Home

NeatWork: A user guide

1. Diameters Orifices Diameters list Nominal SDR Diameter Costim MaxPre Type Roughness 1 2 13 6 0 0182 0 25 76 8 0 0015 a 3j 17 0 0 0235 0 38 40 8 0 0015 1 17 0 0 0285 0 7 40 8 0 0015 1 260 0 0304 0 48 89 6 0 0015 114 170 0 0372 0 98 40 8 0 0015 114 26 0 0 0388 0 75 89 6 0 0015 gt Hardware databa x 114 325 0 0381 0 57 70 4 0 0015 se x 112 170 0 0426 1 28 40 8 0 0015 112 260 0 0446 1 03 89 6 0 0015 Diameters Orifices 112 325 0 0453 0 74 704 0 0015 E 112 41 0 0 0459 071 56 32 0 0015 Orifices list 2 s 17 0 0 0532 2 01 40 8 0 0015 20 26 0 0 0557 1 48 89 6 0 0015 t3 B 2 32 6 0 0566 1 28 70 4 0 0015 2 41 0 0 05739 0 89 56 32 0 0015 Diameter 21 2 17 0 0 06445 3 55 40 8 0 0015 21 2 260 0 0674 2 34 89 6 0 0015 0 0030 212 32 5 0 06856 1 81 70 4 0 0015 0 0040 a 17 0 0 0784 441 140 8 0 0015 3 260 0 082 3 35 89 6 1 0 0015 0 0050 3 325 0 0834 2 97 70 4 1 0 0015 0 0060 3 41 0 0 08456 251 56 32 1 0 0015 4 17 0 0 1008 754 1408 1 0 0015 7 m l i Chose Close The pipe database lists for each pipe wall thickness working pressure in meters of water price category PVC 1 others 2 These two categories arise because a special friction law is assumed for PVC category 1 which has a very small equivalent sand grain roughness For other category 2 pipe wall material the roughness needs to be
2. Hardware The window exhibits the current database the default one or the one you have edited 33 x Hardware Parameters Constraints Load Factors Diameters Orifices All None All None _ Nominal SDR Diam Cost Max Pres Diameter E 142 13 50 0182 0 25 176 8 lt E 0 0030 J 3 4 17 0 0 0235 0 38 140 8 0 0040 1 17000295 07 140 8 m 0 0050 1 26 0 0 0304 0 48 89 6 2 J 0 0060 11 4 17 0 0 0372 0 98 140 8 11 4 26 00 0389 0 75 89 6 114 32 60 0391 0 57 704 11 2 17 0 0 0426 1 28 140 8 11 2 26 0 0 04468 1 03 89 6 11 2 325 00453 074 70 4 11 2 410 00458 071 56 32 J 17 0 0 0532 8 Abort l Design This is where you specify which pipes and which orifices are in fact available to you The faster way to proceed is to click the box All and to un click the entries not available For this reason it is desirable not to load your data base with generally unavailable pipes and orifices Parameters This opens an 8 item display x Hardware Parameters Constraints Load Factors Qd Reset f Apply Jaj Start Design Parameters Fraction of open faucets 0 4 Service Quality 0 6 Target Flow l s 0 2 Limit on budget 1 0E9 Physical con
3. The output of the simulation module is presented as statistics derived from individually computed simulations The various items are Main statistics on faucet flows e Percentiles at faucets e Water velocities in pipe segments Node pressures 13 C Getting familiar with NeatWork 1 Introduction Installation procedure The file neatwork zip contains a windows version of NeatWork3 24 Unzip it into a new folder This folder will contain the following files e NeatWork324 jar executable java file e SOLVER DLL solver library mosek2 1l dllandpthreadvse d11 Mosek solver libraries and folders e db repository of databases projects repository of topographies designs and simulations The distribution also includes a templates folder containing An Excel topography template o The topography template is used as an expeditious way of entering a new topography in the NeatWork324 jar topography tables An Excel design template o The design template provides a rapid way of importing a design into the design tables of NeatWork324 jar design tables An Excel projects summary template o The projects summary template is a convenient accessory which suggests a standard terminology but of your choice for recording topographies designs and simulations and allows a rapid differentiation of the various topographies designs and simulations for a given project e AWord Template Guide to instruct in t
4. Begin End offaucets Suggested load factor Modified load factor A 93 3 A 3 3 93 B 2 53 2 53 2 06 2 06 2 06 2 06 1 59 R A A Cc B P2 C P4 B P1 1 0 1 0 2 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 P2 P2_a P2 P2_b P2 P2c C P3 P4 P4 a P4 P4b e 3 e 3 Sus wo EE ae Abort Design For the initial branch the load factor dropped to 3 18 You can edit the load factors individually in order to get a more user convenient design Pipe cost and uniformity of delivery It was noted earlier that there is generally a conflict between the requirements of minimum pipe cost on one hand and of a sufficiently uniform faucet output when the maximum number of faucets used is a fraction of the total number available Recall that the total friction loss from the tank to an open faucet is the sum of the friction losses in each segment on the path from the faucet to tank But this sum including the loss at the faucet itself is equal to the elevation differential Thus if the main part of the total friction occurs within the ultimate segment of pipe including the faucet the friction within the upward pipes must be negligible a flow variation in that upstream section will still keep the contribution of upstream section negligible In other words the total friction loss remains roughly unchanged and so is the flow at the faucet the closing or opening of a neighboring faucet will not influe
5. E diametersstd db 2KB DB File 27 06 2003 21 24 2db 1 20 04 il 01 2 46 Total File Size 1 63 KB Cupdatedb 20 04 File Folder 27 01 2004 16 46 orifices db P diameters db F gt 2 object s selected 1 63 KB My Computer Note that the folder db contains additional files and subfolders They contain information stored by the user e g alternate databases 19 However NeatWork recognizes only the two files diameters db and orifices db To make an alternate file e g diametersStd db the active database one must first save the current file diameters db ina separate folder or change its name and rename diametersStd db into diameters db The same is true for the orifice database Menu File Clicking on files opens a ten items menu NeatWork 3 24 File Database Help New Topography Open Topography Open Design Save Save as Delete a Topography Delete a Design Exit Submenu New topography You can create a new topography by Entering items nodes and segments within NeatWork Copying clipboard information provided it is put in an appropriate format Warning a topography must be a tree an arborescent network that is Aconnected network Anetwork without loops You can also create a new project by modifying an existing topography This is done using the submenu Open topography See the next section of the manual N
6. NeatWork A user guide A decision support program for the design of gravity water distribution networks Developed by Agua Para la Vida and LOGILAB for free distribution to non profit water supply agencies Agua Para la Vida http bldgsci07 ced berkeley edu aplv Logilab http logisun unige ch appli neatwork 1 Table of contents A Gravity water distribution networks ecce eee eee eee eene eee reete nenne e ee ttn nans aee ene 4 1 What is new about NentWOoFR ieesceeecousseesczens uunkusevabo tos k u Ee pRE ava p nna CRX eC E Fa uen SERE REES 4 For what kind of distribution networks eerie eee eee eese estne tn eaten sinensis tosta stats enata sin ens 4 The two components of a gravity driven distribution network erecta 4 The challenge of gravity driven distribution networks eeeeeeee eee eerte eene etn ernannt 5 Minimum pipe cost design eres eese eerte eese eese teen tenens enata stas tn senses stas tn senses stas ta senses sens tasco 5 Minimum pipe cost vs uniform faucet output 1 eerie eese eese eese e essen senes etat s etse tas e tas ena 6 2 Apud Para Ea VIda aet te art ple Eee NAI LUE Qu as ai iss eisso iro eise 6 EMI Peg TIE Ta PRA 7 B The essentials of NeutlWork i eeseieee eese setate se eEv ree tR e Ve rev Eee ee vp Ep Vae v ege evo Ep VE ev gu euh 8 1 Gravity Driven Distribution Net
7. 3 do not exceed 796 which translate to less than 3 596 for the flow rates of any pipe segment Typical deviations are far less especially for smooth pipes Now combining all the above we get A1 6 hy E a108 p 8 TL 0 086264 47 so that for a pipe of a given diameter and wall properties and for a given approximate water temperature one gets s Cg BE M 7 61 where a109 B 2 d9 B e d C 0 08626 A c d A1 8 The values for A and B are given by e d B A e d B A 0 0000 0 2180 0 2350 0 0042 0 0830 0 0800 0 0001 0 2100 0 2200 0 0045 0 0805 0 0795 0 0005 0 1780 0 1680 0 0048 0 0782 0 0786 0 0010 0 1430 0 1220 0 0051 0 0760 0 0782 0 0012 0 1350 0 1140 0 0060 0 0715 0 0780 0 0018 0 1140 0 0960 0 0062 0 0705 0 0785 0 0020 0 1085 0 0920 0 0080 0 0630 0 0766 0 0023 0 1032 0 0890 0 0085 0 0608 0 0760 0 0024 0 1015 0 0880 0 0100 0 0565 0 0750 0 0028 0 0958 0 0855 0 0104 0 0550 0 0747 0 0030 0 0942 0 0845 0 0150 0 0420 0 0730 0 0034 0 0903 0 0832 0 0200 0 0350 0 0740 0 0036 0 0885 0 0828 0 0400 0 0270 0 0910 0 0040 0 0850 0 0815 From these values one can use linear interpolation formulae which are adequate to calculate A and B for any value of zd within the range presented This is what Neatwork does For smooth pipes like PVC roughness factors are qui
8. B 150 6 31 0 0 c 20 0 12 27 0 0 P2 12 0 9 95 0 0 P4 19 0 16 97 0 0 P1 9 0 6 89 0 0 e P2 a 12 0 10 0 0 0 P2_b 12 0 10 0 0 0 P2t 12 0 10 0 0 0 P3 18 0 12 7 0 0 P4 a 19 0 16 92 0 0 53 Pah 190 1697 nny Close where you can visualize the anticipated pressure loss at any node under normal operating conditions This pressure loss should be less than the negative of the node height remember that the height is relative to the tank elevation the highest point in the network If the pressure loss is too close or higher than the negative of the height the last column reports a potential problem After clicking on close a new window Save the design either via the menu Save NeatWork will create a new file with the same name as the topography file but with the extension dsg or via the menu Save as in that case you will be asked to give a name to the file Checking performance Remember that because the identity of the open faucets is not known unless they are all open at the same time the design has not satisfied the equations for the flow except in a statistical sense You need to verify both 46 e that the range of flow rates for all the faucets is acceptable by your criterion e that the design is not more expensive than necessary This requires that you use the simulation module which uses the design you have just obtained or any other design and applies the equations that th
9. dependent faucets and a probability r 0 4 per faucet to be open Open faucets 1 2 3 4 5 6 Flow in the branch 26 39 4 56 66 Conditional probability 0 1958 0 3263 0 2900 0 1450 0 0387 0 0043 Cumulative Probability 0 1958 0 5220 0 8120 0 9570 0 9957 1 Which flow should be selected in the design process The conditional mean of the distribution 0 198 x 0 3263 x 2 is a sensible choice In the above example this would give 2 5174 However since a design based on that choice will not necessarily perform well under any circumstances in the search for a good design one may wish to select other values To guide the choice of trial values we have introduced the concept of Quality of service To illustrate the concept suppose that the design flow is set to 39 in the above example Looking at the table displaying the conditional probability distribution we realize that the value 364 is large enough to cover flow demand 81 of the time We say that the quality of service of 39 is 0 81 Since we may want to choose a flow between two discrete values say between 26 and 34 we define the quality of service by linear interpolation Reciprocally if we impose a quality of service 0 7 the corresponding flow will be 0 7 0 522 2d p 2 614 20 0821 0 522 4019 We name oad factor the coefficient 2 61 that appears in the right hand side of the equation The service quality is a user free parameter in NeatWork When
10. or initiate a new design and save it The menu shows up only when the corresponding window tab is activated File Design Database Help Tal Extract topography Nod Report in HTML f Design parameters Design Load Factors TTETNET 37 Extract topography This action retrieves the topography on which the design is based lt opens a corresponding window Report in HTML With this submenu you can generate a report of the design information in HTML format We briefly review the items of this file when opened in Excel The Node List looks like this ID Height X Y Ideal Orifice Commercial Orifice Nature R 0 010 0 0 TANK A 1010 10 0 0 BRANCHING NODE B 15 0 10 0 0 BRANCHING NODE C 20 0 0 0 0 BRANCHING NODE P2 1210 10 0 0 BRANCHING NODE P4 19 0 0 0 0 BRANCHING NODE PI 9 0 0 0 01451968 10000 FAUCET NODE P2_a 12 0 O 0 07377202 10000 FAUCET NODE P2_b 12 0 O 0 07377202 10000 FAUCET NODE P2_c 12 0 O 0 07377202 10000 FAUCET NODE P3 18 010 0 0061899 0 006 FAUCET NODE P4 a 19 0 O 0 01561877 10000 FAUCET NODE P4_b 19 0 O 0 01561877 10000 FAUCET NODE The arc list is as follows Begin End Length Length1 Diam1 Length2 Diam2 R A 400 400 D001 0 0 A B 50 50 D002 0 0 A C 250 250 D002 0 0 B P2 80 68 D0
11. stands with multiple faucets None of the faucets have been equipped with orifices For columns 2 to 6 the fraction of open faucets remains 0 4 a 2 3 4 5 6 T 8 1 83E 07 0 018 8 885 34 69 178 0 337 0 126 6E 08 0 054 8 922 26 34 162 0 323 0 127 1 83E 08 0 18 9 082 1796 20 148 0 307 0 130 6E 09 0 536 10 323 7 796 115 0 238 0 134 3 5E 09 0 918 15 034 6 5 36 0 195 0 153 54 The figures in the columns mean Column a Faucet head loss coefficient The lower its value the higher the head loss 2 Faucet loss at target flow as of minimum faucet elevation differential 3 Cost of network pipes 4 Maximum root mean square variation of flow rate at a single faucet 5 Maximum variation of local average as of target flow 6 Difference between flow at one of 5 faucets sharing same terminal pipe when all faucets are open and when only one is open as of target flow 7 Average over all faucets of flow when only one faucet is open 8 Average flow over all faucets when the fraction of open faucets is raised from 0 4 to 0 6 Faucet Constant The usual distribution network is required to deliver water with substantial pressure Next to that pressure the head loss through ordinary faucets is normally a negligible quantity But unaided gravity networks as a rule deliver water through faucets whose outlet is at atm
12. Quality The first two parameters are easily understood The obvious relation links them to the maximum flow rate out of the tank during peak hours maximum flow rate target faucet flow x fraction of open faucets The third parameter reflects the use of a statistical logic to avoid over sizing while respecting the supply norms The motivation is rather straightforward the initial conduit out of the tank has to handle only the desired maximum flow rate as defined above while the ultimate conduit leading to an individual faucet must provide the full desired flow rate when that faucet is open For intermediate segments of pipes between branching nodes the design must provide for a flow rate capacity that will depend on the average number of faucets supplied by that segment In order to accommodate idiosyncrasies of particular configurations the fairing between the two end values is made part of a one parameter design curve The higher the value of this parameter the more flow is allowed in the intermediate segments and so in general the more abundant the supply at all faucets the less frequent the occurrence of abnormal flow rates at any faucet but the more expensive the system This third parameter is called the Service Quality Its final value can only be chosen after probing with the simulation module Thus the Service Quality might be called a global tuning tool Once it is chosen together with the other two parameters the
13. an existing network The basic condition is that the expanded network be a tree Follow these steps 1 Open or create the topography file of the expanded network 2 Activate the Start Design menu 3 For each existing segment arc use the submenu Constraints to enter the appropriate data lengths and internal diameters of the pipes in place Remember that a segment uses at most two types of pipes and most commonly one 4 Click on Run Start Design 5 NeatWork computes the total material cost including that of the already existing pipes Adding a new arc segment in a tree arborescent network creates a loop A loop makes the network more robust in case of pipe failure However it is hard to predict its impact on the flows The flow variation at faucets served by the loop is not always decreased On the other hand the pipe cost is always increased 49 To create a loop just start with an existing design and add appropriate pipes between pairs of nodes You must specify the length of the pipes As yet NeatWork does not offer graphical information to estimate distances between unconnected nodes You must gather the information from an extraneous source Note that if loops are added in the design no schematic will be available for that modification Remember that one cannot optimize the design when the topography has a loop one cannot even save a topography with a loop NeatWork automatically includes the cost of the
14. are contemplated though they are not used in the optimizing design module However a nominal entry for these coordinates e g 0 is required b Aset of design parameters specifying faucet target flow fraction of open faucets and service quality A rough approximation to the spring water temperature should also be specified C Two parameter files one for pipes available internal diameters maximum working pressure cost and equivalent Moody roughness factor or nature of pipe material and one for orifices available diameters In addition the type of faucets used is specified through a faucet loss constant This unfamiliar constant is important enough to justify an elaboration found in the appendix d Specification of constraints if any such as specified preexisting pipe diameters along some segments Output The output of the design module is a design file that reproduces the topographic data and recommends pipe diameters for each segment and orifice diameters for some terminal segments Typically the design module will choose one type of pipe for the entire length of a segment Occasionally it will suggest two diameters for the same segment with two complementary lengths In that case the length of the longer section is chosen to be an integral multiple of the commercial length of pipe units The length of the shorter section is adjusted to add up to the total length of the segment This rule minimizes the number of
15. current folder BaseExample in the table above x Project Path C Documents and Settings jpv My Documents Neatwork323 projec Browse Your topographies BaseExample Filename NewTopo New Cancel You see that the folder BaseExample contains only one topography file namely BaseExample topo Give a name to the new topography and 23 click on New The folder BaseExample will then contain 2 topography files BaseExample topo and NewTopo topo See suggestions on nomenclature in Project summary template Filling data in new topography There are a number of ways to fill data in the node and segment tables We review the most convenient ones Entering data directly in the table If you choose to create a new topography give a name to it and click on the button New The following window pops up Bi xj File Topography Database Help Tables Tree View Text Node list AS Oh av ID Height XP SENE of faucets Nature Source 0 0 0 0 0 0 0 4 Z Z gt ay Arc list AS OR av Begin End Length Q Welcome The upper table Node list will receive the information relative to the nodes which include the root of the tree initial tank the branching nodes with no faucets and the terminals faucets The lower table lists the segments between the nodes Fill the node table first and the seg
16. dissipation can be represented by the formula oi Ej j m where j is a terminal node equipped with a faucet and i j is the arc leading to this terminal node 3 Potential energy It applies to the flows at the terminal node according to the formula Ej 3 hy dij where h j lt 0 is the node elevation A similar term exists at the source node but the elevation there is zero The total energy to be minimized is thus E gt gt Eliit X Eii Y EX i j EA ud jeN 1 2 The minimization of E is constrained by the law of conservation of masses actually flows and by nonnegativity constraints on the faucet flows We thus have for each non terminal node Qij Oj Vj E Me i j EA j i EA and for each terminal node Pij 20 VIE Ny The formulation allows for multiple sources but this feature is not exploited in NeatWork 68 The problem is convex Mosek a state of the art commercial optimization code for convex programming solves it For readers aware of the mathematical theory of optimization we mention that the first order optimality conditions of the energy minimization problem are just the standard condition that along each path from the source to the faucet the total head loss exactly equals the variation of elevation The minimization formulation is thus an integral version of the familiar condition on flows in a network 5 Optimal design The optimal design module deals with networks
17. in the design module or once a design has been produced by the optimizing module by altering either the design i e some specific pipe or orifice diameters in the neighborhood of deficient faucet or the topography i e by introducing loops and gauging the results of the modifications with new simulations The process may be repeated until the answer is judged optimal One may achieve a satisfactory solution by merely finding through successive simulations the lowest value of service quality consistent with the desired norms of faucet variability It must be pointed out that one may often reach a still lower cost by means of the local manual tuning just mentioned at a somewhat lower value of the service quality parameter These ultimate steps of the design are not automatic The manual tuning requires more skill and experience on the part of the designer though the scope of his search and most often the net economy resulting from it have been much reduced 11 Input and output in the two modules Design module Inputs The design module requires as inputs a The topography of the network i e the origin branching sequence and all end nodes of the network with the elevation of all nodes and the length of all segments between them The number of faucets at each terminal node must also be specified There is an entry for plan form coordinates of the nodes North South and East West These data can be helpful especially if loops
18. named trees that is networks without loop In that case the required flows along the intermediary arcs are uniquely defined by the desired flows at the terminal nodes The design issue is to find which pipes should be used to achieve the head loss associated with the given flows Since the choice of pipes must be made among a finite collection of commercially available pipes the problem boils down to determining the lengths of the various pipes to be used on each arc Since the head loss is proportional the pipe length the whole problem is linear and thus belong to the category of easily solvable linear programming problems This module is also based on Mosek 69
19. pipes on the added segments If one cannot optimize in presence of a loop one can still test the performance of designs with loops with the simulator Limit on budget This feature enables the search of a design with good performance at a limited cost Experience will show whether this feature is useful in practice Service quality and load factor The main issue in making a design is to account for the random process of open close faucets It would be relatively easy to conceive a least cost design for a network in which the faucets are all in the open state To understand the concept of load factor and service quality let us start with the extreme case where the design should meet the requirement that all faucets are simultaneously open It is easy to determine the necessary flow in each branch suppose that there are n faucets downstream that branch and let be the target flow at each faucet then the flow in the branch is n times Given the flow in a branch the friction loss in the branch is a direct function of the diameter of the pipe The problem of finding a least cost design is the one of selecting appropriate pipes so that the total friction loss along the paths from the reservoir to each faucet be equal the loss in potential energy proportional to the drop in vertical elevation minus the friction loss at the faucet The operational conditions are not so simple It is never the case that all faucets are simultaneously open
20. required pipe cuts Terminal nodes involve one or more faucets Private faucets are usually single Multiple faucets are often found on public water stands In NeatWork multiple faucets belonging to the same terminal node P are separately listed as P_1 P_2 P_3 etc For orifices the design recommends two sets of diameters ideal diameters the computed choice and the available diameter resulting in the closest flow Simulation module Input The simulation module requires as an input a Adesign which may be the output of the optimized design module or a modified version of this design or any other design including designs with loops in which the same quantities are specified b The type of orifices to be used ideal vs available C The type of simulation mode The normal mode corresponds to a random selection of closed and open faucets in accordance with the fraction of open faucets that does not have to be the same fraction as the one used in the optimizing module The random selection is done by Monte Carlo sampling The faucet by faucet mode computes the flow out a faucet when this faucet alone is open The computation is carried out for each faucet The custom mode allows the user to specify the set of open and closed faucets d For the records and an easier analysis of the simulation output the user may freely set values for the target flow rates and the minimum and maximum flow rates Output
21. should provide This is not necessarily the average flow obtained in the simulation Limit of the budget If your budget is severely constrained you can impose a constraint on the cost of the design In that case NeatWork will propose a design that will meet that constraint and that will satisfy the operational constraints as closely as possible Naturally the performance will suffer Water temperature This determines the water viscosity which affects the pipe head loss In practice it is enough to use a single value for any given region But the difference between widely different climactic regions e g Nepal and Nicaragua has an appreciable effect on the head losses Pipe commercial length The optimizing module occasionally splits a segment in two assigning a different diameter to the two sub segments In that case one of the two always has a length that is an integral number of pipe lengths Orifice coefficient It is the constant that determines the relation between orifice head loss flow rate through it and orifice diameter In principle that coefficient should be fixed in so far as the ratio of the orifice diameter to the diaphragm diameter is large But the geometry of the small hole may affect that 35 constant Also the designer may trust his value of the constant more than someone else s and the data is still scant The defect value provided should be reasonably close Faucet coefficient This is an important p
22. specified See Technical Appendix C for a discussion of roughness in the friction laws used by NeatWork Warning The columns Nominal wall thickness and SDR correspond to the commercial identification of a pipe The pipe identification in the NeatWork engine relies on the inner diameters exclusively third column The user must make sure that no two pipes have the same inner diameter Were it not the case we suggest that the user introduce a small perturbation in one diameter to make sure that NeatWork correctly identifies the pipe The principle for the orifice database is the same but the database is much simpler an orifice is a diaphragm of fixed diameter characterized by the diameter of a small hole located at its center Editing the databases For convenience NeatWork saves in a local database a set of standard pipes as well as a set of orifices with their characteristics Now the databases supplied with your NeatWork program may not be the one that you want to or can work with The database can be edited using the following buttons 18 Last Insert a new item amp J Delete an item D Copy the database into the clipboard The content of the clipboard may be pasted as a table in Microsoft Excel or Word B Paste the clipboard to produce the new database The clipboard must contain a table with suitable format amp j Load data Reloads the current database j Save the displayed data into
23. the database A Save the current database not the displayed data in html Convenient for exchange of data between users The default diameter database applies to PVC pipes classified according to SDR specifications Suppose you want to work with polyethylene pipes rather than PVC or you may have access to PVC pipes classified according to SAE rather than SDR specifications In that case you must first correct or replace the database supplied You can edit the supplied database using the buttons displayed in the above table Do not forget to validate your changes using the save button 5 Your final database should include all the pipes and orifices you may want to use and have some possibility of obtaining Later before initiating a design you will be asked to select from the database the subset of pipes that are available for the project under study But additions to the database are necessarily introduced here Storing databases The databases are stored in the files diameters db and orifices db that are contained in the subfolder db CO o x File Edit View Favorites Tools Help EI Back gt seach Folders cC Histoy 23 OF X wy Eg Address C db Go p cu Name Size Type Modified Li i diameters db 2KB DB File 02 08 2004 12 23 iIlorifices db 1KB DBFile 02 08 2004 12 20 E diametersNew db 2KB DBFile 23 01 2004 16 19 E diametersRin db 2KB DB File 13 07 2003 16 56 2 items selected
24. will reveal the format of the tables in the Excel sheet Any deviation from this makes it impossible to paste an Excel table into NeatWork In the simple three node example press the button D that stands above the node list open an Excel sheet and use the paste function to get the A B C D E E 1 TANK 0 D 2 BNode 20 D 0 NA 3 FNode 10 1 Note that the nature of the node is represented by the figures O 1 or 2 and not anymore by the words TANK BRANCHING NODE FAUCET NODE We can proceed analogously with the segment list A B c D E I 1 TANK D 0 D 2 BNode 20 0 D 1 3 ENode l 10 0 1 2 4 5 TANK BNode 200 B BNode FNode 50 m The copy and paste functions can be used to move data from an Excel sheet to a new topography file in NeatWork Copy the node list first and paste it in NeatWork using the button B and then the segment list For convenience an Excel topography template is included into the NeatWork distribution Warning e The format of the Excel tables must be exactly the same as the ones displayed above In particular the nature of the nodes must be specified by the numbers 0 1 or 2 However one can use any name for the nodes 26 Some horizontal coordinates X and Y need to be entered in the nodes tables even though NeatWork does not use them e The Excel table must correspond to a tree arborescent network rooted at the tank and whose leafs are the termi
25. 0 0 0 01451968 FAUCET NODE P2_a 12 0 0 0 0 0 0 07377202 FAUCET NODE P2_b 12 0 0 0 0 0 0 07377202 z FAUCET NODE P2c 12 0 0 0 0 0 0 07377202 FAUCET NODE P3 18 0 0 0 0 0 0 0061898 0 006 FAUCET NODE P4 a 19 0 0 0 0 0 0 01561877 z FAUCET NODE Arc segment list AS f ev Begin End Length Lengh i1 Diam1 Length2 Diam 2 R A 400 0 400 0 0 0391 0 0 A B 50 0 50 0 0 0235 0 0 7 A C 250 0 250 0 0 0235 0 0 B P2 80 0 68 0 0 0235 12 0 0 018 C P4 110 0 66 0 0 0235 44 0 0 018 B P1 12 0 12 0 0 0182 0 0 P2 P2 a 1 0 1 0 0 0182 0 0 P2 P2 b 1 0 1 0 0 0182 0 0 P2 P2 t 1 0 1 0 0 0182 0 0 C P3 8 0 9 0 0 0182 0 0 P4 1 0 1 0 0 0182 2 BLES GEESE Design total cost 413 BaseExample BaseExample The node and segment design tables are similar to the topography tables but with extra rows and columns 1 Inthe a Node list The column of faucets is replaced by two columns The ideal orifice is the one that NeatWork computes to meet as closely as possible the specifications In the column commercial orifice NeatWork display the diameter of the orifice selected from the user database which gives the best approximation to the ideal orifice Note that orifices larger than 0 00737 are replaced by no orifice The orifice with the largest diameter 0 006 would induce more distortion than no orifice The rows associated with the terminal nodes supporting multiple faucets are replicated a num
26. 02 12 D003 C P4 110 66 D002 44 D003 B PI 12 12 D003 0 0 P2 P2_a 1 1 DOO3 0 0 P2 P2 b 1 1 DOO3 0 0 P2 P2c 1 1 DOO3 0 0 C P3 9 9 D003 0 0 P4 P4 a 1 1 DOO3 0 0 P4 P4 b 1 1 DOO3 0 0 38 Note that the pipes are identified by D001 D002 etc The program assigns these identifiers The next table shows the correspondence between the new identifier Ref and the type of pipe A Internal Unit Max Rough Total Total Ber Nominal SDR Diameter cost Pressure Type net length cost D001 1 1 4 32 5 0 0391 0 57 70 4 1 0 0015 400 228 D002 3 4 17 0 0235 0 38 140 8 1 0 0015 434 164 92 D003 1 2 13 5 0 0182 0 25 176 8 1 0 0015 82 20 5 The summary section provides the rest of the information Project cost 413 Global parameter settings Water Temperature _C 20 0 Pipe Commercial lengths m 6 0 Design Parameter Fraction of open faucets 0 4 Service Quality 0 6 Target Flow l s 0 2 Limit on budget 1 0E9 Advanced parameter settings Orifice Coefficient 0 59 Faucet Coefficient 2 0E 8 Structure Number of Nodes 13 Branching nodes 5 Faucet nodes 7 with 12 individual faucets Total height change 20 m Number of pipes 12 Total length 916 m Design parameters This menu opens a window that summarizes the parameters used in initiating the design x Parameters used by the MAKE DESIGN operation for building this des
27. 4 f Ro A1 11 for smooth pipes and Re lt 100 00 and the Swamee Jain approximation take the Reynolds number explicitly into account This means not only that the others cannot be used for fluids other than water but also that the temperature variations which say between Nepal and Nicaragua would lead to nearly 1496 variations in head loss for smooth pipes are neglected Their fit to the Moody data usually also require ad hoc choices of a disposable constant On the other hand the Swamee Jain formula f 1 325 5 74 V A 12 log 55 84 which is explicit is somewhat better than our own approximations but it is not easily integrated either and so it is awkward to use in our program It is useful however as a first approximation in the evaluation of f by the Colebrook formula which is used for comparison purposes In Summary NeatWork does not offer a choice of alternate formulas for the calculation of head loss by friction because it considers that nothing is gained by this freedom For smooth pipes these losses are on a Solid experimental footing and for rough pipes these alternate formulations only hide what is a physical indeterminacy 63 Introduction and definition The usual distribution network is required to deliver water with substantial pressure Next to that pressure the head loss through ordinary faucets is normally a negligible quantity But unaided gravity networks as a rule deliver water through faucets whose outlet
28. FNode 10 0 0 0 0 0 1 FAUCET NODE Note that NeatWork changed the nature of the two new nodes This is a temporary assignment until the segment table is completed Filling the segment list The segment table has three fields one for each extremity begin and end of the segment and one for the length of the segment When the table is completed validate it with 9 Node list AS HR av ID Height X y sotfaucets Nature j TANK 0 0 0 0 0 0 D TANK BNode 20 0 0 0 0 0 0 BRANCHING NODE FNode 10 0 0 0 0 0 1 FAUCET NODE Arc list AS hh av End Begin Length BNode TANK 200 0 FNode BNode 50 0 Note that NeatWork has now assigned the correct nature to each node Warning The following rules must be obeyed 25 A segment is an oriented arc The beginning extremity is on the tank side while the end extremity is on the faucet side One must not put a faucet at the tank node or at a branching node Failure to observe the rules may lead to inconsistent results Entering data via the Excel template It must be admitted that the above method is cumbersome and slow A better method is to generate the data first on an Excel sheet and using the copy and paste features The format in the Excel table must be exactly the one displayed in the Excel template Before using the Excel template we suggest you get familiar with the copy and paste functionalities between NeatWork and Excel The proposed exercise
29. In the peak hour one can expect that at most a fraction r of the total number of faucets is open This fraction is interpreted as the probability that a faucet be open at the critical peak hour which makes the open close process a random phenomenon It follows that the flow in each branch is also random To get around that difficulty and to produce a design we have to replace the random flow in the branch by some typical value The mean value is a reasonable choice when there are many faucets downstream the branch But when there are few faucets this choice is inappropriate Take the extreme case of the pipe ending at a faucet If the faucet is open the flow is if the faucet is closed the flow is zero In the latter case there is no 50 friction loss for any pipe Therefore the proper flow to consider is and not the mean r Similarly when there are two dependent faucets there are three possible cases 0 1 or 2 open faucets with probabilities 1 r 2r 1 r and r and flows in the branch 0 and 24 By the same token as in the single faucet case a zero flow is not relevant for the design process and we must consider that at least one faucet is open The probabilities that one or two faucets are open given that at least one faucet is open are the conditional probabilities 2 1 r 2 r and r 2 r To further illustrate the point we display the flow conditional probability distribution in a branch with 6
30. anned almost entirely by Nicaraguan residents has multiple functions It undertakes drinking water projects at the request of villages It endeavors to protect the soils which allow the replenishment of the springs on which the water supply depends It encourages hygiene education and especially babies and infant health monitoring both destined to maximize the benefits of clean water The center also facilitates its own duplication by running a small technical school which forms through a two to three year work study curriculum water projects technicians from among selected rural residents Technical material especially appropriate for use by technicians and engineers engaged anywhere in similar activities as APLV s pilot center is developed by a few engineers and scientists who work in Europe and the United States as well as Nicaragua on a voluntary basis The Nicaragua operation serves as a field trial center both for appropriate design techniques and for special pedagogical tools for the transfer of these techniques but the technical material which results in written form and in computer programs is freely offered to other N G Os as well NeatWork is the second version of a distribution network design tool that has been used and tested by Agua Para La Vida in a large number of gravity flow systems 3 Acknowledgements Conception G Corcos amp J P Vial Implementation and interface F Babonneau Laurent Drouet Optimization t
31. arameter when the faucet exit is at atmospheric pressure the usual case Its value depends on the type of faucets used See Appendix to determine this coefficient if you don t know it The program assumes that all faucets are the same for a given distribution network the appendix indicates how to proceed if that is not the case Constraints If the pipe diameters on some segments have already been fixed as in the case of an extension of an existing distribution network you may want the program to conform to this choice This can be used for instance if one modifies an existing network Select first the segment concerned in the menu and second select the type of constraint you want Hardware Parameters Constraints LoadFactors x Sera Hardware l Parameters Constraints Load Factors v Constraint on Arc R gt A diameter Constraint on Arc R gt A Arc R A Arc A gt B Arc A gt C Arc B gt P41 vi diameter must be greater than Y diameter must be greater than Y diameter must be greater than ldiameter must be lower than diameter must be equal to e two diameters must be diameter Arc B gt P2 Arc C P3 Arc C Pa rc P2 P2 a v ipd Abort Design Add constraint Delete add constraint Delete Design The second menu offers four options In the first three menu items the sam
32. available to you The faster way to proceed is to click the box all and to un click the entries not available There are two types of parameters In the first category you find items you may want to vary in a search for a good design Fraction of open faucets Target flow Service Quality and Limit on budget In the second category you find parameters that generally hold for all projects in a certain area 45 Water temperature Commercial length of pipes Faucet coefficient orifice coefficient You probably want to fix them once and for all There are several strategies for giving good initial values to the design parameters In principle Fraction of open faucets and Target flow should be given the value corresponding to your objectives In that case Service Quality can reasonably be fixed at 0 6 or 0 65 Alternatively one can start with a fairly large value 0 75 and use a target flow rate equal to or slightly smaller than the desired average flow rate A large Service Quality tends to increase the flows In general you do not want to activate the tab Constraints and Load Factors when you create a first design Once you have completed your choice of conditions imposed on the design click on Start Design Prior to delivering the full information on the design NeatWork opens the window Node ID Height Forecast Pressure Loss Potential Problem R 0 0 0 0 0 0 a A 10 0 3 8 0 0
33. ave to try a new design using the optimizing design module or modify the design you obtained without the help of the optimizer There are several avenues to reach an optimal design They are enumerated below To start with you may either a be satisfied with the performance of your design but wonder if you could reach a satisfactory one at lower cost or b be dissatisfied with the performance typically because some faucets have either insufficient or excessive flows Your options are similar in these two cases They are not necessarily in any suggested order 1 Increase or decrease the quality of service number and repeat the design phase This parameter is discussed in more details in the 47 section on advanced features In short the higher the number the more generous is the assignment of all diameters in between the first tank outlet section and the ultimate sections leading to the faucets This will tend to increase the flow rates thru all the faucets though not uniformly If you keep the same open faucet fraction you will need therefore in the same time to decrease the target flow rate Repeat the design phase after having modified the table of load factors This allows you to bias the assignment of diameters in the region where you believe a change is suggested Improve manually The following options are modifications to an existing design which by pass the optimizing module but are tested by simulations 1 In the s
34. ber of times 29 corresponding to the number of faucets associated with them For instance the row corresponding to P2 is replicated 3 times we thus have three new identical rows with ID s P2 a P2 b and P2 c 2 In the segment arc list a Four extra columns give information on which pipes are to be used on a segment In general only one type of pipe suffices but occasionally the arc segment is split into two sub segments with different pipe types Pipes are identified by their inner diameter see the warning statement in the section on the databases b Each terminal node with multiple faucets gives rise to as many new rows segment as there are faucets For instance we have now three additional arcs emerging at P2 and ending at P2 a P2 b and P2 c respectively These arcs conventionally have a one meter length This length is small enough not to induce any significant pressure loss for typical faucet flows Note the two tabs at the bottom of the window They allow switching from the topography and design files that are currently open Note also that the topography tab is marked with a small blue disk while the design tab has a yellow disk Note also the four tabs on top Three of them are already present in the topography window Table Tree View and Text the fourth one Simulation give access to the simulation function This will be detailed in a further chapter Submenus Save and Save as Th
35. ce is ideal or commercial Commercial is chosen to see how the real design would behave It is helpful to choose ideal when you want to find out whether an unsatisfactory performance is primarily due to an The number of combinations of 8 faucets open in a network with 20 faucets is 125970 41 insufficient choice of orifices i e if an additional orifice of a different size might improve the performance markedly Type of simulation This submenu offers three choices Monte Carlo sampling the usual kind individual faucets one faucet at a time and user defined Type of simulation monte carlo sampling x If you choose user defined click on Run simulation and you will be asked to select which faucets will be open by clicking on appropriate boxes xl All None Faucet vi P1 j P2_a vi P2 hb vi P2 c v P3 CJ P4_a You may choose any combination of open and closed faucets Simulation output Clicking the Run simulation button starts the simulation sequence When it is completed four tables resume the results of the simulations They are Flows at Faucets Percentiles at faucets Speed in pipes and Nodes pressures Flows at faucets The flows at faucets table has six columns The first identifies faucets Here note that if an end node is equipped with multiple faucets the table will list separately the statistics of flow when one two three faucets are o
36. dently requires knowledge of the horizontal coordinates of the nodes One may introduce one or more loops for other reasons as well such as to limit the vulnerability of the network to breaks or major leaks in parts of the network In summary the last steps in the design will be a matter of personal choice from the above possibilities 48 2 Advanced features File system Once you have created and saved a topography under a certain name say mytopo tpo in a certain folder say myproject NeatWork will proceed as follows with the subsequent design and simulation files All files will be stored in the same folder as the original topography file unless specified otherwise by the user When NeatWork creates a design out of a topography it will use the name of the topography and add the proper extension dsg We thus have a new file mytopo dsg Further designs originated from this same topography file will be named by NeatWork mytopo 2 dsg mytopo 3 dsg etc unless specified otherwise by the user Similarly the successive simulation runs from a design file say mytopo 2 dsg wil be named by mytopo 2 0 sim mytopo 2 1 sim mytopo 2 3 sim etc NeatWork does not offer an opportunity to give alternative names to the simulation files Of course it is always possible to change the names directly in the Windows operating system Expand a design Loops NeatWork makes it possible to optimize the design of an extension of
37. e flow obeys in all the branches for a sufficient number of alternate random choices of open faucets After having initiated the design and run the simulation corresponding to that design you note the resulting average flow rate and observe the flow variations Service quality If the flow variations are within your acceptable bounds you then lower the service quality and anticipate its effect on the flow rate by adjusting the target flow rate accordingly These changes lead to a new design for which you run a simulation You proceed gradually repeating these steps until the flow variations given by the simulations fall just within your acceptable variation norms the average flow rate being close to the desired one Alternatively you may choose to lower the service quality further so that a relatively small number of faucets exhibit unacceptable flow variations and even at times run dry and you use local adjustments discussed below to correct these deficiencies Target flow The relevant flow rate for the design is the resulting average flow rate calculated by the simulation The target flow rate is the flow rate that the optimizing design seeks to approach While these two may be relatively close the target flow should be treated as a parameter to be adjusted in order to home in on the proper average flow in the simulation Improving the design Use of the Start Design tool Depending on the outcome of the simulation you may h
38. e list in the design file and paste it in an Excel worksheet Edit the column of heights copy the modified table and paste it into NeatWork Don t forget to save the modified design presumably under a different name before attempting to run a simulation The change may be made easier in future versions 57 Question entered the node list and the arc list into a new topography window saved it and activated the Start Design feature NeatWork computes a design reports on Node and Faucet Pressure but does not create the design tables Answer Make sure you have not inserted a faucet at a branching node Faucets can be added at terminal nodes only Question In the design produced by NeatWork a terminal node gets the status of Branching Node Is this an error Answer f the terminal node has multiple faucets NeatWork adds a new node for each faucet in the lot links these node to the previous terminal node and makes the latter a branching node Question My correspondent sent me a design file under the html format How can create the design file in NeatWork to run simulations Answer Open the simulation file with Excel You must first edit the file to make it compatible with NeatWork n Node List convert TANK to 0 BRANCHING NODE to 1 and FAUCET NODE to 2 When the operation is completed check whether the Diameter database of your correspondent is the same as yours If not modify your ow
39. e submenu Save saves the topography under the current name See suggestions on nomenclature in Project summary template The submenu Save as allows you to save the topography under a different name and in a different folder You can create a new project subfolder if you wish Submenus Close and Close all Those submenus let the user close the open windows at will Submenus Delete topography and Delete design Those submenus let the user delete any topography or design file If you delete the file that corresponds to an open topography or design the file is erased but the information remains unsaved in the window Using the save or save as menus will create the file again 30 Menu Topography This menu shows up only when a topography file is open and the corresponding window tab is activated Submenu Quick check Quick check verifies that there is enough head to equal the head loss due only to the faucets at a specified flow rate To use it you need to specify the desired flow rate out of the faucets and the faucet constant that defines faucet losses as a function of flow rate Once these two quantities are entered the table indicates which if any of the faucets might fail a with a chosen minimum flow b with the intended or target flow Clearly this constraint is necessary but not sufficient since additional pipe losses are inevitable unless the pipe diameters are unlimited In the exa
40. e type of pipe is used on the segment with a diameter greater than lower than or equal to most useful option a diameter that you chose among the list of available diameters See figure below The last option allows the use of two pipe diameters in series for the segment BEG storceson x Hardware Constraint on arc R gt A diameter 1 diameter 2 Parameters Constraints Load Factors v the two diameters musthe diam 0 0182 sdr 13 5 PVC Y length 1 tiam 0 0182 sdr 13 54PVC diam 0 0235 sdr 17 0 PVC diam 0 0295 sdr 17 0 PVC diam 0 0304 sdr 26 0 PVC Add constraint Delete diam 0 0372 sdr 17 0 PVC 7 diam 0 0389 sdr 26 0 PVC length 2 Hardware Parameters Constraints Load Factors Constraint on Arc R A the two diameters musthe v diameter 1 diam 0 0182 sdr 13 5 PVC V length 1 20 diameter 2 diam 0 0182 sdr 13 5 PVC V length 2 Add constraint Delete diam 0 0391 sdr 32 5 PVC diam 0 0426 sdr 17 0 PVC v Abort Design Abort Design Fix the length of pipe of type 1 The program will automatically compute the length of the type 2 pipe by simple difference Don t forget to click on Add constraint to validate the choice You may repeat the operation to handle several constraints You may also delete constraints 36 Start Design k Hardware C
41. ess from the neighborhood of minimum cost to true minimum cost require a modicum of prior experience on the part of the designer This access is achieved by a First presenting the designer with an optimized lowest cost solution that respects all the chosen constraints albeit in part in a statistical sense b Second allowing the designer to test that solution by simulating the flow within the designed network under a comprehensively large number of operating scenarios the results of which are conveniently compiled and statistically presented NeatWork allows the designer to modify the proposed design whenever the solution is judged deficient in some particular or if additional constraints such as the redundancy provided by loops are desired The modified design can also be tested in the same way The steps for this retouching of the generally optimum solution are made easier automatically by the use of a parameter called quality of service or manually with the help of an influence diagram The simulation can also be applied to any pre existing design The optimization is carried out for branching or arborescent networks only Any number of loops can then be added to any design since the simulation handles loops conveniently But the relative advantages of loops such as their fail safe features are left to the designer to weigh against the cost penalty since these cannot be incorporated rationally in a cost minimizing scheme Minim
42. evel are V t and h t The flow at the faucet is t The dynamics is given by 64 dV t Vo dh t _ p t dt hro dt e t On the other hand neglecting the friction losses in the pipe the faucet flow satisfies 1 ot h t h Thus dh t JVahro di h t h Vo The formula for a 2V V 2 UE suit Vro Fh vh A2 3 hroT follows by direct integration with the boundary conditions t 0 h 0 hro t T h T 0 This formula neglects both the friction in the discharge tube or pipe and the dynamic head at the faucet exit The error due to viscous loss in tube on the value 1 ais given in the table below for several pipe lengths and diameters error in 1 a due to viscous loss in tube 0 0235 0 0304 0 75 2490000 715000 197000 Tube 1 3330000 954000 263000 length 1 5 4990000 1430000 395000 2 6660000 1900000 526000 Since the typical value for 1 a is 105 we see that tube contribution to 1 a is not too important The error due to the neglect of the dynamic head at the pipe exit is of the same order and so they are both negligible if the exit tube is no smaller than 75 Formula derivation for varying faucet characteristics Let us briefly recall the procedure Let the constant for the faucet with the lowest loss constant largest constant be oo and any other be a Let P be a terminal node equipped with a faucet whose constant is o The idea is
43. ew Topography opens a window that reveals the existing name topo files present in the folder projects in the example below there are none 20 x Project Path C Documents and Settings jpv My Documents Neatwork323 projects Browse Your topographies Filename File not found Clicking on the upper right hand button gives access to the folders contained into the folder projects xl Look In C projects 7 allea BE BaseExample Ci MyfirstNetwork C SampleNetwork C designtempla09 C N80T C SimpleNetwork 7 El bihajua J OnePipe 7 Villas1 C ExtraTopos Cis3 File Name C Documents and Settings jpvimy DocumentsiNeatwork324 projects Files of Type All Files M Browse l Cancel The upper right hand buttons allow I Move one level up Move to the home directory c3 Create a new folder H Display as a list 82 Display details Clicking on lal creates a New Folder that you can rename here we named it MyProject 21 x Look In C projects v rs f i fae BaseExample i MyfirstNetwork C 3 C3 designtempla09 C N80T C SampleNetwork C El bihajua C New Folder J SimpleNetwork Cf ExtraTopos f OnePipe 1 Villas1 File Name Documents and SettingsipviMy DocumentsiNeatwork324 projectsiNew Folder Files of Type All Files 7 Browse Cancel To open the new folder select it as in t
44. fferent sizes and types The Target flow is most often 0 15 l sec for all faucets the Quality of Service is 0 7 the maximum faucet fraction assumed for the design is 0 4 The orifice coefficient is 0 62 Either 200 or 500 simulations were used for each case In the data presented below several types of flow rate variations need to be distinguished 1 Flow rate variations at a given faucet as the combination of other open faucets is changed while the number of open faucets is held fixed 2 Variation of mean flow rate from one faucet to the other for a fixed combination of open faucets 3 Flow rate variations at a given faucet when it shares a terminal with several other faucets and the number of open faucets of that terminal is varied 4 Difference between the overall average flow rate for the design fraction of simultaneously open faucets and a greater or smaller fraction Note that our arguments apply to all categories of variations but unequally so For instance in the third category the faucets compared share exactly the same pipe path to the tank For a faucet that shares only part of its piping with other faucets the opening or closing of these other faucets will have less effect on the total head loss of the faucet than the opening or closing of neighboring faucets on the same pipeline The table below summarizes the results for the network Taptest for which the highest faucet is 7 m below the tank and which has a number of
45. g e g Net5 dsg 7 H e Simulation files with suffix sim e g Net5 1 sim Net5 2 sim The user can open topography and design files from appropriate menus Simulation files are not directly accessible NeatWork creates them when a simulation is performed on a specific design They are automatically saved when the design is saved and receive the name of the design file with a number in parenthesis corresponding to the number of the simulation A design file contains a link to all slave simulation files opening a design file gives direct access to all slave simulation files The project files are contained in the folder project The user may create subfolders within the folder project dedicated to a specific project N B The project folder may be moved to any other place in the user s hard disk 15 Measurement units The following units are in use 1 All lengths heights and diameters are measured in meters 2 The speed in pipes is measured in meters per second 3 Pressures are measured with the height in meter of a column of water One unit meter of this measure corresponds 9810 Newtons per square meter 0 981 Newton per square centimeter Illustrative example The tutorial is illustrated by means of the following simple example Ji P1 P2 C P3 P4 The above figure represents a system with one reservoir R three connecting branches R A A B A C four branches lead
46. g is relatively steady but the demands at the individual faucets vary during the day At peak hours the total demand may far exceed the flow at the spring A tank serves as a buffer it accumulates water during low demand hours e g at night and delivers the requested amount at peak hours Thus a complete distribution network is meant of two different components 1 The first component includes the spring the tank and in between a pipe connection 2 The second component links the tank to the individual faucets This is the distribution system per se The design of each component carries its own challenge NeatWork is concerned with the second component exclusively Tools for the design of the first component have been developed by Agua para la Vida They consist of a manual Air in Pipes and a spreadsheet macro These tools can be downloaded at http 209 61 210 46 aplv Downloads AirlnPipesManual pdf http 209 61 210 46 aplv Downloads Aire4j xls The challenge of gravity driven distribution networks In gravity driven distribution networks the flows at individual faucets are regulated by i friction in the pipes and ii friction in orifices near the faucets and in the faucets themselves Friction in the orifices and in the faucets has a local effect while friction in pipes may influence several faucets at once NeatWork exploits the friction law to design distribution network achieving an efficient compromise between two mai
47. hat the diameter of the perforation is a more critical quantity it is difficult to determine with sufficient accuracy and not well reproducible when these orifices are drilled in plastic Orifices can be combined to achieve greater head losses It is easy to verify that two orifices with respective diameters d and d put in series achieve the same head loss as one orifice with diameter 66 d TH di de while the diameter of a single orifice equivalent to n identical orifices d in series is given by d de mre A3 3 Such orifices should be spaced at least 5 pipe diameters apart The table below shows single diameter equivalent to that of two orifices Orifice 2 0 0020 0 0030 0 0040 0 0050 0 0060 0 0020 0 0017 0 0019 0 0020 0 0020 0 0020 0 0030 0 0019 0 0025 0 0028 0 0029 0 0030 0 0040 0 0020 0 0028 0 0034 0 0037 0 0038 0 0050 0 0020 0 0029 0 0037 0 0042 0 0045 0 0060 0 0020 0 0030 0 0038 0 0045 0 0050 Orifice 1 One observes that combining two orifices with very different diameters is ineffective the diameter of the equivalent orifice is almost the same as the smaller of the two diameters On the other hand if the diameters of n orifices in series are identical as n increases the equivalent diameter is less and less reduced 4 Simulation of flows The stationary flows in a network obey the law of minimum energy Since we talk about dynamic quantities the minimum energy mu
48. he image above and click on the button Browse x Project Path C Documents and Settings jpv My Documents Neatwork323 projects Browse Your topographies Filename New Cancel File not found Type the name of your new topography E new topography x Project Path C Documents and Settings jpviMy Documents Neatwork323 projects Browse Your topographies Filename NewTopo New Cancel File not found and click on New to get the window for the new topography 22 iBix File Topography Database Help Tables Tree View Text Node list BS 3f av ID Height x v soffaucets Nature Source 0 0 0 0 0 0 0 TANK Arc list 5 hh av Begin End Length Q NewTopo in which you can start entering your data Create a new topography within a project Proceed as before but select an existing project folder e g BaseExample x Rechercher dans projects M B E E BaseExample 3 ExtraTopos Ci OnePipe C SimpleNetwork 1 designtempla09 C MyFirstNetwork C S3 1 VillaS1 Ci El bihajua C N80T SampleNetwork Nom de fichier uments and Settings pyimy Documents Neatwork323iprojectsiBaseExampl Fichiers du type Tous les fichiers M Browse l Annuler Clicking on the button Browse displays the list of xxx topo files that are stored in the
49. he new segment P P Choose a pipe with inner diameter d 56 The formula to compute L is 1 1 1 po yas 4 7 76 1074 4 Note that the segment P P is virtual It mimics the difference between the reference faucet and the actual faucet For a derivation of formula 4 see Appendix E2 Example While all other faucets are characterized by a constant a 1 83 1078 three terminals are equipped with faucets with a 5 107 Then C DE 1 454105 Q1 Qo With a standard diameter of 1 2 or 0 0182 m the formula for L is L 7107 x ps a 1 ao With 00015m sec L 71079 x 0 00015 107 x 1 45 108 113m 3 Frequently asked questions Question can t alter the target flow in the Quick Check menu Answer You can modify the minimum flow directly on the table but to alter the target flow you need to go to Start Design click the parameters menu on the ruler and select target flow It is thus more logical to complete the preliminary entries in originate design before Quick check Question Granted that the water level assumed by Neatwork in the tank is always the minimum level may want to check the effect of level variations on the flows Since the node heights are given in relation to tank level would have to modify all of them The operation is rather tedious on NeatWork and also prone to errors Is there an alternative Answer We recommend a detour through Excel Copy the nod
50. he use of the templates For convenience the user may add any subfolder A typical organization of the NeatWork folder may look like the image below 14 ini x Fie Edit View Favorites Tools Help EI Back 2 Search LC Folders c4 History US C OX A EE Address c3 Neatwork323 v Go Type Modified an File Folder 05 05 200 C3 Excel Files File Folder 11 07 2002 Neatwork323 htm File Folder 21 06 2002 ooo C projects File Folder 05 05 2002 db C Templates File Folder 02 08 2004 File Folder C Tutorial File Folder 28 07 2007 Modified 05 05 2003 11 29 Bilizre 1 4 1 02 windows i86 exe 8 080 KB Application 05 05 2002 gt mosek2 1 dll 3 748 KB Application Extension 12 02 2002 Attributes normal FS neatwork323 bat 1KB MS DOS Batch File 17 05 2004 E NeatWork323 jar 344 KB Executable Jar File 03 05 2004 S pthreadvse dll 41KB Application Extension 12 02 2002 README txt SKB Text Document 12 02 2002 S solver dll 68KB Application Extension 12 02 2002 E EEMI gt 1 object s selected My Computer To launch NeatWork double click on the NeatWork324 jar file If launching by double click doesn t work open a command prompt move to the NeatWork directory and type java jar NeatWork jar N B NeatWork requires a version 1 3 0 or above of a Java Virtual Machine File system A project involves three types of files e Topography files with suffix tpo e g Net5 tpo e Design files with suffix ds
51. i Save In C BaseExample gt ies c Eg BZ B BaseExample 0 sim G BaseExample 1 sim G BaseExample dsg B BaseExample tpo B NewTopo tpo File Name Files of Type All Files m Save l Cancel Name the file NeatWork will add the suffix html automatically The html file a part of it looks like this 32 Topography report BaseExample Node List Height X Y Faucets Rpo poooo TANK A Hoo popop BRANCHING NODE 000 BRANCHING NODE J0 BRANCHING NODE npn FAUCETNODE B ooi 002 FAUCET NODE FAUCET NODE FAUCET NODE Summary Number of Nodes 8 e Branching nodes 3 e Faucet nodes 4 with 7 individual faucets Total height change 20 m Number of pipes 7 Total length 911 m Thu Aug 05 19 39 59 CEST 2004 An html file can be opened in any browser It can also be opened in Microsoft Word and Excel The latter is convenient as the two tables appear as Excel tables One can edit them and paste them back into NeatWork If you want to copy the node table of this file and paste it back into NeatWork dont forget to replace Tank Branching node and Faucet node with 0 1 and 2 respectively Submenu Start Design The submenu Start Design opens a window associated with the active topography The window displays four tabs on its upper part Hardware Parameters Constraints and Load Factors We ll review them
52. ign Name ATA value Faucet Coefficient 2 0E 8 Orifice Coefficient 0 59 Critical threshold l s 0 1 Limit on budget 1 0E8 Target Flow lis 0 2 Pipe Commercial lengths m 6 0 Water Temperature C 20 0 Fraction of open faucets 0 4 Service Quality 0 6 Close 39 Simulation This important part of the program is accessed from a design window through the second tab in the tab rule The simulation of faucet flows is performed on the design that is currently open Note that you must save the design before performing simulations The simulation module selects at random a choice of open faucets and computes the flows in all the branches of the network for this configuration of open and closed faucets This operation is repeated the number of times that has been specified by the user Clicking on the simulation tab open the following window 10 xl File Design Database Help Tables Simulation TreeView Text New Simulation 2 a av BaseExample BaseExample The main commands are accessed through the buttons New Simulation Clicking on this button opens a window on the simulation parameters e 9 Deletes a simulation file A Creates an HTML report Simulation parameters Click on New Simulation This opens a table displaying the simulation parameters 40 amp Simulation parameters x Number of simu
53. imple case where an unsatisfactory faucet outcome is chronically too high one can equip it with an orifice or if it already has one one can reduce its diameter If it is chronically too low one can increase the orifice diameter if there is one If you wonder whether the failing performance is due to too small a choice of orifice diameters redo the simulation with the ideal orifices option If that simulation is satisfactory you can usually estimate which one or two additional orifices will allow you to approximate that performance Adding them they are easily improvised in the field may be the cheapest modification Using the Tree View functionality to see what sequence of segments a faucet depends on you may attempt to remedy the faulty performance of one or two faucets by modifying the diameters of neighboring segments on which the faucet depends This is fruitful only after some experience is gained If the problem is identified as a high faucet near the end of a distribution line serving many other faucets you may introduce a by pass segment to that faucet You may or may not complete a loop doing so Clicking the green cross left icon above the segment table introduces a new line at the bottom of the table This is where the end nodes the length of the new segment and its chosen diameter need to be specified Unless the by pass borrows the same trench as the previous line a rarely advantageous change this modification evi
54. ing to a faucet B P1 B P2 C P3 C P4 and four end nodes The topography of the system is described by two sets of basic data The branches from R and between nodes are called segments or arcs and the end points of all ultimate segments are called faucet nodes The end nodes host one faucet at least and possibly more e g P2 and P4 Node data Nb of Node ID Height X Y Facet Nature R 0 0 0 Tank A 10 0 0 Branching Node B 15 0 0 Branching Node C 20 0 0 Branching Node P1 12 0 0 1 Faucet P2 9 0 0 3 Faucet P3 18 0 0 1 Faucet P4 19 0 0 2 Faucet Segment arc data 16 Segment ID From node To node Length 1 R A 20 2 A B 50 3 A C 250 4 B P1 12 5 B P2 80 6 C P3 9 7 C P4 110 2 Brief tutorial by menus Launching NeatWork Double clicking on NeatWork jar launches the software which first display the window io xi File Database Help NeatWork A tool for the design and the analysis of gravity driven water distribution system Neatwork The first NeatWork page you meet offers three menus File Data Base and Help We suggest you consult first the menu Data Base Menu Database NeatWork automatically opens the two database files diameters db and orifices db that are contained in the subfolder db of the folder NeatWork324 The databases look like Hardware database x
55. is at atmospheric pressure The head loss through the faucet then becomes an important factor not only because the required head must necessarily exceed it but also because it has a large effect on the variation of the flow out of any given faucet due to the variation of the combination of other faucets opened in the same time It is therefore important to choose a faucet constant that represents accurately the head loss incurred by the faucets you select The head loss through a wide open faucet is assumed of the form 42 Q hg A2 1 Q where Atau is the head loss in meters is the flow rate in m sec and a is a faucet coefficient whose value depends both on the size and on the construction of the faucet For instance Type Brand Nominal Size a Globe U S Standard v2 1 83E 8 Arrow Globe B amp K va 1 08E 8 Ball Unknown 3 8 hole in 3 8 pipe 1 31E 7 valve fitting Dimensional analysis indicates that for self similar faucets o has the form u J R j4 ghi d A2 2 where fis a non dimensional coefficient a weak function of the non dimensional Reynolds number Re see Appendix A1 for it definition is the flow rate hia is the faucet head loss d its diameter and g the gravity constant This implies that o should be nearly proportional to 1 d Formula determining the faucet coefficient The derivation of the formula is as follows At time time f the volume in the tank and the water l
56. iui 40 D Making itWork cm 45 1 Initiate and improve a design ssssssssssssssssssssoocssssssssoooocessssssssoooosssesssssoososesssssesssssoee 45 Create your own Catabase cccccccsccssccssscssccssccssccssccssscssccssscssscssscsssessscssscssscsescsessssssesescssseses 45 Originating a new design eeeeee sees e eerte eene entes ts sensns tata stas ta sesta sens ta iniisa 45 CHECKING PETIGHMANCE ccccsscsssneccsscecsdssscsncessseosntscatsavessssncedscesocoessscacetscosdenesessecenscessendsebsenctsceessens 46 Improving the ESIGN sscicescasscessessdacescovesersscssacesvecdccntecvoveescesccnseesducesdccvecescensenssvessacescevevensesceaceseses 47 2 Advanced JCATUVES DT oi rss e ai 49 Fille System A E E A E E A A H 49 IS OENDEBITIU MET 49 III A AA A ANA A AE A S 49 Limit on budit esscr a e aa Ee 50 Service quality and load factor sesessesesossesossesossesesossesossesossesosossesossesossesossesoseesosossesossesossesosses 50 Pipe cost and uniformity of delivery eeseeseesoesoesoesoesoesoesossoesoesossossoesossossossossossossossossossossossossose 53 Faucet Constant Sei 55 3 Frequently asked QUESTIONS o csesicssvcnssiovebecsivsenedsiasebosssvbucedecossdossiebecesadessdessoseeseseaesenes 57 E Tehnical Appeudbx i iode ie ee veo Er eee best ve dep se ed Leva Ev dpa ee ern ek dp
57. lations 100 Fraction of open faucets 0 4 critical flows l s low 0 1 high 03 Target Flow l s 0 2 Orifices in use commercial Type of simulation monte carlo sampling hd Close ll Run Simulation Number of simulations This depends on the number of faucets in the system The more numerous the faucets the greater the number of random simulations you should choose Even though the number of possible combinations of open faucets can be enormous you need to select far fewer random combinations to obtain statistically significant results The reason is that while the missing combinations may include some that yield at the given faucet an appreciably different flow that event will be rare and therefore unimportant Fraction of open faucets This fraction does not have to be the one you chose for the design But normally you will start your simulation with the same fraction to see how well your design serves your purpose and you will later enquire how the system works for a different higher or lower faucet fraction Critical flows This does not affect the simulation itself You merely choose an acceptable upper and a lower bound for the faucet flows and the simulation computes the 96 of times these bounds are exceeded Target flow The value chosen is only a reminder of the value chosen for the design It has no effect on the simulation Type of orifices The choi
58. manner of determining its value with reasonable accuracy The method is as follow A tank with uniform cross section and initial water volume V is connected by a vertical pipe issuing from its bottom to the faucet below The difference of altitude between the bottom of the tank and the faucet is h suggested value 1 to 3 meters The initial water level in the tank is h o The faucet is then opened and the time T necessary to empty the tank is measured The parameter o given by the formula zy Vio th Va 3 For a derivation of the formula see Appendix E2 How to deal with different types of faucets Let us assume that one wishes to install faucets with different values of a Since NeatWork only recognizes identical faucets one must simulate the different faucets indirectly incorporate either a fake terminal pipe connected to the terminal pipe or an orifice to induce an incremental head loss The second option is somewhat easier but it can only be used to modify an existing design and not to create a new one The first option is more complicated but can be used in the design phase as it is possible to add segments and fix their inner diameter in the design phase We discuss the first option 1 Create a node P with the same elevation as P 2 Add a segment of length L to link P to P 3 Suppress the faucet at P to make it a branching node and add a faucet at P 4 In the Start Design window introduce a constraint on t
59. ment table after Above each table you see six buttons The first four are already familiar The last two are Undo f Validate Filling the node list We illustrate here how to fill the node list with only three nodes the tank a branching node and a terminal faucet node The table fields are ID node ID Height relative elevation with respect to the tank By convention the tank elevation is 0 The elevation of the other nodes is negative e X and Y horizontal coordinates some values must be entered though they do not affect the design of faucets for terminal nodes exclusively 24 e Nature tank branching node or faucet NeatWork automatically fills this field when the node and segment tables are completed Node list f Ir av ID Height x v offaucets Nature TANK 0 0 0 0 0 0 D TANK To enter the information relative to the other two nodes use the button Node list AS hav ID Height X Y stoffaucets Nature Start 0 0 0 0 0 0 0 TANK BNode 20 0 0 0 0 0 D BRANCHING NODE FNode 10 0 0 0 0 0 BRANCHING NODE When adding a new line NeatWork automatically set default values Don t forget to edit those values To validate the information hit the enter key and click on the validate button We get Node list AS IP ev ID Height Xx Y stoffaucets Nature TANK 0 0 0 0 0 0 TANK BNode 20 0 0 0 0 0 FAUCET NODE
60. mple below no problem is detected with a threshold of 0 1 I s and a faucet coefficient 2 0E 8 xi Critical threshold l s 0 1 Faucet Coefficient 2 0E 8 b Apply FaucetID flow min 0 1 l s target flow 0 2 Iis P1 Ok Ok P2 Ok Ok P3 Ok Ok P4 Ok Ok Close You can change the threshold and or the faucet coefficient Do not omit to press the Apply button The table below reveals an anticipated problem at nodes P1 and P2 for a threshold 0 5 l s a very high value in practice but no problem with the target flow 0 2 l s Quick check xi Critical threshold lis 0 5 Faucet Coefficient 2 0E 8 Apply Faucet ID flow min 0 5 l s target flow 0 2 IJ P1 o Ok P2 No Ok P3 Ok Ok P4 Ok Ok Close Submenu Network summary This submenu summarizes the main features of the topography at a glance 31 f amp Topography information x Topography name BaseExample Number of Nodes 8 Branching nodes 3 Faucet nodes 4 with 7 individual faucets Total height change 20 m Number of Pipes 7 Total length 911 Submenu Report in HTML Report in HTML is used to record wherever you choose on your disk and to attach for transmission in HTLM format the complete specifications of the topography The submenu opens a window giving access to the folder where you wish to save the data in html z
61. n as follows Make a backup copy of your own databases n the Excel worksheet make sure the table of Diameter references is ordered by increasing values in the ID column It should be otherwise use sort to order the table Copy the columns Nominal to Roughness of the Excel table of Diameter references Activate the menu Edit database and tab Diameters in NeatWork and paste the clipboard with the appropriate button Do the same with the orifice table Open now an existing design Use Save as menu save the design under the appropriate name and at the appropriate location e Select the full Arc List by clicking on the first and last rows with the shift key pressed Use the delete button to erase the data Validate e Do the same with the Node List Validate 58 Copy the edited node list in Excel and paste it back in NeatWork Validate Copy the edited Arc List in Excel and paste it back in NeatWork Validate Save your design 59 E Technical Appendix Friction formulae for pipe flow in distribution networks The quantity of interest for calculations by NeatWork of head losses by friction is h L the head loss per unit length of pipes for all the pipes utilized That quantity is related to three others by the relation 9 hs 0 0826 r A1 1 L FE where 6 the flow rate is expressed in m sec and d the diameter of the pipe in meters The nume
62. n criteria a The material cost should be as low as possible b The flow at each individual faucet should vary only within acceptable bounds regardless of which and how many faucets are open in the network If criterion b were the only one one could easily prescribe flows at all faucets independently of which and how many faucets are open by over sizing the network i e using pipes with large diameters that are almost frictionless and regulating flows with appropriate orifices only But the cost of piping material is almost always a major factor in the number of systems that can be built at a given time NeatWork therefore also attempts to regulate the flows at faucets through the friction in the pipes This favors the use of pipes with small diameters and thus dramatically reduces the cost of material Unfortunately friction in the pipes introduces faucet interdependences While it makes it impossible for individual faucet flows to be invariant strict invariance is never required Only bounds on variations are desirable These bounds are chosen by the designer NeatWork provides the user with the tools to reach a satisfactory compromise between requirements a and b Minimum pipe cost design NeatWork gives ready access to a minimum material cost design consistent with a comprehensive set of operating constraints and of restrictions on available material both specified by the designer Only the ultimate steps needed to progr
63. n this ratio while in the case when all these faucets are open simultaneously the ratio will be smaller and occasionally null On the above table the three faucets at P2 have been open simultaneously 4 times only out of one hundred simulations Thus a larger number of random simulations is indicated for such cases The next three columns give respectively the minimum average and maximum flow experienced by the given faucet during the simulation The next column gives the flow variability i e quotient of standard deviation to mean The next two columns indicate the percentage of cases when the flow at the given faucet was respectively smaller than your chosen minimum and higher than your chosen maximum The last column gives the number of times the given faucet gave no flow when open these cases are called failures Percentiles at faucets The first second third and last column repeat information found in the Flows at faucets table The remaining five columns give the maximum flow found for the lowest 10 25 50 75 and 90 of the cases 43 rrtcrcr amp 989 igi xi File Design Database Help Tables Simulation TreeView Text 4 i Flows at faucets New Simulation x a b ercentiles Speed in pipes Node pressures simu 0 ni Number of simulations 100 Fraction of open faucets 0 4 Orifices in use commercial ype of simulation random of accu
64. nal nodes Nodes whose nature is 0 source or 1 branching node cannot support faucets Submenu Open topography The submenu works very much the same as New topography Using the browse button move to the project folder containing the topography file you want to open In the example below we moved to the project folder BaseExample and get the display x Project Path C Documents and Settings jpviMy Doc Browse Your topographies BaseExample NewTopo Filename BaseExample Open Close Select the file BaseExample topo the suffix topo is omitted in the display and press Open We get 27 Tables Tree View Text Node list BS HbR av ID Height x Y of faucets Nature R 0 0 0 0 0 0 0 TANK A 10 0 0 0 0 0 0 BRANCHING NODE 2 B 15 0 0 0 0 0 0 BRANCHING NODE C 20 0 0 0 0 0 0 BRANCHING NODE P1 9 0 0 0 0 0 1 FAUCET NODE P2 12 0 0 0 0 0 3 FAUCET NODE P3 18 0 0 0 0 0 1 FAUCET NODE P4 18 0 0 0 0 0 2 FAUCET NODE Arc list AS 3 av Begin End Length R A 400 0 A B 50 0 A C 250 0 B P1 12 0 B P2 80 0 c P3 9 0 C P4 110 0 Q BaseExample Note that the terminal node P2 has 3 faucets while P4 has 2 The topography can be edited using the buttons in the palette AS OR av The changes become permanent when you validate them with 9 and only then On top of the window yo
65. nce the flow rate at the faucet Therefore to achieve uniform delivery in the absence of pumps pipes with small friction losses need to be used right up to the last individual segment attached to the faucets and pipes with large diameter and thus smaller friction should be used elsewhere But large diameter pipes are more expensive than small diameter ones so that this approach is necessarily more costly than one where the head loss is more gradually accumulated along the network We shall use NeatWork with a simple network called taptest 1 whose topography is displayed below to demonstrate this conflict quantitatively 53 ID Height J X v of faucets Nature E Begin End Length Source 0 0 0 0 0 0 TANK Source N1 790 0 0 N1 120 00 00 0 BRANCHING NODE Ni N2 350 N2 80 00 00 0 BRANCHING NODE MI N3 1208 N3 150 00 00 0 BRANCHING NODE NS N4 TT N4 22 0 0 0 0 0 a BRANCHING NODE N2 N5 1600 N5 32 0 0 0 0 0 BRANCHING NODE i N6 150 00 00 0 BRANCHING NODE N3 NG 810 0 N7 220 00 00 0 BRANCHING NODE N3 N7 60 0 P1 7 0 0 0 0 0 1 FAUCET NODE N4 P1 740 0 P3 10 0 0 0 0 0 3 FAUCET NODE N6 P3 235 0 P5 22 0 0 0 0 0 5 FAUCET NODE N7 P5 342 0 P2 34 0 0 0 0 0 2 FAUCET NODE NS P2 220 P4 10 0 00 00 4 FAUCET NODE N6 P4 12 0 P6 70 00 00 1 FAUCET NODE N7 PB 22 0 Our method is simple we merely alter the head loss caused by the faucets e g the faucet coefficient which we can visualize as using faucets of di
66. onstraint on Arc R gt A v thetwo diameters mustbe v Parameters Constraints Load Factors diameter 1 diam 0 0182 sdr 13 5 PVC Y length 1 20 diameter 2 diam 0 0182 sdr 13 5 pvC Y length 2 Addconstramt Delete Arc gt A diam1 0 0182 with length 20 and diam2 0 0182 Abort Design Load factor Clicking the fourth and last tab leads to the following window x Hardware Parameters Constraints LoadFactors Load factors D Begin End offaucets Suggested load factor Modified load factor R A Y 3 18 3 18 A B 4 1 99 1 99 A C 3 1 52 1 62 B P2 3 1 62 1 62 C P4 2 1 26 1 26 B P1 1 1 0 1 0 P2 P2 a 1 1 0 1 0 P2 P2 b 1 1 0 1 0 P2 P2 t 1 1 0 1 0 C P3 1 1 0 1 0 P4 P4 a 1 1 0 1 0 P4 P4 hb 1 1 0 1 0 Abort Design The first two columns identify each segment by its ends The third gives the number of faucets downstream of the segment in question The fourth column predetermined load factors is that computed by the program from the chosen value of Service Quality while the fifth allows you to modify later after simulation the load factor for each segment individually according to perceived local need so that the service quality becomes segment dependant Menu Design This menu applies to an existing design Either open an existing design file e g BaseExample sim
67. ool Mosek by Erling Andersen EKA Consulting ApS The University of Geneva is thanked for its financial support in the development of Version 2 Charlie Huizenga amp Jim Stacey are thanked for helpful comments and suggestions Contacts Frederic babonneau hec unige ch Laurent drouet hec unige ch Jean philippe vial hec unige ch aplv aplv org B The essentials of NeatWork 1 Gravity Driven Distribution Networks NeatWork is a specialized network designing tool adapted to two realities of water development in poor rural communities e Mechanized and electric power are generally either unavailable or unduly expensive and they raise in addition serious maintenance problems n the building by communities themselves of drinking water systems manpower is usually not compensated Maintenance can be made inexpensive so that the main expense is the initial material cost notably that of network piping Material cost acts as one of the two major limitations on the construction of these systems the other one being the need for the contribution of usually external technical and administrative skills As a result e NeatWork deals only with gravity driven distribution networks e NeatWork puts a high priority on finding both minimum material cost and operationally satisfactory solutions Now as has already been mentioned and as we shall show more quantitatively later the minimum cost requirement conflicts with the requiremen
68. optimizing design tool provides a minimum pipe cost solution 10 The simulation module This part of the program solves for the flow out of open faucets for any combination of open faucets The simulation module applies to any design including the recommended minimum cost solution and for any open faucet fraction including the one chosen in the design module It also allows the calculation of flow rates after loops with pipes of chosen diameters have been added to the design The main simulation parameters are a TheFraction of Open Faucets b The Number of Simulations The simulation program can solve sequentially in a short time for the flow out of a large number typically 500 of specific open faucets configurations chosen at random Its output includes several statistical and non statistical quantities of interest From which are immediately available the 9o of cases for any faucet of flow below and above predetermined values the average flow for each faucet and for the average of faucets the maximum and minimum flow for each faucet the corresponding nodal pressures water velocities and other quantities of interest These can be compared to the designer s norms The comparison allows the designer either to stop when the design has been proven adequate in all respects or to tune the design either by altering the value of the service quality factor or more locally that of the related load factors of each pipe segment
69. ospheric pressure The head loss through the faucet then becomes an important factor not only because the required head must necessarily exceed it but also because it has a large effect on the variation of the flow out of any given faucet due to the variation of the combination of other faucets opened in the same time It is therefore important to choose a faucet constant that represents accurately the head loss incurred by the faucets you select Definition The head loss through a wide open faucet is assumed of the form 42 o Piani 1 Q where Atau is the head loss in meters is the flow rate in m sec and a is a faucet coefficient whose value depends both on the size and on the construction of the faucet For instance Type Brand Nominal Size a Globe U S Standard va 1 83E 8 Arrow Globe B amp K v2 1 08E 8 55 Ball Unknown 3 8 hole in 3 8 pipe 1 31E 7 valve fitting Dimensional analysis indicates that for self similar faucets a has the form 4 Re 02 Tu p hug d 2 where fis a non dimensional coefficient a weak function of the non dimensional Reynolds number Re see Appendix A1 for it definition is the flow rate hi is the faucet head loss d its diameter and g the gravity constant This implies that o should be nearly proportional to 1 d Determining the faucet coefficient This parameter is seldom provided by the manufacturer but there is a simple
70. ot known unless they are all open at the same time the design has not satisfied the equations for the flow conservation at the branching nodes except in a statistical sense For a given configuration of open and closed faucets the actual flows in the network may differ from the ones that were used by the program in the design phase To obtain a satisfactory design you may have to go through some or all the steps summarized in the subsequent subsections Create your own database When activated NeatWork opens the two database files E diameters db 2KB DB File 02 08 2004 12 23 orifices db 1KB DB File 02 08 2004 12 20 that are contained in the folder db Edit these files or create new ones to have them correspond to your needs Remember that NeatWork only recognizes those two names diameters db and orifices db NeatWork ignores any other name with suffix ab If you want to save the current file diameters db and create a new one just modify the name e g diametersbk db Originating a new design To initiate a new design a topography file must be open If several topographies are open the command applies to that in the active window Clicking on Start Design in the menu topography opens a window with four tabs Hardware Parameters Constraints and Load Factors In the window Hardware you recover the pipe and orifice database you have created earlier This is where you specify which pipes and which orifices are in fact
71. pen simultaneously 42 Tables Simulation Treeview Text New simutation CR Flows at faucets Percentiles simu 0 I Speed in pipes Number of simulations i Fraction of open faucets Orifices in use commercial ype of simulation 100 0 4 random Node pressures NeatWork 3 23 j File Design Database Help ay Faucet ID of occuren Min Average Variability of failures Global average 0 2161 0 1 0 P1 39 0 176 0 19 0 2564 B 13 0 1 0 P2 4 55 0 2383 0 2615 0 2876 I 8 0 1 P2 2 35 0 1676 0 1794 0 1835 3 0 1 0 P23 4 0 1448 0 1448 0 1448 0 0 1 0 P3 41 0 1608 0 209 02453 M 19 0 1 0 P4 1 49 0 1982 0 2461 0 2715 13 0 0 P4 2 17 0 1205 0 1524 0 1544 0 0 0 Q BaseExample BaseExample For instance terminal P2 has 3 faucets the identifying column will indicate P2 1 for the cases when 1 only of the 3 P2 faucets is open P2 3 for the cases when two of the P2 faucets are open simultaneously and P2 3 for the cases when three of the P2 faucets are open simultaneously The second column lists the number of times the identified faucets were opened The ratio of that number to the total number of faucets should for single faucets hover around the faucet fraction For multiple faucets the number of simulations when exactly one faucet in the lot is open is likely to be higher tha
72. predictable daily and seasonal demand cycles water quality propagation rates etc They do not deal with barely possible energy potentials and unpredictable use scenarios 2 Key design and simulation features Given that there is a conflict between adequate supply and minimum cost of the distribution system how briefly is this conflict spelled out and resolved in the design System capacity and system uniformity Capacity The capacity is defined by the peak hours use pattern of the communities Note that these patterns are partly the result of the designers perception of the communities living habits and partly the result of the adaptation of the communities to a brand new access to water Ultimately the designer will follow some explicit norms For instance the present norm of APLV is that whatever daily water allocation is chosen for a village half of that allocation should be available during the first two waking hours of the day The design will adjust to this peak usage rate which will be averaged over the two hours period This means in the example that the product of the average faucet flow rate and of the maximum number of faucets open on the average during that time should equal half the daily supply Given the total number of faucets installed this yields the maximum fraction of this total which needs to be open simultaneously once the average faucet flow rate has been selected This gives rise to the first two design pa
73. rameters the average target faucet flow rate and the maximum fraction of simultaneously open faucets Uniformity The peak hours supply averaged over all the faucets has to be associated with a flow rate for all faucets which falls almost always within acceptable bounds An excessively slow flow rate and a complete failure to flow can only be accepted as very infrequent events for any faucet An excessively large flow rate is also undesirable Yet the use pattern i e the combination of open faucets is almost completely random within a narrow time frame But for a network near minimum pipe cost the specific combination of open faucets has a definite influence on the flow of each faucet Note that faucets are treated in NeatWork as either open or closed Practically this is found satisfactory even with faucets that can regulate the output The Design and Simulation tools The optimization module The optimization module generates minimum cost designs respecting the supply norms It acts on arborescent or tree network exclusively It takes the flows in each segment as fixed data The pipe diameters are chosen so that the friction induced by the selected flows produces the correct head loss The main design parameters are a The Target Faucet Flow i e an approximation to the desired average faucet flow rate b The Fraction of Open Faucets i e the maximum number of faucets that can be open at the same time C TheService
74. ren P1 39 0 176 0 176 0 176 0 1811 0 1888 0 2554 0 2554 P2 1 55 0 2383 0 2383 0 2383 0 2411 0 2876 0 2876 0 2876 P2 2 35 0 1676 0 1676 0 1808 0 1808 0 1835 0 1835 0 1835 P2 3 4 0 1448 0 1448 0 1448 0 1448 0 1448 0 1448 0 1448 P3 41 0 1608 0 161 0 161 0 2404 0 2453 0 2453 0 2453 P4 1 49 0 1982 0 1982 0 1997 0 2662 0 2715 0 2715 0 2715 P4 2 17 0 1205 0 1544 0 1544 0 1544 0 1544 0 1544 0 1544 O BaseExample Q BaseExample This table provides richer information than the preceding one For each faucet each column provides a point of the cumulative flow rate distribution function To make a plot of this function you should directly extract the information from the name sim file See the section on advanced use of NeatWork Speed in pipes NeatWork computes the average and the maximum values of the speeds in each pipe over the sample of configurations Node pressures NeatWork computes the minimum the average and the maximum values of the pressure at each node over the sample of configurations Reviewing simulations Clicking anywhere in the box containing simu 0 opens a submenu displaying the list of simulations performed with the same design You can select any one of these simulations and review the statistics 4 New Simulation 2 A b 44 D Making it work 1 Initiate and improve a design Remember that because the identity of the open faucets is n
75. rical constant has the units of acceleration m sec The parameter fis dimensionless and is a function of two other dimensionless parameters the Reynolds number Re and the ratio ed of an apparent pipe wall roughness to the pipe diameter The Reynolds number can be written as Re 10 a A1 2 It is a property of the fluid a non dimensional version of its viscosity Once the fluid has been specified the constant a depends only on its temperature For water here are a few values of a TC 2 5 10 15 20 25 30 35 40 a 0 78 0 85 0 98 1 12 1 27 1 42 1 59 1 77 1 95 An approximate formula for a with T in degrees Centigrade is 293 1 1 1 267 559 exp 4 4700 00 A i T 4 273 895 P l z 273 xx j Ars Note for a given region say either Nepal or Nicaragua it is sufficiently accurate to use only one value a The parameter f is a complicated function see below of the Reynolds number Re and of the so called equivalent roughness zd Its base is 60 empirical The effect of roughness is for simplicity s sake taken to be only a function of the pipe diameter and of the nature of the pipe wall On the other hand roughness is hard to characterize and quantify and if the pipe is subject to corrosion or calcification it varies importantly with time Thus the choice of a value ed for a given pipe is substantially subjective This implies that one must give
76. s aede 60 JE V M M 60 Friction formulae for pipe flow in distribution networks eere esee tntnae 60 2s TP QUCCIS pU 64 Introduction and definition eee eee ee eee eee eese essen etes tn senses stas tns essen sensns essen sensns 64 Formula determining the faucet coefficient eere reete esee nete eene ee sensn tata tnsn 64 Formula derivation for varying faucet characteristics eee eee ettet en enne enne 65 Bi OOPULICES PETRO pq op de Pr edP 66 4 Simulation of LOWS vc coacti drca I PIER se isaisa ssns sosise UE DU ROVS SEE AN PER SURE 67 Ds Optimal design p 69 A Gravity water distribution networks 1 What is new about NeatWork For what kind of distribution networks NeatWork is a computer program specifically fashioned for the design of entirely gravity driven water distribution networks for rural areas These networks need to adapt to two characteristics in particular a The driving level differences between the tank and the faucets may be minimal the practical minimum may be as low as 2 meters and b The consumers generally need no water pressure at the exit of their faucets The two components of a gravity driven distribution network A water distribution system from the spring to the individual faucets always includes a tank The flow at the sprin
77. st be understood as the energy per unit of time The problem or minimizing a function here the total energy per unit of time under constraints here the conservation of flows or masses is a mathematical programming problem The total energy turns out to be a convex function of the flows a nice and important feature that is fully exploited by powerful algorithms We briefly present the mathematical programming problem To this end we need few notations The network is represented by a graph G N A where N is the set of nodes and A the set of arcs The graph is oriented that is an arc ij in A has its origin at node and extremity at node J The node set is partitioned in Ns Np and N where 67 Niisthe set of source nodes e Nis the set of transit or branching nodes Niisthe set sink faucet nodes The graph is not necessarily a tree it may contain loops The flows on the arcs are denoted by the double subscript variable j A positive flow corresponds to a flow from the origin to the extremity j A negative flow corresponds to a flow from the extremity to the origin Energy dissipation is a nonlinear function of the flows It occurs under the following forms 1 Friction in pipes The mathematical representation is o E bb Yii pli I2 p 1 where the coefficient y and p are given by formula A1 7 fij Li Ci and p 2 B e d 2 Friction in faucets and orifices The sum of the resulting energy
78. stants Water Temperature C 20 0 Pipe Commercial lengths m 6 0 Advanced Parameters Orifice Coefficient 0 59 A Faucet Coefficient 2 0E 8 E abort Design Once the values of all the parameters have been chosen you need to click the green check mark for the program to use them instead of the default values Let us review the items 34 Fraction of open faucets This is the maximum fraction of open faucets during peak use on which the design will be based You choose this number together with the desired average flow out of the faucets on the basis of your use norms for peak use see System capacity and system uniformity in the introduction A summary of the key design and simulation features You can later in the simulation phase find out what happens when the open faucet fraction is the same as or either greater or smaller than this number Service quality This is the factor discussed in the introduction The higher this value the more generous the assignation of the diameters of intermediate pipes between the initial and the final segments and so for a fixed target flow the higher the average flow rate in the faucets in general A more comprehensive discussion of the meaning of service quality as well as of load factors will be found in the appendix section A A good approach is to start with a value like 0 6 Target flow This is the average flow out of the faucets that the optimizer
79. t of dependable delivery under all uses scenarios For instance a Ifthe design is adequate for use by all faucets in the same time it will be more than adequate for use by only a fraction of the faucets but then the system is clearly oversized and therefore excessively costly since all faucets need not be used simultaneously b In case of fractional and unspecified or random use of the faucets the sensitivity of a given faucet output to the combination of other faucets opened simultaneously can be systematically eliminated but only by methods which can raise seriously the cost of the network and which are inapplicable if the driving elevation difference is minimal Thus a design which yields acceptable flow rates for all faucets under almost all use scenarios at lowest cost requires both sophisticated control by the designer of the optimizing module and from the simulating module a very comprehensive performance analysis This is what NeatWork offers In general already available distribution system analyzing tools are capable of a number of sophisticated answers corresponding to modern pumping options and delivery requirements but they are not able to help the designer in a search for the rustic characteristics demanded by the simple networks we are addressing They deal with other needs e g the requirements for pressure at the delivery sites externally imposed flow rates at hydrants comprehensive fail safe alternate circuits
80. te small and they increase only moderately with time calcium deposits tend to leave streaks which affect little the effective roughness For that reason while one could derive values of A and B from the data above with amp varying from 0015mm to 0025mm it is also possible to simplify the friction calculations by choosing a composite curve The effect of roughness is minimal for PVC on the other hand in both conduction lines and distribution networks elbows tees sudden expansions and contractions of pipes are relatively frequent Since for smooth pipes the skin friction values are far better known than for rough pipes it is logical in the former case to correct on the average for these additional so called minor losses Thus the chosen values of A and B for PVC take the minor losses globally into account and thus yield friction losses about 4 5 above those for smooth pipes without minor losses according to the Colebrook formula up to a Reynolds number of about 200 000 and a decreasing overestimate beyond that Reynolds number They are A 0 235 and B 0 219 This gives in the end for PVC and for water as a fluid 62 h peres Too TEEN 1 95615 hp N 5615 A1 9 2 2 684 fef a In this formula B 9 42 1074ta 919 A1 10 Notes Most of the alternate pipe friction formulae are less accurate than the above Among the better known in addition to the Colebrook expression only the Blasius approximation 0 316
81. the user of the program the possibility of a choice of e for a given pipe wall But it also means that if f depends importantly on z d its value is uncertain The form of f which is generally accepted as fitting the experimental data summarized by the Moody diagram best is the Colebrook formula 1 2010 7 e 251 VF Cd RI A1 4 All other representations of f should be viewed as approximations to A1 4 In other words A1 4 should serve as a reference The fact that A1 4 is implicit is not much of an obstacle for its computation especially since it has at least one good and explicit first approximation On the other hand the direct use of A1 4 in Neatwork raises problems because instead of satisfying head loss equations for each segment we use an equivalent energy minimum principle which requires that the integral of that equation be made explicit Now A1 4 is not integrable in explicit form As a result we have approximated A1 4 by functions which are easily integrable i e f A c d RB A1 5 Note that the exponent of Re does not depend on the flow rate To each value of ed corresponds a pair of values of A and B These have been chosen to minimize the error made on substituting 4 for 3 within the Reynolds number range which is of interest for small drinking water distribution systems 5000 lt Re lt 500 000 in general but the range is restricted further for each pipe diameter The maximum deviations from
82. to introduce a segment of length L and inner diameter d between P and the faucet Since the goal is to achieve a faucet loss one can fix the inner diameter to be the smallest available one and adjust L in consequence 65 The head loss for the second faucet is o 2 hfau A eT hs A2 4 Q1 where hg is the head loss of Faucet 1 over and above that of Faucet 0 Thus h EA If this loss is thought as due to a pipe it is given h 0 0826 fl 7 76 1074 L9 4 A2 5 according to the Blasius friction law see Appendix Equating the two gives 4 LL 1 1 H 1 n 7 761074 ai ao What is called an orifice by Neatwork is a perforated plastic diaphragm fitting in a pipe or union section whose diameter is normally a nominal 1 2 inch upstream of a faucet The perforation is a small hole in the diaphragm center Normally the head loss through such an obstruction depends on the Reynolds number of approach if the perforation diameter exceeds about 30 of the pipe diameter But the orifices used have almost always smaller hole diameters so that it is permissible to consider the head loss to be solely a function of the hole diameter and of the flow rate The simple expression we use is p h 0a A3 1 Where h is the head loss across the orifice 9 is the flow rate in m s and d the diameter of the orifice in meters The value of is left to the choice of the designer Our current best estimate is 0 0 59 Note t
83. u can see three tabs The first one Tables corresponds to the window on display The second one Tree View gives a schematic of the arborescent topography from the tank green through the internal nodes grey to the end nodes faucets blue Anl x File Topography Database Help Tables Tree View Text Node Search Go 1 2 3 B Q BaseExample BaseExample Clicking on any one of the end faucets shows the path along the tree Note that the end node may have multiple faucets This is not indicated on the schematic The third tab shows the data in a text format This is the way the data are stored in the file name topo 28 Submenu Open design This submenu can be put into effect only if a design file already exists To learn how to create a design file from topography see later in this section One such file is provided in the project subfolder BaseExample File Design Database Help Tables Simulation TreeView Text Node list BS HbR av ID Height X Y Ideal Orifice Commercial 0 Nature R 0 0 0 0 0 0 a TANK a A 10 0 0 0 0 0 a 0 BRANCHING NODE B 15 0 0 0 0 0 o 0 BRANCHING NODE C 20 0 0 0 0 0 0 0 BRANCHING NODE P2 12 0 0 0 0 0 0 0 BRANCHING NODE P4 19 0 0 0 0 0 D 0 BRANCHING NODE P1 8 0 0 0
84. um pipe cost vs uniform faucet output There is generally a conflict between the requirements of minimum pipe cost on one hand and of a sufficiently uniform faucet output when the maximum number of faucets used simultaneously is a fraction of the total number available Qualitatively that relation is clear The total available head from the tank to a faucet is fixed The larger the loss through the faucets the smaller the corresponding friction loss through the network of pipes must be Therefore that part of the total head loss due to opening a given faucet which affects flows through pipes serving more than one faucets is smaller This decreases the variability of flow of any given faucet but to achieve this lower loss on the common part of the network one needs to use there larger pipe diameters and these are more expensive This is quantitatively demonstrated in the appendix to this guide NeatWork handles this conflict by making the desired trade off explicitly and easily accessible to the designer 2 Agua Para La Vida Agua Para La Vida is a N G O with presence in Nicaragua the United States and France Its goal is to assist the cooperative development and protection of their drinking water systems by rural communities To this effect Agua Para La Vida Maintains a pilot center in Nicaragua e Diffuses technical material such as NeatWork that it develops for technical and for pedagogical purposes The pilot center that is m
85. workS sssssssssccccssssssssssscccccccssscsssssssseecsssssssssssoes 8 2 Key design and Simulation featur es ccccccccccsscssscsccccssccsscsssssccccccsscsccsssscecccssssssssssees 9 System capacity and system uniformity e seeseesoesoesoesoesossossoesossoesossoesossossossoesossossossossossossossossoss 9 The Design and Simulation tools eere eese eee eee esee esee en etta etiatn staat es stas es stas ense toae 10 Input and output in the two modules eeeee esee ee eese teet eee etn ene tn senses etas tn senses tosta sta 12 C Getting familiar with NeatWork eee ee eee ee eee ee eee eene ee ttn nennen eee etta a assess eno 14 A Introduction WRITE n 14 lidellbepnmilBodecziliig M 14 File iim CIA 15 ljjIzUqjuji4ji o 16 Ill strative example onines entente ueneno oa entente oven ta oen eaae aUa nna Ue ane aUe ana ad 16 PANI LTD RTATINTINIEUTI Te 17 Launching Neat Work esee eese eese eene eene tensa setas tease east tn stesse esses ees teas tease esses esses seas eoate 17 Menu Database M 17 Midi 20 Menu Topography AER 31 Menu Design ai scicsassestecctsestescsseassccdsecdoscssentsccdsecdocsseassccdsecdeacssenssocdsesdsaneseassocdsendsanesenssondsesdeonedones 37 u
86. you click on Start Design you get the window 51 x Start Design i Hardware Parameters Constraints Load Factors Start Design Parameters Fraction of open faucets Serice Quality Target Flow l s Limit on budget Physical constants Water Temperature C Pipe Commercial lengths m Advanced Parameters Orifice Coefficient Faucet Coefficient x 9 Reset 2 amy Hardware Parameters Constraints Load Factors The chosen service quality is 0 6 NeatWork automatically computes the load factor on each branch It is displayed in the Load Factors window 5 Start Design x Load factors e Begin End 3t of faucets Suggested load factor Modified load factor R A T A B 4 1 88 1 99 A C 3 1 52 1 62 B P2 3 1 62 1 62 C P4 2 1 25 1 26 B P1 1 1 0 1 0 P2 P2 a 1 1 0 1 0 P2 P2 b 1 1 0 1 0 P2 P2 c 1 1 0 1 0 C P3 1 1 0 1 0 P4 P4 a 1 1 0 1 0 P4 P4 hb 1 1 0 1 0 52 For branches leading to a single node the load factor is 1 as predicted For the branch from the reservoir to node A with 7 subsequent faucets the load factor is 3 18 If you increase the service quality you decrease the load factor For instance for a service quality 0 8 you get x Hardware Parameters Constraints LoadFactors Load factors A

NeatWork: A user guide

Contents

Download Pdf Manuals

Related Search

Related Contents