 for lead). Thus if _x_ lb. of steel is added to the rim its weight in computing the value of W in the formula _p_ = 35.333W / _d_^2 should be taken as x - x / 7.78. If the actual weight is 7.78 lb., the weight taken for computing W is 7.78 - 1 = 6.78 lb.

THE PRESSURE GAUGE.--The measurement of gas pressure is effected by means of a simple instrument known as a pressure gauge. It comprises a glass U- tube filled to about half its height with water. The vacant upper half of one limb is put in communication with the gas-supply of which the pressure is to be determined, while the other limb remains open to the atmosphere. The difference then observed, when the U-tube is held vertical, between the levels of the water in the two limbs of the tube indicates the difference between the pressure of the gas-supply and the atmospheric pressure. It is this _difference_ that is meant when the _pressure_ of a gas in a pipe or piece of apparatus is spoken of, and it must of necessity in the case of a gas-supply have a positive value. That is to say, the "pressure" of gas in a service-pipe expresses really by how much the pressure in the pipe _exceeds_ the atmospheric pressure. (Pressures less than the atmospheric pressure will not occur in connexion with an acetylene installation, unless the gasholder is intentionally manipulated to that end.) Gas pressures are expressed in terms of inches head or pressure of water, fractions of an inch being given in decimals or "tenths" of an inch. The expression "tenths" is often used alone, thus a pressure of "six-tenths" means a pressure equivalent to 0.6 inch head of water.

The pressure gauge is for convenience provided with an attached scale on which the pressures may be directly read, and with a connexion by which the one limb is attached to the service-pipe or cock where the pressure is to be observed. A portable gauge of this description is very useful, as it can be attached by means of a short piece of flexible tubing to any tap or burner. Several authorities, including the British Acetylene Association, have recommended that pressure gauges should not be directly attached to generators, because of the danger that the glass might be fractured by a blow or by a sudden access of heat. Such breakage would be followed by an escape of gas, and might lead to an accident. Fixed pressure gauges, however, connected with every item of a plant are extremely useful, and should be employed in all large installations, as they afford great aid in observing and controlling the working, and in locating the exact position of any block. All danger attending their use can be obviated by having a stopcock between the gauge inlet and the portion of the plant to which it is attached; the said stopcock being kept closed except when it is momentarily opened to allow of a reading being taken. As an additional precaution against its being left open, the stopcock may be provided with a weight or spring which automatically closes the gas-way directly the observer's hand is removed from the tap. In the best practice all the gauges will be collected together on a board fastened in some convenient spot on the wall of the generator-house, each gauge being connected with its respective item of the plant by means of a permanent metallic tube. The gauges must be filled with pure water, or with a liquid which does not differ appreciably in specific gravity from pure water, or the readings will be incorrect. Greater legibility will be obtained by staining the water with a few drops of caramel solution, or of indigo sulphate (indigo carmine); or, in the absence of these dyes, with a drop or two of common blue-black writing ink. If they are not erected in perfectly frost-free situations, the gauges may be filled with a mixture of glycerin and pure alcohol (not methylated spirit), with or without a certain proportion of water, which will not freeze at any winter temperature. The necessary mixture, which must have a density of exactly 1.00, could be procured from any pharmacist.

It is the pressure as indicated by the pressure gauge which is referred to in this book in all cases where the term "pressure of the gas" or the like is used. The quantity of acetylene which will flow in a given time from the open end of a pipe is a function of this pressure, while the quantity of acetylene escaping through a tiny hole or crack or a burner orifice also depends on this total pressure, though the ratio in this instance is not a simple one, owing to the varying influence of friction between the issuing gas and the sides of the orifice. Where, however, acetylene or other gas is flowing through pipes or apparatus there is a loss of energy, indicated by a falling off in the pressure due to friction, or to the performance of work, such as actuating a gas-meter. The extent of this loss of energy in a given length of pipe or in a meter is measured by the difference between the pressures of the gas at the two ends of the pipe or at the inlet and outlet of the meter. This difference is the "loss" or "fall" of pressure, due to friction or work performed, and is spoken of as the "actuating" pressure in regard to the passage of gas through the stretch of pipe or meter. It is a measure of the energy absorbed in actuating the meter or in overcoming the friction. (Cf. footnote, Chapter II., page 54.)

DIMENSIONS OF MAINS.--The diameter of the mains and service-pipes for an acetylene installation must be such that the main or pipe will convey the maximum quantity of the gas likely to be required to feed all the burners properly which are connected to it, without an excessive actuating pressure being called for to drive the gas through the main or pipe. The flow of all gases through pipes is of course governed by the same general principles; and it is only necessary in applying these principles to a particular gas, such as acetylene, to know certain physical properties of the gas and to make due allowance for their influence. The general principles which govern the flow of a gas through pipes have been exhaustively studied on account of their importance in relation to the distribution of coal-gas and the supply of air for the ventilation of places where natural circulation is absent or deficient. It will be convenient to give a very brief reference to the way in which these principles have been ascertained and applied, and then to proceed to the particular case of the distribution of acetylene through mains and service-pipes.

The subject of "The Motion of Fluids in Pipes" was treated in a lucid and comprehensive manner in an Essay by W. Pole in the _Journal of Gas Lighting_ during 1852, and his conclusions have been generally adopted by gas engineers ever since. He recapitulated the more important points of this essay in the course of some lectures delivered in 1872, and one or other of these two sources should be consulted for further information. Briefly, W. Pole treated the question in the following manner:

The practical question in gas distribution is, what quantity of gas will a given actuating pressure cause to flow along a pipe of given length and given diameter? The solution of this question allows of the diameters of pipes being arranged so that they will carry a required quantity of gas a given distance under the actuating pressure that is most convenient or appropriate. There are five quantities to be dealt with, viz.:

(1) The length of pipe = _l_ feet.

(2) The internal diameter of the pipe = _d_ inches.

(3) The actuating pressure = _h_ inches of head of water. (4) The specific gravity or density of the gas = _d_ times that of air.

(5) The quantity of gas passing through the pipe--Q cubic feet per hour. This quantity is the product of the mean velocity of the gas in the pipe and the area of the pipe.

The only work done in maintaining the flow of gas along a pipe is that required to overcome the friction of the gas on the walls of the pipe, or, rather, the consequential friction of the gas on itself, and the laws which regulate such friction have not been very exhaustively investigated. Pole pointed out, however, that the existing knowledge on the point at the time he wrote would serve for the purpose of determining the proper sizes of gas-mains. He stated that the friction (1) is proportional to the area of rubbing surface (viz., pi_ld_); (2) varies with the velocity, in some ratio greater than the first power, but usually taken as the square; and (3) is assumed to be proportional to the specific gravity of the fluid (viz., _s_).

Thus the force (_f_) necessary to maintain the motion of the gas in the pipe is seen to vary (1) as pi_ld_, of which pi is a constant; (2) as _v^2_, where _v_ = the velocity in feet per hour; and (3) as _s_. Hence, combining these and deleting the constant pi, it appears that

_f_ varies as _ldsv^2_.

Now the actuating force is equal to _f_, and is represented by the difference of pressure at the two ends of the pipe, _i.e._, the initial pressure, viz., that at the place whence gas is distributed or issues from a larger pipe will be greater by the quantity _f_ than the terminal pressure, viz., that at the far end of the pipe where it branches or narrows to a pipe or pipes of smaller size, or terminates in a burner. The terminal pressure in the case of service-pipes must be settled, as mentioned in Chapter II., broadly according to the pressure at which the burners in use work best, and this is very different in the case of flat-flame burners for coal-gas and burners for acetylene. The most suitable pressure for acetylene burners will be referred to later, but may be taken as equal to p_0 inches head of water. Then, calling the initial pressure (_i.e._, at the inlet head of service-pipe) p_1, it follows that p_1 - p_0 = _f_. Now the cross-section of the pipe has an area (pi/4)_d^2_, and if _h_ represents the difference of pressure between the two ends of the pipe per square inch of its area, it follows that _f_ = _h(pi/4)d^2_. But since _f_ has been found above to vary as _ldsv^2_ , it is evident that

_h(pi/4)d^2_ varies as _ldsv^2_.

Hence

_v^2_ varies as _hd/ls_,

and putting in some constant M, the value of which must be determined by experiment, this becomes

_v^2_ = M_hd/ls_.

The value of M deduced from experiments on the friction of coal-gas in pipes was inserted in this equation, and then taking Q = pi/4_d^2v_, it was found that for coal-gas Q = 780(_hd/sl_)^(1/2)

This formula, in its usual form, is

Q = 1350_d^2_(_hd/sl_)^(1/2)

in which _l_ = the length of main in yards instead of in feet. This is known as Pole's formula, and has been generally used for determining the sizes of mains for the supply of coal-gas.

For the following reasons, among others, it becomes prudent to revise Pole's formula before employing it for calculations relating to acetylene. First, the friction of the two gases due to the sides of a pipe is very different, the coefficient for coal-gas being 0.003, whereas that of acetylene, according to Ortloff, is 0.0001319. Secondly, the mains and service-pipes required for acetylene are smaller, _cateria paribus_, than those needed for coal-gas. Thirdly, the observed specific gravity of acetylene is 0.91, that of air being unity, whereas the density of coal-gas is about 0.40; and therefore, in the absence of direct information, it would be better to base calculations respecting acetylene on data relating to the flow of air in pipes rather than upon such as are applicable to coal-gas. Bernat has endeavoured to take these and similar considerations into account, and has given the following formula for determining the sizes of pipes required for the distribution of acetylene:

Q = 0.001253_d^2_(_hd/sl_)^(1/2)

in which the symbols refer to the same quantities as before, but the constant is calculated on the basis of Q being stated in cubic metres, l in metres, and d and h in millimetres. It will be seen that the equation has precisely the same shape as Pole's formula for coal-gas, but that the constant is different. The difference is not only due to one formula referring to quantities stated on the metric and the other to the same quantities stated on the English system of measures, but depends partly on allowance having been made for the different physical properties of the two gases. Thus Bernat's formula, when merely transposed from the metric system of measures to the English (_i.e._, Q being cubic feet per hour, _l_ feet, and _d_ and _h_ inches) becomes

Q = 1313.5_d^2_(_hd/sl_)^(1/2)

or, more simply,

Q = 1313.4(_hd^5/sl_)^(1/2)

But since the density of commercially-made acetylene is practically the same in all cases, and not variable as is the density of coal-gas, its value, viz., 0.91, may be brought into the constant, and the formula then becomes

Q = 1376.9(_hd^5/l_)^(1/2)

Bernat's formula was for some time generally accepted as the most trustworthy for pipes supplying acetylene, and the last equation gives it in its simplest form, though a convenient transposition is

d = 0.05552(Q^2_l/h_)^(1/5)

Bernat's formula, however, has now been generally superseded by one given by Morel, which has been found to be more in accordance with the actual results observed in the practical distribution of acetylene. Morel's formula is

D = 1.155(Q^2_l/h_)^(1/5)

in which D = the diameter of the pipe in centimetres, Q = the number of cubic metres of gas passing per hour, _l_ = the length of pipe in metres, and _h_ = the loss of pressure between the two ends of the pipe in millimetres. On converting tins formula into terms of the English system of measures (_i.e._, _l_ feet, Q cubic feet, and _h_ and _d_ inches) it becomes

(i) d = 0.045122(Q^2_l/h_)^(1/5)

At first sight this formula does not appear to differ greatly from Bernat's, the only change being that the constant is 0.045122 instead of 0.05552, but the effect of this change is very great--for instance, other factors remaining unaltered, the value of Q by Morel's formula will be 1.68 times as much as by Bernat's formula. Transformations of Morel's formula which may sometimes be more convenient to apply than (i) are:

(ii) Q = 2312.2(_hd^5/l_)^(1/2)

(iii) _h_ = 0.000000187011(Q^2_l/d^5_)

and (iv) _l_ = 5,346,340(_hd^5_/Q^2)

In order to avoid as far as possible expenditure of time and labour in repeating calculations, tables have been drawn up by the authors from Morel's formulæ which will serve to give the requisite information as to the proper sizes of pipes to be used in those cases which are likely to be met with in ordinary practice. These tables are given at the end of this chapter.

When dealing with coal-gas, it is highly important to bear in mind that the ordinary distributing formulæ apply directly only when the pipe or main is horizontal, and that a rise in the pipe will be attended by an increase of pressure at the upper end. But as the increase is greater the lower the density of the gas, the disturbing influence of a moderate rise in a pipe is comparatively small in the case of a gas of so high a density as acetylene. Hence in most instances it will be unnecessary to make any allowance for increase of pressure due to change of level. Where the change is very great, however, allowance may advisedly be made on the following basis: The pressure of acetylene in pipes increases by about one-tenth of an inch (head of water) for every 75 feet rise in the pipe. Hence where acetylene is supplied from a gasholder on the ground-level to all floors of a house 75 feet high, a burner at the top of the house will ordinarily receive its supply at a pressure greater by one-tenth of an inch than a burner in the basement. Such a difference, with the relatively high pressures used in acetylene supplies, is of no practical moment. In the case of an acetylene-supply from a central station to different parts of a mountainous district, the variations of pressure with level should be remembered.

The distributing formulæ also assume that the pipe is virtually straight; bends and angles introduce disturbing influences. If the bend is sharp, or if there is a right-angle, an allowance should be made if it is desired to put in pipes of the smallest permissible dimensions. In the case of the most usual sizes of pipes employed for acetylene mains or services, it will suffice to reckon that each round or square elbow is equivalent in the resistance it offers to the flow of gas to a length of 5 feet of pipe of the same diameter. Hence if 5 feet is added to the actual length of pipe to be laid for every bond or elbow which will occur in it, and the figure so obtained is taken as the value of _l_ in formulæ (i), (ii), or (iii), the values then found for Q, _d_, or _h_ will be trustworthy for all practical purposes.

It may now be useful to give an example of the manner of using the foregoing formulæ when the tables of sizes of pipes are not available. Let it be supposed that an institution is being equipped for acetylene lighting; that 50 burners consuming 0.70 cubic foot, and 50 consuming 1.00 cubic foot of acetylene per hour may be required in use simultaneously; that a pressure of at least 2-1/2 inches is required at all the burners; that for sufficient reasons it is considered undesirable to use a higher distributing pressure than 4 inches at the gasholder, outlet of the purifiers, or initial governor (whichever comes last in the train of apparatus); that the gasholder is located 100 feet from the main building of the institution, and that the trunk supply-pipe through the latter must be 250 feet in length, and the supplies to the burners, either singly or in groups, be taken from this trunk pipe through short lengths of tubing of ample size. What should be the diameter of the trunk pipe, in which it will be assumed that ten bonds or elbows are necessary?

In the first instance, it is convenient to suppose that the trunk pipe may be of uniform diameter throughout. Then the value of _l_ will be 100 (from gasholder to main building) + 250 (within the building) + 50 (equivalent of 10 elbows) = 400. The maximum value of Q will be (50 x 0.7) + (50 x 1.0) = 85; and the value of _h_ will be 1 - 2.5 - 1.5. Then using formula (i), we have:

d = 0.045122((85^2 x 400)/1.5)^(1/5) = 0.045122(1,926,667)^(1/5)

= 0.045122 x 18.0713 = 0.8154.

The formula, therefore, shows that the pipe should have an internal diameter of not less than 0.8154 inch, and consequently 1 inch (the next size above 0.8154 inch) barrel should be used. If the initial pressure (i.e., at outlet of purifiers) could be conveniently increased from 4 to 4.8 inches, 3/4 inch barrel could be employed for the service-pipe. But if connexions for burners were made immediately the pipe entered the building, these burners would then be supplied at a pressure of 4.2 inches, while those on the extremity of the pipe would, when all burners were in use, be supplied at a pressure of only 2.5 inches. Such a great difference of pressure is not permissible at the several burners, as no type of burner retains its proper efficiency over more than a very limited range of pressure. It is highly desirable in the case of the ordinary Naphey type of burner that all the burners in a house should be supplied at pressures which do not differ by more than half an inch; hence the pipes should, wherever practicable, be of such a size that they will pass the maximum quantity of gas required for all the burners which will ever be in use simultaneously, when the pressure at the first burner connected to the pipe after it enters the house is not more than half an inch above the pressure at the burner furthermost removed from the first one, all the burner-taps being turned on at the time the pressures are observed. If the acetylene generating plant is not many yards from the building to be supplied, it is a safe rule to calculate the size of pipes required on the basis of a fall of pressure of only half an inch from the outlet of the purifiers or initial governor to the farthermost burner. The extra cost of the larger size of pipe which the application of this rule may entail will be very slight in all ordinary house installations.

VELOCITY OF FLOW IN PIPES.--For various purposes, it is often desirable to know the mean speed at which acetylene, or any other gas, is passing through a pipe. If the diameter of the pipe is _d_ inches, its cross-sectional area is _d^2_ x 0.7854 square inches; and since there are 1728 cubic inches in 1 cubic foot, that quantity of gas will occupy in a pipe whose diameter is _d_ inches a length of

1728/(_d^2_ x 0.7854) linear inches or 183/_d^2_^ linear feet.

If the gas is in motion, and the pipe is delivering Q cubic feet per hour, since there are 3600 seconds of time in one hour, the mean speed of the gas becomes

183/_d^2_ x Q/3600 = Q/(19 x 7_d^2_) linear feet per second.

This value is interesting in several ways. For instance, taking a rough average of Le Chatelier's results, the highest speed at which the explosive wave proceeds in a mixture of acetylene and air is 7 metres or 22 feet per second. Now, even if a pipe is filled with an acetylene-air mixture of utmost explosibility, an explosion cannot travel backwards from B to A in that pipe, if the gas is moving from A to B at a speed of over 22 feet per second. Hence it may be said that no explosion can occur in a pipe provided

Q/(19.7_d^2_) = 22 or more;

_i.e._, Q/_d^2_=433.4

In plain language, if the number of cubic feet passing through the pipe per hour divided by the square of the diameter of the pipe is at least 433.4, no explosion can take place within that pipe, even if the gas is highly explosive and a light is applied to its exit.

In Chapter VI. are given the explosive limits of acetylene-air mixtures as influenced by the diameter of the tube containing them. If we possessed a similar table showing the speed of the explosive wave in mixtures of known composition, the foregoing formulæ would enable us to calculate the minimum speed which would insure absence of explosibility in a supply-pipe of any given diameter throughout its length, or at its narrowest part. It would not, however, be possible simply by increasing the forward speed of an explosive mixture of acetylene and air to a point exceeding that of its explosion velocity to prevent all danger of firing back in an atmospheric burner tube. A much higher pressure than is usually employed in gas-burners, other than blowpipes, would be needed to confer a sufficient degree of velocity upon the gas, a pressure which would probably fracture any incandescent mantle placed in the flame.

SERVICE-PIPES AND MAINS.--The pipes used for the distribution of acetylene must be sound in themselves, and their joints perfectly tight. Higher pressures generally prevail in acetylene service-pipes within a house than in coal-gas service-pipes, while slight leaks are more offensive and entail a greater waste of resources. Therefore it is uneconomical, as well as otherwise objectionable, to employ service-pipes or fittings for acetylene which are in the least degree unsound. Unfortunately ordinary gas-barrel is none too sound, nor well-threaded, and the taps and joints of ordinary gas-fittings are commonly leaky. Hence something better should invariably be used for acetylene. What is known as "water" barrel, which is one gauge heavier than gas-barrel of the same size, may be adopted for the service-pipes, but it is better to incur a slight extra initial expense and to use "steam" barrel, which is of still heavier gauge and is sounder than either gas or water-pipe. All elbows, tees, &c., should be of the same quality. The fitters' work in making the joints should be done with the utmost care, and the sloppy work often passed in the case of coal-gas services must on no account be allowed. It is no exaggeration to say that the success of an acetylene installation, from the consumer's point of view, will largely, if not principally, depend on the tightness of the pipes in his house. The statement has been made that the "paint" used by gas-fitters, _i.e._, the mixture of red and white lead ground in "linseed" oil, is not suitable for employment with acetylene, and it has been proposed to adopt a similar material in which the vehicle is castor-oil. No good reason has been given for the preference for castor-oil, and the troubles which have arisen after using ordinary paint may be explained partly on the very probable assumption that the oil was not genuine linseed, and so did not dry, and partly on the fact that almost entire reliance was placed on the paint for keeping the joint sound. Joints for acetylene, like those for steam and high-pressure water, must be made tight by using well-threaded fittings, so as to secure metallic contact between pipe and socket, &c.; the paint or spun-yarn is only an additional safeguard. In making a faced joint, washers of (say, 7 lb) lead, or coils of lead-wire arc extremely convenient and quite trustworthy; the packing can be used repeatedly.

LEAKAGE.--Broadly speaking, it may be said that the commercial success of any village acetylene-supply--if not that of all large installations-- depends upon the leakage being kept within moderate limits. It follows from what was stated in Chapter VI. about the diffusion of acetylene, that from pipes of equal porosity acetylene and coal-gas will escape at equal rates when the effective pressure in the pipe containing acetylene is double that in the pipe containing coal-gas. The loss of coal-gas by leakage is seldom less than 5 per cent. of the volume passed into the main at the works; and provided a village main delivering acetylene is not unduly long in proportion to the consumption of gas--or, in other words, provided the district through which an acetylene distributing main passes is not too sparsely populated--the loss of acetylene should not exceed the same figure. Caro holds that the loss of gas by leakage from a village installation should be quoted in absolute figures and not as a percentage of the total make as indicated by the works meter, because that total make varies so largely at different periods of the year, while the factors which determine the magnitude of the leakage are always identical; and therefore whereas the actual loss of gas remains the same, it is represented to be more serious in the summer than in the winter. Such argument is perfectly sound, but the method of returning leakage as a percentage of the make has been employed in the coal-gas industry for many years, and as it does not appear to have led to any misunderstanding or inconvenience, there is no particular reason for departing from the usual practice in the case of acetylene where the conditions as to uniform leakage and irregular make are strictly analogous.

Caro has stated that a loss of 15 to 20 litres per kilometre per hour (_i.e._, of 0.85 to 1.14 cubic feet per mile per hour) from an acetylene distributing main is good practice; but it should be noted that much lower figures have been obtained when conditions are favourable and when due attention has been devoted to the fitters' work. In one of the German village acetylene installations where the matter has been carefully investigated (Döse, near Cuxhaven), leakage originally occurred at the rate of 7.3 litres per kilometre per hour in a main 8.5 kilometres, or 5.3 miles, long and 4 to 2 inches in diameter; but it was reduced to 5.2 litres, and then to 3.12 litres by tightening the plugs of the street lantern and other gas cocks. In British units, these figures are 0.415, 0.295, and 0.177 cubic foot per mile per hour. By calculation, the volume of acetylene generated in this village would appear to have been about 23,000 cubic feet per mile of main per year, and therefore it may be said that the proportion of gas lost was reduced by attending to the cocks from 15.7 per cent, to 11.3 per cent, and then to 6.8 per cent. At another village where the main was 2.5 kilometres long, tests extending over two months, when the public lamps were not in use, showed the leakage to be 4.4 litres per kilometre per hour, _i.e._, 1.25 cubic foot per mile per hour, when the annual make was roughly 46,000 cubic feet per mile of main. Here, the loss, calculated from the direct readings of the works motor, was 4.65 per cent.

When all the fittings, burners excepted, have been connected, the whole system of pipes must be tested by putting it under a gas (or air) pressure of 9 or 12 inches of water, and observing on an attached pressure gauge whether any fall in pressure occurs within fifteen minutes after the main inlet tap has been shut. The pressure required for this purpose can be obtained by temporarily weighting the holder, or by the employment of a pump. If the gauge shows a fall of pressure of one quarter of an inch or more in these circumstances, the pipes must be examined until the leak is located. In the presence of a meter, the installation can conveniently be tested for soundness by throwing into it, through the meter, a pressure of 12 inches or so of water from the weighted holder, then leaving the inlet cock open, and observing whether the index hand on the lowest dial remains perfectly stationary for a quarter of an hour--movement of the linger again indicating a leak. The search for leaks must never be made with a light; if the pipes are full of air this is useless, if full of gas, criminal in its stupidity. While the whole installation is still under a pressure of 12 inches thrown from the loaded holder, whether it contains air or gas, first all the likely spots (joints, &c.), then the entire length of pipe is carefully brushed over with strong soapy water, which will produce a conspicuous "soap- bubble" wherever the smallest flaw occurs. The tightness of a system of pipes put under pressure from a loaded holder cannot be ascertained safely by observing the height of the bell, and noting if it falls on standing. Even if there is no issue of gas from the holder, the position of the bell will alter with every variation in temperature of the stored gas or surrounding air, and with every movement of the barometer, rising as the temperature rises and as the barometer falls, and _vice versâ_, while, unless the water in the seal is saturated with whatever gas the holder contains, the bell will steadily drop a little an part of its contents are lost by dissolution in the liquid.