Donetsk National Technical University
Master DonNTU Lazebnik Yelyzaveta ФИО: Lazebnik Yelizaveta
Theme: "Groundwater level forecasting in Donetsk region with help of neural-network technologies"
Scientific supervisor :Gubenko N.
E-mail: fantagirl@mail.ru

[Home] | [Links] | [Report on search] | [Library] | [Individual task] | [Master's portal]

Abstract

Groundwater level forecasting in Donetsk region with help of neural-network technologies

Русский
Contents
Lemma
Tasks of the work
Methods of research
Development of a system for monitoring groundwater state and usage
Preparation and management of computer of groundwater level
Current and planned results of the work

Lemma

Computer technologies are nowadays widely adopted for different problem modeling and forecasting in ecology and climatology. The characteristic property of these problem areas is small number of well-grounded models matching real data.
Industrialization in Donbass precipitated development of dangerous ecological and geological phenomena that had an influence on the state of environment. The problem came into particular prominence when coal-mining industry became restructuring. Contemporary mining engineers have no practical experience in addressing natural cataclysms, because uninterruptible operation of coal mines was provided by means of regular dewatering.
Therefore forecasting of groundwater level and the development of an information analytical system are of particular importance in Donetsk region.

Purpose of this work is to increase the hydrogeological safety on the territory of the Donetsk region on the base of establishment of system dependencies in the changes of aquifers horizons with the help of neural-network technologies.

Tasks of the work

1. Collection of hydrogeological data on a groundwater level in the Donetsk region, on the basis of information from Donetsk-geologia.
2. Classification of groundwater level data by homogeneous classes and creation of a database on the basis of obtained information.
3. Systems and structural data analysis and establishment of dependencies of changes in groundwater level within the framework of homogeneous classes.
4. Finding dependencies for groundwater level forecasting and estimation.
5. Development of the information analysis system on the base of neural-network models.

Neural-network
Image1 - Neural-network

Methods of research

For the accomplishment of the work the following methods are to be used:
- System and structural data analysis;
- Statistical methods, including cross-correlation and regressive analysis;
- Statistical analysis of data distribution;
- Stochastic methods of analysis of time series including the method of ARIMA & autocorrelation functions;
- Simulation of distributions, methods of neural-network technologies, development of computer models for the estimation of groundwater level.

For the accomplishment of the work the following tools are to be used:
o STATISTICA – for the data analysis and building of neural-network model;
o MySQL database management system – for database creation and maintenance;
o Borland Builder C++ programming environments for creation of the object modules of the information analyses system;
o Systems of object-oriented design for creation of computer models.
The research is based on database of groundwater level, which is watched by different organizations of the Donetsk during the protracted period.

Development of a system for monitoring groundwater state and usage

Groundwater monitoring is the special system of observations, which allows to carry out the tasking of processes, occurring in groundwater under the influence of human activity, to estimate the current state of groundwater and to forecast its changes for the rational use control of water sources.
The system of groundwater monitoring supposes the receipt of systematic objective information about the changes in the underground hydrosphere, the analysis of reasons for these changes and trends in their development. As a result the system allows making hydrogeological forecasting.
Conduct of groundwater monitoring and the up-to-date solution of the set problems is impossible without of creation of an automated information data bank and a system. Therefore during the development of the system it is necessary to foresee expenses on the introduction, installation, adaptation and maintenance of the software.

Preparation and management of computer of groundwater level

All the information accumulated in the process of conducting hydrogeological works is recording in special pre-machine documents for the subsequent input in a computer with the purpose of creation of the automated database which allows solve the problems at the contemporary level.
The purpose of the database creation is operative exposure of the anthropogenic influence on groundwater characteristics groundwater pollution control.
The input of information on groundwater will also provide the creation of a prognosis base.
Creation of the automated information system will allow carrying out direct connection with other subsystems of monitoring naturalof environment by means of creating banks of information exchange fund.

Current and planned results of the work

The following works are to be carried out:
o study of literary sources and normative documents on the questions of hydrogeological the situation on the territory of the Donetsk region;
o analysis of existing methods of forecasting and estimation of groundwater level, classification and identification of data;
o data capture on hydrogeological situation and preparation of files with the collected information;
o creation of a groundwater level database and a system of queries to prepare data for the subsequent analysis;
o development of a plan, choice of methods and formulation of content of the future master's degree work.

A course project on the theme of final master's degree work is the first stage on the stage of development, and will be drawn on his results in the future prosecution of master's degree dissertation.

[Home] | [Links] | [Report on search] | [Library] | [Individual task] | DonNTU << Master's portal