A Method for Determining the Capacity of the Waste-Weir of a Storage Reservoir on a Torrential Stream.

A Method for Determining the Capacity of the Waste-Weir of a Storage Reservoir on a Torrential Stream.

The diagram which Mr. Fteley has just made on the blackboard of the cross-section of the South Fork dam, as originally built, closely resembles the type of earthen dams which were built in 1830 and anterior by the canal engineers of New York, Pennsylvania and Ohio. By correspondence, written and oral, a general plan of such dams was agreed upon among the engineers of that day.

The general features of these dams were a top width of 12 or 15 feet, carried up not less than 10 feet above the bottom of discharge way of the waste, and slopes of 2 to 1 on each side. Where rock was used for the lower section of the dam, the lower slope was often reduced to 1to 1. The earth for the upper section, from the waste slope to beneath the rear top angle of the bank, Was selected material (clay and gravel), put on in layers of 6 to 8 inches, moistened and compacted by the travel of the teams, as heavy, grooved iron rollers were then unknown.

Almost universally there was put in the centre of the dam a puddle-wall of clay and gravel, 6 feet wide at a level 6 feet above the water line, and increasing in widtli by steps of 1 foot on each side of the puddle-wall, for a width of 10 to 12 feet. The tightest material, wet and compacted, was put in, and all of the remainder of the front (or water) slope was made of similar selected materials. Large stones, of more than 5 pounds weight, were taken out of the front part and placed on the rear slope, over which turf, soil and light materials were placed.

A gate-chamber of masonry was generally placed just within the front toe of the bank, and cast-iron discharge pipes were extended from this gate-chamber at the foot of the outer toe of the bank. Sometimes these pipes were carried half-way through the bank and discharged into a conduit of masonry, which conveyed the water through the remaining half of the bank. Mr. Fteley’s diagram shows a similar arrangement. The cast-iron pipes which t used were of 8 to 12 inches diameter, and connected by flanges, and were always laid upon and enveloped in a supporting wall of masonry with cross cut-off walls. Frequently, however, the pipes were laid upon and enveloped in heavy walls of puddle, and wide cut-offs of sheetiron put in at the pipe joints to prevent the water following the smooth exterior surfaces of the pipes.

The puddling of that period was expensive, costing for the labor alone about twenty cents per cubic yard. It was made of selected clay and fine, double-screened gravel in the proportion of one part of clay to two of gravel up to two parts of clay to one of gravel. These materials were thoroughly mixed, moistened (not wetted), and put on in layers of 6 inches, and cut with spades (not shovels), the cuts an inch apart anil then cross-cut, the workmen standing upon boards to guide the cutting and the spading extending into the cut puddle below, liy measurements made at that time this process compacted the material from its natural condition in the earth from twenty to twenty-five per cent, and in a month or two the mass became an artificial hard pan, over which water might run for some time without much abrasion.

Mr. John B. Jervis imported this method of puddling from England, and first used it on the Chenango canal in 1830. He directed me to train some of our inspectors in making the new puddle. One of these trained men, I believe, is now alive and well known on the Erie canal (J. Uasbrouck Decker). Great care was taken to have the spill-way of ample capacity, though there was no rule in practice to determine the size. If the hillsides at the dam were of rock, deep and wide trenches were cut as far away from the end of the dam as possible, to where a fall could be obtained, and then a walled channel built to conduct the waste water beyond the outer pool of the earthen dam. If the hillsides were of earth, the channels were walled with masonry and floored with timber and plank.

Frequently waste-gates were introduced at these spill-ways, but no reliance ivas placed upon them, as their usefulness depended upon uncertain human agency. The crest of the waste-weir was usually a hinged plank on top of the masonry, which could be tripped by a blow of a long pole, and thus add a foot to the depth of the discharge.

The title of the short paper which I shall presently read is “ A Method for Determining the Capacity of the Waste-weir of a Storage Reservoir on a Torrential Stream.”

I have been requested to precede this |aper by an expression of my views in regard to the proper construction of dams for storage reservoirs on torrential streams, and particularly of those which the circumstances of the case require, or render advisable, to be made of earth. I have so recently given pub-

• By Wm. j. McAlpinc, Hon. M. Am. Soc. C. E-j read at the Annual Convention of the Am. Soc. C. K.. at Seabright, N. J,, June, 1889. licity to my opinions on this subject, and heretofore in one of the sub-papers of the society, that I will now only occupy your time with general statements and refer to these papers for details.

The recent terrible disaster of the failure of an earthen dam at Conemaugh, Pa., has started in the public mind a distrust of all earthen dams similarly situated, which has been heightened by the publication of the alleged opinion of an eminent engineer, “that earthen dams should not be built on torrential streams.” I am informed, however, that this distinguished engineer denies having expressed or entertained any such opinion.

On three-fourths of the water and hydraulic works in this country there are existing dams of earth either for storage or reception. The latter, in most cases, are either within or contiguous to the towns and cities, and all at elevations far above that of the adjacent towns. Are these structures unsafe, or ilo any of them threaten disasters similar to that at Johnstown ?

The foundations and circumstances of each case, if we should, as we ought, have regard solely to the safety of the structure, will determine whether a dam for storage upon a stream should be made of masonry or of earth. The considerations of economy should never enter into the question. It may, however, be remarked that the difference in the cost of such dams made of masonry or of earth is not great, when each one is selected to suit the circumstances of the case.

When the bed of the valley and its sides are as high as the proposed crest of the dam and are of water-tight rock, at or near the surface, the dam should be built of masonry, especially if stone of suitable quality can readily be obtained. When the bed and sides of the valley are of water-tight earth, suitable material for the artificial bank is certain to be easily obtainable, and then the dam should be built of earth.

When rock occurs, or is accessible on only a portion of the bed or sides of the valley, it will generally be found that the structure can be most safely made of earth, because in a dam of masonry it is one of the most difficult problems to change the foundation connections with rock to those with earth.

The failures of dams have been chiefly from the following causes:

First—In two-thirds of the cases which I have examined, the failures may be traced directly to an improper construction of the outlet pipes or conduits, leading first to leakage, which, continually enlarged, ultimately causes a rupture.

Second—To an incomplete connection at the bottom and sides of the valley between the artificial and natural earths, both of which may be impermeable, and yet (not being properly united) may allow small leakages between them, first in films, which take up the firm earthy matter in solution, and thus continually enlarges the leak passage until finally the dam is ruptured.

Third—In earthen dams, from the want of sufficient waterwaste plan, so that great floods fill the reservoirs to overflowing, and cause the sudden destruction of the dam. There are other causes of destruction, which I will not now stop to discuss.

Dams of masonry and of earth are liable to be ruptured from all of the above causes, though the last mentioned one is peculiar to those of earth, as destruction is certain, while in masonry it is rarely destructive.

The stability of a well-made earthen dam was remarkably shown at the late disaster at the Conemaugh. The original dam, built by the late W. E. Morris, after long years of neglect was ruptured some twenty years or more ago by leakages along the conduit which contained the discharge pipes. This breach was filled up in a moderately good manner, but the height of the dam was lowered two or three feet to provide a roadway on top.

At the great flood of May 31 the water in the reservoir had risen so as to flow over the earthen dam at ten o’clock A. M., and continued to flow over for four and a half hours before the whole dam gave way. It then yielded by scouring out only that portion of the earthwork which had been put into the former breach by a railroad contractor, and leaving that portion built by Engineer Morris intact. Even this less perfect work resisted the pressure of some seven or eight feet of extra depth of water for’ four and a half hours before it failed.

I may here remark that if this dam had been maintained at its original height, it is probable that the increased depth of the”spill-way passage would have discharged the great flood without allowing the water to overflow the dam, and it would probably be now standing. 1 now return to my paper under its title proper.

It is assumed that the greatest quantity of water which will ever be poured into an artificial lake will be when the surface of the ground is frozen and covered with three feet depth of snow, and a rain of four inches depth, with a temperature sufficient to melt the whole of the snow’, equal to four inches more of water, say in six or eight hours.

That the dam is of earth, with its crest ten feet above the bottom of the waste-way. The waste-way must have a capacity to discharge the above flood without allowing the water in the lake to rise higher than four or five feet below the level of the crest of the earthen dam.

From a good many observations of floods I have noticed that the w’ater in the tributary brooks of the valleys flows at the rate of two to six miles an hour, depending upon their rate of descent, the roughness of the bed, and the sinuosity of the stream ; and that the surface water flows over the land to the smaller water channels at from an eighth to a mile an hour, when it is not too much obstructed by trees, bushes, etc., or absorbed in the thirsty earth.

In calculating the tables which follow, I have assumed that the sinuosities of the small brooks generally double the direct distance, and those of the land surface flowing waters quadruples the direct distances, but have made some exceptions to these rules, when the known circumstances warranted.

To show that these assumptions are approximately correct, I offer the following table of the calculated flow-off from the watershed of the west branch of the Croton, as compared with the gauged flow-off as given by Mr. Croes in No. 87, vol. 3, of the papers of the American Society of Civil Engineers.

I am familiar with the topography of this watershed, and I have divided it into eight sections and fifty sub-sections, and have calculated by the foregoing rates the hours which the water from each of the sub-sections would require to reach the weir, where the flood of August 15, 1867, was gauged.

Within this watershed are several lakes, ponds and swamps, which receive the falling water and retard its overflow longer periods of time. I have made some allowances for this effect, but they are of too uncertain a character to be of much value. They do not, however, have much effect upon the general result.

TAm.r. I.

* That this column may be better understood I will explain that in the first six hours the water from 0.9 square mites had passed over the weir ; in twelve hours the water from one additional square mile had passed over, etc.

The rain of August 15 and 16 was 3.38 inches, and the measured flow-off from August 15 to 20 amounted to 152,435,000 cubic feet, and in nineteen days to 227,352,000 cubic feet, or fifteen per cent of the rainfall. Rains had been falling from August 2, so that the ground was considerably saturated. The maximum flow from several other floods on this watershed has been for six or eight hours after the rains had commenced.

The calculated and total measured flow up to the twelfth hour was as 8.7 to 8.4 millions of cubic feet; up to the twentvfourth hour, as 35 to 30 millions cubic feet; up to the thirtysixth hour, as 57.9 to 58.1 cubic feet; up to the forty-eighth hour, as 75.8 to 82 cubic feet.

I have also applied the same rates of flow to several small areas of watershed (five to ten square miles), and by modifying them to suit the circumstances of the cases, have found that the calculations corresponded approximately with the measured flow.

Twenty years ago I applied rates, nearly like these herein stated, to a watershed o{ 250 square miles, when the crest of the shed rose to 500, 600 and 700 feet above the creek at the place where the dam was proposed to be built. The measurements, or rather the estimates, of the actual flow-off were too vague to render a comparison of value, though I used the table to determine the size of the waste-ways.

In that case I subdivided the watershed in sections of a mile square, and drew on my map rough, approximate, contour lines for each fifty feet of elevation, and also the position of the main stream and the small tributary brooks.

The following is an abstract of these calculations:

About the same time I made a calculation of the flow from sixty square miles of the same watershed, as follows :

In 3 hours the water from 2 square miles would reach the dam site.

By applying any assumed percentage of the flowing off of the rainfall to the above areas, the three-hour flood discharges wi 1 be given in cubic feet. The contents of the reservoir between the level of its top-water line and a level six feet higher (viz.: to four feet below the top of the earth embankment), will provide storeroom for a portion of the flood, and the weir or spill-way must have a length sufficient to discharge the remainder of maximum quantities as indicated by the above tables.

In my practice I have always provided for the waste-water in two places to provide for the partial obstruction of one of them. Merely for the purpose of illustrating my subject, I have recently made similar calculations in regard to the South Fork Conemaugh watershed, based upon a published map, and assuming the crest line to be from 1700 to 2500 feet above tide.*

On a plan of the watershed I have drawn assumed contour lines at elevations of 100 feet apart. If these were approximately correct, they would show the rates of slopes of the brooks, and to a considerable extent the land slopes.

The sections are a mile square, and the estimated hours of flow were marked on the brooks and also near the centres of each mile.

This table of the flood flow to the dam is based upon too much uncertain data to be relied upon, and yet I think its general results are useful, and will serve the object of this paper, which is to furnish the engineer with a method of determining the capacity of waste-ways.


First trial, assuming a waste-weir of about 200 feet length discharging in three hours, is-as follows :

Each foot in depth of the reservoir above top water line will store about 18,000,000 cubic feet.

♦The base of the dam is about 1500 feet above tide.

† For the first three hours the flow-off was taken as equivalent to 3 inches, for the first six hours at 6 inches, and for the first nine hours at 8 inches.


NOTE. AS this table is only given to show the method of calculating the length of the weir, the figures are not carried out accurately.

Note.—-This first trial shows that the weir of the spill-way should have been a little longer than 200 feet, to prevent the assumed flood from rising in the lake to nearer than 5 feet below the crest.

The channel below the weir must, of course, have a capacity equal to that of six feet of water over the weir. If the wasteway is merely a channel, it must have the same capacity as the weir.

No posts to display