Chapter XXXVI: Boilers for Stationary Steam Engines (1)

The boiler for a steam engine requires the most careful usage and inspection, in the first case because a good boiler may be destroyed very rapidly by careless usage, and in the second case because the durability of a boiler depends to a great extent upon matters that are beyond ordinary control, and that in many cases do not make themselves known except in their results, which can only be discovered by careful and intelligent inspection. All that the working engineer is called upon to do is, to use the boiler properly, keep it clean, and examine it at such intervals as the nature of the conditions under which it is used may render necessary.

The periods at which a boiler should be cleaned and inspected depend upon the quality of the water, whether the feed water is purified or not, and to a certain extent upon the design of the boiler; hence these periods are variable under different circumstances.

The horse power of a boiler is estimated in various ways, and there is no uniform practice in this respect. Some makers estimate a boiler to have a horse power for every fifteen square feet of heating surface it possesses, while others allow but 12 square feet.

The heating surface of a boiler of any kind is the surface that is exposed to the action of the fire on one side, and has water on the other; hence the surface of the steam space is not reckoned as heating surface, even though it may be exposed to the action of the heat. The effectiveness of the heating surface of a boiler obviously, however, depends upon the efficiency of the fire, and this depends upon the amount of draught, hence the estimation of horse power from the amount of its heating surface, while affording to a certain extent a standard of measurement or comparison while the boiler is not in use, has no definite value when the boiler is erected and at work.

Thus whatever amount of steam a boiler may produce under a poor or moderate draught, it will obviously produce more under an increased draught; hence the efficiency of the same boiler depends to a certain extent upon the draught, or in other words upon the quantity of fuel that can be consumed upon its fire bars.

The amount of water required in steam boilers varies from 16 lbs. to 40 lbs., per horse power per hour, and it has been proposed to compute the horse power of boilers from the water evaporation, taking as a standard 30 lbs. of feed water at a temperature of 70 degrees, evaporated into steam at a temperature of 212 degrees, at which temperature the steam is assumed to equal the pressure of the atmosphere.

[49]"The strength of the shell of a cylindrical boiler to resist a pressure within it, is inversely proportional to its diameter and directly, to the thickness of the plate of which it is formed.

[49] From "_Steam Boilers_."

"For instance, take three cylindrical boilers each made of 1/2 inch plate, the first one 2 feet 6 inches in diameter; the second twice that, or 5 feet in diameter; and the third twice that again, or 10 feet in diameter; and if the 2 foot 6 inch boiler is fit for a safe working pressure of 180 lbs. per square inch, then the 5 foot boiler will be fit for exactly one-half that amount, or 90 lbs. per square inch; and the ten foot boiler will be fit for half the working pressure of the five foot boiler, hence we have:

-----------------+------------+--------------------------
Diameter of | Thickness |Relative working pressure.
boiler shell. | of plate. |
-----------------+------------+--------------------------
2 feet 6 inches.| 1/2 inch. | 180 lbs. per square inch.
5 " | " " | 90 " " " "
10 " | " " | 45 " " " "
-----------------+------------+--------------------------

"The reverse applies to the thickness of the plate. For instance, if we take two cylindrical boiler shells, each 5 feet in diameter, the first one made of plate 1/2 inch thick, and the second twice that, or 1 inch thick, and if the first is equal to a safe working pressure of 90 lbs. per square inch, then the second is equal to a safe working pressure of twice as much, or 180 lbs. per square inch, providing, of course, that the riveted seams are of equal strength in each case, and that both boilers are allowed the same margin for safety; hence we have:

----------------+------------+--------------------------
Diameter of | Thickness | Safe working pressure.
boiler. | of shell. |
----------------+------------+--------------------------
5 feet. | 1/2 inch. | 90 lbs. per square inch.
5 " | 1 " | 180 " " " "
----------------+------------+--------------------------

"These principles (namely, that the strength of a boiler is, all other things or elements being equal, inversely proportional to its diameter, and directly proportional to its thickness) afford us a groundwork upon which we may lay down rules for determining by calculation the strength of the solid part[50] of any boiler shell, and the bases of these calculations are as follows:

[50] In the case of the riveted joints or seams other considerations
come in, as will be shown hereafter.

"If the shell plate of a cylindrical boiler is 1/2 inch thick, there is one inch section of metal to be broken before the boiler can be divided into two pieces, that is to say there is 1/2 inch on each side of the shell, as shown in Fig. 3237, and the two together will make 1 inch. If we take a ring an inch broad, as, say, at A in Fig. 3238, we shall obviously have a section of 1 square inch of metal to break before the ring can be broken into two pieces.

"The next consideration is, what is the average strength of a plate of boiler iron? Now suppose we have a strip of boiler iron 2 inches wide and 1/2 inch thick, or, what is the same thing, a bar of boiler iron 1 inch square, and that we lay it horizontally and pull its ends apart until it breaks, how many lbs. will it bear before breaking? Now for our present purpose we may assume this to be 47,040 lbs., and if this number of lbs. be divided by the diameter of the boiler in inches, it will give the bursting pressure in lbs. for any square inch in the ring, or any other square inch in the cylindrical shell of the boiler.

"The reason for dividing by the diameter of the boiler is as follows:

"Of course the steam pressure presses equally on all parts of the interior surface of the shell, and may be taken as radiating from the centre of the boiler, as in Fig. 3239, which represents an end view of a strip an inch wide, of one half of a boiler. Now leaving the riveted seam out of the question, and supposing the shell to be truly cylindrical, and the metal to be of equal quality throughout, it will take just as much pressure to burst the shell apart in one direction as it will in another, hence we may suppose that the boiler is to be burst in the direction of arrow _a_, and it is the section of metal at _b_ _b_ that is resisting rupture in that direction.

"Now suppose we divide the surface against which the steam presses into six divisions, by lines radiating from the centre C, and to find the amount of area acting on each division to burst the shell in the direction of arrow _a_, we drop perpendicular lines, as line _e_, from the lines of division to the line _b_ _b_, and the length of the line divided off (by the perpendicular) on the diameter represents the effectiveness of the area of that division to burst the boiler in the direction of arrow _a_; thus for that part of the boiler surface situate in the first division, or from _b_ to line _e_, the area acting to burst the boiler in the direction of _a_ is represented by the length of the line _k_, while the general direction of the pressure on this part of the shell is represented by arrow _m_.

"Similarly, for that part of the shell situate between vertical line _e_ and vertical line _f_, the general direction of the steam pressure is denoted by the arrow _l_, while the proportion of this part that is acting to sever the boiler in the direction of _a_ is represented by the distance _n_, or from the line _e_ to line _f_ measured on the line _b_ _b_.

"By carrying out this process we shall perceive that, although the pressure acts upon the whole circumference, yet its effectiveness in bursting the boiler in any one direction is equal to the boiler diameter. Thus in Fig. 3240, the pressure acting in the direction of the arrows _a_ (and to burst the boiler apart at _b_ _b_) is represented by the diametral line _b_ _b_, while the pressure actually exerted upon the whole boiler shell is represented by the circumference of the boiler.

"To proceed, then, it will now be clear that the ultimate strength of the boiler material, multiplied by twice the thickness of the boiler shell plate in inches or decimal parts of an inch, and this sum divided by the internal diameter of the boiler, in inches, gives the pressure (in lbs. per square inch) at which the boiler shell will burst."

We have here only considered the strength of the solid plate of the shell, and may now consider the strength of the riveted joints, because, as the boiler cannot be any stronger as a whole than its weakest part is, and as the riveted joints are the weakest parts of a cylindrical boiler,[51] therefore the strength of the riveted joint determines the strength of the boiler.

[51] It may be here noted that the riveted joint of a flat plate is
stronger than the flat surface of the plate, because at the joint the
plate is doubled, or one plate overlaps the other.

[52]"The strains to which a riveted joint is subjected are as follows: That acting to shear the rivet across its diameter is called the _shearing_ strain. But the same strain acts to tear the plate apart; hence, when spoken of with reference to the action on the plate, it is called the _tearing_ strain.

[52] From "_Steam Boilers_."

"The same strain also acts to crush and rupture the plate between the rivet hole and the edge of the plate, and in this connection it is called the _crushing_ strain.

"Thus, Fig. 3241 represents a single riveted lap joint, in which the joint at rivets A, B, and C is intact, the metal outside of D has crushed, the rivets E, F have sheared, and the plate has torn at H, leaving a piece J on the rivets K L.

"It is obvious that, since it is the same strain that has caused these different kinds of rupture, the joint has, at each location, simply given way where it was the weakest.

"If a riveted joint was to give way by tearing only, the indication would be that the proportion of strength was greatest in the rivets, which might occur from the plate being of inferior metal to the rivets, or from the rivets being too closely spaced. If the rivets were to shear and the plate remain intact, it would indicate insufficient strength in the rivets, which might occur from faulty material in the rivets, from smallness of rivet diameter, or from the rivets being too widely spaced.

"The object then, in designing a riveted joint is to have its resistance to tearing and shearing proportionately equal, whatever form of joint be employed."

The English Board of Trade recommends that the rivet section should always be in excess of the plate section, whereas, in ordinary American practice, for stationary engine boilers, the plate and rivet percentages are made equal.

The forms of riveted joints employed in boiler work are as follows:

Fig. 3242 represents a single riveted lap joint. Fig. 3243 represents a double riveted lap joint, chain riveted; and Fig. 3244, a double riveted lap joint, with the rivets arranged zigzag.

Fig. 3245 represents a single and Fig. 3246 a double riveted butt joint, so called because the ends of the boiler plate abut together. The plates on each side of joint are called butt straps.

The advantages of the butt joint are, first, that the boiler shell is kept more truly cylindrical, and the joint is not liable to bend as it does in the lap joints, in the attempt of the boiler (when under pressure) to assume the form of a true circle, and second that the rivets are placed in double shear. That is to say, if in a lap joint the rivet was to shear between the plates, the joint would come apart, whereas, in a butt joint, the rivet must shear on each side of the plate, and therefore in two places.

Fig. 3247 represents a form of joint much used in locomotive practice in the United States. It is a lap joint, with a covering plate on the inside of the joint; rivets E and F are in single and rivets D in double shear.

[53]"When we have to deal with comparatively thin boiler plates, there is no difficulty in obtaining a sufficiently high percentage of strength in the joints, by using the ordinary double riveted joint, but when we have to deal with thick plates, as in the case of large marine boilers, as 1 inch or upwards, a more costly form of joint must be employed, in order to obtain the required percentage of strength at the joint; hence the ordinary double riveted joint is replaced by various other forms as follows:

[53] From "_Steam Boilers_."

"First, a triple zigzag riveted lap joint, such as shown in Fig. 3248, or a chain riveted joint as in Fig. 3249, in both of which the third row of rivets enables the rivet pitch to be increased, thus increasing the plate percentage, while the third row of rivets also increases the rivet percentage.

"Second, by employing butt joints with butt straps, either double or treble riveted.

"A double riveted butt joint with double straps is shown in Fig. 3250, and a treble with double straps in Figs. 3251 and 3252.

"Third. By various arrangements of the rivets in conjunction with butt joints and double straps, with which it is not necessary, at this point, to deal.

"One of the great advantages obtained by the use of the double strap is that of bringing the rivet into double shear (or in other words, the rivet must shear on each side of the plate, or in two places, instead of between the plates only, before the joint can give way by shearing), and thus obtaining an increased calculated strength of 1-3/4 times the ordinary or single shear, the rule being to find the rivet strength in the ordinary way (as before explained), and then multiply the result by 1.75.

"The Board of Trade rules for spacing the rivets of these joints are as follows:

"Dimension E is the distance from the edge of the plate to the centre of the rivet hole. Dimension V is the distance between the rows of rivets, dimension _p_ is the pitch of the rivets, which is always measured from centre to centre of the rivets, and dimension _pd_ is the diagonal pitch of the rivets.

"The rule for finding dimension E, whether the plates and rivets are either of steel or iron, is as follows:

"Multiply the diameter of the rivet by 3 and divide by 2, the formula being as follows:

3 × _d_
------- = E.
2

"To find the distance V between the rows of rivets in chain riveted joints. This distance must not be less than twice the rivet diameter, and a more desirable rule is four times the rivet diameter plus 1 divided by 2, thus:

4_d_ + 1
-------- = V.
2

"To find the distance between the rows of zigzag riveted joints:

_____________________________
\/(11_p_ + 4_d_) × (_p_ + 4_d_)
------------------------------- = V,
10

that is, multiply 11 times the pitch plus 4 times the rivet diameter, by the pitch plus 4 times the rivet diameter, then extract the square root and divide by 10.

"To find diagonal pitch _pd_, multiply the pitch _p_ by 6, then add 4 and divide by 10, thus:

6_p_ + 4
-------- = _pd_."
10

Fig. 3253 represents a form of high percentage joint, used upon marine boilers of 10 to 14 feet diameter, and carrying from 100 to 190 lbs. pressure of steam. The rivets are what are termed unevenly pitched, or, that is to say, on each side of the joint, there are three rows of rivets, of which the inner and outer rows are wider pitched than the middle row.

[54]"The advantage gained by this spacing is that the shear of the outer row of rivets is added to the plate section at the narrow pitch, that is to say, if the plate section broke through the line of rivet holes at the narrow pitch, it has yet to shear the outer row of rivets before the plate can separate."

[54] From "_Steam Boilers_."

Fig. 3254 represents a second example of joint with rivets unevenly pitched, this form finding much favor in recent practice. The four inner rows of rivets are spaced at narrow pitch and the two outer rows are wide pitched.

[55]"The strength percentage of this joint is calculated from three points of view, as follows:

[55] From "_Steam Boilers_."

"First. The plate section at the wide pitched rivets.

"Second. The rivet section in one pitch.

"Third. The plate section at the narrow pitch plus half the double shear of the outer or wide pitched rivet."

The steam pressures generally employed in the boilers of stationary engines range from about 60 to 100 lbs. per square inch, and as a result of these comparatively low pressures less perfect forms of construction are employed than would be permissible if higher pressures were used.

The strength of the shell plate of boilers of small diameter is always largely in excess of the requirements, and as a result the strength of the joints may bear a very low percentage to that of the solid plate, and yet give a sufficient factor of safety for the working pressure.

Take, for example, a boiler shell of 36 inches internal diameter with a shell plate 1/4 inch thick, and allowing the strength of the material to be 48,000 lbs. per inch of section, and with a factor of safety of 4, the working pressure will be 166 lbs. per square inch, thus:

Strength Plate thickness
of the material. × 2.

48000 × (.25 × 2)
--------------------------------- = 666-2/3 lbs. = bursting pressure.
36

Diameter of boiler.

By dividing this 666 by the factor of safety 4 we get 166-2/3 lbs. as the working pressure of the shell plate independent of the riveted joint. Usually, however, such a boiler would not be used for a pressure above about 60 lbs. per inch, and this leaves a wide margin for the reduction of strength caused by the riveted joints.

Suppose, for example, that a single riveted lap joint is used, and the strength of this joint is but 50 per cent. of that of the solid plate, and we have as follows:

Strength of % strength Twice
material. of riveted the plate
joint. thickness.

48000 × .50 × (.25 × 2)
---------------------------------- = 83-1/3 lbs. = W.P.
36 × 4

Internal diam. Factor of
of boiler. safety.

Here then we find that the working pressure of the solid plate is double that of the riveted joint, and that the working pressure of the boiler is 83 lbs. per square inch, notwithstanding that the strength of the riveted joints is but 50 per cent. of that of the solid plate. Such a boiler would not, however, be used for a pressure of over 60 lbs. per square inch.

If the above-named boiler was double riveted so as to bring the percentage of joint strength up to say 70 per cent, of that of the solid plate, its working pressure would be 116 lbs. per square inch, thus:

Strength of % strength Twice
material of riveted the plate
joint. thickness.

48000 × .70 × (.25 × 2)
---------------------------------- = 116-2/3 lbs. = W.P.
36 × 4

Internal diam. Factor of
of boiler. safety.

But in practice such a boiler would not be used for pressures above about 75 lbs. per square inch, hence the shell plate thickness is still largely in excess of the requirements, and it may be remarked that plates less than 1/4 inch thick are not used on account of the difficulty of caulking them and keeping them steam tight.

On account therefore of the excessive strength of the shell plates in boilers of small diameter, butt straps are rarely used in stationary boilers, while punching the rivet holes and other inferior modes of construction are employed. We may now consider the circumferential seams of the boilers for stationary engines, such boilers sometimes being of great length in proportion to the diameter.

In proportion as the length of a boiler (in proportion to its diameter) is increased, the construction of the circumferential or transverse seams, as they are sometimes called, becomes of more importance.

The strength of the circumferential seams is so much greater than that of the longitudinal seams that it is often taken for granted that they are sufficiently strong if made with a lap joint and single riveted, but that such is not always the case will be shown presently.

In Fig. 3255 is represented a boiler composed of three strakes (_i. e._, three rings or sections), and it is clear that as the thickness of the shell is doubled at the circumferential seams where the ends of the middle strake pass within the end strakes, therefore the strength of the lapped joint of the shell to resist rupture in a transverse direction, as denoted by the arrows A, B, is actually increased by reason of the lap of the riveted joint. But suppose this boiler to be supported at the ends only, and the weight of the shell and of the water within it will be in a direction to cause the middle of the boiler to sag down, and therefore places a shearing strain on the rivets of the circumferential seams.

Moreover, the temperature of the outside of the boiler cannot be made or maintained uniform, because the fire passing beneath the bottom of the boiler first will keep it hotter, causing it to expand more, and this expansion acts to shear the rivets of the circumferential seams. In proportion as the heat of the fire varies in intensity, the amount of the expansion will vary, and the consequence is that the circumferential seams may get leaky or the joint may work, especially in boilers that are long in proportion to their diameters. It is clear, therefore, that for the very best construction at least a double riveted circumferential joint should be employed.

Leaving these considerations out of the question, however, we may find the amount of stress on the circumferential seams by multiplying the area of the end of the boiler by the working pressure, and dividing by the cross-sectional area of all the rivets in one circumferential seam.

Suppose, for example, that the diameter of the boiler is 36 inches, the working pressure 60 lbs. per square inch, and that there are in each circumferential seam 50 rivets, each 3/4 inch in diameter, and we proceed as follows:

The area of a circle 36 inches in diameter = 1017.87 square inches.

The area of a rivet 3/4 inches in diameter = .4417 square inch.

Then

Area of Working
boiler end. pressure.

1017.87 × 60
------------------ = 2765 lbs. per cross-sectional square inch of rivet.
50 × .4417

Number Area of
of rivets. each rivet.

By multiplying the area of the boiler end by the working pressure, we get the total steam pressure acting to shear the rivets, and by multiplying the number of rivets by the area of one rivet, we get the total area resisting the steam pressure, and then by dividing the one quantity into the other, we get the shearing stress per square inch of rivet section.

In the case of longitudinal seams, we have as follows, the pitch being say 2-1/8 and the rivets 3/4.

Diameter Steam Pitch.
of boiler pressure.
in inches.

36 × 60 × 2.125
------------------------------ = 5196 lbs. per square inch of rivet
2 × .4417 area.

Rivets in Area of
one pitch. rivet.

It is seen, therefore, that the stress placed by the steam pressure on the transverse seam is about one-half of that it places on the longitudinal seam. But, as before remarked, the transverse seam is subject to racking strains, from which the longitudinal seams are exempt; thus, for example, the expansion of the boiler diameter, whether uniform or not, does not strain the longitudinal seam, whereas it may severely strain the transverse seam.

The English Board of Trade rules, in assigning values to the various constructions and qualities of workmanship, assign a certain value, in the form of an addition to the factor of safety, which takes into account the difference in the stress upon the transverse and longitudinal seams, the quantities in each case having been determined both from experiment and from experience. A comparison of the different values may be made as follows:

The rules take a boiler shell made of the best material, with all the rivet holes drilled after the strakes are rolled into shape and put together, with all the seams (both longitudinal and transverse) fitted with double butt straps each at least five-eighths of the thickness of the shell plates they cover, and with all the seams at least double riveted, with rivets having an allowance of not more than 75 per cent. over the single shear, and provided that the boilers have been open to the inspection of their surveyors during the whole period of construction, and say that such a boiler shell shall be allowed a factor of safety (divisor of seam strength) of 5.

But for every departure from this, which they deem the best mode of construction, a penalty in the shape of an addition to the factor of safety is made. These additions to the factors of safety with reference to the longitudinal as compared to the transverse seams, are given in the following table:

-------------------------------------+----------------+---------------
|Addition to the |Addition to the
|factor 5 if the |factor 5 if the
Nature of the deviation in the |deviation is in |deviation is in
construction or workmanship. |the longitudinal|the transverse
| seam. | seam.
-------------------------------------+----------------+---------------
The holes not fair and good | .75 | .2
Holes drilled out of place after | |
bending | .15 | .1
Holes drilled before bending | .3 | .15
Holes punched after bending | .3 | .15
Holes punched before bending | .5 | .2
Joints lapped and double riveted | |
instead of having double butt straps | .2 | .1
Joints double riveted but have single| |
butt straps | .3 | .1
Joints single riveted and have a | |
single butt strap | 1.0 | .2
Joints lapped and single riveted | 1.0 | .2
-------------------------------------+----------------+---------------

An addition of .25 is also made to the factor of safety, when the strakes are not entirely under or over. In Fig. 3256 for example, strake _b_ is within or under strake _a_ at one end and strake _c_ at the other end, hence _b_ is entirely under; strake _c_ is over _b_ and _d_, and therefore entirely over; while strake _d_ is under _c_, and over _e_, and therefore not entirely under nor entirely over.

When the rivet holes are punched they do not match properly, and unless the holes are punched somewhat smaller than the required size and reamed out afterwards, some rivets receive more stress than others, and may consequently shear in detail. It is customary, however, to punch the holes for ordinary stationary boilers, and it is with seams having punched holes therefore that we have at present to deal.

In the United States the rivet diameter and plate percentages are, in the boilers of stationary engines, usually made equal, and the reasons advanced both for and against this are as follows:

First, in favor of a greater plate percentage than rivet section, it is advanced that the plate gets thinner by wear, whereas the rivet does not, hence the wear reduces the plate section; that the plate is weakened by the punching process, and requires a greater percentage to make up its strength as compared to the rivet; that the rivets are usually of better material than the plates.

In favor of a greater rivet section than plate section, it is advanced that the shearing strength of iron is but about four-fifths of the tensile strength, and that with equal plate and rivet sections the rivet is therefore the weakest; that with punched holes the rivets may be sheared in detail, and that the rivets may be sheared gradually by the working of the joint from varying expansion and contraction.

From these premises the assumption is drawn that the weakening of the plate from being punched and from corrosion about offsets the excess of the tensile over the shearing strength, and that it is best therefore to employ such a pitch that the area of the rivet and of the metal left between the rivet holes shall be equal.

In order to do this the diameter of the rivet must be determined, and the following are the proportions given by the various authorities named:

TABLE OF THE DIAMETERS OF RIVETS FOR VARIOUS THICKNESSES OF PLATES WITH SINGLE RIVETED LAP JOINT.

---------+-------------------------------------------------------
| DIAMETER OF RIVETS.
+-------+---------+----------+----------+-------+-------
Thickness|Lloyds'|Liverpool| English |Fairbairn.| Unwin.|Wilson.
of Plate.|Rules. | Rules. |Dockyards.| | |
---------+-------+---------+----------+----------+-------+-------
in. | in. | in. | in. | in. | in. | in.
5/16 | 5/8 | 5/8 | 1/2 | 5/8 | 11/16 | 5/8
3/8 | 5/8 | 5/8 | 5/8 | 3/4 | 3/4 | 11/16
7/16 | 5/8 | 3/4 | 3/4 | 21/32 | 13/16 | 3/4
1/2 | 3/4 | 13/16 | 3/4 | 3/4 | 7/8 | 3/4
---------+-------+---------+----------+----------+-------+-------
9/16 | 3/4 | 13/16 | 7/8 | 27/32 | 7/8 | 7/8
5/8 | 3/4 | 7/8 | 7/8 | 15/16 | 15/16 | 7/8
11/16 | 7/8 | 7/8 | 7/8 | 1-1/32 |1 | 7/8
3/4 | 7/8 | 15/16 | 1 | 1-1/8 |1-1/16 | 1
---------+-------+---------+----------+----------+-------+-------
13/16 | 7/8 | 1 | 1 | 1-7/32 |1-3/32 | 1
7/8 |1 | 1-1/8 | 1-1/8 | ... |1-1/8 | 1
15/16 |1 | 1-3/16 | 1-1/8 | ... |1-3/16 | 1-1/8
1 |1 | 1-1/4 | 1-1/8 | ... |1-1/4 | 1-1/8
---------+-------+---------+----------+----------+-------+-------

From the above it is seen that with thin plates the diameter of rivet employed is about twice the thickness of the plate, whereas as the thickness of plate increases the proportion of rivet diameter decreases, and the reasons for this are, first, that with rivets twice the thickness of thick plates and pitched so as to equalize the rivet and plate sections the pitch would be too great to permit of the seams being caulked steam tight.

The diameter of the rivet having been determined, the rivet area and area of plate left between the rivet holes may be made equal by determining the pitch by the following rule:

_Rule._--To the area of the rivet divided by the plate thickness add the diameter of the rivet, and the sum so obtained is the pitch. The correctness of this rule may be shown as follows:

Suppose the rivet diameter to be 7/8 inch = decimal equivalent .875, and its area will be .6013 square inch. Suppose the thickness of the plate to be 9/16 = decimal equivalent .5625, then by the rule:

Rivet area.
Plate thickness = .5625 ) .6013 ( 1.0689
5625
----
38800
33750
-----
50500
45000
-----
55000
50625

To this 1.0689 we are to add the rivet diameter, thus:

1.0689
.8750 = rivet diameter.
------
1.9439 = pitch of the rivets.

We have thus found the required pitch to be 1.9439 inches, and as the joint is single riveted there are two half rivets or one whole one to one pitch, and if we subtract the diameter of the rivet from the pitch we shall get the width of the metal or plate left between the rivets, thus:

1.9439 = pitch of rivets.
.8750 = diameter of rivet.
------
1.0689 = distance in inches between the rivets.

If now we multiply this distance between the rivets by the thickness of the plate, we shall get the area of the plate that is left between the rivet holes, thus:

1.0689 = width of plate between rivets.
.5625 = thickness of plate.
------
53445
21378
64134
53445
---------
Area of plate = .60125625
between rivets

Here then we find the area of plate left between the rivet holes to be 6.01 square inches, and as the area of the rivet is 6.01 square inches, the two are shown to be equal.

We may now place the various rivet diameters and the pitches that will make the rivet area and plate area in a single riveted joint equal in a table as follows:

TABLE OF RIVET DIAMETERS AND PITCHES FOR SINGLE RIVETED LAP JOINTS.

-------------------+------------------+--------
Thickness of Plate.|Diameter of Rivet.| Pitch.
-------------------+------------------+--------
1/4 | 1/2 | 1-1/4
5/16 | 5/8 | 1-5/8
3/8 | 11/16 | 1-11/16
7/16 | 3/4 | 1-3/4
1/2 | 3/4 | 1-5/8
9/16 | 7/8 | 2
5/8 | 7/8 | 1-7/8
11/16 | 7/8 | 1-3/4
3/4 | 1 | ..
13/16 | 1 | 2
7/8 | 1 | 1-1/8
15/16 | 1-1/8 | 2-1/8
1 | 1-1/8 | 2-1/8
1-1/16 | 1-1/8 | 2-1/8
1-1/8 | 1-3/16 | 2-1/4
1-1/4 | 1-3/16 | 2-1/8
-------------------+------------------+--------

The rivets in double riveted lap joints, and in butt strap joints having a single cover, are spaced alike, because in both cases there are two rivets in one pitch, and the rivets are in single shear.

As there are two rivets in one pitch (instead of only one as in a single riveted joint), therefore the percentage of rivet section is doubled, and the plate section must therefore be doubled if the plate and rivet sections are to be made equal, and the rule for finding the required pitch is as follows:

_Rule._--To the amount of rivet area in one pitch, divided by the thickness of the plate, add the diameter of the rivet.

_Example._--Let the plate thickness be as in the last example 9/16, decimal equivalent = .5625, and the rivet diameter be 7/8 inch = decimal equivalent .875, the area of one rivet being .6013 square inch, and the pitch is calculated as follows:

.6013 = area of one rivet.
2 = the rivets in one pitch.
------
Plate thickness = .5625 ) 1.2026 ( 2.1377
1.1250
------
7760 2.137
5625 .875 = rivet diameter.
---- -----
21350 3.012 = pitch.
16875
-----
43750
39375
-----
43750
39375
-----
4375

We find, therefore, that the pitch is 3.012, or 3 inches (which is near enough for practical purposes), and we may now make it clear that this is correct.

In Fig. 3257 the joint is shown drawn one-half full size, and the length a of plate left between the rivet holes measures (as nearly as it is necessary to measure it) 2-5/32 inches, or 2.156, and if we multiply this by the thickness of the plate = .5625 inch, we get 1.2 square inches as the area of the plate left between the rivet holes.

Now there are two rivets in a pitch (as one-half of B, one-half of C, and the whole of F), and as the area of each rivet is .6, therefore the area of the two will be 1.2, and the plate section and rivet section are shown to be equal.

The area at _a_ is obviously the same as that at A, because the pitches of both rows of rivets are equal, this being an ordinary zigzag riveted joint.

We may now consider the diagonal pitch of the rivets, using the rule below.

The pitch × 6, + 4 times the rivet diameter
------------------------------------------- = the diagonal
divided by 10 pitch _p__{D}.

In this example the pitch has been found to be 3 inches, hence we have

.875 = diameter of rivet.
4 = constant.
-----
3.500

3 = pitch of the rivets.
6 = constant.
--
18
3.5 = rivet diameter multiplied by 4.
----
10 ) 21.5 (2.15 = the diagonal pitch.
20
--
15
10
--
50

The diagonal pitch, that is, the distance _p__{D}, Fig. 3257, is therefore found to be 2.15, or 2-1/8 inch full.

The amount of metal left between the rivets, measured on the diagonal pitch, is twice the dimension H multiplied by the thickness of the plate, and as this (with the diagonal pitch determined as above) always exceeds the pitch A or _a_, therefore if the plate fails, it will be along the line _a_, and not through the diagonal pitch.

We may now consider the total amount that the plates overlap in a double riveted lap joint zigzag riveted, this amount being twice the distance E, added to the distance V between the rows of rivets.

The distance E, Fig. 3257, is usually made one and a half times the diameter of the rivet, this being found to give sufficient strength to prevent the edge of the plate from tearing out and to prevent the rivet from shearing the plate out to the edge, rupture not being found to occur in either of these directions.

The rule for finding the distance V, when the diagonal pitch has been determined by the rules already explained, is as follows:

_Rule._--To the pitch multiplied by 11, add 4 times the rivet diameter, then multiply by the pitch, plus 4 times the rivet diameter. Then extract the square root and divide by 10.

Placed in formula, the rule appears as follows, _d_ representing the rivet diameter, and _p_ the pitch.

__________________________
\/(11_p_ + 4_d_)(_p_ + 4_d_)
---------------------------- = distance V between the rows of rivets.
10

As this rule involves the extraction of the square root of the sum of quantities above the line, and as in determining the diagonal pitch, we have already determined the distance V, it is unnecessary to our purpose to carry out this latter calculation, as it is easier to find the diagonal pitch, and then, after drawing the joint, the distance between the rows of rivets can be measured if it is required, as it might be in finding the length of plate required to roll into a strake for a boiler of a given diameter and having a double riveted lap joint.

We may now consider chain riveted joints in comparison with zigzag riveted joints, which is especially necessary, because it has been assumed by some that the second row of rivets in a chain riveted joint added nothing to the strength of the joint.

Fig. 3258 represents a chain riveted joint, having the same thickness of plate, rivet diameter and pitch as the zigzag riveted joint in Fig. 3257, and it will be seen that the plate sections at a and at _a_ are the same in the two figures, and as there are four half rivets, which are equal to two rivets, in one pitch, therefore the strength of the two joints is equal.

Each joint can be as efficiently caulked as the other, as the rivet spacing is the same and the edge of the plate is the same distance from the rivets in both cases.

The pitch of the rivets is obtained by the same rule as for zigzag riveted joints, and all we have now to consider is the distance apart of the two rows of rivets or distance V in the Fig. 3258, and for this there are two rules, the first being that it shall not be less than twice the diameter of the rivet, which would leave a dimension at H in the figure equal to the diameter of the rivet. The second rule is that a better proportion than the above is to multiply the diameter of the rivet by 3. This makes the dimension at H equal to twice the rivet diameter.

When the joints have double buttstraps, the rivets may be spaced as wide as the necessity for tight caulking will admit, because, on account of the rivets being in double shear, the rivet percentage exceeds the plate percentage.

The allowance for the rivets being in double shear is 75 per cent., or in other words, a rivet in double shear is allowed 1.75 times the area of the same size rivet in single shear.

STATIONARY ENGINE BOILERS.

The simplest form of horizontal boiler is the plain cylinder boiler, an example of which is given in Fig. 3259, and which is largely used in iron works and coal mines.

Boilers of this class are easily cleaned, because the whole interior can be readily got at to clean.

As the bottom of this boiler gets thinned from wear, the boiler is turned upside down, thus prolonging its life.

Fig. 3260 represents an internally fired flue boiler, known as the Cornish or Lancashire boiler. The furnace is at one end of the flues, the fire passing through them to the chimney. There is here obviously more heating surface than in the plain cylinder boiler, but somewhat less facility for cleaning.

The Galloway boiler is of this class, but has vertical water tubes placed at intervals in the flues. These water tubes are wider at the top than at the bottom. They serve to break up the body of heat that passes through the flues, and increase the heating surface while extracting more of the heat and promoting the circulation of the water in the boiler.

A water tube is one in which the water is inside and the fire outside, as distinguished from a fire tube, in which the fire passes through the tube and the water is outside. A water tube is stronger than a fire tube, because the former is subject to bursting pressure and the latter to collapsing pressure.

Vertical boilers are internally fired, and in the ordinary forms have no return tubes or flues, examples of those used for small stationary engines being given as follows.

Fig. 3261 represents an ordinary form with vertical tubes. The upper ends of the tubes here pass through the steam space--a condition that under the moderate pressures and firing that this class of boiler is subjected to is of less importance than it is in boilers having higher chimneys and therefore a more rapid draught, and using higher pressures of steam. Furthermore, the small diameters and lengths or heights in which these boilers are made give them ample strength with shells and tubes of less thickness, while the condition of tube ends with steam on one side and fire on the other is permissible without the injurious effects that ensue under rapid combustion and high pressures.

The crown sheet of the fire boxes or furnaces of this class of boiler is very effective heating surface, first, because of the great depth (and therefore weight) of water resting upon it insuring constant contact between the water and the plate, while there is no danger of the crown sheet burning from shortness of water.

A similar boiler, but with the upper ends of the tubes below the water level, is shown in Fig. 3262.

From the small diameters of these boilers, the flat surfaces are not stayed except to the extent that the holding power of the tubes serves that end.

A return flue vertical boiler is shown in Figs. 3263 and 3264. The whole of the surfaces having contact with the fire also have contact with the water, and the height of the crown sheet removes it from the intense heat of the fire. It is stayed to the top of the boiler. The fire box or combustion chamber being taper increases the effectiveness of its sides as heating surface, since the heat in its vertical passage impinges against it.

The products of combustion pass from the top of the combustion chamber through short horizontal flues, which enter an annular space surrounding the lower section of the boiler, and from this space vertical flues pass to a corresponding space at the bottom of the boiler.

The passage of the steam generated at the sides of the combustion chamber is facilitated by the taper of the chamber, which gives increased room for the steam as it gathers in ascending.

Vertical boilers for high pressures, as from 60 to 120 lbs. per inch, are represented in the figures from 3265 to 3269.

In boilers of this class, a majority contain water tubes, which, when properly arranged, promote rapid evaporation and circulation.

A boiler with _Field_ tubes is shown in Fig. 3265. It consists of an outer shell and a cylindrical fire box, from the crown sheet of which a number of Field tubes are suspended in the fire box or combustion chamber.

Fig. 3266 is a sectional view of a Field tube, the construction being as follows:

The outer tube, which is expanded into the tube plate, is enclosed at its lower end, and has at its upper end in the water space of the boiler a perforated mouth piece, from which is suspended an inner tube that extends nearly to the bottom of the outer tube.

As the outer tube is bathed in the fire, steam is generated very rapidly, and a thorough and rapid circulation is kept up, the water passing down the inner and up the outer tubes, as denoted by the arrows.

The outer tube is spread out at the upper end to a slight cone, so that it cannot be forced out of the tube sheet by the pressure, and as it hangs free, there is no liability for it to loosen or get leaky from expansion and contraction.

From the great amount of heating surface obtained with these tubes, the fire box may be kept at a minimum diameter for the duty, while still leaving a wide space for the water leg, which facilitates the circulation.

The damper, which is suspended in the uptake, spreads the fire sideways.

Fig. 3267 represents the arrangement of Field tubes in a boiler.

A boiler of this form may for a given capacity be made lighter and smaller than in any other of the ordinary forms, while the rapid circulation acts to keep the tubes clean.

The inner tubes may be thin, because they are under pressure both inside and out, while the outer tubes may be thin, because they are under a bursting strain, whereas a fire tube is under collapsing pressure.

A design of high rate boilers, in which the uptake does not come into contact with the water, and water tubes are employed, is shown in Fig. 3268. In the fire box is an inclined tube which promotes the circulation, and is very effective heating surface, and in the combustion chamber are a number of vertical water tubes.

Two manholes give access for cleaning purposes.

The efficiency of the heating surface in this class of boiler is increased from the fact that, as the heat does not pass direct through the boiler, it impinges against the surface. In Fig. 3269, for example, the exit from the spherical fire box is on one side of the boiler, and the uptake on the other, the heat passing from the fire box into a combustion chamber, and thence through the horizontal fire tubes to the uptake.

The crown sheet is here stayed by gusset stays, but if made spherical, as in Fig. 3270, the stays may be omitted.

Fig. 3271.]

Figs. 3271, 3272, and 3273 illustrate a 60-inch horizontal return tubular boiler constructed by the Hartford Steam Boiler Inspection and Insurance Company. This class of boiler has found much favor in the United States. It is an externally fired, return tube boiler, the fire passing beneath the boiler and returning through the tubes to the front end of the boiler, whence it passes through the drum to the chimney.

The boiler is supported on the brackets B, B´, the front one, B, resting on an iron plate imbedded in the brickwork, and the back ones on rollers which rest on the plates P´ imbedded in the brickwork. This allows the boiler to expand and contract endways under variations of temperature without racking the brickwork.

A, A, etc., are for holding the brickwork together. The blow-off pipe C is for emptying or blowing down the boiler. The feed-pipe F enters the front end of the boiler, passes along it, and then crosses over. A pipe H from the steam space of the boiler supplies steam to the steam gauge G, and to the upper end of the gauge glass, which is on the casting K. The lower end of the gauge glass receives water from a pipe which passes into the water space of the boiler; at J are the three gauge cocks for testing the height of the water in the boiler.

The manhole affords ingress into the boiler for inspecting and for scaling or cleaning it, the nozzles being for a safety valve. At E is a hand-hole for washing out and cleaning the boiler. P is a damper in the fire door for admitting air above the fire bars, and R is a damper for regulating the draught.

In the brick walls that support the boiler there are air spaces to prevent the conduction of the heat through and prevent cracking of the brickwork. The tubes are arranged in vertical and horizontal rows and are equally spaced throughout.

Fig. 3274 represents the front end, and Fig. 3275 a longitudinal sectional view of the front end of a boiler of this class. In this case, however, the pipes for the water gauge pass direct into the boiler.

In some practice the tubes are arranged as in Fig. 3276, being wider pitched or spaced in the middle of the boiler to increase the circulation of the water in the boiler.

Another arrangement is shown in Fig. 3277, the tubes being _staggered_ or arranged zigzag. This permits of the employment of a greater number of tubes, but does not afford such free circulation of the water.

Fig. 3278 represents an arrangement where the tubes are in rows both vertically and horizontally.

Fig. 3279 represents a boiler by the Erie Iron Works, the details of the setting being as follows:

Fig. 3280 is an end view of the setting with the brickwork in section.

Fig. 3281 side view of the boiler and setting.

Fig. 3282 a front end view of the boiler, and Fig. 3283 a ground plan of the brickwork. When the front plate of the boiler setting extends above the middle of the boiler, as in Fig. 3279, it is said to have a "full arch front." Whereas when this plate or casting extends to the middle only of the boiler, it is said to have a half arch front.

Figs. 3284, 3285, 3286, and 3287 show the setting for a half arch front boiler, the dimensions of the settings of both these boilers being given in the following tables:

MEASUREMENTS FOR SETTING TUBULAR STATIONARY BOILERS WITH FULL ARCH FRONTS.

REFERENCE LETTERS ON DIAGRAMS.