Home

Casses Software Version 2.0.0 User Manual

image

Contents

1. Copy top table Copy bottom table From the File menu the Save and Save as functions can be accessed Save as opens a dialog box which allows the filename and location to be chosen The extension for a saved Project is ksp Choose a filename to save the current project EN Look In LA casses D cos 5 C3 B8 2 EK File Name Files of Type casses file ksp gt save Cancel Save directly saves a Project that has already been saved without changing its name CassesManual_2 0 0_b doc 15 06 2010 Page 66 of 88 A Cemagref GE Sciences eaux amp territoires X1 2 Properties of a Project S Casses software ExempleCassesMono2 0 E Job File Creation Construction Tools Pipe set Comment Elaborati Pipe num Pipe leng Lot de tr ALL 1203 MAT IN 288 MAT IN 790 Diaz IDR EI Obs87 95 Va In R s 2C cHo1 S 7 1190 R s Valida e A 140 1 R s C 1996 CSB al 1180 Res CI SP87 96 2 C CH01 il gt Copy top table Copy bottom table E Project properties Example Creation date 3 19 2010 Pipes without diameter removed Break recording starting date 21411987 Break recording ending date 11111997 Pipe data f
2. IX 3 ii Model page The Model page describes the different components of the model results of the last calibration cf Erreur Source du renvoi introuvable after which the Progression was finalised CassesManual 2 0 0 b doc 15 06 2010 Page 55 of 88 amag ref eaux H territoires Qs DAR S Casses software ExempleCassesMono2 0 0 ksp File Creation Construction Tools CA Pipe set IN All pipes IN tron D Plastic C1 Break set IN Al breaks C 87 95 C3 Environments CI Default Eny Obs87 e G1 SP87 gt cH IN Pro E o CJ Iron DN a 87 95 EJ Covariates o CI Qualitative Dj MAT IN ior DIA2 e Cl Quantitativi D LNG Dj Dia Dj por J Break data Configuration Model Description Progression CHO1 Sub project SP87 96 Observation period starting date 2 5 1987 Observation period stopping date 12 31 1996 Environment Default Environment Pipe set All pipes Break set All breaks Progression completed The calibration converged after 187 iterations Last calibrated model Letz 6 9435E2 0 0 Delta 4 men 1 000060 eist 3 0000E1 3 0000E1 Zeta 0 0000E0 0 0000E0 Bett 0 0000E0 1 3012E2 0 0 ILNG 2 0144E1 7 1816968997273 DIA AAT Oh ITO 1 6396E1 5 1405468397658 Ln Likelihood 738 472 IX 4 Exploring a prediction S Casses software Ex
3. Vill 3 Create a Sub project A Sub project is the association of an Environment and a period of observation Only the breaks occurring during this period are taken into account CassesManual_2 0 0_b doc 15 06 2010 Page 37 of 88 Gm GC Sciences eaux amp territoires By combination of the recording period for breaks in the network the period of observation of the Sub project the pipe installation dates and removal dates if applicable the software determines the window of observation for each pipe Recording period for breaks in network F t Period of observation Installation r gt Pipe 1 Window of observation T1 Installation Pipe 2 Window of observation T2 Installation Removed Pipe 3 Window of observation T3 Having selected an Environment a Sub project is created in the Construction mode either on the right side of the Environments screen or the left side of the Sub projects screen by clicking on the button Create a Sub project Create a sub project Title Environment break recording period Start date 2 4 1987 End date 1111997 Actual breaks First break 2151987 Last break 12 19 1995 Observation period Start date 21511 987 End date 1219 1995 Validate model Validation period Comment Validate Cancel CassesManual_2 0 0_b doc 15 06 2010 Page 38 of 88 A C
4. Z i Abbreviated name of parameter Theta Value of parameter Ref Initial value of parameter CassesManual_2 0 0_b doc 15 06 2010 Page 44 of 88 A Cemagref GE Sciences eaux amp territoires Std Standard deviation Chi2 Chi2 value of the parameter Pval p value of the parameter Construction Environments Sub projects Progressions Calibrations Validation Predictions Progression cmo M Calibration Calibration 3 Computation mode automatic mode Log Likelihood 738 472 Finalised Calibration status The calibration converged after 187 iterations Calibrations Ee Sid Chiz Dal Alpha 1 0484E0 1 4035E 1 6 9435E2 0 0 Calibration 1 Delta 1 0000E0 1 0000E0 NaN NaN NaN Cakbralor 2 Zeta0 1 3 0000E1 3 0000E1 Nan Halen Calibration 3 Zeta 0 000DED 0 0000E0 NaN NaN Betao 9 9763E 1 0 0 8 7458E 2 13012E200 LNG 4 J 2 3588E 5 2 0144E1 7 1816968 DA or 4 1 6396E1 5 1405468 MATIFONTE o NaN NaN MATIPVC 0 0000E0 0 000D0E0 NaNl Hannah MAT Amian 1 6246E 1 0 0000E0 8 0619E 2 4 0611E0 0 0438830 Advice Jan parameters are significant reate new calibration Quit Construction mode The rows of the table are as follows Alpha parameter that takes into account the influence of previous breaks when it is not significant its value is
5. 1 1 1996 End date 12 31 1996 Validate Select the dates for the start and end of the validation period By default the Validation start date is the day after the end of the period of observations and the end date the end of the recording period for the Environment VIII 4 Create a Progression A Progression is a selection of covariates and a list of constraints applied at the Sub project level The Progression contains all the information necessary to obtain a model by successive calibrations Having selected a Sub project a Progression is created in the Construction mode either on the right side of the Sub projects screen or the left side of the Progressions screen by clicking on the button Create a Progression CassesManual_2 0 0_b doc 15 06 2010 Page 40 of 88 LE QSD Sciences eaux amp territoires S Progression creation Comment Covariate set selection Eligible covariates Covariate set Name Name Reference modality DDP DIA DIAZ LING MAT MATSP vv Constrained model parameters etal fixed etal fixed Constrain the model Validate Cancel Firstly the Progression must be named The name can be chosen freely the only constraint being that two Progressions in the same Sub project cannot share the same name A text box for optional additional comments is available fo
6. 1 1 Menu Casses software File Creation Construction Tools Version System License Credits Version displays information concerning the version of Casses installed on your computer as well as information relevant to the protection key that you are using System informs you of the version of Java installed on your computer and on the memory allocated to running Casses Licence displays the user licence of Casses to which you have agreed Credits mentions the freeware used by the software and their appropriate licences 111 2 Preferences Casses software File Creation Construction Tools Show gt Preferences CassesManual_2 0 0_b doc 15 06 2010 Page 12 of 88 Sciences eaux 8 territoires Goma QSD Preferences Linguistic parameters Language english v Export format Date Decimal Default directories Projects directory Temporary directory Validate Casses is by default in the language of your operating system However you have the possibility to select another language The modification requires restarting Casses The display of Casses uses the date and decimal separator formats defined by the operating system of your computer However the input files can use any of the accepted formats so long as they are specified in the
7. CJ Pipe set Pipes Breaks SP Covariates IN All pipes DY tron IN Plastic J Break set IN All breaks C 87 95 E Environments o CI Default En CJ Obs e CI SP87 9 cI CH IN Pro e CJ Iron IN a 87 95 Description Sub project 5P87 96 Observation period starting date 2 5 1987 Observation period stopping date 12 31 1996 Environment Default Environment Pipe set All pipes Break set All breaks Short name Long name Typ Date de pose DAT Diam tre QUANTITATIYE Diameter groups QUALITATIVE Identifiant r seau QUALITATIVE Maximum 711993 400 000 Minimum DEEN 40 000 Modality number ariates o CJ Qualitative Ly MAT Cj IDR IN Gaz o CJ Quantitative Ci LNG D Dia C opp E3 Break data Longueur Materiau QUANTITATIVE m QUALITATIVE 8850 000 issue de la covar Created by mergi QUALITATIVE QUALITATIVE Minimum Maximum PN MOBR DMOBR 71111927 2 4 1934 170 2141934 9101940 0 0 0 126 0 000 9101940 4711947 41751947 0 0 0 000 11211953 13 5 0 042 1112111953 6 27 1960 29 6 27 1960 2111967 44 5 0121 0 047 211967 a973 77 0102 971973 4 13 1980 9 6 173 000 0 058 Copy top table Copy bottom table CassesManual 2 0 0 b doc 15 06 2010
8. Calculation 2006 Example from SIROCO IDR NRE BRSD BRED Network ID Network name Break record start date Break record end date DATE DATE e O m y A m y Barj2006 Commune of Barjols 01 06 2000 31 12 2006 Cast2006 Commune of Castellane 01 01 2000 31 12 2006 Yss2006 SIAEP 01 07 2001 31 12 2006 Canc2006 Nice 01 03 2004 31 12 2006 For Help press F1 IS Remarks concerning the creation of csv files The csv files are text files and can therefore also be read and modified with software such as Notepad WordPad Microsoft Word OpenOffice org Writer etc Most often data derive from databases possibly linked to GIS and are presented in the form of tables Several programs allow the creation opening or modification of csv files notably Microsoft Excel OpenOffice org Calc and Microsoft Access Note OpenOffice org Base doesn t recognise the csv format In practice the creation of an input file may require the creation of an intermediate file of dbf format for example for data stored in the ArcView GIS Experience shows that the repeated manipulation of files can lead to formatting errors or the alteration of data Therefore it is best to be careful and avoid wherever possible using different software for the same file Among other reports The opening of a csv file with Microsoft Excel does not give the same results depending on whether it is opened by double clickin
9. December of that year is used as the removal date This characteristic is not compulsory when it is not present all the pipes are considered to still be in service The labels of additional characteristics They are unlimited in number but the following constraints must be respected Upper or lowercase characters from the Latin alphabet are authorised without accents as well as numbers and underscore _ The first character cannot be a number No spaces Cannot be the same as any compulsory data label The number of characters for short names is limited to 8 this label will be use for writing functions allowing covariates to be created from characteristics as well as for column titles in the tables 8 To be able to be used for the creation of a new covariate involving a mathematical expression a quantitative covariate must have a short name different from the formula names below Functions The short name DIA is reserved for the diameter If an additional characteristic uses DIA as a label then it is imperative that the values are quantitative and greater than zero The short name TCM is reserved for the characteristic maximum failure rate and is proposed to be used in future developments of the software When additional data is a date given by the year the day taken into account is the 1 of July of that year CassesManual_2 0 0_b doc 15 06 2010 Page 6 of 88 v The se
10. Edit progression constraints Parameter Constrain Constraints Zetal Inf Zetal 0 Continue The following constraints are possible Constrain Alpha The influence of previous breaks is not considered unchecked by default Constrain Delta The influence of ageing is not considered unchecked by default Constrain Zeta0 No selective survival bias checked by default Constrain Zetal No correction of the time dependent part of the selective survival bias checked by default Constrain a covariate The covariate is forced i e it will be kept in the model even if it is not significant unchecked by default Vill 5 Calculate a model VIILS i Launching a calculation Having selected a Progression the calculation of a model can be made in the Construction mode by clicking on one of the buttons Automatic computations or Semi automatic computations situated on the right side of the Progressions screen CassesManual 2 0 0 b doc 15 06 2010 Page 43 of 88 A Cemagref CD Sciences eaux amp territoires The calculation kernel LEYP integrated in Casses performs the model calibration cf Erreur Source du renvoi introuvable Other than calculate the parameters of the model a test is performed on each of them to evaluate their significance From these test results an advice module integrated
11. Page 53 of 88 Selecting a Sub project in the Exploration window gives access to three pages on the right under the tabs Pipes Breaks and SP Covariates The Pipes and Breaks pages are organised in the same way as with the Environments It is important to note that the pipes and breaks of a Sub project and those of the Environment in which it belongs are not necessarily the same The Sub project doesn t consider Breaks outside the observation period Pipes for which the window of observation is empty The SP Covariates page only displays covariates that were present at the finalisation of one or more Progressions in the Sub project The page displays from top to bottom The description of the Sub project A table of the covariates involved A section indicating the values of the selected covariate A section indicating how the selected covariate was created IX 2 ii Sub project with Validation DER SS Casses software ExempleCassesMono2 0 0 ksp File Creation Construction Tools CC Pipe set IN All pipes D Iron D Plastic Pipes Actual observed breaks Actual breaks SP Covariates Description Sub project Obs87 95 Val96 J Break set IN All breaks C 87 95 E Environments oe ci Default En CI 0bs87 oe C 9P87 9 9 CICH IN Pro Short name Observation period starting date 2 5 1987 Observation period stopping date 12 31 1995 V
12. Program Files Cemagref CassesMulti according to the version bought For the protection key C Program Files ithea and C WINDOWS system32 For the information files home user Casses created at installation For Sun JVM Java Virtual Machine C Program FilesVava 4 Minimum configuration requirements OS Windows XP Memory 512 Mb Hard disk space 96 Mb Sun JVM Java Virtual Machine 89 Mb CassesManual_2 0 0_b doc 15 06 2010 Page 88 of 88
13. eaux amp territoires X 2 Exporting the predictions Each Prediction made with the help of Casses can be exported in csv format By selecting it in the Exploration window and then right clicking Export By selecting the Prediction in the Exploration window and then selecting the Prediction tab and clicking on the button Export results in CSV format The filename is chosen by the user UserName csv The first lines are as follows Project name Environment name Sub project name Progression name Prediction name Casses filename ksp Pipes filename csv Breaks filename csv Pipe set name Break set name Recording period start date Recording period end date Observation period start date Observation period end date Prediction period start date Prediction period end date IDT PBN PBR Dates are in the format d m y IDT is the pipe ID PBN is the number of breaks predicted over the prediction period PBR is the predicted break rate in breaks per kilometre per year Then there is one row per pipe with the different values being separated by a semi colon The exportation is carried out in the same way for a Validation the prediction period being replaced with the validation period CassesManual_2 0 0_b doc 15 06 2010 Page 62 of 88 A CCemagret Gef Predictions export WordPad File Edit View Insert Format Help Cel ER e B Exemple Default Environ
14. 166 169 0 000 0 114 0 083 0 000 0 084 12 42 31 19 54 57 711 40 PVC 42 135 Res Break data MO 133 166 166 169 0 000 0 052 0 000 0 000 0 173 51 81 72 7 086 32 36 7M 40 PVC 42 300 Res MO 133 166 166 169 ST 108 140 140 144 CS 4237 7488 7489 7854 MIE 8254 115 115 118 TC 1680 4931 4932 5297 0 000 0 115 0 045 0 000 0 046 12 39 30 43 76 79 7M 40 PVC 75 250 Res 0 000 0 103 0 160 0 000 0 147 14 48 34 8 833 34 40 71 40 Ami 80 700 Res 0 000 0 034 0 000 0 000 0 043 89 100 95 45 B1 81 7M 4D PVC 53 800 Res 0 000 0 079 0 374 0 000 0 263 26 60 50 3 518 17 25 711 40 PVC 42 300 Res 0 00010 0311 0 000 0 00010 055 94 100 98 35 73 73 7M 40 PVC 42 550 Res ol lol l lsl lslslslslslslslslsls slslslslsls Copy table Export results in CSV format CassesManual 2 0 0 b doc 15 06 2010 Page 60 of 88 Gann CD Sciences eaux amp territoires X Exporting the results X 1 Exporting intermediate results All the tables in the software can be partially o
15. 1989 search 4 100016 19 05 1993 disruption 4 100016 09 03 1995 sectorisation 4 100017 17 08 1987 signaled 4 20 06 1987 100017 21 08 1989 signaled 4 18 08 2006 100024 14 06 1993 disruption 2 100025 15 12 1989 disruption 4 100033 09 02 1987 signaled 4 01 02 1987 100033 30 12 1988 disruption 4 100033 14 11 1989 search 6 100035 01 12 1985 disruption 4 For Help press F1 ING 1 4 Networks file This file is only necessary for Casses Multi IL4i Structure Title free text Comment 1 free text Comment2 free text Comment n free text Bari2002 Network of 01 06 2000 31 12 2002 The rules are generally the same as for the pipes and breaks files The short names imposed for the four compulsory data fields are IDR Network identification NRE Name of Network DDE Date of start of break records for the network If the date given is a year the 1 January will be used CassesManual 2 0 0 b doc 15 06 2010 Page 9 of 88 A Cemagref QSD Sciences eaux amp territoires DFE Date of end of break records for the network If the date given is a year the 31 December will be used This date must be after the data recording start date The networks file shouldn t contain any additional data when present they will be ignored II 4 ii Example B Network ex WordPad File Edit View Insert Format Help Cee 64 A R Siroco Open database
16. 3 Missing pipe observation window I RT 4 Break outside of observation period I CassesManual_2 0 0_b doc 15 06 2010 Page 17 of 88 NG ari CD Sciences eaux amp territoires Remark For Casses Multi IDR is compulsory and anomalies D 2 D 4 and D 7 apply S Importation successful remarks Pipe data file E Pipes csv Break data file E Breaks csv LIB IDX Pipe with missi 105430001 Pipe with missi 105430002 Pipe with missi 105430004 Pipe with missi 105430011 Pipe with missi 105430016 Missing data va Break on a pipe REM_28 1 29 1991 Break on a pipe REM_41 11 12 2002 Break on a pipe Rems 2ez004 Break on a pipe REM_25 2 19 2006 Break on a pipe R2212430313 1 511 977 Break on a pipe REM_40 1 30 1996 Break on a pipe Jo R1307430100 RIDER EE Break on a pipe REM 45 1 11 1988 204 Break on a pipe REM_29 12 30 1992 216 Break on a pipe REM 27 4 19 2000 219 Break on a pipe REM 60 12 5 2002 225 Break on a pipe REM 11 6 19 1986 242 e Del L MINANAINNAG A4 444 001 SE Copy report as Continue By means of the button Copy report as you can save the anomaly report in csv format the fields being separated by semi colons The first row contains the titles CFT CAN LIB SEV NCG IDX DDC NLI Then one row pe
17. Do you want to continue XII Configuration of system preferences In the installation directory of Casses the file systempref ini contains certain parameters used in the software CassesManual 2 0 0 b doc 15 06 2010 Page 68 of 88 Gem QSD Sciences eaux amp territoires EB systempref ME File Edit View BEEN Format Help DS Shk A Nombre magiques lnepsilon 30 0 maxiter 2000 maxcyc le 2 prec 1 0E 6 seuilpvalue 0 05 longueurmaxid 8 nbmaxmodalites 20 For Help press F1 For normal use of Casses these parameters do not need modifying and it is not recommended for a user to do so without advice from the assistance service The parameters are as follows Inepsilon threshold for calculation precision maxiter maximum number of LEYP iterations maxcycle maximum number of LEYP cycles prec relative precision of parameters seuilpvalue threshold for p val for the advice module longueurmaxid maximum number of characters for short names nbmaxmodalites maximum number of modalities for qualitative covariates To be taken into account modification of the parameters must be made before running Casses CassesManual_2 0 0_b doc 15 06 2010 Page 69 of 88 Cemagref GC Sciences eaux amp territoires XIII Annexes XIII 1 Glossar Actual number of breaks By convention the number of actual breaks concerns the real breaks occurring during the prediction window when it is included in the reco
18. Longueur QUANTITATIVE m Imported DIA Diam tre QUANTITATIVE mm Imported For Help press F1 it The first rows are as follows Project name Environment name Sub project name Progression name Casses filename ksp Pipes filename csv Breaks filename csv Pipe set name Break set name Recording period start date Recording period end date Observation period start date Observation period end date Constraints A row starting with for each constraint Short name Long name Type Unit Source Formula Then there is one row per covariate with the different values being separated by a semi colon Dates are in the format d m y CassesManual_2 0 0_b doc 15 06 2010 Page 64 of 88 A X 3 ii Exporting a model Information on the model parameters of a finalised Progression can be exported in csv format B Model export WordPad File Edit View Insert Format Help DS RAA e amp Exemple Default Environment Obs87 95 Val96 CHO1 Model ExempleCassesMono2 0 0 ksp Exemple Trongons csv Exemple Casses csv All pipes All breaks 4 2 1987 1 1 1997 5 2 1987 31 12 1995 Ln Likelihood 673 014 2 i theta ref std chi2 pval Alpha 0 9113354341063973 0 0 13566695012598057 552 1585414563399 0 Delta 1 1 NaN NaN NaN Zeta0 30 30 NaN NaN NaN Zeta1 0 0 NaN NaN NaN BetaO 1 097289466083 4864 0 0 09335177191714579 138 16476050524633 0 DIa 0 00325267497030818
19. Seege Sciences eaux amp territoires e e Casses Software Version 2 0 0 User Manual Assistance casses cemagref fr Website https casses cemagref fr A Cemagref QD Sciences eaux amp territoires Contents I INSTALLING CASSES E 4 Ul ADVISED CONFIGURATION pecisan i eraa 28 2D AA 4 I 2 INSTALLATION UNDER WINDOWS OPERATING SYSTEM 4 II FORMATTING THE DATA TO BE IMPORTED cccccccccssssssssssscccssssscsccsccccsessssssrcccecesssscsscecceseseces 5 ILI GENERAL eeh ed A Seege 5 II 2 PIPES PILE eege 5 II 2 i tee 5 TT 7 II 3 BREAKS PIDE i3302 a EE 8 IER te 8 TEE 8 II 4 NETWORKS FILE EE 9 IL4 i SI EEN 9 IZA Example se NALANG ANA GD RB ALABANG 10 II 5 REMARKS CONCERNING THE CREATION OF CSV FILES sssssssssssesssenenenesenesenesenesenesenesenesenesenennenee 10 TIT CONFIGURING CASSES is scccsscssssseesescscsssnecsiseasseoscsessscacasesscdsseooeseassess soeessasosesesacseces sesdaeseaseassonessbsdeseasen 12 HI 1 MENU AE AEN EE EE TEE EEE E E E EE 12 IIL 2 PREFERENCES EE 12 IV STARTING A PROJECT NN 14 IV 1 CREATING A NEW PROJECT ccccccsccesecececececececececscseseeeeeeceeececeeseeeeeeeecececeeeceeececeeececeeeceeeceeseeeeeeeeseseeess 14 IV li Casses MOO EEN 14 IV 1 ii CSS E 19 IV 2 OPEN A PROJECT EE 20 V EXPLORING THE DATA 0 GRRR BET GURNA UNA esse 21 VI ORGANISATION OF THE EXPLORER WINDOW a 21 V2 SOME USEFUL FUNCTIONALITY 00000 S SENSES SENSES SENSE SNN NANANA 22 VI CREATING PIPE SETS OR BREA
20. file For exportation these formats can be chosen in the Preferences dialog Preferences concerning both the choice of default folder for opening and saving projects and for temporarily storing importation and calculation reports can also be made It should be noted that files stored in the temporary directory are deleted and replaced after each new importation and each new calculation These files can be accessed directly by the Tools menu Casses software File Creation Construction Tools Show gt Last calculation report Preferences Importation report CassesManual 2 0 0 b doc 15 06 2010 Page 13 of 88 LE QSD Sciences eaux amp territoires IV Starting a Project IV 1 Creating a new Project S Casses software File Creation Construction Tools New Open Close Properties IV 1 i Casses Mono A dialog box asks you to identify the pipes and breaks files that you wish to use as well as their locations Selecting files to import Pipe data file Path Break data file Path Validate Cancel An initial series of importation tests is performed If needs be dialog boxes appear inviting corrective action By selecting Yes the blank values are replaced by the Value Empty By selecting No the data will not be used in the Project CassesManual 2 0 0 b doc 15 06 2010 Page 14 of 88 A Cemagref CD S
21. in the software indicates 1f the obtained model is satisfactory or otherwise proposes a modification of the data entered with a view to a new calibration In Automatic computation mode the software carries out a succession of calibrations without intervention from the user until a model with all the parameters significant 1s created In Semi automatic computation mode the software carries out one calibration at a time allowing the user to intervene When a calibration is being made a window allowing the interruption of the calculation is displayed Performing calculation Press the Cancel button to stop the calculation This action resets the project to the situation preceding the calculation Cancel VIIL 5 ii Consulting the results The results of the calculations are shown in the Calibrations screen in the Construction mode On the left side of the screen a scrolling list allows the selection of the Progression whose results are to be visualised The right side of the screen indicates the results of the selected calibration from the list on the left side Log Likelihood is a global statistical calibration indicator The quality of the model is better the higher its value is Calibration status indicates if the model converged and if so the number of iterations necessary for the calibration A table displays the main results of the calculations It contains the following columns
22. of 88 Gem CD Sciences eaux amp territoires Recording window A period delimited by two dates during which the pipe is in service and the breaks on it are recorded Stratification The process of sub dividing groups of pipes for which a qualitative covariate imported or created has the same value The stratification of a collection of breaks is also possible Sub project A part of a Project regrouping the operations made and results obtained with the aid of the software from a Pipe set and a Break set over a fixed observation period A Sub project can belong to only one Environment Validation The comparison of break predictions with actual breaks that have occurred during a defined period In Validation mode the recording period is divided into two successive periods the calibration period and the validation period Validation mode The context of the calculations allowing the calculation of validity indicators Validation mode is only applicable on Sub projects where the observation period ends before the recording period for breaks in the collection of networks Validation period In Validation mode this is the prediction period It immediately follows the calibration period and is included in the period for which breaks have been recorded for all networks in the Project Validity indicator Values calculated from a Prediction made in Validation mode and giving a measure of the predictive performance of the model Ca
23. of the pipe at the time of a break t DDC DDP BWSD Break recording window start date for pipe BWSD max BRSD DDP BWED Break recording window end date for pipe BWED min BRED DHS ESD Environment break record period start date EST min BRSDi for the network participants EED Environment break record period end date EED min BREDi for the network participants OPSD Observation period start date Fixed by the user OPED Observation period end date Fixed by the user OWSD Observation window start date for pipe OWSD max OPSD BWSD a Age of pipe at observation start date a OWSD DDP OWED Observation window end date for pipe OWED min OPED BWED b Age of pipe at observation end date b OWED DDP PPSD Prediction period start date Fixed by the user PPED Prediction period end date Fixed by the user PWSD Prediction window start date for pipe PWSD max PPSD DDP c Age of pipe at prediction start date c PWSD DDP PWED Prediction window end date for pipe PWED min PPED DHS d Age of pipe at prediction end date d PWSD DDP VPSD Validation period start date equivalent to Fixed by the user PPSD for Validation VPED Validation period end date equivalent to Fixed by the user PPED for Validation CassesManual 2 0 0 b doc 15 06 2010 Page 76 of 88 Cemagref CD Sciences eaux amp territoires To describe a passage of time the term period is used when a collection of pipes is co
24. or break date DDC the range of values are divided into ten equal intervals and the first two columns display the upper and lower bounds for each interval the intervals include the upper bound value For DHS the data are also subdivided into ten intervals but an extra interval 101 01 2998 01 01 2999 is added regrouping all the pipes in service DHS imported blank CassesManual_2 0 0_b doc 15 06 2010 Page 52 of 88 For DDC the first column contains the break year and the table has one row per year The last row is the row TOTAL which concerns the whole Environment The signification of the column titles of the lower table is as follows PN Number of pipes PN Percentage of the total number of pipes LNG km Length of the pipes concerned in kilometres LNG Percentage of the total pipe length OBN Number of breaks observed on the pipes concerned OBN Percentage of the total number of breaks observed MOBR Mean observed break rate in breaks per kilometre per year Jo DMOBR Percentage difference from mean overall observed break rate IX 2 Exploring a Sub project The Exploration screens for Sub projects differ depending on whether there is a Validation or not In the navigation tree the icon symbolising a Sub project with Validation is shaded grey IX 2 i Sub project without Validation DER SS Casses software ExempleCassesMono2 0 0 ksp File Creation Construction Tools
25. prediction period must be defined By default the start date is the day after the end of the observation period The start date must be after the observation period The end date must be after the start date A text box for optional additional comments is available for specifying the nature of the Prediction CassesManual_2 0 0_b doc 15 06 2010 Page 50 of 88 A Cemagref CD Sciences eaux amp territoires The right part of the Predictions screen consists of a table displaying the results for each pipe present in the prediction period selected in the left section PBN is the number of breaks predicted in the prediction period MPBR is the mean predicted break rate in breaks per km per year The table can be copied by clicking on the column headers It can be partially copied select with the mouse and then lt CTRL gt C or copied in its entirety click on the table and then lt CTRL gt A and then lt CTRL gt C CassesManual_2 0 0_b doc 15 06 2010 Page 51 of 88 A Cemagref ee IX Exploring the results IX 1 Exploring an Environment Casses software ExempleCassesMono2 0 0 ksp TER File Creation Construction Tools CJ Pipe set Pipes Breaks IN All pipes tron Description D Plastic Environment Default Environment Cl Break set Pipe set All pipes TN All breaks Break set All breaks C 87 95 E Environments Default Eny Shortname Long name Type Minimum Maximurn Modality nu
26. rate of breaks observed and predicted and the values of a b c defined thus a age of the pipe in days at the start of the observation period b age of the pipe in days at the end of the observation period c age of the pipe in days at the start of the prediction period d age of the pipe in days at the end of the prediction period CassesManual_2 0 0_b doc 15 06 2010 Page 57 of 88 A Cemagret Gef Casses software ExempleCassesMono2 0 0 ksp EEK File Creation Construction Tools CA Pipe set 4 Pipes Breaks Prediction Dalipipeg C tron Description D pisete Prediction 1997 CI Breakset Prediction period starting date 1 1 1997 D ane Prediction period stopping date 1213111997 D 87 95 Progression CHOT Sub project SP87 96 J Environmer S Observation period starting date 2 5 1987 4a E ve S Observation period stopping date 12 31 1996 obs Environment Default Environment Copa Pipe set All pipes a CH b Break set All breaks D Nkovo Da LNG IDR amp CH Iron KS Amia Amia 150 0 300 Rese IN a 87 9 e EJ Covariates Mess d 0 000 0 015 71 1 40 1 o 0 027 0 000 0 180 nn Copytable Export results in CSV format On the Breaks page when selecting Break date DDC the table at the bottom includes an extra row entitled Prediction containing the nu
27. the breaks on it are observed and recorded Pipe set A collection of pipes selected from those in the Project It only contains pipes with valid data Prediction A part of a Sub project regrouping the operations made and results obtained with the aid of the software for calculating break predictions from the data in the Sub project for a defined set of covariates and a fixed prediction period Prediction period A period during which break predictions are made The start date of the prediction period is after the end date of the observation period Prediction window The period delimited by two dates for which the pipe is in service and a prediction of breaks is calculated Progression A succession of calibrations for which each new calibration with the exception of the first is determined by modifying the set of covariates in respect of the results of the previous calibration Project A project regroups all the operations and results obtained using the software from a collection of source data from one or more networks Qualitative covariate A covariate with a limited number of numeric or alphanumeric values modalities Note The terms quantitative and qualitative can be used in the same way for break data or pipe characteristics Quantitative covariate A covariate with a measured value expressed as a value with a unit It can be used directly in a numerical calculation CassesManual_2 0 0_b doc 15 06 2010 Page 72
28. the proportion of the number of actual breaks during the validation period can be expressed as a function of the number of pipes CassesManual_2 0 0_b doc 15 06 2010 Page 84 of 88 Gann CD Sciences eaux amp territoires breaks avoided as a function of replaced pipes by number 100 3 00 4 0 10 20 30 40 50 60 70 80 90 100 Pipes sorted by decreasing predicted break number X axis of number of pipes Y axis red curve actual breaks during the validation period Y axis blue curve predicted breaks during the validation period A random ranking of pipes corresponds closely to that described by the function y x green curve Two indicators are defined An Area under the red curve Cbn Percentage of actual breaks during the validation period on 5 of the number of pipes sorted by descending number of predicted breaks More generally Cxn is the percentage of actual breaks during the validation period on x of the number of pipes sorted by descending number of predicted breaks Typical values are 0 1 0 5 1 5 For a random ranking An is close to 0 5 and C5n is 5 The prediction is therefore more satisfying when An and CSn are greater In all cases An and C5n are less than 1 100 If a significant proportion of long pipes make up the pipes most at risk then this might lead to an optimistic vision of the model quality 5 of the number of pipes could for example represent 15 of the n
29. 0 Delta parameter that takes into account the influence of ageing when it is not significant its value is 1 Zeta0 parameter which corrects the fixed part of the selective survival bias when it is not significant its value is in practice 30 Zetal parameter which corrects the time dependent part of the selective survival bias when it is not significant its value is 0 A row for each quantitative covariate A row for each modality of qualitative covariates The reference modality and the modalities identified as being non significant by the advice process have the values 0 0 NaN NaN NaN The advice given is based on the probability value Pval A parameter is considered significant if Pval is less than 0 05 It is a null hypothesis test Less than 5 chance of error by rejecting the hypothesis that the parameter has no effect It is possible that a calibration doesn t converge notably if there is a linear relationship between covariates In such cases it is necessary to create a different Progression CassesManual_2 0 0_b doc 15 06 2010 Page 45 of 88 Gere 5s5 Sciences eaux 8 territoires It is possible to access the detailed results of the last calibration carried out by LEYP S Casses software ExempleCassesMono2 0 0 ksp C Ex File Creation Construction Tools CL Pipe set lt Show gt Last calculation report 3 All pipes Preferences Importation repo
30. 0 Pipes sorted by decreasing predicted break number actual breaks during validation period predicted breaks during validation period Random CassesManual_2 0 0_b doc 15 06 2010 Page 48 of 88 A Cemagref CD Graph by length of pipes for which the X axis represents the percentage of cumulative pipe length sorted by descending predicted annual breaks rate breaks avoided as a function of replaced pipes by length 100 10 2 J 3 J 40 50 6 H 70 80 90 10 0 Pipes sorted by decreasing predicted break rate actual breaks during validation period predicted breaks during validation period Random For each of the graphs three curves are represented The red curve that allows the indicators to be calculated represents the percentage of actual breaks as a function of x The blue curve represents the percentage of predicted breaks as a function of x The green curve y x simulates random behaviour Each graph can be copied saved in png format or printed It is possible to zoom in to a part of the graph by clicking then dragging towards the bottom right To zoom out click and drag to the top left VIIL6 ii Right part of the Validation screen The right section of the Validation screen consists of a table which displays for each pipe concerned The compulsory data IDT DDP MAT LNG PBN yr the annual number of breaks predicted by
31. 00 breaks avoided as a function of FAF FON 160 0 046 0 000 0 029 0 000 MEZI mo 0 044 0 000 0 049 0 000 replaced pipes by number d E 71 J ami 200 0 049 0 oo 0 244 0 000 Im pue 130 0 066 0 000 0 508 0 000 reve 650 10 044 0 000 0 067 0 000 TAW 350 0 040 0 PAH 150 0 041 0 000 0 275 0 000 Da 0 068 0 000 0 072 0 000 0 031 0 000 0 017 0 000 0 052 0 000 0 070 0 000 0 000 0 048 0 000 E DES 000 34 0 000 0 0 000 DROE i 0 000 0 055 000 0 10 20 30 40 50 60 70 80 o 100 Pipes sorted by decreasing predicted break number actual breaks during validation period SY EURE e at 0 ER D SE Oas psa Yo predicted breaks during validation period Random NG 0 000 0 420 0 000 0 044 0 000 0 040 0 000 Quit Construction mode VIIL 6 i Left part of the Validation screen The section to the top left of the Validation screen presents the validation indicators Les variables used are as follows OPSD Observation period start date OPED Observation period end date VPSD Validation period start date VPED Validation period end date An Area under the red validation curve according to number of pipes Al Area und
32. 55 0 0 00080413 41406077472 16 36154583 698320 LNG 0 00009646847797082134 0 0 00002437756110522737 15 659949080049C MAT Amiante Ciment 0 19721980424197755 0 0 0812279373 6064764 5 895 Ju 3 For Help press F1 The first rows are as follows Project name Environment name Sub project name Progression name Casses filename ksp Pipes filename csv Breaks filename csv Pipe set name Break set name Recording period start date Recording period end date Observation period start date Observation period end date Ln Likelihood value of the model Z i theta ref std chi2 pval Then there is one row per model parameter with the different values being separated by a semi colon Dates are in the format d m y CassesManual_2 0 0_b doc 15 06 2010 Page 65 of 88 Cemagref GE wW Sciences eaux amp territoires XI Closing a work session XI 1 Save a project SS Casses software ExempleCassesMono2 0 0 ksp File Creation Construction Tools Pipe set Comment Elaboration Pipe number Pipe length k Lot de trongo ALL MAT IN FON MAT IN PVC o E 0bs87 95 Vala6 7MM975 150 000 110 180 R seaut 300 000 Amiante e CI cHot 7411975 200 000 180 400 Reseaut 900 000 Amiante IB validation 7975 53 000 40 110 R seau 3000 000 PvC SSC 7 1 1975 200 000 180 400 R seaut 2900 000 Amiante o c 5P87 96
33. AN 87 95 Pipe set All pipes Break set 87 95 Sub projects Sub OPSD Create environment Create sub project Quit Construction mode CassesManual 2 0 0 b doc 15 06 2010 Page 36 of 88 A Cemagref QSD Sciences eaux amp territoires To return to the Exploration window click on the button Quit Construction mode Each screen is separated in two parts by a vertical movable bar The display on the right section depends on the object selected in the left section VIII 2 Create an Environment An Environment is an association of a pipe set and a break set The default Environment is a pipe set with all the pipes and a break set with all the breaks and is created automatically The creation of a new Environment is made from the Environments screen in Construction mode by clicking on the button Create an Environment Environment creation Pipe set All pipes Break set Comment Validate Cancel Firstly the Environment must be named The name can be chosen freely the only constraint being that two Environments in the same Project cannot share the same name Next use the scrolling lists to select a pipe set and a break set from those available in the Project A text box for optional additional comments is available for specifying the nature of the Environment
34. D Proj Imo 8620 118 118 122 gt CJ Iron CS 6428 9679 9680 100 D All 87 95 MO 8620 118 118 122 MIE 8254 115 115 118 MIE 8254 115 115 118 MA 217 250 250 253 Covariates CI Qualitative D MAT MO 8620 118 118 122 0 000 0 044 0 000 0 000 0 067 64 89 81 26 62 65 71 1140 PVC 42 650 R s IN IDR MO 8620 118 118 122 0 000 0 040 0 000 0 000 0 115 77 93 89 12 46 47 757 140 PVC 53 350 Res IN Diaz MO 8620 118 118 122 0 000 0 041 0 000 0 000 0 275 74 93 87 3 261 17 24 711 40 PVC 42 150 Res o CI quantitative MO 133 166 166 169 0 000 0 068 0 000 0 000 0 072 32 65 56 24 59 63 74 140 ami 60 950 Res D LNG TC 1680 4931 4932 5297 0 000 0 031 0 000 0 000 0 017 93 100 97 83 98 97 71 40 PVC 75 185 Res MO 133 166 166 169 0 000 0 053 0 000 0 000 0 070 50 81 71 25 59 64 7M1 40 PVC 53 750 Res D Dia ST 108 140 140 144 0 000 0 046 0 000 0 000 0 048 59 85 78 41 7B 78 7M1 11 Ami 125 950 Res IN DDP MO 1133 166
35. K SETS 0000010000000000000000 000000000000 00s0e0e0o0see0coneea0c0saessocnsae0coenasescossa00s 23 VLI WHATIS ASSET eege Ce eege a 23 VL2 THE FUNCTIONING OF FILTERS EE 24 VI21 QualitativefilteriNg a GANAN ANAK ca vtec thas risa BANNA en er ane nt eme te 25 VI 2 ii Quantitatve FILET ania tin tege NANANG NAKA Eeer nr 25 VI 2 iii Combination of manyfilters saa tienne nr esse 25 VI 2 iv Particular case of filtering by DHS ann ANAN EE 26 VI 3 EXPORTING SETS see AET E ET A AE ETA E AE T E E EAE T 26 VII CREATE MODIFY OR DELETE COVARIATES e ssesssecsesssecesseoeesoeessoseessoocesseceessessssoesssosoeessee 28 VII 1 COVARIATES CREATION MERGING ccc0cscsesssesessssssssesseecssscsssssessesecesesscscssssseesesesseeseseseseseseseeeeees 29 VII2 COVARIATES CREATION DISCRETISATION s s scssssesscsssesnscsnscsnscsnscsnscsnscsnsesesesesesesesesesesesesnsees 30 VIL3 COVARIATES CREATION QUANTIFICATION 211520000000005 000 00S e 31 VII 4 COVARIATES CREATION FORMULA um sanan ANNA 32 VII S COVARIATES CREATION COMBINATION cscsesesesssesssessssscssscsessstsesessereseeeseeeseseseeeseeeseeesereseeenees 33 VII 6 MODIFY OR DELETE COVARIATES snnnnnnnnnnnrrenenenenenenenenenenenenenenenenenenenenenenennneneneneennnnnnnnnnnnnnnnneneneee 34 VI or Modify covariate aNG ANGAT 34 VII om DG Le ATEN ATEN TRE BARA ARA ANA 33 VIII CONSTRUCTING A PREDICTION sseesesseessosecssocecessecssscecssoceessoceeseecesc
36. OPED lt PPSD Rules of consistency of observation and recording periods OPSDZESD OPED lt EED Rules of the existence of windows see above BRSD lt DHS and DDP lt BRED OPSD lt BWED and BWSD lt OPED PPSD lt DHS and DDP lt PPED expanded to DDP lt PPSD CassesManual_2 0 0_b doc 15 06 2010 Page 77 of 88 Sciences eaux amp territoires C Cemagref 5s5 Rules of consistency of recording and validation periods ESD lt VPSD VPED lt EED SUL Am Duration of windows The duration of a window in days is equal to the difference of the dates 1 Dfx EDFx SDFx 1 CassesManual_2 0 0_b doc 15 06 2010 Page 78 of 88 A Cemagref GE Sciences eaux amp territoires XIII 4 Description of the LEYP model LEYP model for recurrent failures of water mains Cemagref Research Unit Networks Purification and Water Quality 50 avenue de Verdun 33612 Cestas Cedex France May 2010 Abstract This document aims at presenting the theoretical bases of the Linear Extension of the Yule Process LEYP designed to model recurrent failures of water mains The LEYP model is implemented in the computation code of the Casses software 1 Introduction The Linear Extension of the Yule Process LEYP allows to give a parametric representa tion of the process of recurrent failures a pressure main is likely to undergo The LEYP model enables to compute the possible number of failures that may af
37. TE 40 110 PYC 40 110 Amiante Ci 110 180 FONTE 110 180 PvC 110 180 Amiante Ci 180 400 FONTE 180 400 PVC 180 400 Amiante Ci Modality 2 Frequency FONTE PVC Amiante Ciment FONTE PVC Amiante Ciment FONTE PVC Amiante Ciment Modality 1 40 110 40 110 40 110 1110 180 1110 180 1110 180 1180 400 1180 400 1180 400 Covariate to be created Short name DIAMAT Long name Material with diamer classes Validate Using the scrolling menus select two existing qualitative covariates to serve as the sources of the new covariate 15 06 2010 Page 33 of 88 CassesManual 2 0 0 b doc Gene GC Sciences eaux amp territoires If the combined number of modalities is greater than 20 the covariate cannot be created The new covariate must be identified with a short name This must contain no more than eight characters and respect the constraints for short names described in the data importation section cf IN An optional long name may be used to further identify the covariate The description of how the new covariate is created is memorised this information is accessible on the right part of the screen when the covariate is selected in the explorer window VII 6 Modify or Delete Covariates It is possible to modify or delete covariates These functionalities are accessible by right clicking the appropriate co
38. a recording period RT 1 and RT 2 only concern Casses Multi CAN is the code for the anomaly CassesManual_2 0 0_b doc 15 06 2010 Page 15 of 88 A Cemagref Sciences eaux amp territoires LIB is the label for the anomaly SEV is the severity of the anomaly with one of two possible values B if the anomaly is blocking critical I if the anomaly is not blocking information NCG is the short name of the data characteristic concerned if applicable otherwise empty IDX is the ID of the network if the anomaly concerns a network Multi network version only otherwise it is the ID of the pipe or empty DDC is the date of the break or empty NLI is the row number of the data in the treated file or empty List of anomalies treated CAN LIB Label of anomaly SEV Severity of anomaly Code anomaly E1 Missing file B F 2 Unreadable file B F3 Error in data specification area B F4 Error in data area B E5 File contains no records B F6 Unable to match any break with any pipe B E7 All breaks outside the record periods of the networks B D 1 Invalid or missing short data name B if compulsory I if additional D 2 Predefined mandatory or additional data with non B if compulsory compliant type I if additional D 3 Additional data type not recognized I D 4 Missing mandatory data B D 5 Quantitative data value invalid B if compulsory I if additional D 6 Categorical
39. alidation period starting date 1411996 Validation period stopping date 12 31 1996 Environment Default Environment Pipe set All pipes Break set All breaks Long name Type Unit Minimum Maximum _ Modality number e CJ Iron IN all_87 95 EJ Covariates o CI Qualitative CI MAT IN ior Ly Gaz oe CJ Quantitative IN LNG IN DIA IN opp C3 Break data Date de pose DATE 7NN927 71993 JUAN mmm 40 000 Mdr Diam tre QUANTITATIVE QUALITATIVE ldentifiant r seau Ge 8950 000 Mat riau QUALITATIVE IVE Longueur QUANTITATIVE m lissue de la covar QUALITATIVE Minimum Maximum LING km MOER 36 DMOBR 71111927 2411934 119 630 0 131 2141934 9 10 1940 9 10 1940 441711947 4171947 1112111953 11121 1953 6 27 1960 81271960 21967 211967 9i711973 100 850 Copy top table Copy bottom table Selecting a Sub project in the Exploration window gives access to four pages on the right under the tabs Pipes Observed breaks Actual breaks and SP Covariates CassesManual 2 0 0 b doc 15 06 2010 Page 54 of 88 Geen CD Sciences eaux amp territoires The Pipes and SP Covariates pages are the same as for a Sub project without Validation and the Observed breaks page is similar t
40. ciences eaux amp territoires Example of a corrective dialog box Warning Pipe data file The qualitative data LINING has 5 empty values Do you want to give them a modality value Yes No If the initial series of tests concludes that the files are invalid then the importation is interrupted and the importation report is displayed Otherwise the following dialog box is displayed Parameterization of break record Break record starting date 1 12 1975 Break record end date 5 5 2007 Continue You must indicate the period during which the breaks presented in the breaks file have been recorded on the network By default the software proposes the day before the first break and the day after the last break in the breaks file A new series of tests is carried out If no anomalies are detected during the various tests then you will be able to access the main screen Otherwise the importation report is displayed Each anomaly detected is displayed in a table with the following column labels CFT CAN LIB SEV NCG IDX DDC NLI CFT is the code of the file s treated It can contain the following values 7 T for the pipes file C for the breaks file TC for the linking of the pipes and breaks files R for the networks file Casses Multi only RT for the linking of the pipes and networks files or in the case of mono network the coherence of pipes and breaks with the dat
41. cond row salmon coloured row contains a long name associated with the short name given in the preceding row The long name can be chosen freely with any characters except semi colons and may contain spaces It should be intelligible and can be used in the results v The third row blue row contains the type of data chosen from one of three possibilities QUAL for qualitative QUAN for quantitative or DATE must be uppercase This can be omitted for compulsory data v The fourth row pink row specifies the type of data For quantitative data the unit is given For numeric data both the comma and full stop can be used as decimal separators Spaces between figures are allowed but the presence of two separators or a currency symbol is forbidden For data in the form of a date it specifies the format of the date from one of the following a or y the data is a year with 4 figures m a or d m y the date is expressed in order by day month and year in figures separated by slashes The year has four figures m j a or m d y the date is expressed in order by month day and year in figures separated by slashes The year has four figures a m j or y m d the date is expressed in order by year month and day in figures separated by slashes The year has four figures For qualitative data the values are empty Qualitative data values may contain all characters wi
42. data with more than 20 modalities B if compulsory Lif additional D 7 Missing data value B if compulsory I if additional D 8 Date format not recognized B if compulsory I if additional D 9 Short data name already used in the dataset B if compulsory I if additional CassesManual 2 0 0 b doc 15 06 2010 Page 16 of 88 A Cemagref Sciences eaux amp territoires CAN LIB Label of anomaly SEV Severity of anomaly Code anomaly TI Missing pipe identifier B T2 Pipe identifier not unique B T 3 Missing or incorrect installation date B T 4 Installation date after removal date B T 5 Pipe length not positive B T 6 Pipe data invalid B if compulsory I if additional T 7 Pipe data missing B if compulsory I if additional Cl Installation date after replacement date I C2 Pipe length not strictly positive I C3 Pipe with invalid data value B C 4 Pipe with missing data value I Co Break without pipe identifier I TC 1 Missing or incorrect break date I TC 2 Pipe with duplicated break date s I TC 3 Break with invalid data value I R 1 Break with missing data value B R2 Break occurrence before pipe installation B R 3 Break occurrence strictly after pipe installation B R 4 Break on a pipe unidentified in the pipe file B R 5 Unidentified network B RT 1 Pipe with network identifier absent from network file B RT 2 Network contains no pipes I RT
43. e Sciences eaux amp territoires Five methods are available allowing covariates to be created Merging Discretisation Quantification Formula Combination Vil 1 Covariates creation Merging Merging allows a new qualitative covariate to be created from an existing one by regrouping existing modalities into fewer groups S Merging New modalities Source covariate Iron Other Modality allocation Original modality Identifier Frequency New modality FONTE FONTE 288 Iron PVC PVC 790 Other Amiante Ciment ETERNIT Covariate to be created Short name Long name MAT2 Grouped Materials Validate Cancel Firstly use the scrolling list to select the source for the merging operation from the existing qualitative covariates You then must create the new modalities of the new covariate with the aid of the Add button The next stage involves associating each of the source covariate modalities with the new covariate modalities using the dropdown list CassesManual 2 0 0 b doc 15 06 2010 Page 29 of 88 A The new covariate must be identified with a short name This must contain no more than eight characters and respect the constraints for short names described in the data importation section cf II An optional long name may be used to further identify the covariate The d
44. e is also similar to that for a Sub project but also contains information relating to predicted breaks when DDC is selected in the top table The Validation Results page presents the indicators and graphics of the Validation in an identical fashion as the left part of the Validation tab in Construction mode cf VIIL 6 i CassesManual 2 0 0 b doc 15 06 2010 Page 59 of 88 A Cemagref CD Sciences eaux amp territoires The Validation Data page is equivalent to the Prediction page associated with a Prediction but contains in addition the following information ABN actual number of breaks during the validation period ABR actual break rate during the validation period xn yno and ynp the x and y co ordinates o for observed and p for predicted permitting the construction of the graph according to pipe number xl ylo and vip the x and y co ordinates o for observed and p for predicted permitting the construction of the graph according to pipe length File Creation Construction Tools Pipe set Pipes Actual observed breaks Actualbreaks Validation Results Validation Data IN All pipes A Iron 3 Description IN Plastic Progression CHO1 Validation Break set Sub project Obs87 95_Val96 D All breaks Observation period starting date 2 5 1987 D 87 95 Observation period stopping date 12 31 1995 s E Validation period starting date 1 1 1996 Env
45. e of the validation The basic principle of the validation is to compare the break predictions with the actual breaks for a period when breaks were observed To perform the validation two distinct periods are defined from the break recording period a calibration period and a subsequent validation period In the case of multiple networks the validation period is a period during which all the networks were subject to break observations Break Network 3 recording period sp Network 2 Network 1 gt t Validation period Calibration period The validation applies to a prediction for which the period of observation coincides with the calibration period and the prediction period coincides with the validation period The calibration and validation periods are defined by the recording period of the Environment which is delimited by ESD Environment break record start date It is the earliest recording start date of the networks with at least one pipe present in the Environment min BRSDi FEED Environment break record end date It is the minimum recording end date of the networks with at least one pipe present in the Environment min BREDi In the case of a mono network Project for all environments ESD coincides with BRSD and EED coincides with BRED XIIL S ii Ranking according to the number of predicted breaks After having sorted the pipes by descending number of predicted breaks per year
46. eate modify filter CassesManual_2 0 0_b doc 15 06 2010 Page 24 of 88 Cann CD Sciences eaux amp territoires VI 2 i Qualitative filtering If the data are qualitative the filtering consists of selecting the modalities to be kept with the aid of the arrows or by double clicking Only the pipes or breaks possessing one of the chosen values will be included in the set S Qualitative filtering Available Selected FONTE PVC Amiante Ciment Continue VI 2 ii Quantitative filtering If the data are quantitative the filtering consists of defining intervals Only the pipes or breaks whose value lies between the intervals will be included in the set S Quantitative filtering Intervals Minimum Minimum included Maximum Maximum included 40 vi 400 Y Delete interval Add an interval Continue VI 2 iii Combination of many filters You can define many filters they are taken into account in a cumulative fashion logical AND CassesManual 2 0 0 b doc 15 06 2010 Page 25 of 88 A Cemagref CD Sciences eaux amp territoires The set is characterised by a logical signature that is indicated in the Elaboration field For example for iron pipes greater or equal to 100m and less than 1000m the signature of the lot is LNG IN 100 1000 MAT IN IRON The number of records in the set is indicated at t
47. ececseseaeseeececseseaseeececscceseesseseeeeeesesnaaeseeseeenenseaees 61 X 2 EXPORTING THE PREDICTIONS suisse seereernsseceecerennene secs enennene eee enennene secs esennene seen 62 X 3 EXPORTING THE RESULTS OF A PROGRESSION sise aaa 63 XF Exporting a configurations EEEE EEEE ENR SEA EEEE EAEE Eii 63 Xadi Exporine d ModeL na monte ins E A E E EENE 65 XI CLOSING A WORK SESSION eeessseessoseesseceessecesosesssoceessocescoessseceessoceesseeeessecescoessseoeessecessseesssosessesese 66 XI 1 SAVE PROJECT NEE 66 XL2 PROPERTIES OF A PROJECT 3250 20 NLNG AN UNANG ADA UG EE GIA KALA Ee Ae Eege de 67 XL3 CLOSING A PROJECT 2320 GANUN See EA de PALAD E S GG NON 68 XII CONFIGURATION OF SYSTEM PREFERENCE G ccsssccsssssscesssscccssssccsessccccscsscccesssscccssscoece 68 XIII ANNEXES cssssiscscssasssscsssscecsassdasssesonsecisosssesscaadesscsseacssoscesesesssesessasseceeasass orosei st ascasessasaasidzscsesesssaesseayes 70 XIIL1 GLOSSARY EEN 70 XIIL2 SCHEMATISATION OF CALCULATION PROCESS aaa T4 XIIL 3 DEFINITIONS AND RULES RELATIVE TO DATES aaa 75 SI 3 Denominations and calculations of dates and age 75 XUL3 1i Principal rules Of dates ma NAA AKBAR 77 X3 DUP AMON Of WINDOWS cans DANG GG eiii tte en AABANG 78 XIII 4 DESCRIPTION OF THE LENPwMoppt aa 79 XIILS CONSISTENCY OF THE VALIDATION uses aaa 84 XIII 3 Principle Of the validalo nsaan NN GANG NG AGAD ler tnt 64 XII S ii Ranking according to the number of predicted breaks 64 XIII 3 Ranki
48. ef B555 Sciences eaux 8 territoires XIII 5 iv Number de breaks predicted To measure the validity of s prediction in terms of the number of breaks predicted the following indicators are defined PBN Total number of predicted breaks for all pipes ABN Total number of actual breaks for all pipes Rn Ratio between the number of predicted and actual breaks for all pipes during the validation period Rxn Ratio between the number of predicted and actual breaks for x of the pipes sorted by descending number of breaks predicted during the validation period Rxl Ratio between the number of predicted and actual breaks for x of the cumulated length of pipes sorted by descending predicted break rate during the validation period The prediction is therefore more satisfying when Rn is close to 1 The same goes for Rxn and Rxl CassesManual 2 0 0 b doc 15 06 2010 Page 87 of 88 A Cemagref CD Sciences eaux amp territoires XIII 6 Computing information 1 Software installed CASSES software ITHEA Actikey driver for the protection key and management software for the keys 2 Types of files created Casses files with KSP extension Tabular data export file with CSV Comma separated value extension Information files Lanceur log Rimp txt Rcal txt fichier des pr f rences utilisateur casses ini 3 Relevant directories For CASSES C Program Files Cemagref CassesMono or C
49. emagref CD Sciences eaux amp territoires Firstly the Sub project must be named The name can be chosen freely the only constraint being that two Sub projects in the same Environment cannot share the same name Next choose the dates for the beginning and end of the observation period By default the dates of the first and last breaks in the Environment are proposed A text box for optional additional comments is available for specifying the nature of the Sub project At this stage of the Sub project the user can choose to carry out validation calculations by clicking the box Validate Model The general principle of the Validation is to compare the predictions of the model with the actual breaks observed cf XIII 5 Recording periods Network 3 of breaks NE Network 1 Calibration period Validation period To perform a Validation it is therefore necessary to choose a date for the end of the period of observations that is before that of the recording period for the Environment CassesManual_2 0 0_b doc 15 06 2010 Page 39 of 88 Gm Sciences eaux amp territoires Create a sub project Title Obs87 95_Val96 Environment break recording period Start date 2 4 1987 Enddate 1 1 1997 Actual breaks First break 2 5 1987 Lastbreak 12 19 1995 Observation period Start date 2 5 1987 End date 12 31 1995 Validate model Validation period Start date
50. ematical functions and operators and the eligible covariates The formula can also be entered directly in the area intended for this purpose CassesManual 2 0 0 b doc 15 06 2010 Page 32 of 88 Gen Sciences eaux amp territoires The new covariate must be identified with a short name This must contain no more than eight characters and respect the constraints for short names described in the data importation section cf II An optional long name may be used to further identify the covariate The units applicable to the new values can also be specified optional The description of how the new covariate is created is memorised this information is accessible on the right part of the screen when the covariate is selected in the explorer window VIL5 Covariates creation Combination Combination allows a new qualitative covariate to be created from two existing qualitative covariates The modalities of the new covariate result from the combination of those of the source covariates Only modalities applying to at least one pipe will be created S Combination Covariate 1 DIA2 Covariate 2 MAT Modality Identifier Frequency Modality Identifier Frequency 40 110 40 110 1050 0 FONTE FONTE 288 0 1110 180 1110 180 111 0 PVC PVC 790 0 42 0 Amiante Ci ETERNIT 125 0 180 400 180 400 Combined modalities Modality 40 110 FON
51. empleCassesMono2 0 0 ksp File Creation Construction Tools CJ Pipe set D All pipes IN tron IN Plastic J Break set IN at breaks Done CA Environments e CI Default En gt E Obs871 e C1 8P87 9 OCH Di Dy eral o gt CJ Iron E Na ez oe CC Covariates c Qualitative Oma CDR Dorz CI Se Ci LNG IN DIA CC Break data Pipes Breaks Prediction DER Description Prediction 1997 Prediction period starting date 1 1 1997 Prediction period stopping date 12 31 1997 Progression CH01 Sub project SP87 96 Observation period starting date 2 5 1987 Observation period stopping date 12 31 1996 Environment Default Environment Pipe set All pipes Break set All breaks Long name Type Unit Minimum Maximum Modali care de pose DATE Midivyyy 7111927 711993 number Diam tre QUANTITATIVE 40 000 400 000 QUALITATIVE QUANTITATIVE DIA2 Diameter groups QUALITATIVE IDR identifiant r seau LNG 1 QUALITATIVE Minimum Maximu 711927 2 4 1934 2 4 1934 9110 1 900 1 41711 4751 11 211 117214 612711 Copy top table Copy bottom table CassesManual 2 0 0 b doc 15 06 2010 Page 56 of 88 Cemagref 5s5 Sciences eaux amp territoires Selecting a Prediction in the Exploration window gives access to three pages on the right under the tab
52. er the red validation curve according to network length PBN Predicted break number during the validation period ABN Actual break number during the validation period CassesManual_2 0 0_b doc 15 06 2010 Page 47 of 88 A Cemagref QSD Rn Ratio PBN ABN Con of breaks at the point on the red curve corresponding to x of pipes by number Cxl of breaks at the point on the red curve corresponding to x of pipes by length Rxn Ratio between the number of breaks predicted and the actual number of breaks for the first x of pipes sorted by descending annual number of breaks Rxl Ratio between the number of breaks predicted and the actual number of breaks for the first x of pipes sorted by descending annual break rate The last row of the table corresponds to a value of x chosen by the user The values of the indicators affected by the choice of x are updated in the table by clicking on the button Calculate The tables are not modifiable They can be copied to the clipboard by means of the buttons at the side The area at the bottom left of the screen shows a graphic visualisation of the Validation A button allows the user to select Graph by number of pipes for which the X axis represents the percentage of pipes sorted by descending predicted annual break number breaks avoided as a function of replaced pipes by number 100 90 0 10 20 30 40 50 60 70 80 90 10
53. ering failure rate The limiting value of the average failure rate when h gt 0 EdN is a pivotal object of the counting process theory called process intensity In the literature dedicated to the reliability of pressure water mains all reported studies seem to agree that the failure rate e tends to increase with the number of failures already undergone e tends to increase with the age of the mains e can significantly vary according to characteristics of the mains such as material size joint type internal or external coating cathodic protection etc and of their environment soil corrosivity traffic intensity service pressure etc These technical evidences have lead to define the LEYP intensity function with parameter 67 a 6 P7 related to a main the characteristics of which compose the vector Z vector of covariates as the conditional expectancy Eg AN 0 NG Z 1 aN t dt exp ZB di The LEYP intensity is then built as the product of three factors CassesManual_2 0 0_b doc 15 06 2010 Page 80 of 88 A Cemagref GE Sciences eaux amp territoires e the so called Yule factor 1 aN t linear function of the number of failures under gone until just before the scalar parameter o gt O measures thus the tendency of the failures to accumulate on the the same mains e time power factor d2 with 6 gt 1 which models ageing e the so called Cox factor exp ZTB making
54. escription of how the new covariate is created is memorised this information is accessible on the right part of the screen when the covariate is selected in the explorer window VII 2 Covariates creation Discretisation Discretisation allows a new qualitative covariate to be created from an existing quantitative one The modalities of the new covariate are intervals defined by the values of the source covariate Discretisation Source covariate Modalities to be created Label Lower bound Upper bound Frequency 40 180 40 180 Upper bound e 1161 1180 290 180 290 Ca 34 290 400 290 400 Ca Covariate to be created Short name JDIA2 Long name Diameter in 3 groups Validate Firstly use the scrolling list to select the source for the discretisation operation from the existing quantitative covariates You can then create the modalities of the new covariate Split After you have selected one row in the table this button splits the selected interval into two halves Merge After you have selected several adjacent rows in the table this button merges the intervals into one covering the whole range of the rows selected Automatic This button opens a dialog box in which you must specify the number of intervals between 2 and 20 to create The intervals crea
55. etwork length For this reason an alternative ranking method is proposed in complement CassesManual_2 0 0_b doc 15 06 2010 Page 85 of 88 Sciences eaux 8 territoires Gene GC XTII 5 iii Ranking according to predicted break rate After having sorted the pipes by descending predicted break rate the proportion of the number of actual breaks during the validation period can be expressed as a function of the cumulative length of pipes breaks avoided as a function of replaced pipes by length 100 7 OO 4 0 10 20 30 40 SH 60 20 80 90 100 Pipes sorted by decreasing predicted break rate actual breaks during validation period Yo predicted breaks during validation period Random X axis of length of pipes Y axis red curve actual breaks during the validation period Y axis blue curve predicted breaks during the validation period A random ranking of pipes corresponds closely to that described by the function y x green curve Two new indicators are defined Al Area under the red curve C5l Percentage of actual breaks during the validation period on 5 of the total length of pipes sorted by descending predicted break rate More generally Cxl is the percentage of actual breaks during the validation period on x of the total length of pipes sorted by descending predicted break rate Typical values are 0 1 0 5 1 5 CassesManual_2 0 0_b doc 15 06 2010 Page 86 of 88 A Cemagr
56. fect a main with known char acteristics within any time interval even in the future The water mains that make up a water supply network can consequently be ranked according to their failure risk hence allowing to build annual pipe renewal programmes and to compare in the medium and long term asset management strategies 2 The counting process N f Repeated failures may affect a single main at random times 7 Nr Their cumulated number define the so called counting random function also known as the counting process denoted N namely aright continuous left bounded step function defined for any gt Oand incremented by one unit at each failure occurrence N O 0 NTj 2 j LNT j Equivalently the differential dN NG dt Nu takes the value O everywhere except at failure times where it takes the value 1 We R T jeN AN 0 Wee T jeN dN 1 Fig 1 illustrates the construct of the counting process NO and of its differential d CassesManual_2 0 0_b doc 15 06 2010 Page 79 of 88 A CCemagret Ge Sciences eaux 8 territoires Figure 1 Counting process 3 The LEYP intensity function The average failure rate within the time interval of length h gt O can be defined as PONG h NG h i e the average failure number within that interval divided by the interval length When divided by the pipe length this rate is commonly used in the field of infrastructure engine
57. g name Corrosivity indicator Unit None e Firstly use the scrolling list to select the source for the quantification operation from the existing qualitative covariates The next stage consists of attributing a numeric value to each of the source covariate modalities CassesManual_2 0 0_b doc 15 06 2010 Page 31 of 88 A Cemagref CD Sciences eaux amp territoires The new covariate must be identified with a short name This must contain no more than eight characters and respect the constraints for short names described in the data importation section cf II An optional long name may be used to further identify the covariate The units applicable to the new values can also be specified optional The description of how the new covariate is created is memorised this information is accessible on the right part of the screen when the covariate is selected in the explorer window VII 4 Covariates creation Formula The formula tool allows a new quantitative covariate to be created by applying a mathematical formula involving one or more existing quantitative covariates S Formula Formula Basic operations Functions Covariates DIA Covariate to be created Short name JLNLNG Long name Lo H Le ngth Unit None Validate Cancel The formula for calculating the new covariate is obtained by clicking on the various math
58. g on the file lines of file as text in the first column of the worksheet or by opening it from Excel table with a different field per column The second method is advised A csv file created with OpenOffice Calc from a dbf file doesn t open in the correct format with Excel the problem can be resolved by inserting a comment row at the beginning of the csv file If a dbf file is created in Excel the field names are truncated at 11 characters CassesManual 2 0 0 b doc 15 06 2010 Page 10 of 88 A Cemagref CD Sciences eaux amp territoires With OpenOffice Calc whatever the regional options selected in Windows if the language option is French the numeric separator is always the comma to have the full stop it is necessary to change to English For the opening or saving of csv files OpenOffice Calc systematically shows a dialog box agreeing to select the semi colon as field separator and delete the proposed text separator Regional options of the machine can lead to incorrect reading of csv file in Excel For example in Norway the date separator is the full stop meaning that certain decimal numbers are interpreted as dates e g 02 10 interpreted as 10 February CassesManual_2 0 0_b doc 15 06 2010 Page 11 of 88 Gari Sciences eaux 8 territoires lll Configuring Casses Launching Casses results in a window with 5 menu items being displayed S Casses software File Creation Construction Tools
59. he Sub projects created from a Pipe set and a Break set Environment recording period The period delimited by the earliest recording start date and the earliest recording end date of the networks for which at least one pipe is included in the Environment CassesManual_2 0 0_b doc 15 06 2010 Page 70 of 88 A Cemagref Sciences eaux amp territoires Finalising a Progression Forced covariate Likelihood Mean observed break rate Modality indicators of a qualitative covariate Network recording period Observation period Pertinent covariate modality CassesManual 2 0 0 b doc An action carried out on a Progression enabling predictions to be made from the last calibration as well as the creation if necessary of new covariates associated with the calibration It is not possible to make further calibrations on a finalised Progression An eligible covariate from a Sub project that the user wished to include in the model even if it is not significant The maximum value for the likelihood function of a model For each model this value is calculated by the LEYP kernel the result shown is the logarithm A model has a closer fit the greater the value of the log likelihood The sum of the number of pipe breaks during their window of observation divided by the sum of the product of the pipe lengths and their observation window duration A quantitative covariate deriving from a qualitative one that takes the value 1 f
60. he bottom of the creation window VI 2 iv Particular case of filtering by DHS The data DHS date removed from service has the particularity of being the only data for which blank values are allowed in the importation pipes in service To take into account this particularity the quantitative filter includes a checkbox Exclude working pipes Quantitative filtering Intervals Minimum Minimum included Maximum Maximum included 12 31 1979 Y 1213111999 ri _ Exclude working pipes Delete interval Add an interval Continue In practice Casses attributes a distant removal date 01 01 2999 to pipes in service thus checking the box excludes the interval 1 1 2998 1 1 2999 VI 3 Exporting sets Sets can be exported as csv files by clicking on the set and on the button Export at the bottom of the screen or by right clicking on the set CassesManual_2 0 0_b doc 15 06 2010 Page 26 of 88 A L Cemagref S Casses software ExempleCassesMono2 0 0 ksp Tools File Creation Construction J Pipe set IN All pipes C9 tron Pipe set Comment Elaboration MAT IN PVC DIA 711975 53 000 R seaut 3000 000 PvC 2650 000 P C 711975 53 000 R seaul Ti A745 Copy top table Copy bottom table Export The sets exported in this way are in the Casses format thus they can be imported int
61. hich breaks to consider The functionality is accessible by the menu Creation or by right clicking in the navigation tree on one of the objects Pipe set or Break set SS Casses software ExempleCassesMono2 0 0 ksp Creation Construction Tools Create a pipe set Create a break set Pipe set Covariate b Comment Elaboration TN All breaks i Let de tron on ALL 1203 Ty 87 95 MAT IN FON 288 EJ Environments o CH Default Environment o Of Obs87 95 Val96 2 DI CH DDP DIA IDR Validation CSBBO150 7MM975 150 000 R8seaut 300 000 Amiante Ci IN 1996 CSBBO200 7111975 200 000 R seau 900 000 Amiante Ci gt v 4 SP87 96 2C CH01 T1997 ipag Copy top table Copy bottom table The creation of a set is made in a new window CassesManual_2 0 0_b doc 15 06 2010 Page 23 of 88 Gm GC Sciences eaux amp territoires S Pipe set creation Covariates Filter selected covariate CreateModify filter Remove filter Elaboration Pipe number 1203 Create set Cancel The set must be named and a space for optional comments is provided to allow a more precise description of the set VI 2 The functioning of filters Sets are created by filtering You must therefore select a covariate from which the filter is performed and then click on the button Cr
62. ile Exemple_Tron ons csy Break data file Exemple Casses csy Validate Cancel You can choose any Project name the default being the first comment line in the Pipes file The Comment box is also freely defined the default being the second comment line in the Pipes file CassesManual_2 0 0_b doc 15 06 2010 Page 67 of 88 Ga 5s5 Sciences eaux 8 territoires The other properties are saved at the moment the data are imported They are not modifiable X1 3 Closing a Project SS Casses software ExempleCassesMono2 0 L If File Creation Construction Tools Pipe set Comment Elaborati Pipe num Pipe leng Allpipes Lotde tr ALL 1203 Iron MAT IN 288 MAT IN 790 au IDR EI Obs87 95 va Ge 1501110 R s Gi Cup 7 200 180 Res Valida M 53 000 40 1 R s C 1996 a 200 180 R s o G4 SP87 96 oe G4 CH01 4 D Copy top table Copy bottom table Close closes the current Project but keeps the Casses application open Quit closes both the current Project and the application In both cases if the Project has not been saved a dialog box invites confirmation of the closure Ask for confirmation Warning there are unsaved changes in the project These changes will be definitively lost if this action is performed
63. ironment TN All breaks Pipe set All pipes C 87 95 Break set All breaks CT Environments CI Default Environment CI Obs87 95 Valg6 E cHot Validation Short name Longname Type Unit Minimum Maximum Modality numbei Datedepose DATE Mida 71111927 711993 Diam tre QUANTITATIVE mm 40 000 400 000 C 1996 Identifiant r se QUALITATIVE CI 5P87 96 Longueur QUANTITATIVE Im 30 000 8950 000 2 OCH Materiau QUALITATIVE C 1997 Ed Iron c 88 96 CJ CH01 Minimum Maximum LNG km OBN O o7 71111927 24411934 119 530 7 19 9 Obs87 94 Val95 96 24934 9 10 1940 0 0 0 000 D 0 0 Da 67 95 9M 011940 4117 1947 0 0 0 000 2 0 0 4M7M947 11 21113 13 5 184 730 149 10 3 1172149 6 27 1960 2 9 28 350 2 3 4 5 6 27 1960 21 1967 44 5 593130 D i 36 6 Ei Covariates 9 CI Qualitative covariates C DR C MAT CI Quantitative covariates Copy top table Copy bottom table PFA nnan J CassesManual_2 0 0_b doc 15 06 2010 Page 21 of 88 A Cemagref QSD Sciences eaux amp territoires The families of objects present in the navigation tree on the left are The pipe sets The break sets The Environments The covariates The break data The networks only in the multi network version V 2 Some useful functionality A certain number of ergonomic rules are applicable for the exp
64. ironments Validation period stopping date 12 31 1996 Default Eny Environment Default Environment CJ 0bs87 Pipe set All pipes HCH Break set All breaks Bi D C3 sPe7 d IDT a b el d OBN REN PEN MR MO MP8R xn yno ynp x yio yip DDP DIA2 MAT DIA LNG IDR 0 000 0 033 0 000 0 000 0 013 91 100 96 90 98 98 711 11 Ami 175 250 Res 0 000 0 033 0 000 0 000 0 066 91 100 96 27 64 66 7 1 40 PVC 75 500 Res 0 000 0 030 0 000 0 000 0 061 94 100 98 31 67 69 741 40 PVC 42 500 Res 0 000 0 0 31 0 000 0 000 0 010 94 100 97 95 100 99 711 18 FO 200 310 Res 0 000 0 029 0 000 0 000 0 058 95 100 98 33 70 71 771 111 PVC 113 500 Res 0 000 0 046 0 000 0 000 0 029 60 85 78 63 85 9O 711 40 FO 60 160 R s 0 000 0 045 0 000 0 000 0 049 62 89 80 41 76 77 7M1 40 PVC 42 900 Res 0 000 0 049 0 000 0 000 0 244 56 81 75 4 096 21 27 74 140 ami 60 200 Res 0 000 0 066 0 000 0 000 0 508 34 69 57 0 972 6 250 11 71 40 PVC 53 130 Res CHL CS 4237 7488 7489 7854 D CS 5698 8949 8950 9315 TC 1680 4931 4932 5297
65. ks Actual breaks Validation Results Validation Data All pipes Description D Iron Progression CH01 Validation D Plastic Sub project Obs87 95 Val96 EJ Break set Observation period starting date 2 5 1987 IN All breaks Observation period stopping date 12 31 1995 D 87 95 Validation period starting date 1 1 1996 2 Validation period stopping date 12 31 1996 Environments Environment Default Environment CJ Default Eny Pipe set All pipes E Obs87 Break set All breaks CI CH Results o CJ P87 9 breaks avoided in regard of replaced pipes by a asi EACH number number of pipes D 100 7 z Graph by length C Pro g of pipes c Iron 90 7 IN All 87 95 80 Copyimage ariates 1 o CH Qualitative al Save Cy MAT 0 IDR Sp i C Gaz o CC Quantitative Ci LNG Cj DIA IN ppp C Break data 20 30 40 60 70 80 Pipes sorted by decreasing predicted break number actual breaks during validation period Yo predicted breaks during validation period Random Selecting a Validation in the Exploration window gives access to five pages on the right under the tabs Pipes Observed Breaks Actual breaks Validation Results and Validation Data The Pipes page is identical to that for a Prediction The Observed breaks page is similar to that for a Sub project The Actual breaks pag
66. l in service and are used in connection with the beginning and stopping observation dates of the network to calculate the ages a and b between which the main was observed The main description is required to include the length and optionnally other important characteristics potentially explanatory of the failure rate such as the material the diameter and also depending on their availability the soil corrosivity the service presure the location under roadway versus sidewalk the traffic intensity etc The characteristics kept as failure risk factors make up the covariate vector Z The failure dataset lists for every main 7 which was observed to fail at least once the event times Gre Lai bi j E Las DCH The information available to estimate the model parameters is so formalised as the following set O O Zi ai Bis tj j Lys Mi i La an The natural logarithm of the likelihood function of the parameter vector given the obser vation set O is then written as follows In L 6 O 3 m Ina InT a m lnr a El Lo m In Lab Zi ula Zi 1 gt in AG Zi aA t 2 ja with A t Z dt exp ZTA The LEYP model parameters are estimated by the vector that maximises In L O 6 Consideration of main abandonment and selective survival bias Ensuring the practical relevance of the LEYP model when applied to a set of mains that comprises a notable proportion of very old pipes requires conside
67. llowing covariates ensuing from progression are about to be created Short name Long name Reference modality INKOVO Created by merging following modalities from covariate MAT FONTE PVC FONTE PVC Create covariates Do not create covariates CassesManual_2 0 0_b doc 15 06 2010 Page 46 of 88 Gann CD Sciences eaux amp territoires A new covariate is created when certain modalities of a qualitative covariate are significant whilst others are not The non significant modalities are merged with the reference modality The new covariate created is therefore only useable in the Sub Project in which the particular Progression belongs VIII 6 Consult a validation When a Sub project includes a validation period the finalisation of the Progression launches the Validation calculations and the results are accessible from the Validation screen cf XIII 5 Construction Environments Sub projects Progressions Calibrations Validation Predictions EE Prediction Validation Predictions p sen T T T T e DT DP MAT LNG En kee validations OPSD OPED DDPY DFPV Ces I m met 10 033 0 000 o 013 0 000 a GE Feb 5 1987 Dec 31 1995 ant 1996 Dec 31 1996 css lm 0 033 0 000 0 066 0 000 TCT 7ft Pve 500 0 030 0 000 0 061 0 000 nes 7H1 FON 310 0 031 0 000 0 010 0 000 7AL 500 0 029 0 000 0 088 0 0
68. loration windows The size of the different sections is adjustable by moving the separating borders The columns in the tables are adjustable in size and moveable by actions in the header row shaded grey The tables can be sorted according to each column by simply clicking 99 DI the column header according to the sequence sort ascending sort descending no sort All or part of each table can be copied to the clipboard by selecting the cells and pressing lt Ctrl gt C The copy includes the header row and the exact data values At the bottom of the right section the buttons allow the entire table to be copied and if necessary exported in csv format Right clicking on an element of the navigation tree accesses the menu of possible operations for that object CassesManual 2 0 0 b doc 15 06 2010 Page 22 of 88 NG aari CD Sciences eaux amp territoires VI Creating pipe sets or break sets VI 1 What is a set A pipe set is a collection of pipes selected from those present in a Project In the same way a break set is a collection of breaks selected from those present in a Project After importing the data Casses automatically creates a pipe set containing all the pipes and a break set containing all the valid breaks You have the possibility of creating other sets This is particularly useful if you only wish to study one category of pipes or if you wish to be selective in terms of w
69. mber EI Obs87 Date de pose DATE 711927 7111993 CI sPa7 9 jam QUANTITATIVE 40 000 400 000 Don DIA2 Diameter groups QUALITATIVE IN Pro lIdentifiant r seau QUALITATIVE gt Cf iron Gale Longueur QUANTITATIVE IM 8950 008 Er COTE IN a groe Mat riau QUALITATIVE ariates o CJ Qualitative Ly MAT Cj IDR Minimum Maximum Ph PN LING km LNG OBN bOBN MOBR DMOBR Dow 40 000 76 000 3 797 982 io 76 000 112000 157 300 CH Quantitative 112 000 18 000 ER 98 360 L LNG 148 000 184 000 50 4 2 79730 oi 184 000 61 200 opp 220 000 256 000 C Break data 292 000 328 000 Copy top table Copy bottom table When you select an Environment in the Exploration window you have access to two pages on the right accessed by the tabs Pipes and Breaks At the top of each of these pages is a section describing composition of the Environment A second section the top table displays a list and description of all the covariates for the Pipes tab and all the data for the Breaks tab The bottom table describes the element selected in the top table The first row of the table contains the column labels For qualitative data the first column displays the modalities For quantitative data other than the date removed from service DHS
70. mber and rate of breaks PREDICTED asse O are empleCasse ono2 0 0 D CJ File Creation Construction Tools Pipe set ialipipes IN tron Description DPlastic Prediction 1997 Break set E Prediction period starting date 1 1 1997 D All breaks Prediction period stopping date 12 31 1997 D 87 95 E Progression CHOT 3 Sub project 5P87 96 Enviranments Observation period starting date 2 5 1987 CJ Default Em Observation period stopping date 12 31 1996 E Obse7 Environment Default Environment CI SP87 9 Pipe set All pipes e OCH Break set All breaks Prediction D z D Pro C3 iron S Short name Long name Type Minimum Maximum Modality number Danita H pe Date de casse DATE Imany 5987 1213111995 Covariates C Qualitative Omar C DR Doa cI Quantitative Dino Year OBN MOBR DMOBR Doa C ppp Break data DIS It S Copy top table Copy bottom table CassesManual 2 0 0 b doc 15 06 2010 Page 58 of 88 Sciences eaux 8 territoires A Cemagref GC IX 5 Exploring a Validation Validations are located at the same level as Predictions in the Exploration window and are distinguished by a grey icon SS Casses software ExempleCassesMono2 0 0 ksp File Creation Construction Tools J Pipe set Pipes Actual observed brea
71. ment 0bs87 95 Val96 CHO1 1996 ExempleCassesMono2 0 0 ksp Exemple Trongons csv Exemple Casses csv All pipes All breaks 4 2 1987 1 1 1997 5 2 1987 31 12 1995 1 1 1996 31 12 1996 IDT NCP TXCP Identifiant trongon Nombre de casses pr vu Taux de casses pr vu QUAN QUAN susu km an CSSTGFE175 0 03306971039091125 0 01320077783 637195 CSSTILU75 0 032844689712 606394 0 06555477004114472 TCTCLN42 0 030397141210580174 0 060669703973576 MOLBMES200 0 030705878920239085 0 00988482485071155 CSLGBO0113 0 02918238475861443 0 05824516957969356 MOCELP60 0 045887943 762046225 0 028621194431501677 HIECM42 0 04456347396184175 0 04941350596406406 MIECCHD60 0 048831960372392576 0 243659474399131 For Help press F1 X 3 Exporting the results of a Progression X 3 i Exporting a configuration Information on the configuration of a Progression can be exported in csv format CassesManual_2 0 0_b doc 15 06 2010 Page 63 of 88 A Cemagref GE Sciences eaux amp territoires B Progression export WordPad File Edit View Insert Format Help DoH SR AB Exemple Default Environment 0bs87 95 Val96 CHO1 Configuration ExempleCassesMono2 0 0 ksp Exemple Tron ons csv Exemple Casses csv 411 pipes 411 breaks 4 2 1987 1 1 1997 5 2 1987 31 12 1995 Constraints ZETAO Inf ZETAL 0 Short name Long name Type Unit Source Formula MAT Mat riau QUALITATIVE Imported LNG
72. n be added after an initial character the text is free but in order to be correctly interpreted it must not contain any semi colons or inverted commas The first comment row is used as the default value for the Project name v The first row without yellow row contains the short name for each of the data associated with the pipe It allows the data and characteristics in the file to be identified therefore there is a uniqueness constraint for each value It comprises CassesManual 2 0 0 b doc 15 06 2010 Page 5 of 88 A Cemagref CD Sciences eaux amp territoires The labels of compulsory or predefined data that are imposed The order of these data is unimportant IDT Pipe Identification Can be numeric or alphanumeric DDP Installation date This can be either a precise date or a year In the case of the latter the 1 of January of that year is used as the installation date LNG Length Quantitative data expressed in metres Decimal value MAT Material Qualitative data IDR Network identification The network identification is only compulsory in the multi network version When it is present in the mono network version it is considered as an additional qualitative characteristic DHS Date removed from service This usually results from the abandonment of a pipe but may also correspond to an important rehabilitation This can be either a precise date or a year In the case of the latter the 31 of
73. ncerned and the term window is used when it concerns a single pipe For pips still in service DHS is not defined therefore it can be ignored in the rules it appears for example min BRED DHS BRED The window of a pipe relative to the period only has a sense if the pipe respects certain constraints Pipes with DHS lt BRSD or DDP BRED don t have a break recording window and a refused in the importation Pipes with BWED lt OPSD or BWSD gt OPED don t have a break observation window and are excluded in the calibrations LEYP 1 Pipes with DHS lt PPSD or DDP PPED don t have a break prediction window Pipes with DHS lt PPSD or DDP gt PPSD are excluded from the prediction calculations LEYP 2 Note Pipes with PPSD lt DDP lt PPED can have a prediction window but are still excluded from the calculations Dates are at the day precision When imported dates are expressed in years they are converted according to the following rules DDP 1 January of the year DHS 31 December of the year DDC 1 July of the year Ages are expressed in decimal years calculated by dividing the number of days by 365 25 XTI1 3 ii Principal rules of dates Rules of the existence of periods DDP zx DHS BRSD lt BRED OPSD lt OPED PPSD lt PPED Rules of recorded breaks BWSD lt DDC sBWED DDP lt DDC t5 0 Rule of succession of observation and prediction periods
74. ng according to predicted break rate 66 XII S iv Number de breaks predicted ss 87 XIIL 6 COMPUTING INFORMATION 88 CassesManual 2 0 0 b doc 15 06 2010 Page 3 of 88 A Cemagref GC Sciences eaux amp territoires Installing Casses There are two versions of the software Casses Casses Mono and Casses Multi In this manual the term Casses is used without distinction and applies to both versions 1 1 Advised configuration Casses uses the virtual Java machine to operate Therefore you also need to install Java version 1 5 or higher software can be freely downloaded at http www java com fr Memory 512 Mb minimum 1024 Mb or more recommended For more details please refer to annex XIII 6 1 2 Installation under Windows operating system The software is protected by a USB key Actikey Important Install the software BEFORE inserting the key Launch by double clicking Cassesaaaa_x y z exe aaaa meaning Mono or Multi and x y z the number of the software version and follow the instructions of the installation program Choose the installation language English or French Accept the licence agreement Choose the installation directory by default C Program Files Cemagref Cassesaaaa Choose the Start up menu directory by default Cassesaaaa Click on lt Install gt to confirm the choices and install Casses and thea Click on lt Finish gt Casses can then be la
75. o a new Project without modification CassesManual 2 0 0 b doc 15 06 2010 Page 27 of 88 LE ER Sciences eaux amp territoires VII Create modify or delete covariates Covariates are the data attributed to pipes and can be distinguished as quantitative or qualitative Quantitative covariates represent a quantity numeric or a date Qualitative covariates have a limited number of values maximum 20 modalities After the importation of the data all the covariates present and useable in the pipes file are available to be used in a Project In the Exploration window they are regrouped as an object called Covariates with a Quantitative Covariates branch and a Qualitative Covariates branch It is possible to create new covariates from existing ones This functionality is accessible from the Creation menu or by right clicking on the Covariates object SS Casses software ExempleCassesMono2 0 0 ksp Creation Construction Tools P Create a pipe set Create a break set Covariates Merging Discretisation Type Unit Source Formula E Break set POSAS DATE Midiyyyy Imported Quantification Var QUANTITA mm Imported Formula i QUALITATI Imported J Environments C Covariates E Break data Combination Earliest date 7 1 1927 Latest date 7 1 1993 Copy table CassesManual_2 0 0_b doc 15 06 2010 Page 28 of 88 en
76. o the Breaks page The Actual breaks page presents information concerning breaks occurring during the validation period and is in the same form as the Observed breaks 1X 3 Exploring a Progression Selecting a Progression in the Exploration window gives access to two pages on the right under the tabs Configuration and Model IX 3 i Configuration page The Configuration page describes the different components of the Progression cf VII 4 IS Casses software ExempleCassesMono2 0 0 ksp EEK File Creation Construction Tools CJ Pipe set Configuration Model DY All pipes tron Description D Plastic Progression CHO1 Break set Sub project 5P87 96 TY all breaks Observation period starting date 2 5 1987 TS 87 95 Observation period stopping date 12 31 1996 Sai Environment Default Environment CT Environments Pipe set All pipes Default Eny Break set All breaks CI Obs87 oe CI SP87 9 e CH Cu Selected covariate set D Pro Short name Long name Type Unit Reference modality Source Formula amp CJ Iron Lat 87 95 DIA Diam tre QUANTITATIVE imported ariates LNG Longueur QUANTITATIVE i imported CJ Qualitative MAT Materiau QUALITATIVE FONTE Imported Cy MAT j Dj IDR D DIA Constraints o CI Quantitative IN LNG ZETAO Inf Dj Dia ZETA s0 Cl Break data Copy table jl Export
77. oessosesssoceessecessseeessssessseseesee 36 VIIL 1 THE ASSISTANT FOR CONSTRUCTING A PREDICTION Aa 36 VIIL2 CREATE AN ENVIRONMENT sn ma nee nn made ne naa te te en 37 CassesManual 2 0 0 b doc 15 06 2010 Page 2 of 88 A Cemagref GE Sciences eaux amp territoires VIIL3 CREATE A SUB PROJECT u 00000000000 0 aaa 3T VIIL4 CREATE A PROGRESSION EE 40 VIIL5 CALCULATE A MODED EE 43 VIII 3 E 43 VUS Consulting EE 44 VUES i Finalise a Progression ossos nn a E R TEE E ER RE ERR 46 VIIL6 CONSULTA VALIDATION sieneen an A ATA GG 47 VII or Left part of the Validation Cegedel 47 VII ot Right part of the Validation screen 49 VIIL 7 CREATE A PREDICTION gengen IT Bala see E GA AA GG 50 IX EXPLORING THE RESULTS cisscssssctens siscescssssscesccssiscsssssxtcsastosscsssebededacssdessscsesdecastessssasssscvccseccessssssvanates 52 IX 1 EXPLORING AN ENVIRONMENT 0 aa 52 IX 2 EXPLORING A SUB PROJECT 200000000000000 09 Naa 53 IX2 i Sub project without Validation sise 53 IX 2 ii Sub project with Moaltdoatton ss 54 IX 3 EXPLORING A PROGRESSION sise reernesse a 55 IX 31 Configuration E 55 IX 3 ii Model page rissin NIY aa GANG AIDA 55 IX 4 EXPLORING A PREDICTION SE aaa 56 IX 5 EXPLORING A VALIDATION 0a 59 X EXPORTING THE RESULTS sisssssssescesssssaccssessseescees socsoasessossssecesesssesi sasesesasescanassvscssieceaesesssasessseseseaeseeses 61 X 1 EXPORTING INTERMEDIATE RESULTS ccccccccscecsessssssec
78. or pipes with the modality considered or else the value 0 In the Statistical model a qualitative covariate with n modalities is represented by n 1 indicators associated with n 1 modalities with the non represented modality being referred to as the reference modality The period delimited by the start and end dates for which breaks associated with pipes in the network have been observed A past period during which break observed on the studied networks are taken into account in the calculations A covariate modality is pertinent for a collection of pipes if at least one of the pipes has this modality as a value Only pertinent covariate modalities can be included in calibrations A collection of adjacent pipeline segments for which all the characteristics other than length have the same value 15 06 2010 Page 71 of 88 Cemagref Qs Pipe break rate The number of breaks during a window of time divided by the pipe length and by the window duration It is expressed in breaks per kilometre per year 5 0 Lx Df Pipe characteristic Data attached to a pipe describing its physical nature its environment or its function Pipe network or network A collection of pipes used for distribution in a geographic zone and for which homogenous information is available In this application service pipes and fittings are not included Pipe observation window A period delimited by two dates during which the pipe is in service and
79. r anomaly detected the information being delivered in the same order as the first row and also separated by semi colons If at least one anomaly is blocking code SEV B the Continue button is disabled Otherwise clicking on this button allows the main screen to be accessed and for the Project to be started The last importation report produced by Casses is accessible by S Casses software File Creation Construction Tools Show gt Last calculation report Preferences Importation report It is also available as a text file in the directory specified in the Preferences under the name Rimp txt CassesManual 2 0 0 b doc 15 06 2010 Page 18 of 88 IV 1 ii Casses Multi The following dialog box appears Selecting files to import Pipe data file Path Ei Break data file Path E Multi network mode Network data file E Validate Cancel As well as the breaks and pipes files it is necessary to specify the name and location of the networks file Unchecking the box Multi network mode reverts to mono network mode Multi network mode proceeds as with mono network mode other than the dialog concerning the break recording period which does not appear CassesManual_2 0 0_b doc 15 06 2010 Page 19 of 88 LE Sciences eaux amp territoires IV 2 Open a Project S Casses software File Creation Construction Tools Close P
80. r specifying the nature of the Progression The next stage consists of selecting covariates to be included using the arrows or by double clicking Only covariates that don t have a unique value for every pipe in the Sub project are available To be treated by the calculation kernel a qualitative covariate with n modalities represented by the pipes in the Sub project is represented by n 1 indicators CassesManual 2 0 0 b doc 15 06 2010 Page 41 of 88 Gm CD Sciences eaux amp territoires One of the modalities is considered as the reference modality Each of the other modalities is represented by a quantitative covariate the indicator that has the value 1 for the pipes with this modality or otherwise the value O The reference modality is chosen with the help of a scrolling list to the side of the list of selected covariates Progression creation Comment Covariate set selection Eligible covariates Covariate set Name Name Reference modality DDP DIA DIA2 LNG MATSP MAT S Constrained model parameters eta0 fixed etal fixed Constrain the model Validate If required constraints for the calculations can be fixed by clicking on the button Constrain the model The following dialog box is displayed CassesManual 2 0 0 b doc 15 06 2010 Page 42 of 88 A Cemagref 5s5 Sciences eaux amp territoires
81. r totally copied by selecting the required cells and the lt CTRL gt C or using the contextual buttons intended for this purpose The importation report can be saved as a csv file with the help of the button shown during the importation phase It is also accessible via the menu Tools Show Importation report Pipe sets and Break sets can be exported in the Casses format by selecting them in the Exploration window and then right clicking Export The detail of the last calibration is accessible via the menu Tools Show Last calculation report S Last calculation report 7 0 Qualitative covariate MAT modality Amiante Ciment 1 Quantitative covariate LNG 2 Quantitative covariate DIA otal nb Failures 748 Total nb Replacements 0 Replacement rate 0 0000 _ thetaD 1 0000000000e 00 1 0000000000e 00 3 0000000000e 01 0 0000000000e 00 0 0000000000e 00 0 0000000000e 00 0 00 00000000e 00 0 0000000000e 00 Precision reached Precision reached LEYP Version 2 0 4 Nb Segments 1203 Nb Failures 748 77 1 2500e 02 1 5218e 05 1 2630e 04 1 5218e 05 2 4305e 09 1 1103e 08 1 2630e 04 1 1103e 08 9 0446e 06 Z7 validity true Initial theta 1 000000e 00 1 000000e 00 3 000000e 01 0 000000e 00 0 000000e 00 0 000000e 00 0 000000e 00 0 000000e 00 Convergence achieved in 187 iterations CassesManual_2 0 0_b doc 15 06 2010 Page 61 of 88 Gar QD Sciences
82. rding window as distinct from the number of breaks observed which only concerns breaks during the recording period Break or failure A rupture or leak on a pipe that leads to a repair Break set A collection of breaks selected from those in the Project It only contains breaks with valid data Calibration A part of a Sub project corresponding to an execution of the LEYP kernel applied to a set of covariates with view to calibrating a model LEYP1 calculation A calibration corresponds to a single execution for a fixed set of covariates Calibration period In Validation mode it is the observation period used for calibrating the model Collection of Pipes A group of pipes imported for use in a Project Whether they belong to one or several networks all the pipes are described with the same set of characteristics Covariate Data attached to a pipe that is the function of one or more pipe characteristics this function can be the ID A covariate has a unique value for a single pipe It can potentially but not compulsory be used in a model Covariate modality Any particular value of a covariate Covariates set A collection of covariates selected from those eligible associated with a particular pipe set Eligible covariate The covariates that do not have a unique value for the pipe set concerned Only eligible covariates can be part of a set of covariates linked to a calibration Environment A part of a Project regrouping t
83. ree text Commentn free text QUAN e IL DATE Empty if Unit if Format if qualitative quantitative date wxv987 17 12 1998 Signalled 4 gt 12 12 1998 The rules are generally the same as for the pipes file The short names imposed for the two compulsory data fields are IDT Pipe identification The pipe on which the break occurred DDC Break date In fact this is generally the day of the repair For calculation purposes no pipe can have more than one break on the same day If the break date is only given by the year unadvisable the 1 of July of that year is used in the calculations The breaks file can also include additional quantitative or qualitative data The constraints of uniqueness and for the short name must be respected II 3 ii Example CassesManual_2 0 0_b doc 15 06 2010 Page 8 of 88 amag ref ER Lee eaux igrel B Break ex WordPad File Edit View Insert Format Help Do 8hA re B CUB7 35 Original File from PHM format Additional data invented IDT DDC GENER DUR DSIGNAL Pipe ID Break date How generated Intervention time Date signalled DATE QUAL QUAN DATE d m y h d m y 100003 08 10 1992 search 4 100003 17 03 1993 signaled 4 12 03 2003 100004 01 05 1985 search 3 100004 08 10 1992 sectorisation 4 100007 19 12 1991 disruption 4 100010 13 09 1996 pertubation 4 100014 17 02 1988 search 4 100015 03 01 1987 search 7 100016 08 08
84. ring main abandonments i e most often replacements The study of such datasets frequently shows indeed that the oldest pipes oddly undergo very few failures This seems to be due to selection mains installed a long time ago but having undergone repeated failures are likely to have been abandoned for that reason before observation starts Observation is consequently subjet to the so called selective survival bias This leads to introduce the random service time T and the function z 0 1 which gives the probability that the main be repared following a failure undergone at age t whereas the probability that the main be abandoned following this failure is 1 en CassesManual_2 0 0_b doc 15 06 2010 Page 82 of 88 A CCemagret GE Sciences eaux amp territoires Assuming f 1 for any gt b i e beyond the observation window the conditional distribution of the number of failures within the prediction window remains negative binomiale N d No Nb NG m T gt b wb U otwidumg NB at m 1M jd ple b L oe duce The Casses software uses the model s t exp exp lo At 0 ER ER where the probability that the main be kept in service following a failure decreases with the age of the main gt O CassesManual_2 0 0_b doc 15 06 2010 Page 83 of 88 Cemagref CD Sciences eaux amp territoires XII1 5 Consistency of the validation XIIL 5 i Principl
85. roperties A dialog box asks you to identify the Project file that you wish to open and its location The files have the extension ksp SSelect a project file Lookin 3v2 z asa D ExempleCassesMono2 0 0 ksp IN ExempleFichierCasses2 0 0 DHS ksp D test1 ksp File Name Files of Type casses file ksp M Cancel Casses Multi allows files that were created with Casses Mono to be opened The inverse is not possible CassesManual_2 0 0_b doc 15 06 2010 Page 20 of 88 A V Exploring the data Casses comprises two modes Exploration that allows visualisation of the Project data and calculations and Construction that allows the calculations to be performed and predictions to be made V 1 Organisation of the explorer window The Exploration window is divided as follows On the left side a navigation tree that acts as a file manager On the right side one or more pages accessible by tabs at the top the contents of these pages is adapted to the object selected on the left side In many cases the right window is itself divided into top and bottom sections The information in the bottom section depends on the selection made in the top section SS Casses software ExempleCassesMono2 0 0 ksp File Creation Construction Tools C3 Pipe set Pipes Breaks IN All pipes IN Iron Description E Break set Environment Default Env
86. rt IN Iron Plastic Sub project SP87 96 Breakset Observation period starting date 2 5 1987 TN All break Observation period stopping date 12 31 1996 D 97 95 Environment Default Environment _ Pipe set All pipes LJ Environment Break set All breaks C Default g Short na Long na Minimum MaximumModality The corresponding text file is accessible in the directory specified in Preferences under the name Rcal txt VULSm Finalise a Progression Once the calibrations of a Progression are finished the Progression needs to be finalised in order to be able to make predictions In the case of an Automatic computation all the calibrations occur in succession until all the parameters are significant or they lead to non convergence In the case of a Semi automatic computation as long as advice can be applied a new calibration can be performed by clicking on the button Create new calibration situated on the right side of the Calibration screen in Construction mode To finalise the Progression click on the button Finalise situated on the right side of the Calibration screen in Construction mode If necessary a dialog box appears indicating the creation of new covariates that you can rename The short name must contain no more than eight characters and respect the constraints for short names described in the data importation section cf ID Fo
87. s Pipes Breaks and Prediction The Pipes and Breaks pages are generally organised in the same way as for the Environment or Sub project but contain additional information concerning the predicted breaks The significance of the abbreviated titles in the tables is as follows PN Number of pipes PN Percentage of the total number of pipes LNG km Length of the pipes concerned in kilometres LNG Percentage of the total pipe length OBN Number of breaks observed on the pipes concerned OBN Percentage of the total number of breaks observed PBN Number of breaks predicted on the pipes concerned PBN Percentage of the total number of breaks predicted MOBR Mean observed break rate in breaks per kilometre per year DMOBR Percentage difference from mean overall observed break rate MPBR Mean predicted break rate in breaks per kilometre per year DMPBR Percentage difference from mean overall predicted break rate It is important to note that the pipes and breaks of a Prediction and those of the Sub project in which it belongs are not necessarily the same Pipes with an empty prediction window and the breaks associated with them are not considered The Prediction page contains a top section describing the Prediction and its filiations The bottom section displays a table showing for each pipe in the Prediction its ID the covariate values the number and
88. s previously created from the deleted covariate CassesManual_2 0 0_b doc 15 06 2010 Page 35 of 88 LE CD xN Sciences eaux amp territoires VIII Constructing a prediction VIII 1 The assistant for constructing a prediction From the Exploration window the prediction constructing assistant is accessible from the Construction menu or by right clicking on the Environment object or any Environments within it S Casses software ExempleCassesMono2 0 0 ksp Bis File Creation Construction Tools C Pipe set Environment 3 All pipes 7 environments AN Iron B Plastic Environment Pipe set Break set C3 Break set Environnement p All pipes All breaks D All breaks Fonte lron All breaks C 87 95 C Environments E Covariates o c Qualitative covaria IN MAT Pipe set elaboration C IDR pia2 o CI Quantitative covar C LNG a _ pano D Copy table The Construction mode comprises six screens Environments Sub projects Progressions Calibrations Validation and Predictions Comment Break set elaboration ALL These screens are accessible sequentially using the gt gt and lt lt buttons or by clicking on the tabs Construction Environments Sub projects Progressions Calibrations Validation Predictions Environments Default Environment EE Iron J
89. ssesManual 2 0 0 b doc 15 06 2010 Page 73 of 88 err ref CC Sciences eaux H territoires XIIL2 Schematisation of calculation process SUB PROJECT Inherited covariates Own covariates gt y PROGRESSION Set of covariates zeta or forced covariates Constraints On alpha delta AUTOMATIC n calibrations Application of advice Creation of own covariates created by Progression LEYP1 without intervention F of the user E ROVE A VV FINALISE PROGRESSION he a CALIBRATION N SEMI AUTO p calibrations p Application of lp controlled by user v PREDICTION CassesManual 2 0 0 b doc 15 06 2010 Page 74 of 88 A Cemagref GC Sciences eaux amp territoires XIIL3 Definitions and rules relative to dates XIIL 3 i Denominations and calculations of dates and ages CassesManual 2 0 0 b doc 15 06 2010 Page 75 of 88 A Cemagref Sciences eaux amp territoires Abbreviation Definition Calculation BRSD Break record start date for network 1 January if imported in years BRED Break record end date for network 31 December if imported in years DDP Installation date of pipe 1 January if imported in years DHS Removal date of pipe 31 December if imported in years DDC Break date 1 July if imported in years t Age
90. ted have the same size range and include the upper bound CassesManual_2 0 0_b doc 15 06 2010 Page 30 of 88 A Cemagref QSD Sciences eaux amp territoires For each interval created you can modify its upper bound so long as it remains consistent with the adjacent intervals The inclusion or not of the upper bound of the interval must be chosen The lower bounds are deduced from the upper bounds A default label describes the interval and serves as the name for the modality of the created covariate This label is modifiable The new covariate must be identified with a short name This must contain no more than eight characters and respect the constraints for short names described in the data importation section cf II An optional long name may be used to further identify the covariate The description of how the new covariate is created is memorised this information is accessible on the right part of the screen when the covariate is selected in the Exploration window VII 3 Covariates creation Quantification Quantification allows a new quantitative covariate to be created from an existing qualitative one A numeric value is attributed to each of the source covariate modalities C Quantification Source covariate MAT v Modality scoring Modality Identifier Frequency FONTE FONTE 288 0 PVC PVC 790 0 Amiante Ciment ETERNIT 125 0 Covariate to be created Short name COR Lon
91. th the exception of semi colons and may even contain exclusively numbers which will be considered as text Qualitative data can contain up to 20 different values modalities II 2 ii Example E Pipe ex WordPad File Edit View Insert Format Help Cee 464 Aa By B IDT LNG MAT DDP C7 C8 C9 C10 Pipe ID Length Material Installation year Internal lining External lining Joint type Wall thickness QUAN QUAL DATE QUAL QUAL QUAL QUAN ele es Yee e wm 100001 290 FGLC 1958 black black mech 7 2 100002 160 FGLM 1878 black black lead 11 9 100003 30 FGLN 1900 black black lead 9 6 100004 170 FGL M 1924 black black lead 9 6 100005 120 FG5 1983 cement zinc auto 6 4 100006 90 FGS 1983 cement zinc auto 6 3 100007 120 FGL M 1878 black black lead 9 6 100008 140 FGS 1991 cement zinc auto 6 1 100009 100 FGL 1878 black black lead 9 6 100010 100 FG3 1991 cement zinc auto 6 1 100011 70 FG5 1982 cement zinc auto 6 2 100012 90 FGS 1988 cement zinc auto 6 1 100013 950 FGLM 1934 black black lead 19 7 100014 180 FGLM 1932 hlack black lead 9 6 100015 440 FGLC 1937 black black mech 7 2 100016 490 FGLC 1956 black black mech 8 100017 350 FGLC 1947 black black mech 7 2 100018 110 FGLM 1907 hlack black lead 10 9 For Help press F1 CassesManual_2 0 0_b doc 15 06 2010 Page 7 of 88 11 3 Breaks file II 3 i Structure Title free text Comment 1 free text Comment2 f
92. the LEYP model belong to the class of Pro portional Hazard Models parameterized by the regression coefficient vector 4 Counting process distribution The negative binomial distribution of the counting process constitutes a pivotal property of the LEYP model Tia m Tia doi exp A amp Z 1 exp a AG Z with A t Z A exp Z P du P exp ZTP 0 Pg N t Mm Z This explicit distribution property makes the counting process expectancy easy and quick to compute exp aA Z 1 a Bam This result can without difficulty be extended to the computation interesting from a practical point of view of the number of failures likely to occur in a prediction interval c d given the number of failures within the observation interval a b et Naay o Ks 0 C d t The conditional distribution of the process N d N c given N b Na mis negative binomial NG NG N b NG m Z NB GG om AXES MSE d Z ple Z ub Z ula Z 1 avec p t Z exp aAft Z 5 Parameter estimation procedure The Casses software expects two input datasets e one related to the mains e the other related to failures CassesManual 2 0 0 b doc 15 06 2010 Page 81 of 88 A CCemagret Gef A The mains are nin number and indexed by 1 n The installation date of each main must be documented as well as the abandoned date left empty if the main is stil
93. the model over the validation period ABN yr the annual number of actual breaks during the validation period PBR and ABR the predicted and actual break rates in breaks per km per year The table can be sorted by clicking on the column headers It can be partially copied select with the mouse and then lt CTRL gt C or copied in its entirety click on the table and then lt CTRL gt A and then lt CTRL gt C CassesManual_2 0 0_b doc 15 06 2010 Page 49 of 88 A Cemagref QSD Sciences eaux amp territoires VIIL 7 Create a Prediction To create a Prediction a finalised Progression has to be associated with a prediction period The prediction period is a period of any length that occurs after the observation period of the Sub project The creation of a new Prediction is carried out from the left of the Predictions screen in the Construction mode A scrolling list allows the selection of the finalised Progression of the model on which the prediction is based Click on the button Create new prediction Prediction creation Starting date of prediction Mangga n1996 period Stopping date of prediction period Comment Validate Cancel Firstly the Prediction must be named The name can be chosen freely the only constraint being that two Predictions in the same Progression cannot share the same name Next the dates for the start and end of the
94. unched from the Start up menu or by double clicking the file lanceur exe in the chosen installation directory The language used by Casses depends on the regional and language options selected by the user in the Windows Control Panel If the language option is French or German Casses is installed in French or German respectively If the language option is neither French nor German Casses is installed in English CassesManual 2 0 0 b doc 15 06 2010 Page 4 of 88 LE QSD Sciences eaux amp territoires Il Formatting the data to be imported IL1 General The imported files are of text csv Comma separated values format with the semi colon as the data separator Whatever the version of Casses at least two files are necessary The pipes file The breaks file For Casses Multi an additional file is needed The networks file Fach of these three files is structured in the same fashion An area for optional comments in the first rows e Four rows used to describe the data present and their format The correctly formatted data one row per record II 2 Pipes File IL 2 i Structure Title free text suggested to be used for the Project name Comment1 free text Comment2 free text Commentn free text al a QUAL ESCH CDe au a D TS DATE SE LE wxv987 i8barjols 1932 24 10 2006 34 87 Cl Urban 10 23 gt 28 06 1996 y At the beginning of the file green area comments ca
95. variate in the exploration window S Casses software ExempleCassesMono MIE File Creation Construction Tools J Pipe set All pipes B Iron IN Plastic E Break set J Environments LL Covariates Name Identifier Quantitative covariate 3 modalityfies c Qualitative covariates D MAT 40 11 0 40 110 Dior 1110 180 0110 180 C DIA Goen Belete ku Variable DIA discretised according to interval a Modify ariaple Iscretised according to intervals Break data 40 11 0 40 110 110 180 110 180 gt Copy table VIL 6 i Modify a covariate For quantitative covariates only the long name can be modified Covariate modification Short name LNG Long name Length Validate CassesManual_2 0 0_b doc 15 06 2010 Page 34 of 88 For qualitative covariates it is also possible to change the modality labels and the order in which they appear Modification of quantitative covariate Short name MAT Long name Material Modalities Identifier FONTE IFONTE PVC DVC Amiante Ciment ETERNIT Validate Cancel Two distinct modalities of the same covariate cannot share the same name VIL 6 i Delete a covariate Covariates that are not compulsory and that are not used in any Progression can be deleted This action is irreversible and does not affect the pipe sets nor covariate

Download Pdf Manuals

image

Related Search

Related Contents

Manual de instalación y mantenimiento Actuador electrico  Kota Kuru 14 and 20 sauna heaters  LED Festoon Lamp Holder User Manual  图片EN-2316SR 即将上市  

Copyright © All rights reserved.
DMCA: DMCA_mwitty#outlook.com.