Monte Carlo Simulation for Probabilistic Modeling of the Input DataUncertainties in a Thermal Power Plant Site Selection Process

Spatial position of a power plant can have an important role on its efficiency, as well as on its socioeconomical and environmental impacts. GIS is an efficient and powerful tool for integration of the related data and criteria for selection of a proper site. Because of the importance of considering uncertainties in GIS-based data integration, effects of errors in input data for thermal power plant site selection, in Bushehr province has been considered in this paper. Maps of the important factors in site selection including the slope, elevation, distance to rivers, cities and gas pipes have been prepared andtheir weights have been defined by using the expert's knowledge. Geometric uncertainties of the related data layers have been quantified by using the knowledge of errors due to the scale and digitizing errors. The extracted information from the maps of 1:1000000 and 1:250000 scales and Monte Carlo simulation approach have been used to generate the probability maps of uncertainties of each factor. Index Overlay and Boolean models have been applied to combine the factor maps and to produce the corresponding output maps. The frequency of selection of each site in different runs of the Boolean model with different input maps has been used to compute the probability or reliability ofselection of each candidate site. By integration of the uncertainty and index overlay models the suitability values and uncertainty of each candidate site has been determined. Results of the research have shown that valuable data about the reliability of different sites can be achieved and the resulting maps of uncertainties are very useful for more precise discrimination of reliable sites in a site selection problem. In addition, by evaluation of the results it is very straightforward to determine which data layers and where needs higher precision to improve the reliability of the site selection process.