Date

2014

Author

Eratak, Özlem Deniz

Metadata

Show full item record

Item Usage Stats

384
views

119
downloads

Cite This

Safety in underground coal mining has become an important issue because of increasing number of accidents. There are many different hazards may cause these accidents and the most efficient method for coping with risks is the use of risk management techniques. In this thesis, accident data including workday losses, age of the injured, organ affected by accident, season, shift and reason of the accident was collected from TKI mines (ELI- Soma Eynez and GLI Tuncbilek) and TTK mines. Those variables were initially analysed by using basic statistics to have the general information about the most hazardous conditions. Comparison was made between these mines. Then, a risk analysis study was performed using severity, probability and exposure components. Risk matrices were developed and the most hazardous places were determined together with comparison of those three mines. Probability analysis was performed to understand the expected accident frequencies in each mine and reliability in a time period between accidents. The study was also targeted to develop a model for severity component using three different methods which are regression, neural network and fuzzy logic techniques. These techniques applied to every mines data and decision analysis was made to choose the most suitable technique by comparing the results. Finally, future accident estimation models were developed with regression and neural network techniques based on the data such as number of accidents, deaths, injured, total working hours, total workers and total raw coal production of those mines.

Subject Keywords

Coal mines and mining, Coal mines and mining, Regression analysis., Fuzzy logic.

URI

http://etd.lib.metu.edu.tr/upload/12616827/index.pdf
https://hdl.handle.net/11511/23287

Collections

Graduate School of Natural and Applied Sciences, Thesis