Self-Purification in Sewers
303
Fig. 7 shows a plot of e vs. su. The regression line indicates q = 0.51. The standard
error of the exponent is about 0.11. The data are not free from internal correlations; the
uncertainty of the exponent, therefore, is greater than indicated by this standard error.
Taking q to be
X
A and determining CE, the equation becomes
e = 5.3(su)
1/2 (e in m/hr, u in m/sec)
(7)
In filled pipes, s varies approximately as u
2 . Hence (su)
1/2 is proportional to u
3/2 . Heat
transfer to or from the wall of a pipe is generally held to vary as u
2 .
Eq. 7 applies only where the slime is an effective oxygen sink. In a well aerated stream
of low oxygen demand, the substrate may be entirely aerobic, in which case the rate of
usage by the slime layer would not be proportional to the oxygen concentration.
One should not think of the slime layer as a bulky mass. Where velocities are good, the
submerged pipe wall usually appears clean, but the slime layer always gives the surface a
characteristically slippery feel unless it is cleaned off by rubbing. It is significant that the
highest oxygen flux to the pipe wall, 1.4 grams/m
2 -hr, was in line 18J at a velocity of 7.1
fps (2.2 m/sec). The finding of a highly reactive slime layer where the stream is so swift is
in a sense an extension of the finding, previously reported [3], that a velocity of 4 fps did
not diminish sulfide production in a force main. Let no one suppose that high velocity
will strip a pipe wall free of biologically active material, unless it be at some velocity
greater than 7 fps, or unless much abrasive material is present.
Condition of Deficient Oxygen Supply
Three of the reaches tested (48D, 27D, and 27J) had very low oxygen concentrations
at the downstream ends. It was calculated in all of these cases that the oxygen
concentrations had come close to the steady state values, where d(0 2 )/dt contributed
little or nothing to the balance. The rates of oxygen consumption equaled the rates of
supply.
By using Eq. 7 to predict slime usage and Fig. 2, together with pot tests, to estimate
R r , predictions may be made of the steady state concentrations. Seventeen tests in the
three lines mentioned showed oxygen concentrations ranging from 0.03 to 0.27 mg/1.
Predicted concentrations were in the same range, but the standard error of the predictions
for individual runs was 0.09 mg/1. In view of the errors of measurement and the fact that
Fig. 2 represents only an average trend, closer agreement would not be expected; it is
sufficient that the observed range is rationally accounted for.
Deductions
The significance of these findings for the problem of septicity in sewers will be treated
elsewhere.
Wherever the slime layer acts as an efficient oxygen sink, and with normal surface
aeration providing the only source of oxygen, concentrations will drop to low levels. The
transfer coefficient to the slime layer is in general about 2Vi times as great as the
coefficient for absorption of oxygen from the air. If no oxygen were used in a stream of
sewage, but the slime layer were active, and if the stream were broad and shallow, so that the
surface area was practically equal to the slime area, the steady-state oxygen concentration
would be 1/3.5 or 28% of saturation. In a half-filled sewer it would be 20% of saturation.
In a very small sewer, with swift flow, oxygen consumption in the stream may be small
compared to the through-flux, and these concentrations might be approached, but the
303
Fig. 7 shows a plot of e vs. su. The regression line indicates q = 0.51. The standard
error of the exponent is about 0.11. The data are not free from internal correlations; the
uncertainty of the exponent, therefore, is greater than indicated by this standard error.
Taking q to be
X
A and determining CE, the equation becomes
e = 5.3(su)
1/2 (e in m/hr, u in m/sec)
(7)
In filled pipes, s varies approximately as u
2 . Hence (su)
1/2 is proportional to u
3/2 . Heat
transfer to or from the wall of a pipe is generally held to vary as u
2 .
Eq. 7 applies only where the slime is an effective oxygen sink. In a well aerated stream
of low oxygen demand, the substrate may be entirely aerobic, in which case the rate of
usage by the slime layer would not be proportional to the oxygen concentration.
One should not think of the slime layer as a bulky mass. Where velocities are good, the
submerged pipe wall usually appears clean, but the slime layer always gives the surface a
characteristically slippery feel unless it is cleaned off by rubbing. It is significant that the
highest oxygen flux to the pipe wall, 1.4 grams/m
2 -hr, was in line 18J at a velocity of 7.1
fps (2.2 m/sec). The finding of a highly reactive slime layer where the stream is so swift is
in a sense an extension of the finding, previously reported [3], that a velocity of 4 fps did
not diminish sulfide production in a force main. Let no one suppose that high velocity
will strip a pipe wall free of biologically active material, unless it be at some velocity
greater than 7 fps, or unless much abrasive material is present.
Condition of Deficient Oxygen Supply
Three of the reaches tested (48D, 27D, and 27J) had very low oxygen concentrations
at the downstream ends. It was calculated in all of these cases that the oxygen
concentrations had come close to the steady state values, where d(0 2 )/dt contributed
little or nothing to the balance. The rates of oxygen consumption equaled the rates of
supply.
By using Eq. 7 to predict slime usage and Fig. 2, together with pot tests, to estimate
R r , predictions may be made of the steady state concentrations. Seventeen tests in the
three lines mentioned showed oxygen concentrations ranging from 0.03 to 0.27 mg/1.
Predicted concentrations were in the same range, but the standard error of the predictions
for individual runs was 0.09 mg/1. In view of the errors of measurement and the fact that
Fig. 2 represents only an average trend, closer agreement would not be expected; it is
sufficient that the observed range is rationally accounted for.
Deductions
The significance of these findings for the problem of septicity in sewers will be treated
elsewhere.
Wherever the slime layer acts as an efficient oxygen sink, and with normal surface
aeration providing the only source of oxygen, concentrations will drop to low levels. The
transfer coefficient to the slime layer is in general about 2Vi times as great as the
coefficient for absorption of oxygen from the air. If no oxygen were used in a stream of
sewage, but the slime layer were active, and if the stream were broad and shallow, so that the
surface area was practically equal to the slime area, the steady-state oxygen concentration
would be 1/3.5 or 28% of saturation. In a half-filled sewer it would be 20% of saturation.
In a very small sewer, with swift flow, oxygen consumption in the stream may be small
compared to the through-flux, and these concentrations might be approached, but the