American Society of Civil Engineers.

Transactions of the American Society of Civil Engineers (Volume 81) online

. (page 69 of 167)
Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 69 of 167)
Font size
QR-code for this ebook

reached by the writers, that wells, properly located, drilled, and de-
veloped, would yield continuously, for a sufficient period, all the water

discussion: watee supply of parkersburg, w. va. 835

to be reasonably demanded by the City of Parkersburg; at the same Messrs.
time, the various imits could be designed so as to be readily adapted, ^°^nd^^
with a minimum monetary loss, to use in connection with a filter plant, Stevenson,
to provide against the remote possibility of failure of the wells.

The estimated cost of the construction recommended by the writers
was $220 865. Of this sum, 54% was for distributing reservoir, rein-
forcing mains, etc., and would be equally useful in connection with
the water-works then existing, or with any other that might be
designed; 33% was for pumping stations and equipment — structures
equally useful, either with a filtration plant, or with a well supply;
and only 13% was allotted to equipment useful for a well supply alone.
Thus, it is evident that, even if the wells were to be abandoned at
some future time, the loss on the investment would be trivial, and
fully met in such times by the lesser cost of operating the well system.

Among the various developments studied in connection with the
writers' investigation at Parkersburg was that known as the Smith
infiltration system. The four principal reasons advanced at that time
by the promoters for its adoption were: Economy in installation and
operation; purified river water and not hard ground- water ; ease of
repairs and replacements; and reliability of supply. Instead of con-
firming any one of these points, the investigation of the writers de-
cidedly discredited all of them and, further, led to the conclusions:
that the infiltration system, being submerged in the bed of the river,
is subjected to the shifting of the river bottom, and to uncontrolled
variation in the thickness of the covering over the strainers; that such
variation, whether increase or decrease, would be detrimental to the
process, and that the infiltration system, being below the low-water
level, cannot be subjected to the same accurate control and careful
observations and studies of operation as are possible with a modem
filter; furthermore, that the only method of cleaning the beds is by
back-flushing, and that such a procedure, without controlling devices
for guaranteeing an even distribution of the wash-water, may do
decidedly more harm than good; and, finally, that the danger of the
introduction of polluted river water into the system was ever present.
A special report, dated January 15th, 1909, therefore, recommended
that it was unwise to introduce such a system. It is now evident that
several of the difficulties foreseen have already occurred; the author's
account of the replacements during the summer of 1916 point to
this fact.

The investigations of the Pittsburgh Filtration Commission in
1907-09, with which one of the writers was connected, showed that no
system of filtration, without a sufficient period of subsidence, will
effectively treat the Allegheny Hiver water; and that, if the rate of
filtration is increased to more than the normal moderate rate of slow
sand filtration, a coagulant must be used. The investigations at Cin-

836 discussion: water supply of parkersburg, w. va.

Messrs. clnuati fixed an average of 125 parts per million as a conservative

^"""a^nd*^^ estimate of the suspended matter in the unsubsided Ohio River water,

Stevenson, ^yj^ich could be treated regularly and fairly satisfactorily by slow sand

filters. On this basis, a slow sand filter without a period of subsidence

would be a failure 65% of the time, if treating this river water.

For all these reasons and because of the many points of inferiority
of the Smith infiltration system, when compared w-ith the modern slow
sand filter, designed, constructed, and operated in accordance with
accepted engineering practice, a negative recommendation on the
Smith system was made.

There is, however, nothing new or novel in the Smith infiltration
system. Although it was claimed to be an improvement on and sub-
stitute for the established method of slow sand filtration, in reality it
resembles more nearly, both in principle and in practice, the "filter
cribs" which have been used for the supply of water to communities
along the Allegheny and Upper Ohio Rivers, and, no doubt, on many
other streams, for more than 20 years. In most cases, the construction
of the latter consists of a large wooden crib, open at the sides and
bottom, placed in an excavation in the river bed, and covered on toj)
with from 5 to 7 ft. of selected sand and gravel.

At the time of the investigations, by the Pittsburgh Filtration
Commission, of possible sources of water supply for that city, con-
siderable study was given to the results obtained through filter cribs.
Analyses of the river water and of efiluents from the cribs indicated
that, in general, the effluents from the cribs were clear and free from
suspended matter, except at extreme high stages of the river, and that
the number' of bacteria in the effluents Avere usually less, but the hard-
ness materially greater, than that of the water from the river itself.
The Pittsburgh investigations showed clearly that, early in the life of
the cribs, much or all of the water is drawn from the sand and gravel
beds lying under and around them; but that, with age, the sand and
gravel beds become clogged, and larger and larger quantities of water
are drawn down from the river through the top covering. In many cases
this downward flow is increased by raking the sand covering, and
although the top covering may be broken up in this way or even re-
placed, it is impossible to clean efficiently the material under and
around the crib. With the increased infiltration direct from the river
through the top, there is a corresponding deterioration in quality.
These statements are fully confirmed by the experience of the numerous
water plants with crib supplies which have been obliged to supplement
these with filter plants within the past few years.

The author has expressed the opinion that a large percentage of
the water drawn is taken from the sand beds, possibly a larger part
than that from the open river. This opinion is quite in accord with
the experience of the writers, and it would appear not unlikely that


during the early life, while the strainers are new, and before the strata Messrs.
at the bottom and sides become clogged, 70% or more of the total "and^
water may be drawn from this lower bed. The so-called filtered water, Stevenson.
therefore, is in large part a ground-water, similar in chemical char-
acteristics to that which would be had from wells; and in so far as
satisfactory results with respect to quality are secured, they are with-
out doubt due more to the quality of the ground-water than to the effi-
cient filtration of the river water. Nor would it be safe either to
predict how long the situation might continue, or to expect other than
unsatisfactory results with the increase of the proportion of river
water. Back-flushing, if forceful enough to be successful, invariably
results in disturbing the sand layers to a dangerous degree and fre-
quently forms craters or troughs of washed material, leaving the rest
of the bed undisturbed and uncleansed.

It is worthy of note also that, at Parkersburg, the decision between
a well supply and a supply by mechanical filtration was a very close
one. Either system was considered certain to provide a safe, abundant
water supply for the community, and both systems had the endorse-
ment of the best engineering practice and were sanctioned by the most
exacting Boards of Health in adjoining States. The system adopted
had none of these assurances, and was closely akin to a practice which
has been condemned, and is rapidly being abandoned in neighboring
States, for water supply for domestic purposes. It would seem that
the time is fast approaching when most of the expenditure of $80 700
will be required to be renewed or replaced by some other system.

If the story of the development of the Parkersburg Water- Works
might be permitted to point a moral, it would seem to be that of the
advantage to be gained by every community in having a sovereign
State organization, the duties of which should include investigations
and experimental studies on all problems concerning water supply and
sewage disposal. In this manner a large fund of valuable information
might be collected, placing the supervising body in a position to
advise local authorities in these matters in an impartial way and with
freedom from the pressure of local controversial opinion. Fortunately,
this has since been recognized in West Virginia, by the re-organization
of the State Board of Health, and the addition to its staff of a com-
petent sanitary engineer. Had the State had such an organization at
the time of the adoption of the present water supply system, and had
the City been obliged to have that Health Department approve the
installation, the community might have been spared the expense of
having many different engineers make separate investigations, only to
disregard at the end the recommendations of most of them, and to put
in a system which was largely experimental and in the adoption of
which the theory and practice of water-works engineering was probably
not the sole consideration.

838 discussion: water supply of parkersburg, w. va.

Mr. Edward Mayo Tolman,* Jun. Am. Soc. C. E.f (by letter). — This

To man. coj^prgj^gnsive and valuable paper, together with the discussion on
the so-called Smith system of filtration which it has evoked, has been
of special interest to the writer, as that system has been made the sub-
ject of several bitter political controversies, into which one side or the
other has endeavored to drag him, ever since his appointment as Engi-
neer for the State Health Department, with demands that he either
praise the system to the skies, or condemn it as utterly worthless. It
was so evident that both sides, for the most part, had so far lost all
sight of the engineering principles involved, that a careful report on
the system would in no way change the opinion of those interested,
but might serve as political capital, that the writer, heretofore, has
refrained from an expression of opinion. Nevertheless, the Division
of Sanitary Engineering of the State Department of Health of West
' Virginia has gathered considerable data on the system, some of which

may prove of interest.

Beginning in January, 1917, the Department, once a week, has made
complete chemical examinations of samples of the Ohio River from
directly over the surface of the filters and from a tap in the City Hall.
The results of five pairs of samples are given in Table 5.

These results would be of greater value if analyses of water from
wells along the river bank were available for comparison. The series
of samples were collected primarily for the purpose of showing the per-
centage of ground-water in the city supply, but the well samples showed
such high contamination that the results are of no value in the present
discussion. The State Chemist, Mr. Alton A. Cook, in commenting on
these results, states :

"That practically no ground-water enters the supply is shown by
the figures for ammonia, nitrates, hardness, and alkalinity, since typical
pure ground-water would make a greater decrease in the figures for free
ammonia than is shown here, and the analytical figures for nitrates,
hardness, and alkalinity would tend to be considerably increased vmder
these conditions."

This opinion is interesting, as it is generally thought that from 20
to 40% of the water collected in the Smith system at Parkersburg is
ground-water. In the writer's opinion, the samples represent very
special conditions which can occur only when there has been a rapid
rise in the river following a considerable period of low stage, during
which the ground-water has had a chance to run out. At Gallipolis,
Ohio, and Owensboro, Ky., where somewhat similar systems have been
adopted, it is claimed that they are filtering river water, not ground-
water, and it may be that such is the case at Parkersburg. However,
at Newell, W. Ya., where the water supply is obtained from submerged

* Charleston, W. Va.

t Now Assoc. M. Am. Soc. C. E.

discussion: water supply of parkersburg, w. va.


" ' (ti r ■ ::r m « S5 2 03 Q- •

' a M

-•» ^ - '"
3-1 o

a a

(t en
tr. c:




^ orq

r«tS-} W


bocni&.inbsVoo'i-'ift.' ' iwT2.

OOIOSt-'ODlOOo'o* * ^ ^




Cft to 05 I-' _ -' ^ C



005 05

-'35 — o

totocotooioooooog S
^ jq

tOOT -JOOtOOl-"-' tCfi

to — — oi a S.

— o to 05 — B 3

O0D00^-*OCOOOOOq o

Oct' to '-I ^ o o k-' ' ' B 5

































1— '•










840 discussion: water supply of parkeksbueg, w. va.


cribs, 4 ft. below the bed of the river, the contrary is certainly true,
as shown by the three analyses in Table 6.

It is evident that, at this town, from 60 to 80% of the water is
derived from the underground-water stratum, and is not filtered river
water. Similar conditions obtain at Chester, Wellsburg, and other
West Virginia towns along the Ohio Eiver, which derive their water
supplies from cribs or wells in the river bed. The underground waters
along the Ohio in this State are very hard, and it is probable that
sufficient underground water is entering the Parkersburg system to
raise the hardness of that supply, during a large portion of the year,
above that of a filtered river water, such as would be obtained by rapid
sand filtration.

TABLE 6. — Results of Chemical Analyses of Water, Newell

(Results, in parts per million.)


March 19th, 1917.



Drilled well.









Total ash













Next, considering the character of the water from the standpoint of
turbidity and color : No analyses of the tap water made at the State
Laboratory have ever shown either color or turbidity. The highest
turbidity recorded by Dr. L. O. Rose, of Parkersburg, who made daily
tests throughout 1913 and frequent analyses since then, was 15. This
followed a flood in the Miskindum River which empties into the Ohio
a few miles above Parkersburg. At the time of the flood. Dr. Rose
informs the writer, the results of ten very careful determinations of
the turbidity of the Ohio at the filtration plant showed 25 000 parts per
million, almost, if not quite, "straight mud." In the fall of 1916, the
writer had one of his assistant engineers, Mr. Ellis S. Tisdale, visit
some 200 residences and places of business in Parkersburg and inquire
whether any turbidity had ever been noticed in the city water and
whether the water had always been satisfactory from all other stand-

discussion: water supply of parkersbukg, w. va.


points. Not a single person interviewed had anything but praise for Mr.
the supply. '^'°'"^'^"-

That the filtered water has not been highly turbid at times is
remarkable, when some of the details of construction of the plant and
of operation are considered. With the present design there is no
method of telling what proportion of the water is passing through each
filter unit. This is particularly dangerous, when it is considered that
the pumps suck the water from the collecting manifolds. Thus, if the
cover over any one filter should happen to be very thin, this unit would
be operated at too great a rate. The water should flow by gravity from
the collecting system to the pump-well, and each filter unit should be
provided with a rate controller or a recording Venturi. Such devices
would furnish data for the intelligent control of the filters. As it is,
back-flushing is now conducted with no rhyme or rhythm, as can be seen
clearly from an examination of Table 7, which shows a range of from
only 1 300 000 gal. pumped per flushing to more than 92 000 000 gal.



1915. 1916.


No. of


No. of


No. of







103 2S0 000

118 180 000
95 690 000

92 460 000
95 .350 000

115 170 000

119 600 000
105 690 000
140 830 000

91 940 000

93 770 000

104 560 000


113 030 000

100 600 000

86 910 000*
88 630 000

115 880 000
99 760 Oi)0*
98 760 000

126 800 000

101 500 000
108 690 000

87 610 000
91 500 000






122 830 000
97 560 000
87 350 000

113 710 000


97 970 000

92 530 000


125 570 000






104 300 000

114 960 000*
97 880 000
92 540 000

115 220 000

* Meter stopped.
+ Filters cleaned.

The frequency with which the filters are flushed at present depends
in no way on the quantity of water pumped or the character of the
water. It is the writer's opinion that, in the absence of any control-
ling devices on the units, daily tests of the turbidity of the river should
be made, which, combined with the stage of the river, whether rising
or falling, and the quantity of water pumped, would probably serve as
a basis for more intelligent filter operation. As it costs approximately
$5.80 per million gallons to pump the water, and as between 40 000 and
50 000 gal. are used in a single flush, it will be seen that back-flushing
has cost as much as $100 per month, a figure which could be reduced
materially if more thought were given to the problem.


842 DISCUSSION : watek supply of parkersbukg, av. va.

Mr. As to the purity of the water obtained by the so-called Smith sys-

tem, at Parkersburg: The analyses made by Dr. Rose, up to the
spring of 1916, showed a bacterial efficiency ranging from 95 to 100%,
with an average of about 98.8 per cent. In the spring of 1916, the
analyses made each month by the State Hygienic Laboratory began
to show high bacterial coimts and the presence of the colon organism
in the filtered water. Dr. Rose then started making check analyses,
and obtained similar, though less pronounced, results. To correct this
condition, the city, in June, 1916, erected, in the pump-house, a Wallace
and Tiernan chlorinating plant. The public has never been advised
of this fact for fear that all manner of tastes and odors would be
detected immediately. The results of the bacteriological analyses have
been more satisfactory since chlorination was begun, but the count is
still higher than it should be, and the colon organism is frequently
encountered in 10-cu. cm. samples. It is possible that the coli found
are entering at the reservoir. The water in the reservoir, owing to the
location of the inlet and outlet pipes, has very little circulation, and
probably some of it has stood there for many months. Dr. Rose informs
the writer that samples taken at a depth have always shown a marked
colon content.

In commenting on the results of the chemical analyses (Table 5),
Mr. Cook states:

"In general, the purification is not all that can be desired, as shown
by the figures for ammonia (free), nitrates, and oxygen consumed in
one or more of the last three samples. The city drinking water, on
the whole, is of fair quality, but there is room for improvement, before
it can be considered a first-class drinking water."

It may be said, however, that Parkersburg. has a satisfactory and
safe water supply which is being delivered to the consumer at very
low cost.

The writer has frequently heard the so-called Smith system of filtra-
tion condemned, on the ground that it cannot produce a safe or satis-
factory water. This, he feels, is a mistake. At Owensboro, Ky., Galli-
polis, Ohio, Chester and Newell, W. Va., and at other places using
systems in many ways similar to that at Parkersburg, the water is
satisfactory, both from chemical and bacteriological standpoints. At
Chester and Newell, the colon organism has never been encountered,
to the writer's knowledge, and yet no disinfectant is used.

It is the writer's opinion that the Smith system of filtration can
furnish a satisfactory water at a low operating cost, provided nothing
happens to injure the under-drains in the river, the problems of proper
back-flushing being largely those of design and operation. It would
appear then that, other things being equal, the choice of the Smith
system over a rapid sand filter hinges on a gambling chance, the engi-
neer betting that nothing will happen to the collecting drains against

discussion: watek supply of parkersbueg, w. va. 843

the high cost of chemicals. If he wins, the profit to the city is large, Mr.
but, if he loses, the city may pay in life, as well as in money. Toiman.

An interesting example of what may happen is afforded by experi-
ence at Wellsburg, W. Va., in 1916. The water supply of this town is
derived from wells in the Ohio Eiver, about 100 ft. from the shore.
These wells are constructed of 60-in. cast-iron pipe set vertically in
the river bed, with their upper ends 3 ft. below the surface of the
sand. The bottoms of the pipes were left open, but the tops were
sealed with iron caps leaded into the bells of the pipe.

The main suction line from the wells to the pump was provided
with four branches, each branch extending through the cap of one of
the wells. This system had been in use for a number of years, and had
furnished a satisfactory drinking water. In the spring of 1916, how-
ever, one of the wells began to settle, which caused a crack in the
suction main, and raw water entered the system. A typhoid outbreak
of several cases and one death followed shortly.

WTien the supply was first used, it was jKtssible to go back and
forth in rubber boots over the surface of the sand-bar in which the
wells were placed, consequently, the cost of the original construction
was comparatively small. Since then, however, the Federal Govern-
ment has built a dam across the Ohio just below "Wellsburg, increasing
the depth of the water over the bar by several feet. Investigation has
shown that the cost of repairs to the well system would be almost pro-
hibitive, and the town contemplates abandoning the present supply
entirely and looking for a new source.

That a similar accident might happen to the Parkersburg system
is shown clearly by the finding last fall (1916) of the broken strainer
pipe which nearly caused a blow-out in the sand bed.

However, it must be stated that considering the lack of skilled
attention given to this plant since it was put in commission, it has
shown remarkable results. The plant is virtually in charge of the man
at the pump, who, although very able, does not have the intimate engi-
neering knowledge necessary for the proper operation of a practically
untried form of filtration. The Water Commissioners are not trained
engineers, and their supervision extends primarily to the financial side
of the water- works problem. Parkersburg has adopted a very short-
sighted policy in not placing a trained man in charge of the operation
of the plant, and the fruits of this policy are now being noticed. It
is probable that, if the plant had been operated scientifically, instead
of in the hit-or-miss fashion of the past, the results would have been
more to the credit of the system.

William M. Hall,* M. Am. Soc. C. E. (by letter). — The writer is Mr.
gratified with the extent of the discussion, and wishes to express his
appreciation to each of those who have contributed to it.

* Parkersburg, W. Va.

844 DISCUSSION : water supply of parkersburg, w. ya.

Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 69 of 167)