Umeritev modela toka podzemne vode na Dobovskem polju s pomočjo metode »Pilotnih točk«

Koren, Eva (author), Vižintin, Goran (mentor)

Na območju Čateškega in Dobovskega polja sta bila do sedaj, za potrebe gradnje hidroelektrarne Mokrice (HE Mokrice), izdelana dva matematična modela. Oba temeljita na metodi končnih diferenc. Prvi model je osnova celotnega modeliranja, drugi pa je nadgradnja prvega. Oba modela sta bila izdelana z orodjem ModFlow in ročno umerjena s spremembami območij različnih K-jev. Drugi model je vključeval nova dognanja in predstavlja končno sliko stanja podzemnih vod ter je osnova za projektiranje. V okviru modeliranja sta bili najprej modelirani srednja in visoka gladina podzemne vode za primer brez HE Mokrice, temu je sledilo modeliranje obeh omenjenih stanj po zapolnitvi akumulacije. Oba modela modelirata stacionarno stanje. Modela pokažeta, da so vplivi pri srednjem stanju večji kot pri visokem stanju gladine podzemne vode. V magistrski nalogi sem ponovila modeliranje srednjih in visokih gladin podzemne vode za stanje pred zapolnitvijo akumulacije HE Mokrice. Pri tem sem uporabila program FeFlow, ki deluje na osnovi metode končnih elementov. Umeritev modela pa sem izvedla s pomočjo pilotnih točk. V času izdelave naloge sem najprej ugotovila, da popoln prenos modela iz ModFlowa v FeFlow ni mogoč, glavni vzrok je v geometriji in načinu generiranja mreže, ki sta pri modelih povsem različna. Zato sem pri prenosu iz ModFlowa v FeFlow pazila, da sem se v FeFlowu geometrijsko čim bolj približala obliki v ModFlowu, pri čemer sem žal žrtvovala del kakovosti modela, ki izhaja iz boljšega geometrijskega opisa problema metode s končnimi elementi. Kljub temu sem po moji oceni izdelala geometrijsko dovolj podobna modela in še ohranila »dobre« lastnosti metode končnih elementov. S primerjavo sem ugotovila, da kljub podobnim vrednostim NRMS modela zlasti v severnem delu dajeta nekoliko drugačne rezultate. Razloga za to sta po mojem mnenju izbrana metoda umerjanja in uporaba dodatnih podatkov iz dveh piezometrov, PM-13/07 in PM-15/07, v FeFlowu, ki v ModFlowu nista bila vključena. Hkrati pa se je še enkrat pokazalo, da ima vsaka metoda umerjanja in modeliranja svoje prednosti in slabosti.

• Language: Slovenian
• Keywords: modeliranje toka podzemne vode, metoda končnih elementov in končnih diferenc, kalibracija modela, linearna parameterizacija in pilotne točke
• Work type: Master's thesis/paper
• Organization: NTF - Faculty of Natural Sciences and Engineering
• Year: 2018
• PID: 20.500.12556/RUL-101351
• COBISS.SI-ID: 1728351
• Publication date in RUL: 30.05.2018

Secondary language

• Language: English
• Title: Calibration of groundwater flow model on Dobovskem field by using the method of “pilot points

Abstract

In the area of Čateško and Dobovsko field for the needs of the construction of the hydroelectric power plant Mokrice (HPP Mokrice) until now two mathematical models have been made. Both models were made on the basis of final difference method. The first model is a base of entire modeling, the second one is an upgrade of the first. Both models were made in program ModFlow. They were calibrated manually with the modification of areas with different values of k. The second model was including new events which were represented a final picture of groundwater status which is a base for the design. In the modeling process, the average and high levels of groundwater were modeled for the example without the Mokrice HPP followed by the modeling of the two before mentioned states after the accumulation. Both models are modeling stationary condition. The models show that influence is higher at the average condition than at high condition of groundwater level. In my master's thesis, I have had repeated the modeling of average and high levels of groundwater for the state before filling the accumulation HPP Mokrice. I repeated the modeling by using a program FeFlow which is basing on the finite element method. For calibration, I used a pilot points method. At the time of my work, I first find out that the perfect transfer from ModFlow to FeFlow is not possible. The main reason is the differences in geometry between the programs and the difference at the generating the mesh which is very different between the programs. While transferring the data from ModFlow to FeFlow I was trying to retain the geometry from ModFlow as close as possible to FeFlow. During the procedure, I have unfortunately sacrificed a part of the quality of the model which results from a better geometric description of the problem with the finite element method. Despite that I asses that in my estimation I have produced a geometrically similar model and still retained the »good« properties of the finite element method. Based on the comparison I found that despite similar values of NRMS models especially in the northern part give slightly different results. In my estimation, the reason is in the selected calibration method and the use of additional data from two piezometers PM-13/07 in PM-15/07 in FeFlow which were not included in ModFlow. At the same time once again it has been shown that each calibration and modeling method has its advantages and disadvantages.

• Keywords: groundwater flow modelling, finite element end finite difference method, model calibration, linear parameterization and pilot points