us stacks that would meet the
requirements of a particular case there must be one which can be
constructed more cheaply than the others. It has been determined from
the relation of the cost of stacks to their diameters and heights, in
connection with the formula for available draft, that the minimum cost
stack has a diameter dependent solely upon the horse power of the
boilers it serves, and a height proportional to the available draft
required.

Assuming 120 pounds of flue gas per hour for each boiler horse power,
which provides for ordinary overloads and the use of poor coal, the
method above stated gives:

For an unlined steel stack--

diameter in inches = 4.68 (H. P.)^{2/5} (28)

For a stack lined with masonry--

diameter in inches = 4.92 (H. P.)^{2/5} (29)

In both of these formulae H. P. = the rated horse power of the boiler.

From this formula the curve, Fig. 33, has been calculated and from it
the stack diameter for any boiler horse power can be selected.

For stoker practice where a large stack serves a number of boilers, the
area is usually made about one-third more than the above rules call for,
which allows for leakage of air through the setting of any idle boilers,
irregularities in operating conditions, etc.

Stacks with diameters determined as above will give an available draft
which bears a constant ratio of the theoretical draft, and allowing for
the cooling of the gases in their passage upward through the stack, this
ratio is 8. Using this factor in formula (25), and transposing, the
height of the chimney becomes,

d^{1}
H = ----- (30)
.8 K

Where H = height of stack in feet above the level of the grates,
d^{1} = available draft required,
K = constant as in formula.

Losses in Flues--The loss of draft in straight flues due to friction and
inertia can be calculated approximately from formula (26), which was
given for loss in stacks. It is to be borne in mind that C in this
formula is the actual perimeter of the flue and is least, relative to
the cross sectional area, when the section is a circle, is greater for a
square section, and greatest for a rectangular section. The retarding
effect of a square flue is 12 per cent greater than that of a circular
flue of the same area and that of a rectangular with sides as 1 and 1-1/2,
15 per cent greater. The greater resistance of the more or less uneven
brick or concrete flue is provided for in the value of the c