Transient Based Leak Detection
Content
Available online at www.sciencedirect.com
Procedia Environmental Sciences 19 (2013) 814 – 822
Four Dec cades of Pro ogress in Monitoring M an nd Modelin ng of Proces sses in the S Soil-PlantAt tmosphere System: S Ap pplications and a Challen nges
Eff fectivene ess asses ssment o of pipe sy ystems by b mean ns of transien nt test-based tec chniques
B. Brunonea*, M. Ferrantea, S. Menicon nia, and C. Massaria
a
Dipa artimento di Ingeg gneria Civile ed Ambientale, A Univ versity of Perugia a, via G. Duranti 93, 06125 Perug gia, Italy
Abstract
In this paper, the t reliability y of transien nt test-based techniques for f detecting pipe system m s shown by means m of num merical and ex xperimental tests t – both in n laboratory a and real pipe e anomalies is systems – w with particular regard to the case of a part tially closed in-line i valve. With W regard to the analysis s of pressure s tion is focused d on the time domain appro oach. Moreover, the optima al modality of f signals, attent transient tes st execution is s discussed. References R con ncerning the diagnosis d of other anomalie es (e.g., leaks, , partial block kages, and ille egal connectio ons) by means s of transient tests are also included for the interested d reader.
A Authors.Published Publishedby byElsevier Elsevier r B.V. © 2013 2013 The The Authors. © B.V Selection and/or and d/orpeer-review peer-review wunder under respons sibility of the Sc cientific Comm mittee the con nference. Selection responsibility of the Scientific Committee of of the conference
Keywords: pipe e diagnosis; trans sient test; anomali ies.
1. Introduc ction In this pa aper the impo ortance of fas st, reliable an nd cheap techn niques for the e diagnosis of f supply pipe e systems – he ereafter referr red to as SPS – is discussed d by consideri ing the presen nt scenarios in n which water r is scarcer an nd scarcer as well w as energy y costs rise un npredictably. The T efficiency y of SPS can increase i) by y reducing lea aks that are ar re costly in ter rms of lost wa ater, potential lly adverse wa ater quality ef ffects, and the e energy cons sumed in supp plying them; ii) i checking th he status of in n-line valves as a well as det tecting partial l blockages d due to depositi ion of sedime ents, fouling p processes, and d corrosion. In n fact, neglige ently partially y
nding author. Tel l.: +39-075-58536 617 * Correspon E-mail addr dress: brunone@u unipg.it
1878-0296 © 2013 The Authors. Published by Elsevier B.V Selection and/or peer-review under responsibility of the Scientific Committee of the Conference doi:10.1016/j.proenv.2013.06.090
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
815
closed in-line valves and partial blockages cause undesirable local head losses that reduce the carrying capacity of SPS and increase energy costs, particularly in elevatory mains. Because of limited resources for maintenance of SPS, an efficient program has to be established in which the budgets should be allocated to the most vulnerable system components on the basis of the results of as frequent as possible surveys. Frequent surveying of SPS, that is essential to detect any anomaly as quickly as possible, is a complicated and time-demanding task. In fact current practice in water industry does not provide a comprehensive method for SPS surveying. In addition there is not enough research in this area since, for both economic and cultural reasons, water industry has never been stimulated to accept efficient management rules. The comparison with the gas and oil industries is severe: well-instrumented pipelines and large investment in instrumentation as a routine make the difference. In such a scenario, that is slightly changing since the value of the drinkable water is still surprisingly underestimated, there is an extreme need of simple, reliable and cheap techniques for SPS survey and diagnosis. An impulse to research activity in the SPS management derives from the unexpected reliability of transient test-based techniques with respect to more consolidated procedures in which pressure and flow measurements are carried out in steady-state conditions. For the sake of brevity, in this paper – as an example of SPS diagnosis –, the case of a partially closed in-line valve is considered and the results of numerical, laboratory and real pipe system experiments are shown. For the sake of clarity, the mechanism of interaction between pressure waves and a partially closed in-line valve is analyzed pointing out differences with respect to the case of a single pipe (i.e. a constant diameter pipe with no anomalies). Moreover, attention is focused on the first phases of the transients when the presence of an anomaly can be revealed by considering the reflected pressure wave. In such a context, the reflection coefficient of the anomaly, C, defined as the ratio between the reflected, h, and incoming pressure wave, h:
C
h h
(1)
plays an important role. In addition, references address the interesting reader towards detection and sizing of other anomalies as well as within other approaches for the analysis of transient test results: leaks [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11], illegal connections [12], and partial blockages [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]. 2. Numerical experiments Numerical simulations concern a frictionless elastic pipe with internal diameter D = 93.30 mm, length L = 164.93 m, and pressure wave speed a = 372.00 m/s supplied by a constant head reservoir. At the downstream end section of the pipe the maneuver valve is installed, discharging into the atmosphere. The partially closed in-line valve is installed at a distance LIV = 75.87 m upstream of the maneuver valve. Pressure signal, h – i.e. the pressure time-history –- is acquired immediately upstream the maneuver valve. Figure 1 shows numerical pressure signals during transients with a steady-state discharge Q0 = 2.57 l/s and different values of the local head loss coefficient of the in-line valve, IV, defined by:
0, IV 2 0 2
IV
Q 2 gA
(2)
816
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
d loss at the in n-line valve IV V, g = gravity acceleration, and A = where 0,iV is the steady-state local head pipe area. M More precisely, the cases of IV = 109, 399 9, and 916 are e considered.
Figure 1: Num merical pressure si ignals in the case of a pipe with a p partially closed in n-line valve for Q0 = 2.57 l/s. As u useful reference, the case of a pip pe with no anomaly (single pipe) is also included.
Each sha arp variation in i the pressur re signal (Fig gure 1) corres sponds to a pressure p wave e crossing the e measuremen nt section: at tV = 0 s, the wave w generate ed by the clos sing of the end valve passe es through the e measuremen nt section, ca ausing the fir rst sharp pres ssure rise h = 14.29 m and proceed ds toward the e reservoir. W When at t = LIV / a = 0.204 s the pressure w wave generate ed by the man neuver reaches s the partially y V closed in-lin ne valve, it is partially p reflected and then n it returns to the t measurement section at tIV = 2LIV/a = 0.408 s wel ll before the pressure wav ve that, conti inuing beyond d the in-line valve, is refl flected by the e upstream res servoir. At tR = 2L/a = 0.885 s, this latter r wave returns s to the measurement section n. By evalu uating times at t which press sure waves pa ass through th he measurement section, it is possible to o localize the partially close ed in-line valv ve by means o of the followin ng equation:
L IV t IV tR tV tV
(3)
section to the nks the distanc ces from the measurement m e in-line valve e, LIV, and the e reservoir, L, which lin to times tV, tIV, and tR. Moreover, by analyzin ng Figure 1 it can be observ ved that, for a given Q0, the e larger IV, i. e. the smaller r , the larger the e increase of the t pressure si ignal at tIV; no o pressure rise e the opening degree of the in-line valve, erved for the single s pipe. Furthermore, th he latter incre ease is double e the wave ref flected by the e can be obse in-line valve e, hIV because the end va alve is compl letely closed at a tIV. The ref flection coeff ficient, CIV, is s given by the e following rel lationship:
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
817
C IV
hIV h
1
1
IV
0, IV
Q0 2 aA
g IV V 2a2
(4)
yzing Figure 1, for IV = 916, as an e example, hIV nd the resulti ing reflection n By analy V = 3.14 m an coefficient i is CIV = 0.22. By using Equ uation 4 inver rsely it is possible to derive the value of f IV and then n the valve op pening degree. . 3. Laborato ory and real pipe p system experiments e Laboratory tests have e been execut ted on a hig gh density polyethylene (H HDPE) pipe w with nominal l N110, internal l diameter D = 93.3 mm, a and thickness e = 8.1 mm. The anomaly y is a partially y diameter DN closed in-lin ne ball valve [22] [23]. As the numerica al tests, the di istance betwee en the maneuv ver valve and d the anomaly y (reservoir) is s LIV = 75.87 m (L = 164.93 3 m) and the st teady-state dis scharge is Q0 = 2.57 l/s.
Figure 2: Exp perimental pressu ure signals in the e case of a pipe with a partially y closed in-line valve v for Q0= 2.5 57 l/s. As useful l reference, the c case of a pipe with h no anomaly (sin ngle pipe) is also o included.
To highli ight pressure waves w clearly in Figure 2, t the difference between the transient, t h, an nd the steadystate value, h0, is plotted vs. the time t relative to t the beginning g of the maneu uver. As for t the numerical l s, the effect of f a partially clo osed in-line v valve on the pr ressure signal is a sudden in ncrease, yIV, at t experiments tIV. The larg ger IV, the lar rger yIV, with yIV = 0 in the case of fully opened valve e. For a partia ally closed inline valve, t the entity corr responds to th he value of the l coefficien nt, IV, that de epends on the e e local head loss characteristi ics of the in-li ine valve (e.g g., type and siz ze). In Figure e 3, the experi imental mean n values of IV V vs. the relati ive area, AIV/A – with AIV be eing the in-lin ne valve openi ing area – are shown. An exam mple of a trans sient test-based d diagnosis pr rocedure in a real pipe syst tem is reported d in Figure 4. . Precisely, by y means of a Lagrangian L model m [5] [19], the clear pres ssure rise in th he pressure sig gnal of Figure e
818
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
4b – due to t the closure ex xecuted manua ally at valve V of Figure 4a a – has been as scribed to the presence of a partially closed in-line va alve.
Figure 3: Hydr raulic characteristics of the in-valv ve.
Figure 4: Rieti i supply system: a) a skeletonization n; b) pressure sign nal at section M due d to the comple ete closure of valv ve V.
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
819
4. Locating g and sizing an nomalies As for a any diagnosis procedure, th he following two requirem ments have to o be fulfilled d: i) to locate e anomalies (i i.e., to determ mine LIV in the case of the pa artially closed d in-line valve) ) with the high hest precision n to limit the costs of the in ntervention (e e.g., those due e to excavatio on), and ii) to give the large est number of f aracteristics (i.e., kind and d entity; the valve v opening degree in th he case of the e information about its cha partially closed in-line va alve) to evalua ate properly if f any action is needed. st requirement t, according to o Eq. (3), an anomaly a can be b located by y detecting the e With regard to the firs s at the measu urement sectio on of the pres ssure waves generated g by the t maneuver and reflected d arrival times by the anom maly itself. As s a consequence, the higher r the precision n in the evalua ation of such arrival times, , the higher th he quality of location estima ate. As shown n in [4] [5] [6 6] [10] [24] [25], arrival tim mes in Eq. (3) can be detecte ed reliably by y means of the e wavelet tran nsform [26] [2 27]. An exam mple of the per rformance of the wavelet a analysis is giv ven in Figure 5b, where eve en if the time e domain anal lysis of the sig gnal in Figure e 5a, in which IV = 109, do oes not reveal any evident si ingularities at t tIV, the wav velet analysis properly poi ints out at th his time the presence p of a very small s step variation n attributable to the passag ge of the pre essure wave r reflected by the t partially closed c in-line e valve. More e precisely, sc crutiny of Figu ure 5b reveals s the singulari ities occurring g at tV = 0.0156 s, tIV = 0.42 277 s, and tR = 0.9131 s. In n this case, the e distance LIV rom Equation (3) is equal to t 75.73 m, w with a relative e V calculated fr error equal t to 0.32 %. With W respect to time domain n analysis, using wavelet fu unctions not o only improves s the precision n in evaluatin ng times in Eq quation (3) and d consequentl ly in the localization of the e singularities, , but also furt thers the impa artiality of the pressure sign nal analysis. With regard to the sec cond requirem ment – i.e., to give information about the e anomaly – w within a short t period analy ysis the evalua ation of the re eflection coeff ficient can allo ow to size the e anomaly. Of f course, more e precise information can be b obtained within w a Inver rse Transient Analysis A (ITA A), as propose ed by Liggett t 28]. and Chen [2
Figure 5: (a) P Pressure signal of f Figure 2 in the fi irst characteristic time with
IV
= 109, and (b) corr responding wavellet analysis.
820
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
5. Generati ion of "safe" pressure wav ves As any te echnique, the implementati ion of transien nt tests in real systems may y encounter so some practical l problems tha at are discusse ed below. When de ealing with re eal pipe systems, accessing g and generat ting transients s are often ch hallenging. In n many cases, valves are installed in locations l in w which pressur re measureme ents are diffi icult or valve e lly for valves in large-diam meter pipes. Another A complication is that t many valves s response is slow, especial mit a sufficien ntly precise de etermination o of pressure su urges caused by b the maneu uver. Thus, in n do not perm many condit tions introduc cing transients s by means oth her than valve e manipulation n would be of f great benefit. . However, to o date, a port table and eas sy-to-use devi ice that does not entail ov verly sophistic cated support t instrumentat tion beyond th hat required fo or registering the pressure signal s has not t been availabl le. To fill this s need, the po ortable pressu ure wave mak ker (PPWM) d device (Figur re 6) has been n refined and d tested in the e Water Engi ineering Labo oratory in th he Departmen nt of Civil and a Environm mental Engine eering of the e University o of Perugia. Th he PPWM device (Figure 6 6) consists of f a steel vesse el filled with water and air r that can be placed under r pressure by means of a standard com mpressor. The PPWM and test pipe are e gth conduit wit th a small diam meter valve at a the end secti ion (hereafter referred to as s connected by a short leng valve). At the e beginning of o the test, aft ter disinfectin ng the device and refilling it with water r connection v from the tes st pipe, the ins side pressure is fixed larger r of a given amount a than th he pressure in nside the pipe. . Opening of the connectio on valve cause es a pressure w wave that trav vels along the e pipe detectin ng anomalies. . onnection valv ve can be actu uated very fas st and the pres ssure wave ge enerated by th he PPWM can n Since the co be controlle ed with reason nable precision n both in term ms of amplitu ude and stabili ity, by using s such a device e e weak point ts of transien nt test-based techniques ca an be eliminated. In fact,, it has been n most of the demonstrate ed [3] [29] [30 0] that the enti ity of the pres ssure wave can be fixed reliably – and th hen the risk of f unsafe overp pressure is con ntrolled – and d sharp pressur re waves can be b generated.
Figure 6: The P PPWM (Portable e Pressure Wave Maker) M device an nd the test pipe.
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
821
Acknowledgements This research has been supported by Fondazione Cassa Risparmio Perugia under the Project ‘‘Leaks and blockages detection techniques for reducing energy and natural resources wastage’’. References
[1] Brunone, B. Transient test-based technique for leak detection in outfall pipes. J Water Resour Plann Manage 1999;125:302-306. [2] Brunone, B., and Ferrante, M. Detecting leaks in pressurised pipes by means of transients. J. Hydraul. Res.2001;39:539-547. [3] Brunone, B., Ferrante, M., and Meniconi, S. Portable pressure wave-maker for leak detection and pipe system characterization. J. of AWWA 2008;100:108-116. [4] Ferrante, M., Brunone, B., and Meniconi, S. Wavelets for the analysis of transient pressure signals for leak detection. J. Hydraul. Eng. 2007;133:1274-1282. [5] Ferrante, M., Brunone, B., and Meniconi, S. Leak detection in branched pipe systems coupling wavelet analysis and a Lagrangian model. J. Water Supply: Research and Technology – AQUA 2009;58:95-106. [6] Ferrante, M., Brunone, B., and Meniconi, S. Leak-edge detection, J. Hydraul. Res. 2009;47:233-241. [7] Lee, P. J., Vítkovský, J. P., Lambert, M. F., Simpson, A. R., and Liggett, J. A. Frequency domain analysis for detecting pipelines leaks. J. Hydraul. Eng. 2005;131:596-604. [8] Meniconi, S., Brunone, B., Ferrante, M., and Massari, C. Numerical and experimental investigation of leaks in viscoelastic pressurized pipe flow. Drinking Water Engineering and Science 2013;6:11-16. [9] Mpesha, W., Gassman, S. L., and Chaudhry, M. H. Leak detection in pipes by frequency response method. J. Hydraul. Eng. 2001;127:134-147. [10] Stoianov, I., Karney, B. W., Covas, D., Maksimovic, C., and Graham, N. Wavelet processing of transient signals for pipeline leak location and quantification. Proc., Int. Conf. on Computing and Control for the Water Industry (CCWI), B. Coulbeck and J. Rance eds. 2001;1:65-76. [11] Wang, X. -J., Lambert, M. F., Simpson, A. R., Liggett, J. A., and Vitkovsky, J. P. Leak detection in pipelines using the damping of fluid transients. J. Hydraul. Eng. 2002;128:697-711. [12] Meniconi, S., Brunone, B., Ferrante, M., and Massari, C. Transient tests for locating and sizing illegal branches in pipe systems. J. of Hydroinformatics 2011;13:334-345. [13] Brunone, B., Ferrante, M., and Meniconi, S. Discussion of “Detection of partial blockage in single pipelines” by P. K., [1] Mohapatra, M. H., Chaudhry, A. A., Kassem, and J., Moloo. J. Hydraul. Eng. 2008;134:872-874. [14] Lee, P. J., Vítkovský, J. P., Lambert, M. F., Simpson, A. R., and Liggett, J. A. Discrete blockage detection in pipelines using the frequency response diagram: numerical study. J. Hydraul. Eng. 2008;134:658-663. [15] Meniconi, S., Brunone, B., and Ferrante, M. Water hammer pressure waves at cross-section changes in series in viscoelastic pipes. J. of Fluids and Structures 2012;33:44-58. [16] Mohapatra, P. K., Chaudhry, M. H., Kassem, A. A., and Moloo, J. Detection of a partial blockage in single pipelines. J. Hydraul. Eng.2006;132:200-206. [17] Mohapatra, P. K., Chaudhry, M. H., Kassem, A. A., and Moloo, J. Detection of partial blockages in a branched piping system by the frequency responce method. J. Fluids Eng. 2006;128:1106-1114. [18] Sattar, A. M., Chaudhry, M. H., and Kassem, A. A. Partial blockage detection in pipelines by frequency response method. J. Hydraul. Eng. 2008;134:76-89. [19] Meniconi, S., Brunone, B., Ferrante, M. and Massari, C. Potential of transient tests to diagnose real supply pipe systems: what can be done with a single extemporary test. J Water Resour Plann Manage 2011;137:238-241. [20] Wang, X. -J., Lambert, M. F., and Simpson, A. R. Detection and location of a partial blockage in a pipeline using damping of fluid transients. J Water Resour Plann Manage 2005;131:244-249. [21] Contractor, D. N. The reflection of waterhammer pressure waves from minor losses. J. Basic Eng.1965;445-451. [22] Meniconi, S., Brunone, B., and Ferrante, M. In-line pipe device checking by short period analysis of transient tests. J. Hydraul. Eng. 2011;137:713-722. [23] Meniconi, S., Brunone, B., Ferrante, M., and Massari, C. Transient hydrodynamics of in-line valves in viscoelastic pressurised pipes. Long period analysis. Experiments in Fluids 2012;53:265-275. [24] Ferrante, M., and Brunone, B. Pipe system diagnosis and leak detection by unsteady-state tests. 1 Harmonic analysis. Advances in Water Resources. 2003;26:95-105. [25] Ferrante, M., and Brunone, B. Pipe system diagnosis and leak detection by unsteady-state tests. 2 Wavelet analysis. Advances in Water Resources. 2003;26:107-116.
822
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
[26] Mallat, S. G., and Hwang, W. L. Singularity detection and processing with wavelets. IEEE Trans. on Information Theory 1992;38:617-643. [27] Mallat, S. G., and Zhong, S. Characterization of signals from multiscale edges. IEEE Trans. on Pattern Analysis and Machine Intelligence 1992;14:710-732. [28] Liggett, J. A., and Chen, L.C. Inverse transient analysis in pipe networks. J. Hydraul. Eng. 1994;120:934-955. [29] Meniconi, S., Brunone, B., Ferrante, M., and Massari, C. Small amplitude sharp pressure waves to diagnose pipe systems. Water Resources Management 2011;25:79-96. [30] Brunone, B, and Meniconi, S. Discussion of "Case studies of leak detection and location in water pipe systems by inverse transient analysis" by D. Covas and H. Ramos. J Water Resour Plann Manage 2013;139:126-127.
Procedia Environmental Sciences 19 (2013) 814 – 822
Four Dec cades of Pro ogress in Monitoring M an nd Modelin ng of Proces sses in the S Soil-PlantAt tmosphere System: S Ap pplications and a Challen nges
Eff fectivene ess asses ssment o of pipe sy ystems by b mean ns of transien nt test-based tec chniques
B. Brunonea*, M. Ferrantea, S. Menicon nia, and C. Massaria
a
Dipa artimento di Ingeg gneria Civile ed Ambientale, A Univ versity of Perugia a, via G. Duranti 93, 06125 Perug gia, Italy
Abstract
In this paper, the t reliability y of transien nt test-based techniques for f detecting pipe system m s shown by means m of num merical and ex xperimental tests t – both in n laboratory a and real pipe e anomalies is systems – w with particular regard to the case of a part tially closed in-line i valve. With W regard to the analysis s of pressure s tion is focused d on the time domain appro oach. Moreover, the optima al modality of f signals, attent transient tes st execution is s discussed. References R con ncerning the diagnosis d of other anomalie es (e.g., leaks, , partial block kages, and ille egal connectio ons) by means s of transient tests are also included for the interested d reader.
A Authors.Published Publishedby byElsevier Elsevier r B.V. © 2013 2013 The The Authors. © B.V Selection and/or and d/orpeer-review peer-review wunder under respons sibility of the Sc cientific Comm mittee the con nference. Selection responsibility of the Scientific Committee of of the conference
Keywords: pipe e diagnosis; trans sient test; anomali ies.
1. Introduc ction In this pa aper the impo ortance of fas st, reliable an nd cheap techn niques for the e diagnosis of f supply pipe e systems – he ereafter referr red to as SPS – is discussed d by consideri ing the presen nt scenarios in n which water r is scarcer an nd scarcer as well w as energy y costs rise un npredictably. The T efficiency y of SPS can increase i) by y reducing lea aks that are ar re costly in ter rms of lost wa ater, potential lly adverse wa ater quality ef ffects, and the e energy cons sumed in supp plying them; ii) i checking th he status of in n-line valves as a well as det tecting partial l blockages d due to depositi ion of sedime ents, fouling p processes, and d corrosion. In n fact, neglige ently partially y
nding author. Tel l.: +39-075-58536 617 * Correspon E-mail addr dress: brunone@u unipg.it
1878-0296 © 2013 The Authors. Published by Elsevier B.V Selection and/or peer-review under responsibility of the Scientific Committee of the Conference doi:10.1016/j.proenv.2013.06.090
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
815
closed in-line valves and partial blockages cause undesirable local head losses that reduce the carrying capacity of SPS and increase energy costs, particularly in elevatory mains. Because of limited resources for maintenance of SPS, an efficient program has to be established in which the budgets should be allocated to the most vulnerable system components on the basis of the results of as frequent as possible surveys. Frequent surveying of SPS, that is essential to detect any anomaly as quickly as possible, is a complicated and time-demanding task. In fact current practice in water industry does not provide a comprehensive method for SPS surveying. In addition there is not enough research in this area since, for both economic and cultural reasons, water industry has never been stimulated to accept efficient management rules. The comparison with the gas and oil industries is severe: well-instrumented pipelines and large investment in instrumentation as a routine make the difference. In such a scenario, that is slightly changing since the value of the drinkable water is still surprisingly underestimated, there is an extreme need of simple, reliable and cheap techniques for SPS survey and diagnosis. An impulse to research activity in the SPS management derives from the unexpected reliability of transient test-based techniques with respect to more consolidated procedures in which pressure and flow measurements are carried out in steady-state conditions. For the sake of brevity, in this paper – as an example of SPS diagnosis –, the case of a partially closed in-line valve is considered and the results of numerical, laboratory and real pipe system experiments are shown. For the sake of clarity, the mechanism of interaction between pressure waves and a partially closed in-line valve is analyzed pointing out differences with respect to the case of a single pipe (i.e. a constant diameter pipe with no anomalies). Moreover, attention is focused on the first phases of the transients when the presence of an anomaly can be revealed by considering the reflected pressure wave. In such a context, the reflection coefficient of the anomaly, C, defined as the ratio between the reflected, h, and incoming pressure wave, h:
C
h h
(1)
plays an important role. In addition, references address the interesting reader towards detection and sizing of other anomalies as well as within other approaches for the analysis of transient test results: leaks [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11], illegal connections [12], and partial blockages [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]. 2. Numerical experiments Numerical simulations concern a frictionless elastic pipe with internal diameter D = 93.30 mm, length L = 164.93 m, and pressure wave speed a = 372.00 m/s supplied by a constant head reservoir. At the downstream end section of the pipe the maneuver valve is installed, discharging into the atmosphere. The partially closed in-line valve is installed at a distance LIV = 75.87 m upstream of the maneuver valve. Pressure signal, h – i.e. the pressure time-history –- is acquired immediately upstream the maneuver valve. Figure 1 shows numerical pressure signals during transients with a steady-state discharge Q0 = 2.57 l/s and different values of the local head loss coefficient of the in-line valve, IV, defined by:
0, IV 2 0 2
IV
Q 2 gA
(2)
816
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
d loss at the in n-line valve IV V, g = gravity acceleration, and A = where 0,iV is the steady-state local head pipe area. M More precisely, the cases of IV = 109, 399 9, and 916 are e considered.
Figure 1: Num merical pressure si ignals in the case of a pipe with a p partially closed in n-line valve for Q0 = 2.57 l/s. As u useful reference, the case of a pip pe with no anomaly (single pipe) is also included.
Each sha arp variation in i the pressur re signal (Fig gure 1) corres sponds to a pressure p wave e crossing the e measuremen nt section: at tV = 0 s, the wave w generate ed by the clos sing of the end valve passe es through the e measuremen nt section, ca ausing the fir rst sharp pres ssure rise h = 14.29 m and proceed ds toward the e reservoir. W When at t = LIV / a = 0.204 s the pressure w wave generate ed by the man neuver reaches s the partially y V closed in-lin ne valve, it is partially p reflected and then n it returns to the t measurement section at tIV = 2LIV/a = 0.408 s wel ll before the pressure wav ve that, conti inuing beyond d the in-line valve, is refl flected by the e upstream res servoir. At tR = 2L/a = 0.885 s, this latter r wave returns s to the measurement section n. By evalu uating times at t which press sure waves pa ass through th he measurement section, it is possible to o localize the partially close ed in-line valv ve by means o of the followin ng equation:
L IV t IV tR tV tV
(3)
section to the nks the distanc ces from the measurement m e in-line valve e, LIV, and the e reservoir, L, which lin to times tV, tIV, and tR. Moreover, by analyzin ng Figure 1 it can be observ ved that, for a given Q0, the e larger IV, i. e. the smaller r , the larger the e increase of the t pressure si ignal at tIV; no o pressure rise e the opening degree of the in-line valve, erved for the single s pipe. Furthermore, th he latter incre ease is double e the wave ref flected by the e can be obse in-line valve e, hIV because the end va alve is compl letely closed at a tIV. The ref flection coeff ficient, CIV, is s given by the e following rel lationship:
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
817
C IV
hIV h
1
1
IV
0, IV
Q0 2 aA
g IV V 2a2
(4)
yzing Figure 1, for IV = 916, as an e example, hIV nd the resulti ing reflection n By analy V = 3.14 m an coefficient i is CIV = 0.22. By using Equ uation 4 inver rsely it is possible to derive the value of f IV and then n the valve op pening degree. . 3. Laborato ory and real pipe p system experiments e Laboratory tests have e been execut ted on a hig gh density polyethylene (H HDPE) pipe w with nominal l N110, internal l diameter D = 93.3 mm, a and thickness e = 8.1 mm. The anomaly y is a partially y diameter DN closed in-lin ne ball valve [22] [23]. As the numerica al tests, the di istance betwee en the maneuv ver valve and d the anomaly y (reservoir) is s LIV = 75.87 m (L = 164.93 3 m) and the st teady-state dis scharge is Q0 = 2.57 l/s.
Figure 2: Exp perimental pressu ure signals in the e case of a pipe with a partially y closed in-line valve v for Q0= 2.5 57 l/s. As useful l reference, the c case of a pipe with h no anomaly (sin ngle pipe) is also o included.
To highli ight pressure waves w clearly in Figure 2, t the difference between the transient, t h, an nd the steadystate value, h0, is plotted vs. the time t relative to t the beginning g of the maneu uver. As for t the numerical l s, the effect of f a partially clo osed in-line v valve on the pr ressure signal is a sudden in ncrease, yIV, at t experiments tIV. The larg ger IV, the lar rger yIV, with yIV = 0 in the case of fully opened valve e. For a partia ally closed inline valve, t the entity corr responds to th he value of the l coefficien nt, IV, that de epends on the e e local head loss characteristi ics of the in-li ine valve (e.g g., type and siz ze). In Figure e 3, the experi imental mean n values of IV V vs. the relati ive area, AIV/A – with AIV be eing the in-lin ne valve openi ing area – are shown. An exam mple of a trans sient test-based d diagnosis pr rocedure in a real pipe syst tem is reported d in Figure 4. . Precisely, by y means of a Lagrangian L model m [5] [19], the clear pres ssure rise in th he pressure sig gnal of Figure e
818
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
4b – due to t the closure ex xecuted manua ally at valve V of Figure 4a a – has been as scribed to the presence of a partially closed in-line va alve.
Figure 3: Hydr raulic characteristics of the in-valv ve.
Figure 4: Rieti i supply system: a) a skeletonization n; b) pressure sign nal at section M due d to the comple ete closure of valv ve V.
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
819
4. Locating g and sizing an nomalies As for a any diagnosis procedure, th he following two requirem ments have to o be fulfilled d: i) to locate e anomalies (i i.e., to determ mine LIV in the case of the pa artially closed d in-line valve) ) with the high hest precision n to limit the costs of the in ntervention (e e.g., those due e to excavatio on), and ii) to give the large est number of f aracteristics (i.e., kind and d entity; the valve v opening degree in th he case of the e information about its cha partially closed in-line va alve) to evalua ate properly if f any action is needed. st requirement t, according to o Eq. (3), an anomaly a can be b located by y detecting the e With regard to the firs s at the measu urement sectio on of the pres ssure waves generated g by the t maneuver and reflected d arrival times by the anom maly itself. As s a consequence, the higher r the precision n in the evalua ation of such arrival times, , the higher th he quality of location estima ate. As shown n in [4] [5] [6 6] [10] [24] [25], arrival tim mes in Eq. (3) can be detecte ed reliably by y means of the e wavelet tran nsform [26] [2 27]. An exam mple of the per rformance of the wavelet a analysis is giv ven in Figure 5b, where eve en if the time e domain anal lysis of the sig gnal in Figure e 5a, in which IV = 109, do oes not reveal any evident si ingularities at t tIV, the wav velet analysis properly poi ints out at th his time the presence p of a very small s step variation n attributable to the passag ge of the pre essure wave r reflected by the t partially closed c in-line e valve. More e precisely, sc crutiny of Figu ure 5b reveals s the singulari ities occurring g at tV = 0.0156 s, tIV = 0.42 277 s, and tR = 0.9131 s. In n this case, the e distance LIV rom Equation (3) is equal to t 75.73 m, w with a relative e V calculated fr error equal t to 0.32 %. With W respect to time domain n analysis, using wavelet fu unctions not o only improves s the precision n in evaluatin ng times in Eq quation (3) and d consequentl ly in the localization of the e singularities, , but also furt thers the impa artiality of the pressure sign nal analysis. With regard to the sec cond requirem ment – i.e., to give information about the e anomaly – w within a short t period analy ysis the evalua ation of the re eflection coeff ficient can allo ow to size the e anomaly. Of f course, more e precise information can be b obtained within w a Inver rse Transient Analysis A (ITA A), as propose ed by Liggett t 28]. and Chen [2
Figure 5: (a) P Pressure signal of f Figure 2 in the fi irst characteristic time with
IV
= 109, and (b) corr responding wavellet analysis.
820
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
5. Generati ion of "safe" pressure wav ves As any te echnique, the implementati ion of transien nt tests in real systems may y encounter so some practical l problems tha at are discusse ed below. When de ealing with re eal pipe systems, accessing g and generat ting transients s are often ch hallenging. In n many cases, valves are installed in locations l in w which pressur re measureme ents are diffi icult or valve e lly for valves in large-diam meter pipes. Another A complication is that t many valves s response is slow, especial mit a sufficien ntly precise de etermination o of pressure su urges caused by b the maneu uver. Thus, in n do not perm many condit tions introduc cing transients s by means oth her than valve e manipulation n would be of f great benefit. . However, to o date, a port table and eas sy-to-use devi ice that does not entail ov verly sophistic cated support t instrumentat tion beyond th hat required fo or registering the pressure signal s has not t been availabl le. To fill this s need, the po ortable pressu ure wave mak ker (PPWM) d device (Figur re 6) has been n refined and d tested in the e Water Engi ineering Labo oratory in th he Departmen nt of Civil and a Environm mental Engine eering of the e University o of Perugia. Th he PPWM device (Figure 6 6) consists of f a steel vesse el filled with water and air r that can be placed under r pressure by means of a standard com mpressor. The PPWM and test pipe are e gth conduit wit th a small diam meter valve at a the end secti ion (hereafter referred to as s connected by a short leng valve). At the e beginning of o the test, aft ter disinfectin ng the device and refilling it with water r connection v from the tes st pipe, the ins side pressure is fixed larger r of a given amount a than th he pressure in nside the pipe. . Opening of the connectio on valve cause es a pressure w wave that trav vels along the e pipe detectin ng anomalies. . onnection valv ve can be actu uated very fas st and the pres ssure wave ge enerated by th he PPWM can n Since the co be controlle ed with reason nable precision n both in term ms of amplitu ude and stabili ity, by using s such a device e e weak point ts of transien nt test-based techniques ca an be eliminated. In fact,, it has been n most of the demonstrate ed [3] [29] [30 0] that the enti ity of the pres ssure wave can be fixed reliably – and th hen the risk of f unsafe overp pressure is con ntrolled – and d sharp pressur re waves can be b generated.
Figure 6: The P PPWM (Portable e Pressure Wave Maker) M device an nd the test pipe.
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
821
Acknowledgements This research has been supported by Fondazione Cassa Risparmio Perugia under the Project ‘‘Leaks and blockages detection techniques for reducing energy and natural resources wastage’’. References
[1] Brunone, B. Transient test-based technique for leak detection in outfall pipes. J Water Resour Plann Manage 1999;125:302-306. [2] Brunone, B., and Ferrante, M. Detecting leaks in pressurised pipes by means of transients. J. Hydraul. Res.2001;39:539-547. [3] Brunone, B., Ferrante, M., and Meniconi, S. Portable pressure wave-maker for leak detection and pipe system characterization. J. of AWWA 2008;100:108-116. [4] Ferrante, M., Brunone, B., and Meniconi, S. Wavelets for the analysis of transient pressure signals for leak detection. J. Hydraul. Eng. 2007;133:1274-1282. [5] Ferrante, M., Brunone, B., and Meniconi, S. Leak detection in branched pipe systems coupling wavelet analysis and a Lagrangian model. J. Water Supply: Research and Technology – AQUA 2009;58:95-106. [6] Ferrante, M., Brunone, B., and Meniconi, S. Leak-edge detection, J. Hydraul. Res. 2009;47:233-241. [7] Lee, P. J., Vítkovský, J. P., Lambert, M. F., Simpson, A. R., and Liggett, J. A. Frequency domain analysis for detecting pipelines leaks. J. Hydraul. Eng. 2005;131:596-604. [8] Meniconi, S., Brunone, B., Ferrante, M., and Massari, C. Numerical and experimental investigation of leaks in viscoelastic pressurized pipe flow. Drinking Water Engineering and Science 2013;6:11-16. [9] Mpesha, W., Gassman, S. L., and Chaudhry, M. H. Leak detection in pipes by frequency response method. J. Hydraul. Eng. 2001;127:134-147. [10] Stoianov, I., Karney, B. W., Covas, D., Maksimovic, C., and Graham, N. Wavelet processing of transient signals for pipeline leak location and quantification. Proc., Int. Conf. on Computing and Control for the Water Industry (CCWI), B. Coulbeck and J. Rance eds. 2001;1:65-76. [11] Wang, X. -J., Lambert, M. F., Simpson, A. R., Liggett, J. A., and Vitkovsky, J. P. Leak detection in pipelines using the damping of fluid transients. J. Hydraul. Eng. 2002;128:697-711. [12] Meniconi, S., Brunone, B., Ferrante, M., and Massari, C. Transient tests for locating and sizing illegal branches in pipe systems. J. of Hydroinformatics 2011;13:334-345. [13] Brunone, B., Ferrante, M., and Meniconi, S. Discussion of “Detection of partial blockage in single pipelines” by P. K., [1] Mohapatra, M. H., Chaudhry, A. A., Kassem, and J., Moloo. J. Hydraul. Eng. 2008;134:872-874. [14] Lee, P. J., Vítkovský, J. P., Lambert, M. F., Simpson, A. R., and Liggett, J. A. Discrete blockage detection in pipelines using the frequency response diagram: numerical study. J. Hydraul. Eng. 2008;134:658-663. [15] Meniconi, S., Brunone, B., and Ferrante, M. Water hammer pressure waves at cross-section changes in series in viscoelastic pipes. J. of Fluids and Structures 2012;33:44-58. [16] Mohapatra, P. K., Chaudhry, M. H., Kassem, A. A., and Moloo, J. Detection of a partial blockage in single pipelines. J. Hydraul. Eng.2006;132:200-206. [17] Mohapatra, P. K., Chaudhry, M. H., Kassem, A. A., and Moloo, J. Detection of partial blockages in a branched piping system by the frequency responce method. J. Fluids Eng. 2006;128:1106-1114. [18] Sattar, A. M., Chaudhry, M. H., and Kassem, A. A. Partial blockage detection in pipelines by frequency response method. J. Hydraul. Eng. 2008;134:76-89. [19] Meniconi, S., Brunone, B., Ferrante, M. and Massari, C. Potential of transient tests to diagnose real supply pipe systems: what can be done with a single extemporary test. J Water Resour Plann Manage 2011;137:238-241. [20] Wang, X. -J., Lambert, M. F., and Simpson, A. R. Detection and location of a partial blockage in a pipeline using damping of fluid transients. J Water Resour Plann Manage 2005;131:244-249. [21] Contractor, D. N. The reflection of waterhammer pressure waves from minor losses. J. Basic Eng.1965;445-451. [22] Meniconi, S., Brunone, B., and Ferrante, M. In-line pipe device checking by short period analysis of transient tests. J. Hydraul. Eng. 2011;137:713-722. [23] Meniconi, S., Brunone, B., Ferrante, M., and Massari, C. Transient hydrodynamics of in-line valves in viscoelastic pressurised pipes. Long period analysis. Experiments in Fluids 2012;53:265-275. [24] Ferrante, M., and Brunone, B. Pipe system diagnosis and leak detection by unsteady-state tests. 1 Harmonic analysis. Advances in Water Resources. 2003;26:95-105. [25] Ferrante, M., and Brunone, B. Pipe system diagnosis and leak detection by unsteady-state tests. 2 Wavelet analysis. Advances in Water Resources. 2003;26:107-116.
822
B. Brunone et al. / Procedia Environmental Sciences 19 (2013) 814 – 822
[26] Mallat, S. G., and Hwang, W. L. Singularity detection and processing with wavelets. IEEE Trans. on Information Theory 1992;38:617-643. [27] Mallat, S. G., and Zhong, S. Characterization of signals from multiscale edges. IEEE Trans. on Pattern Analysis and Machine Intelligence 1992;14:710-732. [28] Liggett, J. A., and Chen, L.C. Inverse transient analysis in pipe networks. J. Hydraul. Eng. 1994;120:934-955. [29] Meniconi, S., Brunone, B., Ferrante, M., and Massari, C. Small amplitude sharp pressure waves to diagnose pipe systems. Water Resources Management 2011;25:79-96. [30] Brunone, B, and Meniconi, S. Discussion of "Case studies of leak detection and location in water pipe systems by inverse transient analysis" by D. Covas and H. Ramos. J Water Resour Plann Manage 2013;139:126-127.
