Προσομοίωση υπόγειας ροής με τη χρήση ελεύθερου λογισμικού ανοικτού κώδικα

Κολτσίδα, Ευγενία; Koltsida, Evgenia

dc.contributor.author	Κολτσίδα, Ευγενία	el
dc.contributor.author	Koltsida, Evgenia	en
dc.date.accessioned	2018-01-29T11:14:10Z
dc.date.available	2018-01-29T11:14:10Z
dc.date.issued	2018-01-29
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/46325
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.7454
dc.description	Εθνικό Μετσόβιο Πολυτεχνείο--Μεταπτυχιακή Εργασία. Διεπιστημονικό-Διατμηματικό Πρόγραμμα Μεταπτυχιακών Σπουδών (Δ.Π.Μ.Σ.) “Επιστήμη και Τεχνολογία Υδατικών Πόρων”	el
dc.rights	Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα	*
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/gr/	*
dc.subject	Αριθμητικά μοντέλα	el
dc.subject	Προσομοίωση υπόγειας ροής	el
dc.subject	Ελεύθερο λογισμικό ανοικτού κώδικα	el
dc.subject	Διαχείριση υδατικών πόρων	el
dc.subject	Ανάλυση ευαισθησίας	el
dc.subject	Groundwater modelling	en
dc.subject	Freewat	el
dc.subject	Modflow	el
dc.subject	Ucode	el
dc.subject	Free and open source software	el
dc.title	Προσομοίωση υπόγειας ροής με τη χρήση ελεύθερου λογισμικού ανοικτού κώδικα	el
heal.type	masterThesis
heal.classification	ΜΟΝΤΕΛΑ ΠΡΟΓΡΑΜΜΑΤΙΣΜΟΥ	el
heal.language	el
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	23-10-2017
heal.abstract	Η χρήση αριθμητικών μοντέλων στην προσομοίωση υδατικών πόρων συμβάλλει στην αποτελεσματικότερη κατανόηση των σύνθετων υδρολογικών και υδρογεωλογικών συνθηκών και ενισχύει τη διαδικασία λήψης αποφάσεων και στρατηγικών σχεδιασμού. Η προσομοίωση της υπόγειας ροής της παράκτιας περιοχής του Νομού Ροδόπης πραγματοποιήθηκε με την πλατφόρμα FREEWAT, η οποία ενσωματώνει τους κώδικες αριθμητικών μοντέλων προσομοίωσης ροής υπόγειων νερών (MODFLOW, USGS) στο ελεύθερο και ανοικτού κώδικα λογισμικό QGIS. Το υπόγειο σύστημα της περιοχής προσομοιώθηκε και ρυθμίστηκε για την περίοδο Απρίλιος - Οκτώβριος 2003. Η αναπαράσταση του εννοιολογικού μοντέλου επιτεύχθηκε με τις οριακές συνθήκες: σταθερής εισροής, ποταμού, κατείσδυσης από ατμοσφαιρικά κατακρημνίσματα, εξατμισοδιαπνοής και άντλησης. Μετά την πρώτη ρύθμιση του μοντέλου, εφαρμόστηκε ο κώδικα UCODE_2014 για τις διαδικασίες της ανάλυσης ευαισθησίας και της παραμετροποίησης. Για τη σύγκριση των προσομοιωμένων τιμών στάθμης με τις πραγματικές μετρήσεις υπαίθρου, χρησιμοποιήθηκε το Head Observation Package(HOB) το οποίο δημιουργήθηκε με τη βοήθεια του Observation Analysis Tool (OAT) κατά το οποίο 41 πηγάδια χρησιμοποιήθηκαν ως πηγάδια παρακολούθησης. Η αξιολόγηση των αποτελεσμάτων πραγματοποιήθηκε με τις μεθόδους Fit independent statistics (Dimensionless scaled sensitivity, Composite scaled sensitivity, Parameter correlation coefficient, Leverage και τους δείκτες Nash-Suttclife και mass balance error. Οι παράμετροι που εμφάνισαν τη μεγαλύτερη ευαισθησία ήταν η υδραυλική αγωγιμότητα και ο εμπλουτισμός από ατμοσφαιρικά κατακρημνίσματα, οι οποίες επιλέχθησαν να εκτιμηθούν κατά την παραμετροποίηση του μοντέλου. Μετά τη ρύθμιση των εν λόγω παραμέτρων, οι υπολογισμένες τιμές στάθμης παρουσίασαν ταύτιση με τις μετρημένες τιμές στο μεγαλύτερο τμήμα της περιοχής έρευνας, με εξαίρεση μικρές αποκλίσεις στο κέντρο της περιοχής όπου εμφανίζεται ο κώνος κατάπτωσης του υδροφόρου συστήματος. Η παρούσα εργασία αποτελεί μια παρουσίαση των δυνατοτήτων του λογισμικού FREEWAT και αποδεικνύει περαιτέρω τη δυνατότητα εφαρμογής του σε θέματα προσομοίωσης υδατικών πόρων.	el
heal.abstract	Water resources management can be enhanced by numerical models to achieve a more effective understanding of complex natural systems and improve planning strategies. Integrating GIS and hydrological codes, facilitates the use of complex modeling environments, allowing to store, manage and visualize large spatial datasets. FREEWAT (Free and open source software tools for water resource management; Rossetto et al., 2015) is a public domain, QGIS-integrated platform, developed to simulate several hydrological processes in order to address decision making in water resource management (De Filippis et al., 2016). The groundwater flow simulation of a sedimentary aquifer, located in Northern Greece, at the south-west coastal part of the Prefecture of Rhodope, was achieved by the application of the FREEWAT platform and the MODFLOW-2005 code (Harbaugh, 2005). The modules for data pre-processing and model implementation, which were used in the present postgraduate thesis are mentioned below.  The AkvaGIS module provides several tools for the analysis and interpretation of hydrogeological data and generation of thematic maps for implementation of conceptual models  The Observation Analysis Tool (OAT) is a pre-processing tool allowing to import, analyze and visualize time-series data which are then can be used for model calibration  MODFLOW-2005 (Harbaugh, 2005) is a code developed by USGS and is implemented to perform groundwater flow simulation in the saturated and unsaturated zones  Sensitivity analysis and parameter estimation are performed using the UCODE-2014 (Poeter et al., 2014) which improves the model fit, by reducing the difference between model-simulated heads and flows and the observed data The conceptual model consists of two layers: an upper layer, which includes impermeable formations and provides the confined conditions of the aquifer system and the underlying formation which is composed of permeable alluvial deposits (approximately 50 m). The bottom of the aquifer system is composed of a grey-green clay layer, which extends up to the whole area of investigation. The recharge of theaquifer system is achieved by direct infiltration from precipitation, irrigation returns, percolation from surface water bodies of the area and the lateral hydraulic contact with the northern alluvial cone of Kompsatos river (Kallioras, 2008). The simulation involves the period between April-October 2003 and the model is divided into two stress periods. The 1st SP (steady-state) lasts one day, during which no stresses (recharge, pumping wells, etc.) affect the hydrogeological system. The 2nd SP (transient) lasts 181 days, during which 411 pumping wells penetrating the deepest layer are activated. The simulation was accomplished by the use of the following packages:  General head boundaries (GHB) for the parts of the aquifer which are in hydraulic connection with the alluvial cone of Kompsatos river, at the parts of Vistonida lagoon inlet and at the southern part of Lake Ismarida.  River boundaries (RIV) across the Vozvozis river  Well Package to simulate wells that withdraw water from the aquifer at a constant rate during a stress period (WELL)  Recharge Package (RCH) to simulate the direct infiltration from precipitation  Evapotranspiration Package (EVT) to simulate the conjunctive effects of plant transpiration and direct evaporation from groundwater  No flow boundaries at the parts of the aquifer where no flow occurs (inflow or outflow) In order to evaluate the results and compare the simulated values with observed values, the Head Observation Package (HOB) was developed using the Observation Analysis Tool (OAT) where 41 wells were used as monitoring wells. After the first run of the model, a sensitivity analysis was performed using the UCODE_2014 code (Poeter et al., 2014) to gain information on the quantitative relationship between parameters and observations. The parameters taken into consideration are: hydraulic conductivity (HK) which was divided in 4 zones, specific storage (SS) and direct infiltration from precipitation (RCH). The statistical analysis of the results was carried out using a series of fit independent statistics (Hill and Tiedman, 2007) such as: dimensionless scaled sensitivity, composite scaled sensitivity, parameter correlation coefficient and leverage. The most important parameters with the highest css values affecting the results were the first zone of hydraulic conductivity of the aquifer (HK1), followed by the direct infiltration from precipitation (RCH) and the second zone of hydraulic conductivity (HK2). A parameter estimation was conducted with the same code, to obtain the set of parameters which give the best fit between observed and simulated values of the square weighted residuals regression (Foglia et al., 2017). Parameter estimation refers to the process of optimizing a selected set of parameters using a nonlinear regression which minimize the objective function. The objective function was defined as the sum of squared weighted residuals. In the parameter estimation process, the three parameters with the highest css values were used. The parameters SS, HK3 and HK4 were excluded from further estimation because of their low sensitivity. The high correlation between HK1 and RCH was addressed by imposing constraints based on existing data. After the parameter estimation, the Nash-Sutcliffe coefficient and the mass balance error accounted for 0.74 and 0.8 respectively, indicating that the simulated level values are slightly greater than the observed values. The calculated values were compared with the measured values of 13 October 2003 and it was concluded that at the eastern, NE, western, NW and southern boundaries of the study area a coincidence of measured and calculated groundwater levels occurred. On average, the model results were within 4.15 m of the measured heads. Small declinations between the calculated and the measured groundwater levels were observed at the central part of the area at the point of formation of the cone of depression (wel_33), at the south-west part of Vistonida lake (wel_44, wel_11) and at the south part of Aspropotamos river (wel_43).	en
heal.advisorName	Καλλιώρας, Ανδρέας	el
heal.committeeMemberName	Μπαλτάς, Ευάγγελος	el
heal.committeeMemberName	Μακρόπουλος, Χρήστος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Πολιτικών Μηχανικών	el
heal.academicPublisherID	ntua
heal.numberOfPages	140 σ.	el
heal.fullTextAvailability	true