You are on page 1of 8

Generalized Method for Storm-Water Pumping

Station Design
S. David Graber, P.E., F.ASCE1

Abstract: The present paper presents a generalized solution for the hydrologic and hydraulic design of small to medium-sized storm-
water pumping stations. The solution enables determination of pump capacity and the corresponding storage capacity. The benefits of the
solution include: 共1兲 the ability to consider the effects of storm duration rather than the fixed hydrology assumed in other methods; 共2兲 a
Downloaded from ascelibrary.org by Cambridge University on 06/24/15. Copyright ASCE. For personal use only; all rights reserved.

simpler and more intuitive design procedure than other available methods; and 共3兲 enabling the calculation of storage requirements
independent of the geometry of the wet well and any inundated upstream area. The relation of the method presented herein to several other
available methods is presented and the other methods critiqued conceptually and by example. The example is favorably tested using
commercial software, and the constant-pumped-discharge assumption of the example is further tested using that software for an actual
pump discharge-total dynamic head and system flow-head relationship. Extension of the methods presented in this paper to more complex
situations is discussed. The importance of considering different storm durations is emphasized.
DOI: 10.1061/共ASCE兲HE.1943-5584.0000268
CE Database subject headings: Dimensional analysis; Pumping stations; Rainfall duration; Rainfall intensity; Stormwater
management; Surface drainage.
Author keywords: Dimensional analysis.

Introduction dS
=I−Q 共1兲
dt
Small to medium-sized pumping stations are defined by Water
in which I = volumetric inflow rate; Q = volumetric outflow rate;
Environment Federation 共WEF兲 共1993兲 as those for which a
and dS / dt = time rate of increase of effective storage volume.
single pump provides design capacity 共with one or more others Second is the inflow relationship, which is assumed to have
for low flows, alternating sequences, and/or backup兲; such sta- the form of the trapezoidal hydrograph depicted on Fig. 1 with
tions are of less than 200 L/s 共3,200 gpm兲 capacity 共WEF 1993兲. peak inflow I p, linear rising limb to the time of concentration tc,
Larger submersible pumps have become available 共ASCE/WEF duration tr, and linear receding limb of duration Rtc; additional
1992兲, which allow larger storm-water pump stations to rely on a terms on Fig. 1 are defined below. The trapezoidal inflow hy-
single submersible pump for design capacity. The present paper drograph shown on Fig. 1 is piecewise continuous, having one
presents a generalized solution for the hydrologic and hydraulic continuous relationship applicable for t 艋 tc, another applying
design of small to medium-sized storm-water pumping stations for tc ⬍ t 艋 tr, and another for tr ⬍ t 艋 共tr + Rtc兲. Mathematically
and discusses the practical aspects of its use. The benefits of the 共Graber 2009a兲
solution include: 共1兲 the ability to consider the effects of storm I
duration rather than the fixed hydrology assumed in other meth- = t/tc, t 艋 tc 共2a兲
Ip
ods; 共2兲 a simpler and more intuitive design procedure than other
available methods; and 共3兲 enabling the calculation of storage
I
requirements independent of the geometry of the wet well and = 1, tc ⬍ t 艋 tr 共2b兲
any inundated upstream area. Ip

冉 冊
The solution can be derived by first specializing four basic
relationships on which a generalized solution for detention basin I 1 t tr
=1− − , tr ⬍ t 艋 共tr + Rtc兲 共2c兲
design are based 共Graber 2009a兲. The first is the conservation of Ip R tc tc
mass relationship given by The third and fourth relationships are simplified forms of the
storage and outflow terms in Eq. 共1兲, expressed, respectively, as
follows:
1
Consulting Engineer, 118 Larson Rd., Stoughton, MA 02072. E-mail:
sdavidgraber@cs.com S = NHᐉ 共3兲
Note. This manuscript was submitted on October 2, 2009; approved
on April 26, 2010; published online on October 15, 2010. Discussion Q = MHm 共4兲
period open until April 1, 2011; separate discussions must be submitted
for individual papers. This paper is part of the Journal of Hydrologic in which H = elevation measured upward from the elevation at
Engineering, Vol. 15, No. 11, November 1, 2010. ©ASCE, ISSN 1084- which effective storage and outflow begin and N, M, ᐉ, and m
0699/2010/11-901–908/$25.00. = coefficients and exponents defined by Eqs. 共3兲 and 共4兲.

JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / NOVEMBER 2010 / 901

J. Hydrol. Eng. 2010.15:901-908.


I
1 1
Q Inflow hydrograph, I Sⴱ = Mts + 共I p − M兲共tc − ts兲 + 共I p − M兲共tr − tc兲
2 2
Ip Outflow 1
hy drograp h, Q
+ 共I p − M兲共tⴱ − tr兲 共9兲
2
M
in which ts, the time of pump starting, is given, from Eqs. 共2a兲
and 共7兲, by
R tc
ts M
= 共10兲
ts tc tr t* te t tc I p

Fig. 1. Inflow and outflow hydrograph characteristics Substituting Eqs. 共8兲 and 共10兲 into Eq. 共9兲, rearranging, and non-
dimensionalizing

冉 冊 冋 冉 冊册
Downloaded from ascelibrary.org by Cambridge University on 06/24/15. Copyright ASCE. For personal use only; all rights reserved.

Sⴱ tr M 1 M M 1
Eq. 共3兲 is presented only for comparison with related formu- = 1− + 共R − 1兲 − R− 1+ R 共11兲
I pt c t c Ip 2 Ip Ip 2
lations 共Graber 2009a,b兲; for present purposes no restrictive as-
sumption is made at this stage regarding the nature of S. Eq. 共4兲 is The nondimensional inflow volumes are derived by integrating
specialized for the case of constant pumped outflow 共when the Eqs. 共2兲, giving

冉冊
pump is on兲 so that m = 0, Hm = 1, and Q = M. Constant pumped 2
V 1 t
outflow is a useful preliminary design assumption, also made by = , t 艋 tc 共12a兲
others 关e.g., Burton 1980; Baumgardner 1983; ASCE/WEF 1992; I pt c 2 t c
Federal Highway Administration 共FHWA兲 1982, 2001; Froehlich
1994兴. The constant-pumped flow is assumed to start at ts = time V t 1
of intersection of M with the rising limb of the hydrograph, as = − , tc ⬍ t 艋 tr 共12b兲
I pt c t c 2
also assumed by Burton 共1980兲 and depicted on Fig. 1.

Generalized Solution
V t
= − −冉 冊
1 t tr
I ptc tc 2R tc tc
2
1
− , tr ⬍ t 艋 共tr + Rtc兲
2
共12c兲

The total inflow volume Vi is given by Eq. 共12c兲 for t = 共tr + Rtc兲
Based on the above, the applicable differential equation is given
by Vi R + 1 tr
= + −1 共13兲
I pt c 2 tc
dS
= I共t兲 − M 共5兲 The ratio of maximum storage volume to total inflow volume is
dt
obtained from Eqs. 共11兲 and 共13兲

冉 冊 冋 冉 冊册
in which the more generalized storage term S is retained for the
moment. Eq. 共5兲 can be nondimensionalized with respect to the tr M 1 M M 1
1− + 共R − 1兲 − R− 1+ R
peak inflow I p, time of concentration tc, and total inflow volume Sⴱ tc Ip 2 Ip Ip 2

冉 冊
Vi = 共14兲
Vi R + 1 tr

冋 册
+ −1
2 tc
d共S/Vi兲 I共t兲 M
= B⬘ − 共6a兲
d共t/tc兲 Ip Ip The following ratio of Sⴱ / Vi for tr / tc → ⬁ is obtained by taking
the limit of Eq. 共14兲

B⬘ =
I pt c
Vi
共6b兲 lim
Sⴱ
tr/tc→⬁ Vi
= 1−冉 冊
M
Ip
共15兲

One wants to solve for the maximum storage and concurrent Eqs. 共14兲 and 共15兲 are plotted on Fig. 2 for R = 1.67; the selection
maximum outflow, which occur when dQ / dt = d共Qᐉ/m兲 / dt = 0 and of this R value is explained in Graber 共2009a兲. The values of
I = Q 共Graber 2009a兲 共see Fig. 1兲. Denoting the magnitude of the Sⴱ / Vi are equal 共independent of tr / tc兲 at M / I p = 共1 + R兲 / 共2 + R兲 at
maximum outflow and the time at which it occurs by Qⴱ and tⴱ, which Sⴱ / Vi = 1 / 共R + 2兲; the numerical values for R = 1.67 are
respectively, one obtains M / I p = 0.728 and Sⴱ / Vi = 0.272.

I共tⴱ兲 = Qⴱ 共7兲
Noting that tⴱ, as depicted on Fig. 1, must fall within the range Other Methods
tr ⬍ tⴱ 艋 共tr + Rtc兲, the above equation together with Eq. 共2c兲 gives
Burton 共1980兲 similarly applied the modified rational method
tⴱ
tr
= +R 1−
tc tc
M
Ip
冉 冊 共8兲
with a trapezoidal inflow hydrograph, and also recognized the
importance of storm duration: “We must find the storm duration,
which…produces the maximum required storage volume.” His
The maximum storage Sⴱ is then given by the area between the equations are limited to the case of R = 1, a specific form of rain-
inflow hydrograph and constant pumped flow 共the cross hatched fall intensity-duration relationship of limited applicability, and
area on Fig. 1兲 other restrictive assumptions.

902 / JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / NOVEMBER 2010

J. Hydrol. Eng. 2010.15:901-908.


1

0.9

0.8
R = 1.67 except as noted
0.7

0.6
S */Vi

0.5

0.4
t r /t c = 1.0

0.3
1.5
Downloaded from ascelibrary.org by Cambridge University on 06/24/15. Copyright ASCE. For personal use only; all rights reserved.

0.2 2.0

ASCE/WEF (1992) for R = 1 3.0


0.1 5.0

0 Infinite

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
M /I p

Fig. 2. Sⴱ / Vi versus Qⴱ / I p and tr / tc

Froehlich 共1994兲 expanded on Burton 共1980兲 by using a form problem: the storage is maximized at an absolute maximum as
of rainfall-intensity relationship of broader applicability and im- M → 0. This is supported by Fig. 3, which is referenced for the
proved upon Burton’s assumptions for the case of a prescribed example discussed below.
value of constant pump discharge M 共using the present paper’s Also plotted on Fig. 2 is the relationship for estimating the
notation兲. He then attempted to apply calculus of variations 共e.g., required storage given in ASCE/WEF 共1992兲 共although attributed
Hildebrand 1952兲 to the trivariate 共two independent variables兲 therein to the Federal Highway Administration 共FHWA 1982兲, the
problem of maximizing storage volume for both variable I p and method does not appear in that publication兲. That relationship is
variable M. However, the writer’s review of that derivation re- based on fitting a triangular hydrograph to the peak-flow portion
veals an error in the mathematics: the correct solution for one of of a Soil Conservation Service 共SCS兲 共1986兲 关Soil Conservation
the variables 共I p兲 is an imaginary algebraic expression. That Service, now National Resource Conservation Service 共NRCS兲兴
means that no relative maximum exists over the real range of type inflow hydrograph, and deriving an equation given, in terms
values of that variable. This comports with the physics of the of present notation, by Sⴱ / Vi = 共1 − M / I p兲2. The triangular hy-

M (cfs)
0.0 1.0 2.0 3.0 4.0 5.0
7000

180
6000
160

140 5000

120
4000
S* (cu ft)
S* (m )
3

100

3000
80

60
2000

40
1000
20

0 0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
3
M (m /s)

Fig. 3. Example maximum storage volume versus constant pumped outflow

JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / NOVEMBER 2010 / 903

J. Hydrol. Eng. 2010.15:901-908.


drograph corresponds on Fig. 2 to tr / tc = 1; the approximation can Example
be seen to follow the curve for tr / tc = 1 up to M / I p ⬵ 0.15 and to
significantly underpredict the required storage for larger values of A drainage area containing homes, appurtenant structures, and
M / I p 关if the writer’s curves are plotted for R = 1, their distance pervious surfaces is protected from external flooding by a berm.
above the ASCE/WEF 共1992兲 curve becomes even greater兴. Even However, runoff from the drainage area itself requires a perma-
for tr / tc = 1 and R = 1 as assumed in the ASCE/WEF 共1992兲 nent solution to address flooding within the area behind the berm
method, Eq. 共14兲 does not reduce to the ASCE/WEF equation; the during times when the external water level is too high to permit
method fails to consider what Baumgardner 共1983兲 refers to as one-way gravity discharge through the berm. At a time of con-
“storage below pump-on elevation” and which, for the case con- centration of 6 min, the peak design 共50-year兲 rate of runoff using
sidered here, is the storage associated with the area to the left of the SCS graphical peak discharge method 共SCS 1986兲 is
tc 共see Fig. 1兲 which becomes more significant as M / I p gets 0.182 m3 / s 关6.43 cubic feet per second 共cfs兲兴. An essentially iden-
larger. This is further addressed in an example presented below. tical value of peak runoff 共within 1% in this case兲 can alterna-
FHWA 共2001兲 properly cautions against using long-duration tively be obtained by converting the curve number value to a
events, such as SCS/NRCS 24 h hyetographs: “Shorter duration Rational Method runoff coefficient 共Graber 1992兲 and using the
Downloaded from ascelibrary.org by Cambridge University on 06/24/15. Copyright ASCE. For personal use only; all rights reserved.

storms that compare with the estimated time of concentration rational method. Table 1 gives values of tr, tr / tc, and rainfall
for the drainage area are usually more appropriate for pump sta- intensity i. The latter are based on published rainfall data for
tion design.” A method which accomplishes that and is related to Topsfield, Massachusetts 共Graber 1992兲, with linear interpolation
the method discussed in the previous paragraph is given in ASCE/ for intermediate values of i. Successively larger storm durations
WEF 共1992兲. It entails a triangular approximation to the maxi- are considered; the peak rate of runoff decreases in proportion
mum intensity portion of the SCS/NRCS hyetograph. A to the decrease in rainfall intensity as the storm duration increases
comparative example of that method is also given below. 共Graber 2009a,b兲. Values of Vi / 共I ptc兲 are calculated using
Eq. 共13兲; values of Vi are also given, which shows their increase
with increasing tr.
Pump Operation Different values of M are then considered, with the remaining
columns of Table 1 based on one such value. For each value of
From Fig. 1 and Eq. 共10兲, the pump start volume is given by M / I p, values of tⴱ / tc and Sⴱ / Vi are calculated using Eqs. 共8兲 and
共14兲, respectively. The values of Sⴱ / 共I ptc兲 are calculated from
V p,start = 共1/2兲tsM = 共1/2兲M 2tc/I p 共16兲 共Sⴱ / Vi兲关Vi / 共I ptc兲兴; Eq. 共11兲 could alternatively be used. Values of
The start volume is measured upward from the dead-storage Sⴱ are then calculated from I ptc关Sⴱ / 共I ptc兲兴. The procedure is easily
elevation in the wet well, which may generally be taken to be repeated for other values of M in the same spreadsheet 共not
equal to the pump-stop elevation. Water below that elevation must shown兲. The maximum values of Sⴱ all correspond to values of
be assumed to remain in the wet well. The dead-storage elevation M / I p, tr / tc, and Sⴱ / Vi that comport with Fig. 2; that figure could
can be based on the minimum required submergence of submers- be used instead of Table 1 to estimate values of Sⴱ / Vi giving
ible pumps or the minimum acceptable submergence of bell in- maximum Sⴱ. It should be understood that values in Table 1 of
lets. Avoidance of air-core vortices is one important consideration tⴱ / tc, Sⴱ / Vi, Sⴱ / 共I ptc兲, and Sⴱ for other than the maximum Sⴱ are
in this regard. For submersible pumps, the minimum submer- only trial values with no physical significance.
gence is also based on cooling requirements. Some submersible Fig. 3 plots the maximum storage volume versus constant-
pumps have a cooling jacket with a glycol coolant that allows the pumped outflow. Based on economic and space considerations,
pump to operate without the motor being submerged; in such the 0.0849 m3 / s 共3.00 cfs兲 pump capacity and corresponding
cases, the minimum submergence may be at the elevation of the 44.6 m3 共1,580 ft3兲 storage capacity were selected. Pertinent val-
top of the volute 共the manufacturer should always be consulted兲. ues of the selected design in relation to Fig. 1 are M
For bell inlets and other configurations, minimum submergence = 0.0849 m3/s 共3.00 cfs兲, I p = 0.138 m3/s 共4.87 cfs兲, M / I p
guidance is provided by American National Standards Institute, = 0.0849/0.138= 0.616, ts = 0.616共6兲 = 3.70 min 关Eq. 共10兲兴, tc
Inc./Hydraulic Institute 共ANSI/HI兲 共1998兲. = 6 min, tr = 14 min, tⴱ = 14+ 1.67共6兲共1 − 0.616兲 = 17.8 min 关Eq.
Referring to Fig. 1, the pump will continue to run until the 共8兲兴, tr + Rtc = 14+ 1.67共6兲 = 24 min, and te = 0.616共6兲 + 关14/ 6
time that maximum storage has been reached. Beyond that time, + 1.67/ 2 − 共1 / 2兲兴共6 / 0.616兲 = 29.7 min. 关Eq. 共18兲兴. Fig. 4 plots the
the pump will continue to run until the difference between the mass curves: cumulative inflow volume based on Eqs. 共12兲 and
cumulative inflow volume and the cumulative pump discharge 共13兲, and cumulative discharge volume based on Eq. 共17兲. The
volume equals zero. Referring again to Fig. 1, this will generally maximum difference between the two curves equals the 44.6 m3
occur sometime after t = tr + Rtc. The cumulative inflow volume 共1 , 580 ft3兲 storage capacity, and the two curves can be seen to
based on Fig. 1 has reached a maximum and remains unchanged intersect at te where the pump stop volume is zero.
for t ⬎ tr + Rtc, and is given by Eq. 共13兲. The pump start volume of 9.43 m3 共333 ft3兲 关Eq. 共16兲兴, maxi-
Denoting the cumulative pump discharge volume by V p, the mum storage volume of 44.6 m3 共1,580 ft3兲, and pump stop
non-dimensional cumulative pump discharge volume is given by volume of zero are relative to a datum which is the mini-
mum submergence. The minimum submergence for the selected

= 冉 冊
V p M t ts

I pt c I p t c t c
共17兲
submersible, glycol-cooled, pumps is 33 cm= 0.33 m 共13 in.
= 1.083 ft兲. The selected underground storage chamber has verti-
cal sides and interior plan dimensions of 2.74 m共9.0 ft兲
Setting Vi = V p using Eqs. 共13兲 and 共17兲, and solving for t / tc, gives ⫻ 4.88 m共16.0 ft兲 = 13.4 m2 共144 ft2兲, for which the corresponding
the pump stop time, denoted by te, as follows: depths from the bottom of the chamber are: 0.33 m 共1.083 ft兲
stop, 0.33+ 9.43/ 13.4= 1.03 m 共3.39 ft兲 start, 0.33+ 44.6/ 13.4
te M tr/tc + R/2 − 1/2 = 3.66 m 共12.0 ft兲 maximum.
= + 共18兲
tc I p M/I p To check pump cycling during more frequent storms with

904 / JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / NOVEMBER 2010

J. Hydrol. Eng. 2010.15:901-908.


Table 1. Storm-Water Pumping Station Design Example
M = 0.0849 m3 / s = 3 cfs
tr 50-year i 50-year i Ip Ip Vi Vi Sⴱ Sⴱ
共min兲 tr / tc 共mm/h兲 i共in. / h兲 共m3 / s兲 共cfs兲 Vi / 共I ptc兲 共m3兲 共ft3兲 M / Ip tⴱ / tc Sⴱ / Vi Sⴱ / 共I ptc兲 共m3兲 共ft3兲
5 0.833 195.6 7.700 0.188 6.658 1.168 79.3 2,800 0.451 1.751 0.353 0.413 28.0 990
6 共tc兲 1.000 188.9 7.436 0.182 6.429 1.335 87.5 3,090 0.467 1.891 0.366 0.489 32.0 1,131
7 1.167 182.1 7.172 0.176 6.201 1.502 94.9 3,352 0.484 2.029 0.372 0.559 35.3 1,247
8 1.333 175.4 6.907 0.169 5.972 1.668 101.6 3,587 0.502 2.164 0.373 0.623 37.9 1,339
9 1.500 168.7 6.643 0.163 5.744 1.835 107.4 3,794 0.522 2.298 0.371 0.680 39.8 1,406
10 1.667 162.0 6.379 0.156 5.516 2.002 112.5 3,974 0.544 2.428 0.365 0.730 41.0 1,449
11 1.833 157.3 6.193 0.152 5.355 2.168 118.3 4,180 0.560 2.568 0.360 0.782 42.7 1,507
12 2.000 152.6 6.007 0.147 5.194 2.335 123.6 4,366 0.578 2.705 0.354 0.827 43.8 1,547
Downloaded from ascelibrary.org by Cambridge University on 06/24/15. Copyright ASCE. For personal use only; all rights reserved.

13 2.167 147.8 5.821 0.142 5.033 2.502 128.3 4,533 0.596 2.841 0.346 0.867 44.5 1,570
14 2.333 143.1 5.635 0.138 4.872 2.668 132.5 4,680 0.616 2.975 0.337 0.899 44.6 1,577
15 2.500 138.4 5.449 0.133 4.711 2.835 136.1 4,808 0.637 3.107 0.326 0.924 44.4 1,567
16 2.667 134.7 5.303 0.130 4.585 3.002 140.3 4,954 0.654 3.244 0.316 0.950 44.4 1,568
17 2.833 131.0 5.156 0.126 4.458 3.168 144.0 5,085 0.673 3.380 0.306 0.969 44.0 1,555
18 3.000 127.2 5.010 0.123 4.332 3.335 147.2 5,201 0.693 3.513 0.294 0.981 43.3 1,530
19 3.167 123.5 4.863 0.119 4.205 3.502 150.1 5,301 0.713 3.645 0.281 0.985 42.2 1,491
20 3.333 119.8 4.717 0.115 4.079 3.668 152.5 5,386 0.736 3.775 0.267 0.981 40.8 1,440
21 3.500 117.2 4.613 0.113 3.988 3.835 155.9 5,506 0.752 3.914 0.257 0.984 40.0 1,413
22 3.667 114.5 4.509 0.110 3.898 4.002 159.0 5,616 0.770 4.051 0.245 0.982 39.0 1,377
23 3.833 111.9 4.404 0.108 3.808 4.168 161.8 5,715 0.788 4.188 0.233 0.972 37.7 1,332
24 4.000 109.2 4.300 0.105 3.718 4.335 164.3 5,803 0.807 4.323 0.220 0.955 36.2 1,278
25 4.167 106.6 4.196 0.103 3.628 4.502 166.5 5,880 0.827 4.456 0.207 0.930 34.4 1,215
26 4.333 104.6 4.118 0.101 3.561 4.668 169.4 5,984 0.843 4.596 0.196 0.913 33.1 1,170
27 4.500 102.6 4.040 0.099 3.493 4.835 172.1 6,080 0.859 4.736 0.184 0.890 31.7 1,119
28 4.667 100.6 3.962 0.097 3.426 5.002 174.6 6,168 0.876 4.874 0.172 0.860 30.0 1,060
29 4.833 98.6 3.884 0.095 3.358 5.168 176.9 6,248 0.893 5.011 0.159 0.823 28.2 995
30 5.000 96.7 3.806 0.093 3.291 5.335 178.9 6,320 0.912 5.148 0.146 0.779 26.1 923
Note: Bold font corresponds to critical duration.

150

140 5000

130

120
4000
110
3
45 m

100

Cumulative Volume (cu ft)


Cumulative Volume (m3)

90
3000
80

70

60
2000
Cumulative Inflow Volume
50
Cumulative Discharge Volume
40

30 1000
20

10

0 0
0 5 10 15 20 25 30 35
Time (min.)

Fig. 4. Example mass curves

JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / NOVEMBER 2010 / 905

J. Hydrol. Eng. 2010.15:901-908.


0.150

0.140 5.0
12.010, 0.138 12.010, 0.138
0.130 12.000, 0.130 12.020, 0.130

0.120
4.0
0.110

0.100
Inflow Rate (m 3/sec)

0.090

Inflow Rate (cfs)


3.0
0.080

0.070

0.060
2.0
Downloaded from ascelibrary.org by Cambridge University on 06/24/15. Copyright ASCE. For personal use only; all rights reserved.

0.050

0.040

0.030 1.0
0.020

0.010
11.846, 0.000 12.174, 0.000
0.000 0.0
11.0 11.2 11.4 11.6 11.8 12.0 12.2 12.4 12.6 12.8 13.0
Elapsed Time (hrs)

Fig. 5. Triangular approximation of inflow hydrograph

lower inflows, advantage can be taken of the demonstration by culated using the method given above 共43 m3 if R = 1 is used兲.
Pincince 共1970兲 that the minimum cycle time for a single pump Application of the ASCE/WEF 共1992兲 method would result in
occurs when the inflow is one-half of the discharge; that leads to upstream flooding under the design storm conditions 共in this case,
the following equation for minimum cycle time tmin: the 50-year storm of critical duration兲.
As an aid in making some of the additional desired compari-
4Vss⬘ sons, the example given above was first run using HydroCAD
tmin = 共19兲
M version 9.00’s modified rational method, dynamic storage-
indication method, and pump modeling features. The input
in which Vss⬘ = volume of the wet well between pump start and
matched that used for the above example, except that constant
stop and M = pump discharge as used above. In this case, Vss⬘
pump discharge was replaced with a pump discharge-TDH rela-
= 13.4共1.03− 0.33兲 = 9.38 m3, giving tmin = 4共9.38兲 / 关0.0849共60兲兴
tionship having miniscule increases in discharge as the TDH de-
= 7.4 min., or 60/ 7.4= 8 starts per hour. This is acceptable for the
creases 共a necessary workaround兲. The resulting HydroCAD
selected pumps which are rated for 15 starts per hour.
output values were within 1% of those in the above example.
An additional benefit of the generalized solution given above
is that no assumption is made in deriving the generalized solution The second of the ASCE/WEF 共1992兲 methods discussed
regarding the wet-well geometry; that assumption did not come in above was tested by first generating an SCS/NRCS Type III
until the chamber geometry was selected in this example. In the 24-h hyetograph based on the Cronshey equations and coeffi-
case considered in this example, it was desired to know the extent cients 共W. Merkel, personal communication, December 11, 2008兲,
of flooding upstream of the pumping station associated with a and multiplying the intensities by a factor giving the peak flow of
100-year rainstorm. An inlet grate was set at an elevation corre- 0.130 m3 / s 共4.58 cfs兲 predicted by the SCS TR-20 method using
sponding to the maximum 50-year chamber water elevation. The HydroCAD with all other variables corresponding to the example
same process as demonstrated above for the 50-year storm was given above. The resulting inflow hydrograph is depicted in Fig. 5
used to determine a 100-year storm Sⴱ = 52.4 m3 共1 , 850 ft3兲. The within the 11 to 13 h elapsed time range also depicted in ASCE/
difference of 52.4− 44.6= 7.8 m3共275 ft3兲 can be stored on the site WEF 共1992兲. The hydrograph generated by HydroCAD, which is
without causing damage. also based on Cronshey, was found to be identical. A triangular
approximation to the hydrograph is depicted on Fig. 5, and can be
seen to give I p = 0.138 m3 / s 共4.878兲 cfs and effective values of
Comparison of Example to Other Methods tc = 9.84 min and R = 1.00. Using again HydroCAD’s features as
and a Refinement described in the previous paragraph except for the changes just
Comparison can be made with the first of the ASCE/WEF 共1992兲 mentioned gave Sⴱ = 20.9 m3 共739 ft3兲. Application of this method
methods discussed above by taking the peak inflow and in- would also result in upstream flooding under the design storm
flow volume at the time of concentration given in Table 1, conditions 共in this case, the 50-year storm of critical duration兲.
i.e., I p = 0.182 m3 / s 共6.43 cfs兲 and Vi = 87.5 m3 共3,090 ft3兲, and Finally, the original example was run again with one change:
M = 0.0849 m3 / s 共3.00 cfs兲. Then, M / I p = 0.0849/ 0.182= 0.466, an actual pump curve and discharge characteristics were used
Sⴱ / Vi = 共1 − M / I p兲2 = 共1 − 0.466兲2 = 0.285, and Sⴱ = 0.285共87.5兲 rather than the 0.0849 m3 / s 共3.00 cfs兲 constant-flow assumption
= 25 m3共880 ft3兲. That is compared to Sⴱ = 45 m3 共1,580 ft3兲 cal- used above. That included a pump discharge-TDH relationship

906 / JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / NOVEMBER 2010

J. Hydrol. Eng. 2010.15:901-908.


for the selected pump and discharge information which, coupled a commercial software package to reproduce this paper’s results
with wet-well depth variations which affect the static head, enable is demonstrated. Additionally, the assumption of constant pumped
calculation of the system flow-head relationship. HydroCAD re- discharge made in this paper’s generalized solution is favorably
sults show a pump discharge varying from 0.0813 m3 / s 共2.87 cfs兲 tested using commercial software for an actual pump discharge-
when the pump starts, to a maximum of 0.0892 m3 / s 共3.15 cfs兲, TDH and system flow-head relationship.
and then decreasing to 0.0796 m3 / s 共2.81 cfs兲 just before the Extension of the methods presented in this paper to more com-
pump stops; this properly reflects the variation in static head. The plex situations is discussed. The importance of considering differ-
maximum computed storage volume is 45 m3 共1,580 ft3兲, which is ent storm durations is emphasized.
identical to the value calculated in the original example 共this is
good, but undoubtedly somewhat coincidental, and it would suf-
fice for those values to be close to each other兲. Acknowledgments

The writer gratefully acknowledges the editors and anonymous


Extension of the Methods
reviewers for comments, questions, and suggestions that signifi-
Downloaded from ascelibrary.org by Cambridge University on 06/24/15. Copyright ASCE. For personal use only; all rights reserved.

The method given above can be readily extended to more com- cantly improved the quality of the paper.
plex situations if certain assumptions are made. For example, as-
sume there are two pumps with constant capacity related by the
factor k1 = 共M 1 + M 2兲 / M 1 in which M 1 and M 2 = capacities of Notation
the first and second pumps, that the first pump starts at ts
= 共M 1 / I p兲tc as assumed above, and that the second pump starts at The following symbols are used in this paper:
time tc. Then the following equation can be derived in a manner B⬘ ⫽ I ptc / Vi, dimensionless ratio;
similar to the derivation leading to Eq. 共11兲 given above for the H ⫽ elevation measured upward from elevation at which
single pump: effective storage and outflow begin;

冉 冊
I ⫽ volumetric inflow rate;
Sⴱ tr M1 1 I p ⫽ peak volumetric inflow rate;
= 1 − k1 + 共R − 1兲
I pt c t c Ip 2 i ⫽ average rainfall intensity;


M1
Ip

k 1R + 1 − k 1 −
M1
Ip
冉 1
1 + k21R
2
冊册 共20兲
k1 ⫽ 共M 1 + M 2兲 / M 1;
ᐉ ⫽ exponent in storage-elevation relationship 关Eq. 共3兲兴;
M ⫽ coefficient in outflow-elevation relationship
which reduces to Eq. 共11兲 for k1 = 1 共i.e., one pump, M 2 = 0兲. 关Eq. 共4兲兴 and constant-pumped outflow;
More generally, for large, multiple-pump installations, inflow M 1 ⫽ constant-pumped outflow for first pump in
hydrographs with varying storm durations should be used to de- multi-pump installation;
fine the critical duration for each condition. FHWA 共2001兲 dis- M 2 ⫽ constant-pumped outflow for second pump in
cusses the importance of considering different storm durations, two-pump installation;
but gives no guidance on how to do so. Nor is such guidance m ⫽ exponent in outflow-elevation relationship 关Eq. 共4兲兴;
given in Highway Hydrology 共FHWA 2002兲. Inflow hydrographs N ⫽ coefficient in storage-elevation relationship
developed by the modified rational method such as shown on Fig. 关Eq. 共3兲兴;
1 are useful for that purpose. The analysis will then require an Q ⫽ volumetric outflow rate;
extension of the trial-and-error or iterative procedures discussed Qⴱ ⫽ peak outflow;
in, e.g., Baumgardner 共1983兲, ASCE/WEF 共1992兲, and FHWA R ⫽ ratio of duration of receding limb of inflow
共1982, 2001兲. Such analyses can be aided greatly by automatic hydrograph to duration of rising limb;
computation. S ⫽ effective storage volume;
Sⴱ ⫽ maximum storage volume;
t ⫽ time from start of rainfall;
Conclusions tc ⫽ time of concentration;
te ⫽ pump stop time;
A generalized solution for the hydrologic and hydraulic design of tmin ⫽ minimum pump cycling time;
small to medium-sized storm-water pumping stations provides a ts ⫽ time of pump starting;
useful aid for the design of such stations. The present paper ex- tr ⫽ duration of rainfall;
presses such a solution and discusses the practical aspects of its tⴱ ⫽ time of peak outflow;
use. The solution enables determination of pump capacity and the V ⫽ inflow volume;
corresponding storage capacity. The benefits of the solution in- Vi ⫽ total inflow volume;
clude: 共1兲 the ability to consider the effects of storm duration V p,start ⫽ pump start volume; and
rather than the fixed hydrology assumed in other methods; 共2兲 a Vss⬘ ⫽ volume of wet well between pump start and stop.
simpler and more intuitive design procedure than other available
methods; and 共3兲 enabling the calculation of storage requirements
independent of the geometry of the wet well and any inundated References
upstream area.
The relation of the method presented herein to several other American National Standards Institute, Inc./Hydraulic Institute 共ANSI/
available methods is presented and the other methods critiqued HI兲. 共1998兲. American national standard for pump intake design, Par-
conceptually and by an example. The other methods are shown to sippany, NJ.
be erroneous, less general, or to underpredict the required wet- ASCE and Water Environment Federation 共WEF兲. 共1992兲. Design and
well storage. The example presented demonstrates the application construction of urban storm-water management systems: ASCE
of the writer’s method to an actual design situation. The ability of manual of practice no. 77, New York, 390–395.

JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / NOVEMBER 2010 / 907

J. Hydrol. Eng. 2010.15:901-908.


Baumgardner, R. H. 共1983兲. “Hydraulic design of stormwater pumping cients and methods.” Catchment runoff and rational formula, B. C.
stations: The effect of storage.” Transportation Research Record. 948, Yen, ed., Water Resource Publications, Littleton, Colo., 60–73.
Transportation Research Board, Washington, D.C., 63–70. Graber, S. D. 共2009a兲. “Generalized numerical solution for detention
Burton, K. R. 共1980兲. “Stormwater detention basin sizing.” J. Hydr. Div., basin design.” J. Irrig. Drain. Eng., 135共4兲, 487–492.
106共3兲, 437–439. Graber, S. D. 共2009b兲. “Rain loads and flow attenuation on roofs.”
Federal Highway Administration 共FHWA兲. 共1982兲. Manual for highway J. Archit. Eng., 15共3兲, 91–101.
stormwater pumping stations, Publication No. FHWA-IP-82-17, Vols. Hildebrand, F. B. 共1952兲. Methods of applied mathematics, Section 2.1,
I and II, Washington, D.C.
Federal Highway Administration 共FHWA兲. 共2001兲. “Highway stormwater Prentice-Hall, Englewood Cliffs, NJ.
pump station design.” Hydraulic Engineering Circular No. 24: Pincince, A. B. 共1970兲. “Wet-well volume for fixed-speed pumps.”
FHWA-NHI-01-007, Washington, D.C. J. Water Pollut. Control Fed., 42共1兲, 126–130.
Federal Highway Administration 共FHWA兲. 共2002兲. Highway hydrology, Soil Conservation Service 共SCS兲. 共1986兲. “Urban hydrology for small
2nd Ed., Publication No. FHWA-NHI-02-001, Washington, D.C. watersheds.” Technical Release No. 55, USDA, Washington, D.C.
Froehlich, D. C. 共1994兲. “Sizing small stormwater pumping stations.” Water Environment Federation 共WEF兲. 共1993兲. “Design of wastewater
and stormwater pumping stations.” Manual of practice no. FD-4, Al-
Downloaded from ascelibrary.org by Cambridge University on 06/24/15. Copyright ASCE. For personal use only; all rights reserved.

Water Resour. Bull., 30共6兲, 1055–1062.


Graber, S. D. 共1992兲. “Relations between rational and scs runoff coeffi- exandria, Va., 36–39.

908 / JOURNAL OF HYDROLOGIC ENGINEERING © ASCE / NOVEMBER 2010

J. Hydrol. Eng. 2010.15:901-908.

You might also like