Page 212 - Water Engineering Hydraulics, Distribution and Treatment
P. 212

190
                                                   Water Distribution Systems: Components, Design, and Operation
                                          Chapter 6
                                            (d) Pipes added: two 10 in. (250 mm) at 0.6MGD = 1.2 MGD (4.5 MLD).
                                               Pipes removed: one 6 in. at 0.2 MGD (one 150 mm at 0.76 MLD).
                                               Net added capacity: 1.2 − 0.2 = 1.0 MGD (3.8 MLD).
                                               Reinforced capacity = 6.5 + 1.0 = 7.5MGD or (24.6 + 3.8 = 28.4 MLD).
                                            (e) The reinforced system equivalent pipe at 7.5 MGD (28.4 MLD) and a hydraulic gradient of 2‰ is 26.0 in. (650 mm)
                                               This will carry 7.6 MGD (28.8 MLD) with a loss of head of 2.1‰.
                                    6.6.2 Relaxation (Hardy Cross)
                                                                                          Two procedures may be involved, depending on whether
                                    A method of relaxation,or controlled trial and error,was
                                    introduced by Hardy Cross, whose procedures are followed  (a) the quantities of water entering and leaving the network or
                                                                                       (b) the piezometric levels, pressures,or water table elevations
                                    here with only a few modifications. In applying a method
                                                                                       at inlets and outlets are known.
                                    of this kind, calculations become speedier if pipe–flow
                                                                                          In balancing heads by correcting assumed flows, neces-
                                    relationships are expressed by an exponential formula with
                                                                                       sary formulations are made algebraically consistent by arbi-
                                    unvarying capacity coefficient, and notation becomes simpler
                                                                                       trarily assigning positive signs to clockwise flows and asso-
                                    if the exponential formula is written:
                                                                                       ciated head losses, and negative signs to counterclockwise
                                                         H = kQ n               (6.1)  flows and associated head losses. For the simple network
                                                                                       shown in Fig. 6.11a, inflow Q and outflow Q are equal and
                                                                                                              i           o
                                    where, for a given pipe, k is a numerical constant depending  known, inflow being split between two branches in such a
                                    on C, d, and L, and Q is the flow, n being a constant exponent  manner that the sum of the balanced head losses H (clock-
                                                                                                                               1
                                    for all pipes. In the Hazen–Williams equation, for example,  wise) and −H (counterclockwise) or ΣH = H − H = 0. If
                                                                                                 2
                                                                                                                           1
                                                                                                                                2
                                                                                       the assumed split flows Q and −Q are each in error by the
                                                                                                                   2
                                                                                                           1
                                           Q = 405 Cd 2.63 0.54  (US customary units)  same small amount q, then
                                                       s
                                                                                                                   n
                                    where Q = rate of discharge, gpd; d = diameter of circular       ΣH =Σk (Q + q) = 0
                                    conduits, in.; C = Hazen–Williams coefficient, dimension-  Expanding this binomial and neglecting all but its first
                                    less; S = H/L = hydraulic gradient, dimensionless; H = loss  two terms, because higher powers of q are presumably very
                                    of head, ft; L = conduit length, ft.               small, we get
                                                                                                              n
                                                                                                       n
                                                             S
                                               Q = 0.278 Cd 2.63 0.54  (SI units)        ΣH =Σk(Q + q) =ΣkQ +ΣnkqQ    n−1  = 0, whence
                                                                                                  ΣkQ n       ΣH
                                                              3
                                    where Q = rate of discharge, m /s; d = diameter of circular  q =−  n−1  =−                     (6.3)
                                    conduits, m; C = Hazen–Williams coefficient, dimension-     nΣkQ        nΣH∕Q
                                    less; S = H/L = hydraulic gradient, dimensionless; H = loss  If a takeoff is added to the system as in Fig. 6.11b, both
                                    of head, m; L = conduit length, m. For the Hazen–Williams  head losses and flows are affected.
                                    using either the US customary units or the SI units, the fol-  In balancing flows by correcting assumed heads, neces-
                                    lowing relationship hold true:                     sary formulations become algebraically consistent when pos-
                                                                                       itive signs are arbitrarily assigned to flows toward junctions
                                                                 ′
                                                       ′
                                                   s = k Q 1∕0.54  = k Q 1.85          other than inlet and outlet junctions (for which water table
                                                   H = sL                       (6.2)  elevations are known) and negative signs to flows away from
                                                   H = kQ 1.85                         these intermediate junctions, the sum of the balanced flows at
                                                                                                   Take-off
                                    Q i            Q 1                  Q i            Q 1            T
                                                                         Inflow
                                      Inflow
                                               Assumed flow Q i
                                               incorrect by + q
                                                                                                    Q 3
                                              Assumed flow Q 2
                                               incorrect by – q
                                                                                                           Figure 6.11 Simple network
                                                   Q 2          Q o                    Q 2         Q  Outflow  illustrating (a) the derivation of the
                                                                                                    o
                                                                 Outflow
                                                                                                           Hardy Cross method and (b) the
                                                     (a)                                (b)                effect of changing flows.
   207   208   209   210   211   212   213   214   215   216   217