Home Articles Gis Integration Method For Thermal Power Plants Siting Case Study: Fars Province...

Gis Integration Method For Thermal Power Plants Siting Case Study: Fars Province Southern Part of Iran

Hossein Yousefi1, Sachio Ehara1, Hossein Yousefi1
1Department of Earth Resources Engineering, Kyushu University

Reza Samadi2, Tika Sohrab2, Hamid Khadem2,
2Iran Energy Efficiency Organization, Shahrak Ghods, Tehran, Iran,
Email: [email protected]

In this study, a Geographic Information System (GIS) was used as a decision-making tool to target potential power plant sites in Fars province, Southern part of Iran. The aims of the study are to identify suitable areas to install a thermal power plant (TPP), as the base study for the future investigations and development. In Iran, the developer must apply for siting and EIA (Environmental Impact Assessment) license from the Department of Environment (DOE) when the construction of a new TPP is over 100MW prior the development. Recent legislations have increased the role of GIS in siting studies in Iran therefore this paper tries to illustrate not only factors that are considered in siting studies, but also the role of GIS to locating some reliable sites with reasonable cost and minimal environmental impacts. Fars province is located at southern part of Iran in a critical situation and because of its unique location; land use planning studies were not successful to find industrial sites for the future. In this research, conventional models for combining factor maps have been investigated and index overlay and fuzzy logic models were selected. Also an integration model using of appropriate models have been proposed. For experimental case study, the suitable potential map of Fars province in the south of Iran, with appropriate methods in different inference networks have been produced and an appropriate inference networks were selected. Results of the selected network are in a good accordance with field tripe observation. Proposed model capability with required variation can be used for other studies in the country. In the case study, the scale of maps for preliminary siting was 1:250,000 and in detailed study was 1:25,000 and results of the study showed four suitable sites for construction of thermal power plant.

Keywords: Iran, Fars, GIS, Site selection, Power plants

In the last few decades, power industries have been developing production plants and transmission systems to catch up with the rapid growth power demand. Meanwhile, suitable sites for new power plants have been getting limited due to the development of countryside (rural areas) and the rising concern over environmental and legal issues. The location of a power plant has significant effects on the efficiency of electricity generation, environmental impacts, price of 2 electricity, transmission and distribution lines. Therefore the selection of the location for a new power plant should be done very carefully and take into account many affecting factors (Yousefi et al, 2007).

The Iranian Ministry of Energy (MOE) plans to develop the electricity generation capacity and distribution network. In this plan the construction of the fossil fuel power plants is the most significant part of the electricity production expansion program. As national regulations in Iran, the power plant developer must apply for site selection study and EIA license from the Department of Environment (DOE) when planning to construct a new power plant over 100MW prior the development. Power plant site selection combines several technical and environmental data layers to locate suitable areas. Many of the factors for site selection are essentially spatial, and the data is from different sources and different scales. Therefore GIS with potential for storing, updating, retrieving, displaying, processing, analyzing and integration of different geo-spatial data, should be used to define the suitability of different locations for construction of power plants (Clarke, 1997). GIS was used as a decision-making tool to determine the spatial association between technical evidence layers including power generations, fuel supply, water supply and environmental evidence layers such as forests, lakes and slopes.

In this study, the necessary conditions for the establishment of thermal power plants are comprehensively studied including geology, climate, water resources, fuel supply, access roads, existing electric power network, topography, and etc. Finally different models and methods for integration of the data layers using ArcGIS in power plants site selection are investigated.

According to the characteristics of factors and their effect on power plant siting, two different types of maps were generated, binary and factor maps. These maps were integrated using Boolean “AND” and index overlay operators, respectively. In the first phase, data were used in the scale of 1:250,000 with binary system and in the second and more detailed phase, 169 data layers in the scale of 1:25,000 with overlay system were used. Fig.1 shows the study framework.

Fig. 1: Framework and study plan in the TPP siting research
The study area is Fars province, one of the 30 provinces of the country which is one of the main and environmentally fragile areas in southern part of Iran with an occupied area of 122,780 Km2. The province has population of, 323,626. Shiraz is the major city of the province, and one of the beautiful, historical cities in the world. The location map of the study area is shown in Figure 2.

Fig.2: Location of the Fars province southern part of Iran

Different models exist for mapping suitable potential. These models are based on data-driven and knowledge-driven. In this section, conventional models for integrating data in site selection study are investigated.

Boolean modeling involves the logical combination of binary maps resulting from the application of conditional “AND” and “OR” operators. In practice, it is usually unsuitable to give equal importance to each of the criteria being combined. Evidence needs to be weighted depending on its relative significance. Expert knowledge cannot interfere in this model.

In Weight of Evidence models limitations recognition criteria by using the known thermal power plant (control points) and statistical methods, were weighted and integrated. This method only applied in regions where the response variable (e.g. distribution of known TPP in the case study) is fairly well known. This method is not always applicable in siting program in detailed stage but this model in the small scale is appropriate method (Yousefi et al, 2006).

In Index overlay method, each class of every map is given a different score, allowing for a more flexible weighting system and the table of scores and the map weights can be adjusted to reflect the judgment of an expert in the domain of the application under consideration. At any location, the output score, S, is defined as (equation 1) (Bonham-Carter, 1991)

Where the Wi is weight of i-th factor map, Sij is the i-th spatial class weight of j-th factor map and S is the spatial unit value in output map.

In the Fuzzy Logic method, total of sheet maps (fuzzy membership) based on the significance distance of features are weighted (for each pixel or spatial position particular weight between zero to one is appointed). Five operators that were found to be useful for combining site selection datasets are the fuzzy AND, fuzzy OR, fuzzy algebraic product, fuzzy algebraic sum and fuzzy gamma operator. These operatore are briefly reviewed here (Valadan Zoej et al, 2005).

The fuzzy AND operation is equivalent to a Boolean AND operation on classical set. It is defined as (equation 2)

Where WA, WB,… is the fuzzy membership values for maps A, B, … at a particular location. This operation is appropriate where two or more pieces of evidence for a hypothesis must be present together for the hypothesis to be accepted.

Evidence map can be combined together in a series of steps, by using an inference network. The inference network an important means of simulating the logical thought processes of an expert (Noorollahi et al, 2007). Concerning the rule of conceptual modeling, the expert knowledge, existing data and characters of the models for combining factor maps, Boolean “AND”, Index Overlay and Fuzzy Logic models were selected in TPP site selection study.

Considering the study area size, the available spatial data in the country and the diversity of parameters, it was decided that the study should be done in two phases by using two different scales of data. Firstly spatial data on the scale of 1:250,000 with precision about 100 m were used and generally suitable areas were selected by applying the technical and environmental evidences.

Secondly, detailed site selection was carried out using data layers in the scale of 1:25,000 with precision about 10 m in the previous selected locations. The model was the same in both phases but in the second phase more data were employed. Figure 3 presents the conceptual model of the Study.

Fig. 3: Conceptual model of the TPP siting study
After comprehensive study about the existence of digital data in the country and required data for the siting study all data categorized into four major classes, 13 classes and several subclasses. For the integration purposes, all data layers are classified into two main evidence data sets including environmental and technological data sets. Table 1 shows the limitation factor maps with their buffer size.For site selection the similar maps must be combined using different models of maps combination. These models are based on data-driven and knowledge-driven method. In this study, conventional models for integrating data in power plant site selection were used, such as Bolean, Index Overlay, and Fuzzy Logic methods. Then, according to the characteristics of parameters and their effect on power plant siting, two different types of factor maps were generated: binary and weighted factor maps. These maps are integrated using Boolean and index overlay operators, respectively in both phases of the study.

In a binary map, the areas with restrictive condition are given the value of zero and the suitable areas are assigned the value of one. For example, the areas with elevation more than 1800 m a.s.l. is represented with zero value (not-suitable) and the areas with elevation less than that are represented with the value of one (suitable). It means, such a map defines and separates the area that cannot be used for the power plant siting.

In this research these binary maps are overlaid where input maps can be integrated by using logical operators. Then all of the limitation maps integrated in one map as a binary map. The limitation data layers and their assigned criteria are presented in Table1. By employing restrictive layers using defined criterion in Table 1, the suitable area based on these layers in the first phase selected. The employed major classes, classes, subclasses and their limitation with buffer size were summarized in Table 1, like faults which surrounded with 1 km buffer size.

Table 1: Evidence data layers and their limitation with buffer size
Generally, both the construction and operation of a thermal power plant requires the existence of some conditions such as water and fuel resources. There are still other criteria that although not required for the power plant, yet should be considered, because some criteria have a positive or negative effect on the suitability analysis, such as land use and population center. The effect of both parameters can be modeled by giving them appropriate weights. For example, consumption center and gas pipe line layers are more important than land use layer. Therefore in this study we have two type of weighting, between important layers and between different classes inside one layer. In this study, factor weights are defined to describe their significance in the selection of proper location for power plants. These factors and their given weight are listed in Table 2. On the other hand these weights show their importance in thermal power plant site selection.

Associated values in a factor map represent both the relative importance of the factors and the relative values corresponding to different locations on the map area. For example, in main road factor map, associated values are decreasing when the distance from existing main road line is increased. In fact, each value represents the suitability of the pixel area for the thermal power plant regarding to the related criteria. In this research, all weighted values are between 0 – 1. As mentioned before there are classifications, inside of the subclasses that also are given a weight. Some of these parameters and their given weights are listed in Table 3. The weights are knowledge- driven and experts in subject are agreed.

Table2: Weight of the important features for three types of thermal power plant

Table 3: Classification of subclasses in factor maps and given weights
As mentioned above, the integration of the resulted factor maps was carried out in two stages:

– Limitation maps are overlaid using the Boolean “AND” operation which resulted in the selection of areas that have value of ‘one’ in all limitation maps.
-Factor maps are integrated with the Index Overlay method.
Finally in the first phase when final binary map and final factor map with precision of 100 m,overlaid, 176 sheet of 1:25000 scale map of Iran for detailed study were selected.

In the second phase the same model runs with more features in precision of 10 m, then the resulted map was investigated and 0.17% of the Fars province was selected as suitable area which covers about 208 km2 in 49 sites. The majority of suitable areas were located around Fasa district, where electrical energy demand is more than other places. For final selection, satellite image (Landsat, 2002) and field observation data were used to compare the characteristics of the sites in the performed matrices. Fig. 4 shows four first priorities of final selection.

Fig. 4: Four final selected sites for installation of 1000 MW, TPP
Flexibility of the GIS method allows the user to apply a variety of data integration methods based on the characteristics of the data parts and the way which effect (support or decline) each other regarding the application (Valadan Zoej et al. 2005). The purpose of this study was not only to find suitable sites for demand of electricity at year 2011 but also make a site selection conceptual model using GIS in the country for future studies. In this study for the first time integration operators to integrate the data layers in GIS for power plant sitting were used in Iran. Binary maps and Boolean operators were utilized to identify limited areas for power plant construction and 71% of the county fell in this limited area. Factor weighted maps were applied to determine suitable locations for power plant construction in the remaining 29% area and for combining of the factor maps and binary maps index overlay method were applied. As the result of the model running, 0.17% of the study area in 49 sites was selected as the suitable area and finally 4 first priority sites with less than 4 Km2 was selected.

The authors would like to thank Ministry of Energy, Iran Power Development Company and Iran Energy Efficiency Organization for their financial support and members of Faculty of GIS of Khaje Nasir Technical University for their data and consultations on this study.

Bonham-Carter, G.F, (1991), Geographic Information System for Geoscientists: Modeling with GIS, Pergamon, Ontario, PP. 319-470.

Clarke, KC, (1997),Getting Started With Geographic Information Systems. Upper Saddle River, NJ: Prentice Hall1.

Noorollahi, Y., R. Itoi, et al. (2007). “GIS model for geothermal resource exploration in Akita and Iwate prefectures, northern Japan.” Journal of Computers & Geosciences 33(8): 1008-1021.

Valadan Zoej, M.J, Samadi, R and Yousefi, H, (2005), The role of GIS in the development of the power network, Indonesia – Jakarta, 22-25 August.

Yousefi. H, Ehara. S, and Noorollahi. Y., (2007), “Geothermal Potential Site Selection using GIS in Iran”, 32nd workshop on Geothermal reservoir engineering, January 22 – 24, 2007, Stanford, CA, USA, https://pangea.stanford.edu/ERE/db/IGAstandard/record_detail.php?id=5061

Yousefi. H, Ehara. S, Samadi. R, Sohrab. T, and Noorollahi. Y., (2006), Thermal power plants Site selection using GIS in Fars province Southern part of Iran, 4th Internationa