VGEST issueshttps://gitlab.irstea.fr/in-wop/vgest/-/issues2020-12-10T20:01:53+01:00https://gitlab.irstea.fr/in-wop/vgest/-/issues/6VGEST crashes for time series longer than 5000 years2020-12-10T20:01:53+01:00Dorchies DavidVGEST crashes for time series longer than 5000 yearsAs @quan-van.dau has reported, a try with a 7000 years time series on Paris station crashes with this message:
```
vgest
Le contenu du fichier MODE.txt entraŒne la lecture des options de calcul
dans le fichier CHOIX.txt.
Rang du calcu...As @quan-van.dau has reported, a try with a 7000 years time series on Paris station crashes with this message:
```
vgest
Le contenu du fichier MODE.txt entraŒne la lecture des options de calcul
dans le fichier CHOIX.txt.
Rang du calcul : 351 ; systeme = PARIS_05[YONNE, SEINE, AUBE, MARNE] ; parametres barrages et reseaux : 1 et 1 ; repartition variable (2
- fonction de V et Tpot(moyen)) ; apports = 7000ys_leap.txt ; objectif = BATCH (Qmax) ; 17/01/2000-22/12/8999
calcul de Vobj en chrono inverse : termin‚ - fonction
$0040787A
$00412631
```Dorchies DavidDorchies Davidhttps://gitlab.irstea.fr/in-wop/vgest/-/issues/1The format of Chrono.txt and PaChrono.txt is not constant2020-10-23T17:37:34+02:00Dorchies DavidThe format of Chrono.txt and PaChrono.txt is not constantThe width of the columns is not constant and since it is a fixed format, it is very difficult to read.
The columns with real data should be all in free format with a 22 characters width.The width of the columns is not constant and since it is a fixed format, it is very difficult to read.
The columns with real data should be all in free format with a 22 characters width.Dorchies DavidDorchies Davidhttps://gitlab.irstea.fr/in-wop/vgest/-/issues/3Succes indicator is wrong when QXdef2 is negative2020-12-14T09:13:10+01:00Dorchies DavidSucces indicator is wrong when QXdef2 is negativeSee the conclusion of #2 for more details.See the conclusion of #2 for more details.https://gitlab.irstea.fr/in-wop/vgest/-/issues/7Split VGEST into several elementary tools2021-03-25T11:41:52+01:00Dorchies DavidSplit VGEST into several elementary toolsAll the outputs of 'VGEST' are not of interest for 'Irmara' so for avoiding bugs and safe time computing it's interesting to target what parts of 'VGEST' are of interest and to be able to run the different parts separately.
In fact the ...All the outputs of 'VGEST' are not of interest for 'Irmara' so for avoiding bugs and safe time computing it's interesting to target what parts of 'VGEST' are of interest and to be able to run the different parts separately.
In fact the only product that is really crucial is the outputs of the reverse chronology simulation which contains all necessary informations and indicators for calculating iso-frequency curves.
As all intermediate data are recorded into files (especially binary files in the "TEMPORAI.RE" folder), it does not seem too difficult to run each step of 'VGEST' separately.
There are two options to proceed this splitting:
- Split the code into several executables
- Add options in the execution to choose which step to run
The first option has the advantage of forcing to clean the code but could be a long process. The second one is more simple to do but force to deal with a big monolithic peace of code with the risk of unexpected interactions.Dorchies DavidDorchies Davidhttps://gitlab.irstea.fr/in-wop/vgest/-/issues/10Set up docker container for testing and deploiment2022-09-01T10:01:42+02:00Dorchies DavidSet up docker container for testing and deploimenthttps://gitlab.irstea.fr/in-wop/vgest/-/issues/8Remove leap year handling2021-03-26T16:54:39+01:00Dorchies DavidRemove leap year handlingReasons:
- All other tools around (streamflow generator) don't use leap years.
- Date handling is OK from 1600 and it's a limitation for using more than 8000 years.
- Conversions are painful and source of bugs.Reasons:
- All other tools around (streamflow generator) don't use leap years.
- Date handling is OK from 1600 and it's a limitation for using more than 8000 years.
- Conversions are painful and source of bugs.Dorchies DavidDorchies Davidhttps://gitlab.irstea.fr/in-wop/vgest/-/issues/9Output a PaChrono.txt with Vobj only2021-03-26T16:54:39+01:00Dorchies DavidOutput a PaChrono.txt with Vobj onlyTo speed up execution and readingTo speed up execution and readingDorchies DavidDorchies Davidhttps://gitlab.irstea.fr/in-wop/vgest/-/issues/5Crash of VGEST on 1000 years flow database2020-11-12T17:36:16+01:00Dorchies DavidCrash of VGEST on 1000 years flow databaseTry with 1000 years generated flows:
```
Rang du calcul : 339 ; systeme = PARIS_05[YONNE, SEINE, AUBE, MARNE] ; parametres barrages et reseaux : 1 et 1 ; repartition variable (2 - fonction
de V et Tpot(moyen)) ; apports = synth_gen...Try with 1000 years generated flows:
```
Rang du calcul : 339 ; systeme = PARIS_05[YONNE, SEINE, AUBE, MARNE] ; parametres barrages et reseaux : 1 et 1 ; repartition variable (2 - fonction
de V et Tpot(moyen)) ; apports = synth_gen_daily_1000ys.tsv ; objectif = BATCH (Qmax) ; 17/01/2000-23/04/2999
calcul de Vobj en chrono inverse (it‚r. en cours sur Vobj initial : 1)Runtime error 216 at $00424125
$00424125
$0042A195
$0040658C
$00412631
```Dorchies DavidDorchies Davidhttps://gitlab.irstea.fr/in-wop/vgest/-/issues/2Can QXdef1 be lower than QXDef0?2020-11-07T20:52:20+01:00Dorchies DavidCan QXdef1 be lower than QXDef0?In some simulations QXDef1 is lower to QXDef0 and QXDef2 was negative.
Regarding the definition of these variables, described in the J-C Bader's report of 2016, this should not be possible:
> To sum up, the best possible satisfaction ...In some simulations QXDef1 is lower to QXDef0 and QXDef2 was negative.
Regarding the definition of these variables, described in the J-C Bader's report of 2016, this should not be possible:
> To sum up, the best possible satisfaction of the downstream objective cannot sometimes avoid certain total failures QXdef1, corresponding to the lack or excess of flow obtained in the river at the downstream station in relation to the objective QXobj. These total failures QXdef1 are the sum of :
> - QXdef0: absolute unavoidable failures due to physical constraints and instructions at intakes and releases, independent of the volumes of water present and reservoir capacities, as well as of the concerted reservoir management method;
> - QXdef2: additional failures, of which an unavoidable part, due to the overall lack of reservoir capacity, is independent of the management method, while the other part, theoretically avoidable, is linked to the distribution of the volume of water present between reservoirs and depends on the management method adopted.Dorchies DavidDorchies Davidhttps://gitlab.irstea.fr/in-wop/vgest/-/issues/4Add new reservoir ruleset: 5th ruleset without gradient limitation on Yonne lake2020-11-12T18:10:58+01:00Dorchies DavidAdd new reservoir ruleset: 5th ruleset without gradient limitation on Yonne lakeThis is the ultime reservoir ruleset! No local constraints except the mandatory ones:
* Minimum flow immediatly downstream the inlets (and Yonne outlet)
* Minimum and maximum flow in inlets and outlets (physical constraint)This is the ultime reservoir ruleset! No local constraints except the mandatory ones:
* Minimum flow immediatly downstream the inlets (and Yonne outlet)
* Minimum and maximum flow in inlets and outlets (physical constraint)Dorchies DavidDorchies David