THE USE OF RECORDING GAUGES IN A FILTRATION PLANT
The use of recording gauges in the various operations about a filtration plant has been, in the past, extremely limited.
The plant at Harrisburg, the writer believes, is the first one in which they were used to any extent. in this plant being used only on the coagulant feed.
At Lorian, Ohio, the coagulant-feed tanks were equiped with recording gauges, also the clear well and the loss of head gauges on the filters.
The charts illustrated here give a very clear idea of the value of these gauges on all parts of the plant, and their importance cannot be overestimated. There is probably no apparatus in connection with a filtration plant, that gives more definite knowledge of the mechanical operation of the plant than these charts, nor is there any manner, in the writer’s knowledge, of replacing them or securing a record that is comparable with that given on the chart records. Not only is it valuable in having definite knowledge of the operation of the plant and a permanent record for future reference in comparison, but it is an absolute check on the operators who have charge of the plant, and a slightest neglect is immediately apparent. If, therefore, there is any difficulty in the operation of the plant, it is an easy matter to definitely place the responsibility.
Chart No. i is a record of coagulant-feed. This is taken in parallel for two days’ run, in order to illustrate the perfect uniformity of the feed. Commencing at point A you notice the curve to point B and point A-2 and point li 2 are as nearly a perfect curve as could be secured. From point 1) to point A indicates the time when the alternate tank is in use. From point C to point D illustrates filling the tank.
Turning to chart No. 2: Curve A to B indicates the filling of the tank, B to C the period during which the tank is out of operation, C to A the period during which the tank is in operation. You will note the parallel line 1) covering a peri od of about one hour while the tank is about half empty. This was made by reason of the flow of the solution having stopped, owing to disarrangement of the motor-pump. Without a definite record of the time when this solution tank was nit of service, a careless operator would possibly ake his own time to make the necessary adjustnents. As there would be no record of the time he solutions were not in service, any explana ion he might make must be accepted With, lowevcr, this record of the time when his motor was not working, together with the character of the adjustments necessary, is given the superintendent an opportunity of definitely judging whether or not the diligence wa* used in getting the apparatus back into working condition
Charts No. i and No. 2 indicate the feed on iron-solution tanks.
Chart No. 3 is a record of the feed of the limesolution tank. The curves from A to B, A-x to A-2 and B-t to B-2 indicate the filling of the solution tank. From B to C, B-i to C-t, B-2 to C-2, the time when the alternate tank was in use and from C to A-t and C-t to A-2 and C-2 to A indicate the time during which the solutions were being supplied, or the tank was being used to perform its proper functions.
You will note the offset in the line from A-2 > B-2 near the top, at which the tank was tilled, here was evidently some slight derangement or ioppage in the filling for a few moments, which as almost immediately adjusted.
In the line from C to A t there will also he oted a break in the curve showing an interrupon about to minutes. This was undoubtedly due to the dogging of the orifice. The balance of the curves are practically perfect.
These gauges being located in the main operating room, while the dissolving tanks are located at another portion of the building, and the orifice in an upper floor give the operator instant notice of derangement in either case, and the few minutes interruption has evidenced that they were quickly adjusted.
Without these gauges it is not only possible but very likely that in neither case would the interruption have been noticed for some considerable period of time, possibly resulting in the case of the orifice stoppage of a material breaking in the coagulant feed to the water, which might affect the working of the filters, on account of lack of properly coagulated water.
Giart No. 4 is the chart of the loss of head on the filter. This curve, it will be noted, covers a period of about 2t!^’hours, which is a fair run for a filter. Point A indicates the initial loss of head in starting the filter in operation.
The curve to the point B, it will he noted, is a gradual increase somewhat towards the latter end of the run, when the loss of head will naturally increase a little more rapidly, owing to the coating which the filterhed has accumulated, in the meantime.
This chart, however, is almost a perfect curve. The slight irregularities shown are due to a slight difference and variation in the water-level of the sedimentation basin; which level, of course, is transferred to the level of the water in the filters and at points A and A-2 appa rently amount to several inches. This may be du e and probably is to a sudden letting in of the wai ter to the surface of the filterbed which has ju ist been washed, preliminary to starting it in serv ice again, thereby lowering the water-level in all the beds by reason of taking a greater quantit :y away than the pumps have been gauged to su ipply for a short time.
The time at which the filter is entirely out of commission is indicated by the short inward line from A to point C. The initial loss of head indicated on this chart is somewhat higher than on the others. This may be due to the fact that in the previous washing a greater amount of deposit was left on the bed, increasing, therefore, the surface friction in passing through.
Chart No. 6 is given merely to illustrate an abnormally long run, being a run of 35 hours, and even with this length of run the filter still had several feet of available operating head before reaching point of reduction in rate. In fact, on this plant there have been runs made as high as 39 hours. …. ,
The apparent variations m this curve are due entirely to the variations in the water-level of the basin and on account of the slight variations in the pumpage or filling of other filters after washing.
This, of course, might be controled by either pumping against an absolute cut-off level or an overflow, but would entail either a pumpage against pressure or a wastage of water, for which there seems to be no adequate excuse.
It will be noted on the charts that the initial loss of head on these filters is rarely 12 in. immediately after washing, the average loss of head being about 6 in., which is remarkably low. It may be stated, however, that the rate at which these filters operate is only 100,000,000 per acre instead of 125,000,000, as usually designed for mechanical filters, which is, of course, accountable for considerable of this difference between what is usually calculated for initial loss of head and that found in the operation of this plant.
Chart No. 7 illustrates some comparatively short runs due to an adverse condition of the water. It also illustrates an imperfect wash and an attempt to increase the run by blowing or partial washing, which should be avoided, except in the case of an emergency call for a quantity of water, and it is necessary to get the utmost capacity in your plant to supply.
The line from A to B is the curve for the run, apparently at the offset C, between A and B the operator for some reason concluded to wash and allowed his surplus water above the trough to run through the sewer instead of filtering through. The parallel curve D for an hour or so previous to this offset would indicate that there was some derangement in the outlet of the filter, possibly a valve partially shut, which did not allow it to filter at the rated capacity, thereby showing an increase loss of head as it should have done.
He, therefore, concluded to wash the filter and let off the surplus water from his troughs. From this point C to B indicates the time consumed in filtering the water from the top of the trough down to the point of the sandbed. Very large initial loss of head immediately after washing indicates that the washing was very improperly and imperfectly done.
Note the filter starting in operation again at point E, with loss of head of only about 2 ft. iess than when it was placed out of commission for washing. From this point to F we have a very even good uniform curve showing the filter doing excellent work. At this point, in order to increase the run still further, the filter was given a blow, as indicated by the perpendicular line covering another space of time—i. e., the filtered water outlet was cut off and water-pressure applied just the same as it would be in washing for a few minutes time to loosen up the sandbed and the surface accumulation, then immediately placed back in operation without wasting any water.
You will note this again reduced the loss of head about 2 ft. and increased the length of run about two hours to point where a sudden peak at G indicates the closing up of the bed when the filter was shut down and washed. This chart shows a very unsatisfactory run.
Chart No. 8 is a chart showing the water level in the clear-water basin, the drops shown and resembling somewhat the teeth of a saw, represent the points and times at which filters were washed. A glance at the 24-hour chart at the clear well shows, incidentally, how many filters were washed during that period of time and, moreover, shows the amount of water that is used in washing them, whether one filter received a more thorough washing than another—for instance, you will note at A, B and C a drop of about 1 ft. in the depth of the clear well. At D, F and F a drop of nearly r8 in., which indicates that on point A, B and C the filters were only washed about one-half the–amount of water that were at D, F. and F.
Of course, such a variation might also indicate that an abnormal amount of water was drawn from the clear well for fire or other emergency purposes of that kind: but, as this would only he at very rare intervals, and a definite knowledge of those intervals would always be available, and the curves would not be so pronounced. they need not be taken into consideration in the reading of this chart.
By a careful observation of the coagulant feed records, filter records and clear water well record each day the superintendent may become familiar with every condition of his plant and the work that it is doing, exclusive of the laboratory work, in a few moments’ observation. He has a daily report of practically everything that occurs therein and an absolute supervisor who informs him in the minutest detail of the action? of his plant and operator,
Comparing these charts with the daily reports which he should receive from his filter operator, he has a perfect check in the supervision of the whole plant, and his daily reports simply become confirmation of his records, with the exception of his chemical and Bacteriological reports which, of course, must be dependent upon the operator entirely, at least, up to the present time.