Professional Documents
Culture Documents
Modeling of Water Pipeline Filling Events Accounting For Air Phase Interactions
Modeling of Water Pipeline Filling Events Accounting For Air Phase Interactions
net/publication/263965679
Modeling of Water Pipeline Filling Events Accounting for Air Phase Interactions
CITATIONS READS
26 1,699
2 authors:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Jose G. Vasconcelos on 16 July 2014.
2 Bernardo C. Trindade1
Jose G. Vasconcelos2
3 ABSTRACT
4 In order to avoid operational issues related to entrapped air in water transmission mains, water refilling
5 procedures are often performed carefully to ensure no pockets remain in the conduits. Numerical models
6 may be an useful tool to simulate filling events and assess whether air pockets are adequately ventilated.
7 However, this flow simulation isn’t straightforward mainly because of the transition between free surface and
8 pressurized flow regimes and the air pressurization that develops during the filling event. This work presents
9 an numerical and experimental investigation on the filling of water mains considering air pressurization
10 aiming towards the development of a modeling framework. Two modeling alternatives to simulate the
11 air phase were implemented, either assuming uniform air pressure in the air pocket or applying the Euler
12 equations for discretized air phase calculations. Results compare fairly well to experimental data collected
14 Keywords: Water pipelines; Pipeline filling; Flow regime transition; Air Pressurization; Numerical modeling
15 INTRODUCTION
16 Transmission mains are important components of water distribution systems and a relevant concern is
17 the safety of operational procedures performed on those. Among the operational procedures one includes
18 what is referred to as pipeline priming, the refilling operations that often follows maintenance tasks that
19 require total or partial emptying of the conduits. During refilling procedures, the air phase initially present
20 in the pipeline may become entrapped between masses of water in the form of air pockets. Entrapped
21 pockets may lead to pressures surges in the system and loss of conveyance when not properly expelled via air
22 valves. However, as it will be shown, there is limited investigation on the development of numerical models
24 Similarly to other applications that involve the transition between pressurized and free surface flows
25 (also referred to as mixed flows), there are certain characteristics on water pipeline filling events that pose
University, 238 Harbert Engineering Center, Auburn, AL, 36849. E-mail: bct0004@auburn.edu
2 A.M. ASCE, Assistant Professor, Dept. of Civil Engrg., Auburn University, 238 Harbert Engineering Center, Auburn, AL,
1
27 1. Pipeline filling events are characterized by the transition between free surface and pressurized flow
28 regimes, and while there are different approaches to simulate such transitions, current models are
29 limited by difficulties in properly incorporating the interaction between flow features (e.g. bores and
31 2. Pipeline filling is a two-phase, air-water flow problem, and models handling the two separate phases
32 need to be appropriately linked. The handling of the interface between air and water is particularly
33 challenging;
34 3. Due to the formation of bores and the large discrepancy in the celerity magnitudes between different
35 portions/phases of the flow, non-linear numerical schemes should be used if bores are anticipated so
37 4. At certain regions of the flow, particularly at the vicinity of curved air-water interfaces, shallow water
38 assumptions are not applicable due to strong vertical acceleration as the problem is intrinsically
40 5. Several different mechanisms may result in the entrapment of air pockets during filling events, yet
41 because conditions leading to such entrapments are still not fully understood, these cannot be properly
43 Related studies on the interference between air and water in closed conduits started as early as Kalinske
44 and Bliss (1943), focusing on steady flows in which a hydraulic jump filled the conduit causing air entrainment
45 through the jump. The work presented an expression relating the amount of air entrained by the jump as in
46 terms of the Froude number of the free-surface portion of the flow upstream the jump. Recent experimental
47 investigation on air-water interactions in water pipelines has led to advances on the understanding of the
48 removal of air pockets by dragging, leading to expressions for the required water flow/velocity that will result
49 in the removal of air bubble and pockets from pipes. Among such works one includes Little (2008) Pothof
50 and Clemens (2010) and Pozos et al. (2010). The removal of these pockets, however, would occur during
51 the operation of these pipelines, and thus is different from the air ventilation process that takes place during
52 pipeline priming.
53 Numerical simulation of the filling of water pipelines required the use of flow regime transition models
54 since conduits start empty and will be fully pressurized by the end of the event. Such models simulate
55 both flow regimes, and one may classify these models in two main types: shock-capturing and interface-
56 tracking models. Shock-capturing models apply a single set of equations to calculate both pressurized and
57 free-surface flow regimes and require the use of a conceptual model to handle pressurized flows using free-
58 surface flow equations. These models are generally simpler to implement, provide seamless representation of
2
59 the interaction between various flow features, but have the drawback of generating numerical oscillations at
60 pipe-filling bore fronts that can be addressed by appropriate flux selection or numerical filtering technique
61 (Vasconcelos et al., 2009). Interface-tracking models are more complex to implement as they require a set of
62 equations for each flow regime and proper interaction between flow features is more complex to represent;
63 however the resolution of pressurization bores is exact (no diffusion) and there are no post-shocks at pipe-
64 filling bores.
65 With regards to air simulation, most models used a framework that is based on the ideal gas law, with
66 fewer alternatives using a discretized framework. To date, most one-dimensional, unsteady flow models that
67 accounted for air pockets in pipelines is of the interface-tracking type and use lumped inertia approach
68 to simulate the water phase, while air is simulated using implementations of the ideal gas law. Possibly,
69 the first such work was proposed by Martin (1976), which presented a model to compute surges caused by
70 the compression of air pockets at the end of an upward sloped pipeline. Such models assume well defined
71 interfaces between air and a water phases, essentially a portion of the system would be in pressurized water
72 regime while the other would have the entire cross section occupied by air. In some instances of these
73 model application an orifice equation is added to the air formulation to account for air ventilation during
74 the compression process. This modeling approach further assumes uniform pressure head for the entire air
75 mass, and is herein referred to as Uniform Air Pressure Head or for brevity UAPH model. An example of
76 other models that uses the same principles of the UAPH model is includes the work presented by Zhou et al.
77 (2002) which studied the compression of air pockets in the context of stormwater systems considering that
79 Two other numerical modeling studies focused in the filling water mains and have also used the interface-
80 tracking and lumped inertia approaches. Liou and Hunt (1996) proposed an simple alternative to simulate
81 the pipeline filling characterized by flow regime transition, but have not including effects of air pressurization.
82 The model avoided the difficulties associated with shock-fitting technique by assuming a vertical interface
83 between air and water, implying in an instantaneous transition between dry pipe and pressurized flow upon
84 inflow front arrival. The second water main filling model using interface tracking was proposed by Izquierdo
85 et al. (1999). The model simulates the rapid startup of pipelines that are partially filled, so that the flow
86 admission generates the entrapment of air pockets. Air phase modeling uses the UAPH approach without
87 ventilation for air pockets. Fuertes et al. (2000) tested that model against experimental data in order to
88 validate the model with good agreement, but the tested inflow rate is possibly too large to be representative
90 As presented, these modeling studies that combine lumped inertia and UAPH approaches assume: 1)
91 well defined interfaces separating air and water phases; 2) high inflow rates as consequence of the elevated
3
92 driving pressure head; and 3) uniform pressure in the air pockets. The first assumption may be invalid in
93 cases when air is not actually intruding into the pressurized flow, as discussed in (Vasconcelos and Wright,
94 2008). The second assumption usually does not hold if the filling is performed gradually. However, the latter
96 Chaiko and Brinckman (2002) presented a comparative study of three modeling alternatives to simulate
97 air/water interactions in a idealized pipe filling problem. The first model solves both phases by applying
98 the method of characteristics (M.O.C.), so that water characteristic lines need to be interpolated to match
99 the grid during simulation; the second model simplifies air phase modeling by applying a type of UAPH
100 model; and the third applied UAPH for air phase and M.O.C. for water phase, however calculating only the
101 unperturbed portion of the water flow so that the characteristic lines have constant slope and match the
102 grid without the need of interpolation. The authors run tests for a vertical set up which consisted of a a
103 cylinder with an air pocket on the upper part which is compressed by the water phase due to an increase in
104 the water pressure at the bottom of the cylinder. It was found that the second model alternative captured
105 all the relevant events as well as the first model, even though the second didn’t capture small oscillations
106 due to the reflection of the pressure wave in the air, which has no major importance in most practical
107 applications. However, the geometry used in that study is idealized, and a question is how the obtained
108 results are translated to a more complex setup where initially stratified flow exists along with moving water
110 Two studies are presented here as instances of shock-capturing models to simulate rapid filling of closed
111 conduits. In the context of stormwater simulation Arai and Yamamoto (2003) presented a model that
112 performs flow regime transition calculation including a discretized air phase calculation approach. The model
113 apply the Saint-Venant equations for water phase and the Preissmann slot to account for pressurization. The
114 model was implemented in a simple, quasi-horizontal geometry, and air was modeled with a set of discretized
115 mass and momentum equations. Model results compared well with experimental results, and indicate the
116 importance of accounting for air phase effects during simulation. The conditions for air pocket entrapment
117 were not focused in this study, and the linear numerical scheme applies in the study (four-point Preissmann
118 scheme) is not appropriate to adequately capture bores when Courant numbers are very low; this is an issue
120 The second study involved a finite volume models using numerical schemes based on approximate Riemann
121 solvers that overcome problems with low Courant numbers were applied in the context of pipeline filling. A
122 model based on the Saint-Venant equations was presented by Vasconcelos (2007) using the TPA approach
123 (Vasconcelos et al., 2006). This model was subsequently used in a study that involved the filling of an actual
124 4.4 km-long, 350-mm diameter water transmission main operated by CAESB, Environmental Sanitation
4
125 Company from Brasilia, Brazil (Vasconcelos et al., 2009). Field measurements of inflow rates, pressure
126 heads indicated that the model was able to capture the general trend of the filling. Yet, some discrepancies
127 between the pressure measurement and predictions were attributed to the inability of the numerical model
128 to incorporate limited ventilation conditions and consequently effects of air pressurization to the flow.
129 OBJECTIVES
130 The present work aims to obtain further insight on air-water interactions during water pipeline filling
131 operations, with the overarching objective of developing a numerical framework that may be used to simulate
132 a priori filling operations in pipelines and detect operational issues related to the entrapment of air pockets.
133 To achieve this objective, two numerical models were proposed that differ in the strategy in which air is
134 modeled. Both alternatives use the variation of the TPA model presented by Vasconcelos and Wright (2009)
135 to describe the water phase. Air phase modeling is performed either by using a discretized framework that
136 applies the Euler equation or by using a type of UAPH model. Another objective was to assess the benefits
138 Associated with the numerical development, an experimental program was conducted using a scale model
139 apparatus that incorporates essential features of a water pipeline filling problem. Key parameters in the
140 problem were systematically varied, including inflow rate, pipeline slope and ventilation degree. Experimen-
141 tal measurements included pressure, pressurization interface trajectories and inflow rates. Both modeling
142 alternatives for air phase were compared to experimental data and to the field data of an actual water main
144 METHODOLOGY
146 Certain flow features of the water main pipeline filling problem were determinant in the model’s formula-
147 tion so that it could describe the filling process adequately. With regards to the water phase, these features
148 include:
149 • Mixed flows: handled by applying the TPA model variation from (Vasconcelos and Wright, 2009);
150 • Post-shock oscillations at pipe-filling bore fronts: use of a numerical filtering scheme (Vasconcelos
152 • Air pocket entrapment and pressurization: used either the Euler equations or UAPH model;
153 • Free-surface and pipe-filling bores: Use of the approximate Riemann solver presented by Roe (Mac-
155 • Solution stationarity: use approach presented by Sanders and Bradford (2011).
5
156 Air phase in the model is represented by a well-defined air pocket that is not significantly fractured.
157 This pocket shrinks due to compression by the water phase that gradually occupies the lowest points in
158 the pipeline profile. Air is displaced and escapes through ventilation orifices located at selected locations.
159 According to Tran (2011) for such flow conditions air compression process may be considered isothermal and
160 this assumption is used in both models approaches used to simulate air phase. Air phase is calculated as
161 if the only atmospheric connections occur at ventilation points, which are treated as orifices for simplicity.
162 Ideal ventilation with negligible air phase pressure head is assumed to exist prior to the formation of an
163 entrapped air pocket, as it will be discussed later. When a pocket forms, it is delimited by a ventilation
164 orifice and a flow regime transition interface, either abrupt (pipe-filling bore) or gradual. In the proposed
165 model, an air pocket is caused by the closure of a downstream valve or by the development of a pressurization
166 interface as water fills the lowest points of the pipeline. Figure 1 presents a sketch of a typical application,
167 whereas Figure 3 presents the overall structure of the proposed model.
169 The TPA model, used in the water phase simulation, modifies the Saint-Venant equations, enabling them
170 to simulate both pressurized flows and free-surface flow regimes. This model has been improved in the past
171 years and the alternative used here was presented in Vasconcelos and Wright (2009). This alternative has a
172 term that accounts for air phase pressure head, so that the modified St. Venant equations are, in divergence
173 format:
∂U ∂F(U)
+ =S (1)
∂t ∂x
174 where
A Q 0
U = , F(U) = 2 , S(U) = (2)
Q
Q A + gA(hc + hs ) + gApipe hair gA(S0 − Sf )
= 0 → Free-surface flow without entrapped air pocket or Pressurized flow
hair (3)
6= 0 → Free-surface flow with entrapped air pocket
0
→ Free-surface flow
hs = (4)
a2 (A − Apipe )
→ Pressurized flow
g Apipe
6
D 3sin(θ) − sin3 (θ) − 3θcos(θ)
→ Free-surface flow
2θ − sin(2θ)
3
hc = where θ = π − arccos[(y − D/2)(D/2)] (5)
D
→ Pressurized flow
2
175 where U = [A, Q]T is the vector of the conserved variables, A is the flow cross sectional area, Q is the
176 flow rate, F(U) is the vector with the flux of conserved variables, g is the acceleration of gravity, hc is
177 the distance between the free surface and the centroid of the flow cross section (limited to D/2), hs is the
178 surcharge head, hair is the extra head due to entrapped air pocket pressurization, θ is the angle formed by
179 free surface flow width and the pipe centerline, D is the pipeline diameter, Apipe is the cross sectional area
180 (0.25πD2 ) and a is the celerity the acoustic waves in the pressurized flow.
181 The numerical solution used in the implementation of the water phase model used the Finite Volume
182 Method and the approximate Riemann solver of Roe, as presented in Macchione and Morelli (2003). This
183 choice was motivated by the significant discrepancy in the celerity values between the free-surface and
184 pressurized flows, and between air and water flows. This discrepancy may be in the order of 2 or 3 orders of
185 magnitude and yields an extremely low Courant number for the free-surface water flow. Non-linear schemes
186 such as the Roe scheme are known to provide accurate bore predictions even in very low Courant number
187 conditions.
188 For dry bed regions of the flow it was assumed that the flow depth would start as a minimum water
189 depth of 1 mm. In such cases, the model would then predict the existence of a non-physical flow of this thin
190 layer down the pipeline slope. To deal with this problem it was used the approach presented in Sanders and
191 Bradford (2011). In this formulation, in order keep a minimum water layer with no motion and at the same
192 time keep the stationarity of the solution, two criteria were followed in order to perform flow calculations
193 at interior finite volume cells: one is based on the ratio between friction forces and the other based on the
194 minimum submerged area of the cell. After computing these criteria to all cells in the domain, only the
195 cells in which at least one of the two criterion is met have the flow variables calculated. Details of this
196 formulations are omitted for brevity, but may be found in the aforementioned paper.
197 Two source terms were considered for the water phase modeling, one accounting for pipe walls friction
198 and the other accounting for pipe slope. Both calculations followed the approach presented in Sanders and
199 Bradford (2011). For the pipe walls friction source term, a semi-implicit formulation based on the Manning’s
200 equation was used, while for the pipe slope a formulation which preserves stationarity of the solution was
201 used.
202 The upstream boundary condition for water phase refers to all which is inside the red dashed box in
7
203 figure 1. It is based on an iterative solution that ensures that local continuity and linear momentum at the
204 pipeline inlet are satisfied, regardless of the flow regime at that location. The local continuity equation for
dHres
= Qrec − Qin (6)
dt
206 where Hres is the reservoir water level, Qrec is the flow rate which is admitted into the reservoir from
207 the recirculation system and Qin is the flow rate which enters the upstream end of the pipe. The calculation
208 of the updated flow velocity at the upstream boundary cell (un+1
1 ) uses an ordinary differential equation
209 representing the linear momentum conservation, which in turn is derived from a lumped inertia approach:
un2 |un2 |
g
un+1
1 = un1 + ∆t n
Hres − Keq − (wdepth 2 + max(0, hs n2 + hair n2 ))
∆x 2g
(7)
un2 |un2 | un 2
−f − 2
2∆x 2∆x
210 where wdepth is the local water depth, n is the time step index, Keq is the overall local loss coefficient in
211 the inlet and f is the friction head loss in the short pipe portion inside the boundary condition right after
213 After the velocity in the cell is obtained, Froude number is calculated with the current wdepth 1 in order
214 to assess if the flow is subcritical. If this is the case, wdepth n+1
1 is updated according to the M.O.C. Hartree
215 for free-surface flows as shown in Sturm (2001). If flow depth at inlet is less than the pipe diameter D then
216 the surcharge head hs is set to zero while hair may be non-zero if there is any entrapped pocket. On the
218 the piezometric head at the upstream end, calculated with the energy equation. With the depth and the
220 The downstream boundary condition used to compare the model with the experiments can be either
221 a fully opened or closed gate valve. For the case in which the valve is fully opened, the approach called
222 transmissive condition presented in Toro (2001) is used. For the case in which the downstream boundary
223 condition is a closed valve the boundary condition is calculated enforcing the relevant characteristic equation
cr
wlevel n+1
No = (Kr ) un+1
No = 0 (8)
g
225 where Kr is a constant factor that depends on the flow conditions in the previous time step (Sturm,
8
227 is set to the value of the pipe diameter and hs is set as the extra head of the cross section minus D, and
230 In the proposed model, air is initially considered as a continuous layer over the water layer (stratified
231 water in free-surface flow mode). During the simulation, air is handled in one of two following manners:
232 1. At the first stage of the filling, when the air within a given pipe reach consists in an entire layer that
233 is connected to atmosphere at the ventilation orifice and at its lowest point within the pipe (ideal
234 ventilation), it is assumed that the air pressure in the entire layer is approximately atmospheric, and
235 air velocity is assumed negligible. This condition persists until a pocket is formed at the lowest point
237 2. At the second stage, once an air pocket is formed, the connection to the atmosphere at its lowest point
238 is lost. Air phase pressure is expected to develop during the filling process, requiring calculations with
239 either one of the two proposed air phase models to determine its pressure and influence in the water
240 flow.
241 An algorithm was developed to track air pocket volume, start and end nodes as it shrinks in order
242 to simulate its behavior with any of the two models. For this, the mechanism considered for air pocket
243 formation is the isolation of an air mass due to the development of a flow regime transition interface or a
244 closed downstream valve. As mentioned, it is assumed that during a pipeline filling event this air pocket will
245 be delimited by a ventilation orifice and a flow regime transition interface. This interface will move mainly
246 towards the air pocket ventilation point, compressing the air pocket and forcing its elimination through the
248 The model alternative that uses UAPH model assumes: a) uniform pressure in the whole air phase; 2)
249 the validity of the ideal gas law; and 3) isothermal air flow. This model may be expressed either as:
n+1
ρn Vpn = ρn+1 Vpn+1 → n
Mair = Mair (9)
250 where Mair is the mass of air within the pocket with volume Vp , and ρ is the specific mass of air. In
251 order to consider the air escape or admission an extra term was added to equation 9, yielding:
9
252 where Mair out is the air mass that escapes through the ventilation orifice in that instant, calculated as
254 The second alternative to model the air phase uses a discretized framework, applying an one-dimensional,
∂U ∂F(U)
+ =S (11)
∂t ∂x
256
ρ ρu Sd1
U = , F(U) = , S(U) = (12)
ρu ρu2 + p Sd2 + Sf a
p = ρα2 (13)
258 where U is the vector of conserved variables, F is the vector of the conserved variables fluxes, S is the
259 vector of source terms, α is the celerity of the acoustic waves in the air, and Sd1,i , Sd2,i and Sf a are source
260 terms.
261 Applying the Lax-Friedrichs scheme - LxF - as presented in Toro (2001) to equation 12, one has the
ρni+1 + ρni−1 ∆t
ρn+1 (ρu)ni+1 − (ρu)ni−1 + ∆tSd1,i
i = −
2 2∆x
(ρu)ni+1 + (ρu)ni−1 un + uni−1
∆t
(ρu)n+1
i = − (ρu)ni+1 − (ρu)ni−1 i+1 (14)
2 2∆x 2
n n
ρ − ρi−1
+α2 i+1 + ∆t(Sd2,i + Sf a )
2∆x
263 The choice for the LxF scheme was based on its simplicity and the lack of shocks in the air phase flow.
264 In pipeline filling problems, the mechanism causing the motion of the air phase is the displacement of
265 air in the cross section caused by changes in the water flow depth underneath the air pocket. This effect is
Sd1 ρ
1 ∂Aair ∂Aair
= + uair (15)
Sd2 A ∂t ∂x ρu
267 where Aair = (π/4)D2 − A, and is calculated only in free surface flow cells. An explicit implementation of
268 source terms Sd led to instability of the numerical solution, so a semi-implicit approach was applied here. The
269 air phase is first calculated without considering changes in Aair returning a preliminary solution Ǔ = [ρ̌, ρu],
ˇ
270 which then needs to be adjusted with a correction factor φ so that a definitive solution is achieved. The
10
271 definitive solution and correction factor φ are represented by:
ρ ρ̌
U = = φ (16)
ρu ρu
ˇ
273 The solution of Sd source terms presented some oscillations at the region of the strongest free surface flow
274 gradients, at the vicinity of the pressurization front. Two approaches were used together to minimize these
275 oscillations. The first one was to limit φ to the range φ = [1.005 : 0.995]. The second was the application of
276 an oscillation filter in the air phase internal nodes, following Vasconcelos et al. (2009) with = 0.05. This
277 approach resulted in a good balance between pressure accuracy, presenting an average continuity error of
278 less than 4% for the air phase in the comparison with the experimental cases and 1% in the comparison
279 with the field measurements. The likely source of the continuity error for the Euler equation model are the
280 orifice boundary condition and the limitation of the correction factor φ (equation 16) to a certain value,
281 which distorts the actual required air compression due to pocket vertical shrinking. However undesirable,
282 air continuity errors did not seem to have an major impact in the overall results, as the comparison between
284 The other source term added to the simulation of the air phase flows was the friction between the air
285 phase and the pipe walls, as described in Arai and Yamamoto (2003):
fa Pa ua |ua |
Sf a = (18)
8gAair
287 For the UAPH model, the boundary condition used at the uppermost point in the pipeline reach (where
288 the ventilation valve was located) was a discharging orifice. The orifice is represented by an equation similar
291 where Cd is the discharge coefficient that is assumed as Cd = 0.65, and Aorif is the orifice area. Equation
s
ρn+1 − ρatm 2
ρn Vpn = ρn+1 Vpn+1 + ∆tCd Aorif ρn+1 2 α (20)
ρatm
11
293 For the Euler equation model, two boundary conditions are required. At the lower point of the pipe,
294 where water pressurization front is displacing the air, a reflexive boundary condition (Toro, 2001) was used.
295 At the uppermost point, the ventilation orifice boundary condition for the Euler equation approach is similar
296 to the UAPH model in the sense that both apply a continuity equation along with the orifice equation. The
298 where Mair out is the air mass discharged through the ventilation orifice, calculated using equation 19.
301 On both models, calculations are stopped when pocket has evacuated 95% of its original volume. When
302 this condition is reached, the average head of the air pocket is assigned to its cells, remaining constant
303 until the end of the simulation. This was motivated to avert calculations instabilities caused by increasingly
304 smaller air pocket volumes, and follow the strategy used in Zhou et al. (2002).
306 An experimental investigation was conducted to gather insights on the characteristics of the pipeline
307 filling problem with limited ventilation, and to validate the proposed numerical model. The apparatus
308 was inspired in the one presented in Trajkovic et al. (1999), yet with modifications that limited ventilation
309 conditions.
311 A sketch of the experimental apparatus is presented in Figure 2. The experimental apparatus included a
312 clear PVC pipeline with length L = 10.96 − m and diameter D = 101.6 − mm, with adjustable slope. At the
313 upstream end, a 0.66 m3 capacity water reservoir supplied flow to the pipeline through a 50 mm ball valve;
314 at the downstream end flow discharged freely through a 101.6 mm knife gate valve into a 0.62 m3 reservoir,
315 and flow was subsequently recirculated with pumps. Right after the inlet control valve, a T junction was
316 installed in the pipe so that different caps with ventilation orifices could be installed. Initial, steady flow
317 conditions were such that free surface flows exist at the whole pipeline, as the downstream gate was fully
318 opened. A sudden closure maneuver (within 0.3 seconds) of the knife gate valve at the downstream end of
319 the pipeline blocked the downstream ventilation, triggered a backward-moving pressurizing interface, and
12
320 resulted in the entrapment of an air pocket. These air pockets became pressurized as water accumulated
321 at downstream end of the pipe pushed the air mass through the ventilation orifice in the beginning of the
322 pipe. Two pressure transducers Meggit-Endevco 8510B-5 were installed at the upstream end of the pipe
323 (x∗ = x/L = 1, measured from the downstream end) and at an intermediate point x∗ = 0.39. Transducer
324 results were calibrated each experimental run with the aid of four digital manometers, with of 3.5-m H2 0
325 maximum pressure head and 0.3% accuracy. Flow rates were measured with a Nortek MicroADV positioned
328 1. With the desired slope set in the pipeline, the pumps were started; valves near the pump were throttled
329 enough to provide the selected steady flow rate to the system;
331 3. When water level at upstream reservoir and pipeline attained steady levels, it was performed readings
333 4. The data logging was started for the pressure transducers, the MicroADV, and manometer at the
335 5. The downstream knife gate valve was rapidly maneuvered and closed entrapping an air pocket and
337 6. Digital cameras (30 FPS) recorded the whole pipe filling process, one of them tracking the bore and
339 7. When the pressurization interface approached the ventilation orifice, it was rapidly closed to avoid
341 8. Pump was shut down and control valves closed so that pressure could attain a hydrostatic conditions;
342 9. Manometers readings were performed and data logging was stopped;
343 The use of two cameras to track the inflow/pressurization front was particularly necessary in the cases
344 where interface breakdown (Vasconcelos and Wright, 2005) occurred; otherwise just one camera tracking
345 the pipe-filling bore front was used. The described experimental program varied systematically flow rates,
346 ventilation orifice diameters and pipeline slopes. Table 1 presents the ranges of the tested experiment
347 variables, with a total of 36 conditions tested. A minimum of two repetitions were performed for each
350 The experimental results are compared with the proposed numerical model using both approaches to
351 simulate the air phase pressure. The numerical model predictions are also compared with the field data
13
352 collected by Vasconcelos et al. (2009), during an actual refilling operation of a 4.4-km long, 350-mm diameter,
353 ductile iron pipeline in Brasilia, Brazil, operated by the waterworks company, CAESB.
355 Figure 4 shows the pressure history close to the ventilation orifice for all tested cases in the experimental
356 program with normalized orifice diameter d∗orif = dorif /D smaller or equal to 0.125. The transducer at that
357 station was located at the pipe crown, so it measured air phase pressures for most of the filling processes.
358 As it would be anticipated, higher pressurization levels were observed for smaller ventilation orifices while
359 the filling time was smaller for higher flow rates.
360 Air phase pressure head results were not significantly different for varying pipeline slopes. On the other
361 hand, there was a slight difference in the filling time between different pipeline slopes for a given ventilation
362 orifice and flow rate. This difference is attributed to the different initial water levels in the apparatus prior
363 to the closing of the knife gate valve at the downstream end. Also as it can be noticed in Figure 4, the
364 smallest ventilation the air phase pressure head kept increasing during the filling process, indicating steady
366 Figure 5 presents the pressure head hydrograph for x∗ = 0.39 (measured at pipe crown) for experimental
367 runs with d∗orif ≤ 0.125. A sudden step up in the pressure values the moment in which the flow regime
368 transition interface reached that station. As in the case of pressure measurements at the ventilation orifice,
369 these pressures kept increasing due to the increase in the air pressure for the smallest ventilation case. The
370 magnitude of the jump in the pressure readings was an indication of the strength of the pipe-filling bore
371 front, which increased for larger inflow rates and ventilation orifices. The absence of this discontinuity was
372 a sign of either gradual pressurization interface and/or the occurrence of interface breakdown feature due to
373 the interaction between the backward propagating pipe-filling bore front and the depression wave generated
374 at the pipeline inlet by the air pressurization. The relevance of the interface breakdown feature is that
375 its occurrence may pose difficulties to the application of pipeline filling models that use well-defined inflow
377 To illustrate the impact of the interface breakdown feature to results, Figure 6 presents two set of
p
378 trajectories of moving pressurization interfaces for normalized flow rates of Q∗ = Q/ gD5 = 0.245 and
379 0.490 and 2% slope, measured for all tested ventilation diameters. All these interfaces start as pipe-filling
380 bore fronts at x∗ = 0 as the gate valve is closed. Such bores lasted until x∗ ≈ 0.28 when they encountered
381 the depression wave originated from the upstream end of the pipe. For both flow rates interface breakdown
382 feature was noticed when the smallest ventilation orifice was used. Upon occurrence of the feature, the
383 pipe-filling bore became an open-channel bore that moved more slowly toward the ventilation orifice, leaving
14
384 an air intrusion on its top. Interestingly, as the backward-moving bores approached the ventilation orifice
385 there was an acceleration on their motion, regardless whether interface breakdown occurred or not. The
386 cause for this change in bore velocity is not determined yet.
387 When interface breakdown occurred, interface measurements included both the position of the open-
388 channel bore and the pressurization front. The latter was a gradual transition, and immediately following
389 the interface breakdown it was noticed that the pressurization front retreated. Soon afterwards, the pres-
390 surization front resumed the motion toward the ventilation orifice, trailing the open-channel bore. Figure 7
391 presents the trajectories for the condition Q∗ = 0.245 and 1% slope, for all four tested d∗orif . The largest one
392 (d∗orif = 0.50) has not generated any sign of air pressurization, and the pipe-filling bore kept its shape as it
393 propagated toward the ventilation point. However, for (d∗orif < 0.50) there was the occurrence of interface
394 breakdown, and the the trajectory of the pressurization fronts (thin lines) are plotted along the trajectories
395 of the pressurization bores. For the cases when d∗orif = 0.125 and d∗orif = 0.0938 the trajectory of the gradual
396 pressurization front was approximately parallel to the backward moving bore, which was moderately slowed
397 by the interface breakdown. For the smallest ventilation (d∗orif = 0.0625), on the other hand, the velocity of
398 the pressurization bore was significantly reduced, with a much larger separation between the pressurization
401 The comparison between experimental results and corresponding numerical predictions is presented in
402 Figures 8 and 9. All calculations were performed assuming that the clear PVC pipeline Manning roughness
403 coefficient was n=0.0085, and the wave celerity assumed for the PVC pipe was 200 m/s. Courant numbers
404 of 0.80, 0.90, and 0.95 were tested and results are shown for the case Cr=0.80.
405 Air phase pressure head predictions for (d∗orif < 0.50) and 1.0% slope are compared with experimental
406 results in Figure 8. Both Euler and UAPH approaches models showed good agreement with experimental
407 data for most of the cases, specially for the lower flow rates (Q∗ = 0.245 and Q∗ = 0.368). At higher flow
408 rates and smaller orifice sizes air pressures were slightly over-predicted. In few of the numerical predictions
409 there were strong, high-frequency oscillations on the air pressurization results as the pocket volume shrank
410 to zero. Results obtained in the intermediate point (x∗ = 0.39) for slope of 0.5% are presented in Figure
411 9, and indicate fair agreement between numerical and experimental results. There is a tendency of pressure
412 over-prediction by the numerical model, which also anticipates the arrival of the pressurization front at the
413 upstream end of the system. In chart with Q∗ = 0.490, d∗orif = 0.0625, there are two zones of with oscillations,
414 but these are due to the flow regime transition that are resulting from the use of the shock-capturing TPA
415 method. Animation of the simulated results for this particular case show that the backward-moving bore
15
416 front stops briefly close to x∗ = 0.4 between T ∗ 0.6 and T ∗ 0.8, resuming its movement towards upstream
417 later on. During this the oscillations are not detected in the simulation results.
418 A limitation of the numerical predictions was related to the prediction of the interface breakdown oc-
419 currences. The proposed numerical model (both air phase model implementations) was able to predict the
420 onset of the interface breakdown as the interaction of the depression wave and pipe filling bore resulted in
421 an open channel bore. However, unlike the experiments, the predicted pressurization front does not retreat
422 following the breakdown. There was a some over-prediction of the pressure head in cases when interface
423 breakdown occurred; results obtained with Euler equation indicate the instant of the breakdown by a second,
424 smaller increase in pressure head at t∗ ≈ 0.2. This discrepancy, however, was not significant and has not
426 Observations during the experimental runs indicate that pressure increases at the upstream reservoir
427 during the filling events. One recalls that prior to the knife gate valve closure, the reservoir head was steady.
428 Considering that the inflow rate into the reservoir was constant, the increase in reservoir head following the
429 knife gate valve maneuver indicates a drop in the inflow rate admitted into the pipeline inlet due to the
430 almost instantaneous air pressurization. Steeper reservoir pressure head increase is linked to stronger air
431 pressurization and decrease in the flow admitted into the system, which resulted in the depression wave that
432 trigged interface breakdown events. The numerical model provides generally agreement with experimental
433 measurements. Results, however, are not presented here for brevity.
434 It can be observed that both models yielded similarly accurate results when compared to experimental
435 data despite the two considerably different approaches to simulate air phase. Yet, an aspect to be considered
436 is the computational effort involved in each air phase modeling alternative. In general, the simulation time
437 using the Euler equation model was over 9 times larger than one required by the UAPH model approach
438 for the comparison with the experimental results. Not only due to the additional model complexity, but
439 the enforcement of the Courant condition for the air phase simulation using resulted in even smaller time
440 steps as the celerity of the air phase was in the order of 300 m/s. The criterion for computation end was the
441 shrinkage of the pocket to a lower volume limit (5% of initial volume).
443 The comparisons between the field data and the numerical predictions for the filling of CAESB ductile
444 iron pipeline are presented below in Figures 11 to 13. This 350 mm diameter transmission main has a pump
445 station, and the filling process occurs in two steps. In the first step, the initial 4.4-km extension line is filled
446 by gravity, throttling the upstream butterfly valve so that the inflow rate is limited to Q∗ = 0.18. In the
447 second step, pumps are turned on and the remainder 2.8 km of the pipeline is filled. The analysis presented
16
448 here focuses in simulation of the initial 1,700 meters of the gravity filling. The air valves positioned at x=400
449 m correspond to a couple of 50-mm, spherical shutter air release valves. The actual discharge area of these
450 valves was not measured, and was calibrated in the numerical model so that the observed air pressure at the
452 The pipeline profile is presented in Figure 10, and the assumed values for the Manning roughness was
453 n = 0.011 and for the celerity was 100 m/s. While the anticipated celerity is probably much larger, the
454 adopted value is adequate considering that the modeling is not focusing on transient pressure issues but
455 instead on pipeline filling. Moreover, the larger celerity helped reduce the computational effort for the
456 simulation. In these simulations, continuity error for the air phase calculation were limited to 1%.
457 The work by Vasconcelos et al. (2009) applied the TPA model that did not incorporate effects of air
458 pressurization to simulate pipeline filling. Figure 11 presents a comparison of the pressure head hydrograph
459 measured downstream from the pump station (x ≈ 380m), and the sample frequency was 4 seconds. One
460 notices that the field measurement signal an increase in the pressure head at about 200 seconds into the
461 simulation and attains a stable level. At about t > 1100 s the pressure begins to steadily rise again and
462 will arrive at 8 m when t > 2700 s. The results obtained with the traditional TPA model indicate a small
463 pressure rise (corresponding to the water depth) until about t = 2300 s, when the pressure rapidly climbs
464 achieving levels over 7.5 m after t = 3500 s. The results obtained with both the proposed model better
465 approximate the field measurements. Pressure begins to climb at t=500 seconds, and will arrive at 8.0 m
466 for T=2900 seconds (Euler equation model). The UAPH model presents fairly good agreement too, but at
467 t = 1700 s it begins to diverge from the solution obtained by the Euler equation, and pressure will attain
469 An analogous comparison, this time however focusing on the measured and predicted inflow rate admitted
470 into the pipeline, is presented in Figure 13. Flow measurements in the water main were performed with an
471 electromagnetic flow meter, with a sampling frequency of 1 minute. The butterfly valve opening was gradual,
472 and took approximately 4 minutes. The simulation performed with the traditional TPA model (presented in
473 Vasconcelos et al. (2009)) reproduced this gradual opening; the results presented here have skipped this for
474 simplicity, assuming the final opening right on the onset of simulation. Flow measurement indicate an initial
475 flow rate slightly above 40 L/s, which will start declining for t > 1600s, stabilizing in 29 L/s when t=2700
476 seconds. Assuming that the flow rate drop is caused by air pressurization (as in the case of the experiments
477 performed in this study), there seems to be a slight inconsistency with the pressure measurements which
478 indicate that pressure begins to climb when for t > 1200s. The cause for this possible inconsistency is not
479 determined. The numerical prediction by the TPA model indicate that the flow rate drop will occur much
480 later, whereas the proposed model indicate the flow rate drop occurring much sooner, as soon as air pressure
17
481 begins to climb in the pipeline.
482 CONCLUSIONS
483 This work presented two model framework alternatives for the simulation of the filling of a water mains
484 considering effects of air pressurization, together with experimental investigations on this problem using
485 an experimental apparatus. While different modeling alternatives to this or related problems have been
486 proposed, some of the hypothesis used in these previous investigations may limit the applicability of these
487 in cases when the filling is performed gradually and ventilation is limited. Thus there is a need of a new
489 In general both numerical models alternatives were successful in capturing the general trend of the
490 pressures heads and flow rates variation for experimental and field data. The model was also able to predict
491 accurately predict the occurrence of flow regime transitions, pipe-filling bores, interface breakdown and
492 interactions between flow features. While the UAPH model presents itself as an alternative that has much
493 smaller computational effort, the results obtained with the model using the Euler equation better approached
494 the field measurements. Future versions of the proposed model frameworks will aim to improve its stability
495 and reduce air phase continuity errors. Moreover, future versions of this model will try to incorporate other
496 mechanisms such as air pocket movement caused by drag and buoyancy forces.
497 Experimental results confirmed the importance that ventilation design has on the maximum pressures
498 observed in a system. A significant increase in the maximum pressure in the system along with a increase
499 in the filling time was observed for the smallest ventilation orifices. Experimental results established the
500 relationship of stronger pipeline slopes with increased pipeline filling time.
501 Future developments for this problem should also address more complex pipeline geometries, including
502 the formation of several air pockets at the same time and a wider range of boundary conditions anticipated
503 in transmission mains. Finally, as more insight is gained in the mechanisms for air pocket formation in closed
504 conduits, these may be included in the formulation of numerical models for the pipeline filling flow problem.
505 ACKNOWLEDGEMENTS
506 The authors thank the data provided by CAESB regarding the monitoring of the filling of the water main
507 presented in this work. The support of Auburn University in conducting this research is also acknowledged.
18
508 NOTATION
D = pipeline diameter
f = friction head loss in the short pipe portion inside the boundary condition right after the inlet
g = acceleration of gravity
hc = distance between the free surface and the centroid of the flow cross section (limited to D/2)
hs = surcharge head
Mair out = air mass that escapes through the ventilation orifice
Qrec = water flow rate which is admitted into the reservoir from the recirculation system
Qin = water flow rate which enters the upstream end of the pipe
Sf = source term that accounts for shear stress between air and pipe walls
19
u = flow velocity
θ = angle formed by free surface flow width and the pipe centerline
512
513 REFERENCES
514 Arai, K. and Yamamoto, K. (2003). “Transient analysis of mixed free-surface-pressurized flows with modified
515 slot model 1: Computational model and experiment.” Proc. FEDSM03 4th ASME-JSME Joint Fluids
517 Benjamin, T. B. (1968). “Gravity currents and related phenomena.” Journal of Fluid Mechanics, 31(02),
518 209–248.
519 Chaiko, M. and Brinckman, K. (2002). “Models for analysis of water hammer in piping with entrapped air.”
521 Cunge, J. A. et al. (1980). Practical Aspects of Computational River Hydraulics. Pitman Advanced Pub.
522 Program.
523 Fuertes, V., Arregui, F., Cabrera, E., and Iglesias, P. (2000). “Experimental setup of entrapped air pockets
524 model validation.” BHRA Group Conference Series Publication, Vol. 39, Bury St. Edmunds; Professional
526 Izquierdo, J., Fuertes, V., Cabrera, E., Iglesias, P., and Garcia-Serra, J. (1999). “Pipeline start-up with
528 Kalinske, A. and Bliss, P. (1943). “Removal of air from pipe lines by flowing water.” Proc. American Society
530 Liou, C. P. and Hunt, W. A. (1996). “Filling of pipelines with undulating elevation profiles.” J. Hydr. Engrg.,
20
532 Little, M. J. , Powell, J. C. , and Clark, P. B. (2008). “Air movement in water pipelines – some new
534 Macchione, F. and Morelli, M. (2003). “Practical aspects in comparing shock-capturing schemes for dam
536 Martin, C. S. (1976). “Entrapped air in pipelines.” Proc. Second BHRA International conference on pressure
537 surges.
538 Pothof, I. and Clemens, F. (2010). “Experimental study of air-water flow in downward sloping pipes air-water
540 Pozos, O., Sanchez, A., Rodal, E., and Fairuzov, Y. (2010). “Effects of water-air mixtures on hydraulic
542 Pulliam, T. H. (1981). “Characteristic Boundary Conditions for the Euler Equations.” Proceeding of Sym-
544 Sanders, B. and Bradford, S. (2011). “Network implementation of the two-component pressure approach for
547 Toro, E. (2009). Riemann solvers and numerical methods for fluid dynamics: a practical introduction.
549 Toro, E. F. (2001). Shock-Capturing Methods for Free-Surface Shallow Flows. Wiley.
550 Trajkovic, B., Ivetic, M., Calomino, F., and D’Ippolito, A. (1999). “Investigation of transition from free
551 surface to pressurized flow in a circular pipe.” Water science and technology, 39(9), 105–112.
552 Tran, P. (2011). “Propagation of pressure waves in two-component bubbly flow in horizontal pipes.” J. Hydr.
554 Vasconcelos, J. G. (2007). “Modelo matemático para simulação de enchimento de adutoras de água.” Proc.
555 24th Brazilian Congress of Environ. Sanit. Engrg., Belo Horizonte, Brazil (in Portuguese).
556 Vasconcelos, J., Wright, S., and Roe, P. (2006). “Improved simulation of flow regime transition in sewers:
558 Vasconcelos, J., Wright, S., and Roe, P. (2009). “Numerical oscillations in pipe-filling bore predictions by
21
560 Vasconcelos, J. G., Moraes, J. R. S., and Gebrim, D. V. B. (2009). “Field measurements and numerical
561 modeling of a water pipeline filling events.” Event Proc. 33rd IAHR Congress.
562 Vasconcelos, J. G. and Wright, S. J. (2005). “Experimental investigation of surges in a stormwater storage
564 Vasconcelos, J. and Wright, S. (2008). “Rapid Flow Startup in Filled Horizontal Pipelines.” J. Hydr. Engrg.,
566 Vasconcelos, J. and Wright, S. (2009). “Investigation of rapid filling of poorly ventilated stormwater storage
568 Zhou, F. F., Hicks, F. E., and Steffler,P. M.(2002). “Transient flow in a rapidly filling horizontal pipe
570 Zhou, L., Liu, D., Karney, B., and Zhang, Q. (2011). “The influence of entrapped air pockets on hydraulic
22
572 List of Tables p
573 1 Experimental variables. Flow rate normalized by Q∗ = Q/ gD5 and ventilation diameter by
574 d∗ = dorif /D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
23
p
TABLE 1. Experimental variables. Flow rate normalized by Q∗ = Q/ gD5 and ventilation diameter by
d∗ = dorif /D
24
575 List of Figures
576 1 Representation of the proposed model key components . . . . . . . . . . . . . . . . . . . . . . 26
577 2 Sketch of the experimental apparatus used in the investigation . . . . . . . . . . . . . . . . . 27
578 3 Flow chart for the model calculation procedures . . . . . . . . . . . . . . . . . . . . . . . . . . 28
579 4 Measured air phase pressure heads for all tested conditions where dorif < 0.5. . . . . . . . . . 29
580 5 Pressure head variation at the pipe crown for x∗ = 0.39 for all tested conditions where d∗orif <
581 0.5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
582 6 Trajectory of moving bore for Q∗ = 0.245 (a) and Q∗ = 0.368 (b) and pipeline slope = 2% . . 31
583 7 Trajectory of moving bore (thick line) and pressurization interface (thin line) for Q∗ = 0.245
584 and slope of 1% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
585 8 Measured and predicted air phase pressures head for all tested conditions where dorif < 0.5
586 and slope 1% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
587 9 Measured and predicted pressures at the pipe crown for x∗ = 0.39, dorif ∗ < 0.5 and slope 1% 34
588 10 Pipeline profile used by Vasconcelos et al. (2009) investigation . . . . . . . . . . . . . . . . . . 35
589 11 Field measurements and predicted heads at the upstream ventilation valve of the water main 36
590 12 Field measurements and predicted heads at the downstream ventilation valve of the water main 37
591 13 Field measurements and flow rates at the upstream ventilation valve of the water main . . . . 38
25
Ventilation orifices
Energy Grade Li
ne
wa
te
rfl
ow
Air pockets
FIG. 1. Representation of the proposed model key components
26
Water inlet from a reservoir
6.69 m
10.96 m
ADV
Recirculation pump
27
FIG. 3. Flow chart for the model calculation procedures
28
7
Q*=0.245, dorif*=0.125 Q*=0.368, dorif*=0.125 Q*=0.490, dorif*=0.125
6
0
7
Q*=0.245, dorif*=0.09375 Q*=0.368, dorif*=0.09375 Q*=0.490, dorif*=0.09375
6
5
hair*=hair/D
0
7
Q*=0.245, dorif*=0.0625 Q*=0.368, dorif*=0.0625 Q*=0.490, dorif*=0.0625
6
0
0 1 2 3 0 1 2 3 0 1 2 3
* 0.5
t =t/(L/(gD) )
FIG. 4. Measured air phase pressure heads for all tested conditions where dorif < 0.5.
29
7
Q*=0.490, dorif*=0.0625 Q*=0.490, dorif*=0.09375 Q*=0.490, dorif*=0.09375
6
0
7
Q*=0.368, dorif*=0.0625 Q*=0.368, dorif*=0.09375 Q*=0.368, dorif*=0.125
6
5
hair*=hair/D
0
7
Q*=0.245, dorif*=0.0625 Q*=0.245, dorif*=0.09375 Q*=0.245, dorif*=0.125
6
0
0 1 2 3 0 1 2 3 0 1 2 3
t*=t/(L/(gD)0.5)
FIG. 5. Pressure head variation at the pipe crown for x∗ = 0.39 for all tested conditions where d∗orif < 0.5.
30
FIG. 6. Trajectory of moving bore for Q∗ = 0.245 (a) and Q∗ = 0.368 (b) and pipeline slope = 2%
31
FIG. 7. Trajectory of moving bore (thick line) and pressurization interface (thin line) for Q∗ = 0.245
and slope of 1%
32
7
Q*=0.245, dorif*=0.125 Q*=0.368, dorif*=0.125 Q*=0.490, dorif*=0.125
6
5
4
3
2
1
0
7
Q*=0.245, dorif*=0.09375 Q*=0.368, dorif*=0.09375 Q*=0.490, dorif*=0.09375
6
5
hres*=hres/D
4
3
2
1
0
7
Q*=0.245, dorif*=0.0625 Q*=0.368, dorif*=0.0625
6
5
4
3
2
1
Q*=0.490, dorif*=0.0625
0
0 1 2 3 0 1 2 3 0 1 2 3
*
t =t/(L/sqrt(gD))
FIG. 8. Measured and predicted air phase pressures head for all tested conditions where dorif < 0.5 and
slope 1%
33
7
Q*=0.245, dorif*=0.125 Q*=0.368, dorif*=0.125
6
5
4
3
2
1
Q*=0.490, dorif*=0.125
0
7
Q*=0.245, dorif*=0.09375 Q*=0.368, dorif*=0.09375
6
5
hres*=hres/D
4
3
2
1
Q*=0.490, dorif*=0.09375
0
7
Q*=0.245, dorif*=0.0625
6
5
4
3
2
1
Q*=0.368, dorif*=0.0625 Q*=0.490, dorif*=0.0625
0
0 1 2 3 0 1 2 3 0 1 2 3
*
t =t/(L/sqrt(gD))
FIG. 9. Measured and predicted pressures at the pipe crown for x∗ = 0.39, dorif ∗ < 0.5 and slope 1%
34
FIG. 10. Pipeline profile used by Vasconcelos et al. (2009) investigation
35
10
Measured
Traditional TPA
8 UAPH
Euler equations
6
head (m)
0
0 500 1000 1500 2000 2500 3000 3500 4000
time (s)
FIG. 11. Field measurements and predicted heads at the upstream ventilation valve of the water main
36
25
Measured
Traditional TPA
20 UAPH
Euler equations
15
head (m)
10
0
0 500 1000 1500 2000 2500 3000 3500 4000
time (s)
FIG. 12. Field measurements and predicted heads at the downstream ventilation valve of the water
main
37
0.06
0.05
0.04
flow rate ($m3/s$)
0.03
0.02
Measured
Traditional TPA
0.01
UAPH
Euler
0
0 500 1000 1500 2000 2500 3000 3500 4000
time (s)
FIG. 13. Field measurements and flow rates at the upstream ventilation valve of the water main
38