Site selection study for fire extinguisher stations using network analysis and A.H.P. Model Case study: city of Zanjan

Mohsen Ahadnejad Reveshti
Dept of geography,
Zanjan University
ahadnejad@gmail.com

Abdolah Heidari
GMsc in Geography and Urban Planning

Abstract
Land use planning and site selection studies for an optimal distribution of different service centers are among those issues which urban planners are dealing with. Fire extinguisher stations are significant infrastructures in every major city. During the present study efforts are made to analyze the location as well as effective coverage of fire extinguisher stations throughout Zanjan. Network analysis is employed to investigate the spatial distribution of stations. Analytic Hierarchy Process(AHP) along with GIS have also been used to suggest new stations in areas beyond the coverage of existing stations according to the world as well as country’s common standards. The results of present investigations revealed that existing stations within the city of Zanjan don’t suffice to cover the whole city and over the half of city are beyond the standard coverage of fire extinguisher stations. Therefore five new stations were suggested to be established to cover the whole city.

Introduction
Land use planning for optimal distribution of different land use classes as well as service centre distributions are important issue which is often considered by urban planners. Population increase resulting urban growth causes many problems as lack of utilities as well as uneven spatial distribution of different land uses. Spatial distribution of fire extinguisher stations is of vital importance due to its major role to establish security in case of an unforeseen accident. No doubt among the other methods for loos reduction caused by fire in urban area, urban planning through establishing of standards and regulations may play a considerable role to reduce life and financial losses within a long period. Off course this is very important for those cities which have a very dense texture and disordered network. In order to take better relief actions to help the people in case of an accident, necessary standards should be established for emergency disastrous cases in a disordered urban texture. City of Zanjan with a very dense urban structure and improper unavailable network is a typical example of such cities. There are not enough fire extinguisher stations to cover the central and southern portions of the city. Based upon the investigations the four existing fire extinguisher stations of the city don’t cover the whole city and over the half of city is out of the reach of these stations.

Study area and location of present stations
City of Zanjan is situated between longitudes of 48 26 39 and 48 32 50 and latitudes of 36 39 00 and 36 42 50. The city is located at the elevation of 1630 meters above the sea level covering about 6000 hectares. Due to the city’s unique location within the province its population has increased from 47159 in year 1956 to 339000 in year2003, but the utility in the city have not been developed with the same speed so the lack of different city services is clearly visible.

One of the most important lacking in the city is fire extinguishing stations throughout the whole city. According to our investigations fire extinguishing organization in Zanjan has only four stations throughout the whole city which don’t cover the whole city. Two stations are located in Tohid and Navab-e-Safavi Avenues covering southern portion, another station in Azadeghan Avenue covers northern portion and final station is covers eastern part of the city (map 1).

Map 1: Location of fire extinguishing stations in Zanjan City

Methodology
This study started by literature work as well as field study and analysis of graphic data from the city emphasizing Network Analysis and Analytic Hierarchy Process.

Network analysis
Network analysis are employed to study the spatial distribution pattern of present fire extinguishing stations and their coverage to find the areas out of the coverage of present stations. All streets and accessible networks of the city were digitized using Auto Cad Map (The real direction of traffic in the streets was used to select the digitizing direction). In order to establish different parameters as passage length, Id, F node and T node, topology was established using ArcGIS. Attributes were prepared through questionnaires and inquiries from drivers in traffic rush hours and speed limit were obtained through these investigations. The other information as the street type (one or two way), street width and cross roads were added and prepared for final analysis. Average car speed were calculated using the following formula (V= D/ T).Therefore the city’s intelligent street map was prepared by this mean.

According to world standards it should not take more than 3 minutes for fire extinguishing busses to reach the accident place while this exceeds 5 minutes in Iran. The average reach time of 5 minutes were applied to consider the four existing stations in the study area. The results revealed that the whole city were not covered by existing stations and over the half of city were left out of the coverage. So it is concluded that there is a need to establish new stations (map 2).

Map 2: coverage of the present stations in Zanjan City

A.H. P. model for site selection studies
Analytic Hierarchy Process is an effective, flexible and simple method to be used in certain cases on which many contradictory criteria make decision making process very difficult. This multi- criteria evaluation method was propose by Thomas L. Saati in 1980 and since then had several applications in different sciences. A principal method to test AHP method is a binary comparison method. This method considerably reduces the conceptual complexity of decision making process because of simultaneous evaluating of two components. The method includes three steps as following:

A - Binary comparison metrics generation
B- Calculating the weights of criteria
C- Estimating of correspondence ratio

Following steps have been applied for site selection of fire extinguishing stations and the results were analyzed to find suitable locations for site selection using GIS technique. Consequently the final site maps were prepared.

A- Binary comparison metrics generation
This method applies a certain scale from 1-9 to evaluate the priority of two different criteria (table 1). In fact we compare the two criteria in order to determine their weight. In this case binary comparison in one metric as n* n (10 * 10 for this case) is being registered which is called binary comparison metric A = (a ij) nn. All parameters of this matrices is positive and considering the “reverse condition principal” in Analytic Hierarchy Process we will have two digits as a ij and 1/ a ij . Table 2 shows binary comparison matrices for the proposed problem. The results of literature works as well as the specialists’ comments have been used to determine effective parameters and their weights for site selection of fire extinguishing stations. The results of this effort have been briefed in table 2 which have been used for GIS analysis.

Description	Weight
1	equal weight
2	equal – moderate
3	moderate
4	Moderate – strong
5	Strong
6	Strong-very strong
7	Very strong
8	Very strong-extremely strong
9	Extremely strong

Table 1: nine class scales for binary comparison

criteria	Accessibility to water pump	Closeness to major road network	Closeness to industerial centres	Closeness to Trade centres	Closeness to Gas stations and warehouses	Closeness to Education and medical centres	Closeness to Arable lands and orchard	Closeness to Administrative centres	Closeness to Libraries and	Closeness to Overpopulated area	Wight criteria
Accessibility to water pump	1	2	3	4	5	5	6	6	7	7	0.271
Closeness to major road network	1/2	1	2	3	4	4	5	5	6	6	0.203
Closeness to industrial centres	1/3	1/2	1	2	3	3	4	4	5	5	0.136
Closeness to Trade centres	1/4	1/3	1/2	1	2	3	4	4	5	5	0.110
Closeness to Gas stations and warehouses	1/5	1/4	1/3	1/2	1	2	3	3	4	4	0.084
Closeness to Education and medical centres	1/5	1/4	1/3	1/3	1/2	1	2	3	3	4	0.066
Closeness to Arable lands and orchards	1/6	1/5	1/4	1/4	1/3	1/3	1	2	2	3	0.044
Closeness to Administrative centres	1/6	1/5	1/4	1/4	1/3	1/3	1/2	1	2	2	0.035
Closeness to Libraries and mosques and other religious sites	1/7	1/6	1/5	1/5	1/4	1/3	1/2	1/2	1	2	0.026
Closeness to Overpopulated area	1/7	1/6	1/5	1/5	1/4	1/4	1/3	1/2	1/2	1	0.025

Table 2: binary comparison metric of evaluation criteria

B-weight calculation of criteria:
This section includes the following steps:

summation of every column
dividing every matrices component by the summation of column
Calculating of the average of components in every line of normalized matrices, in other words dividing the summation of normalized weights by 10 (the number of criteria).

The resultant average value makes the relative weight estimates. The results of this calculation are presented as the weight of criteria. According to this evaluation, fire extinguishing valve accessibility, closeness to the main valve and those streets with over 15 meters width obtain the highest weights respectively and rest of the parameters obtain weights from moderate to weak respectively.

C- Agreement ratio estimation
In this step agreement of comparisons are being compared. This step composes the following steps:

determining of the total weight vector by multiplying the weight of first parameter (accessibility of fire extinguishing valves) by the first column of binary comparison matrices, then multiplying the second parameter (closeness to main valve) by the second column, then the third parameter by the third column respectively and finally multiplying the tenth parameter by the tenth column of the main matrices. Finally all these amounts were added together.
Determination of comparison vector by dividing total weight vector by the weights of parameters which have been determined before. When comparison vector calculated we need to calculate two other parameters of which is equal to the average amount of agreement vector and agreement index (CI) which may be calculated from the following equation:

Also agreement ratio may be calculated by the following equation:

In the above equation R1 is random index which is derived through table 3 for different numbers of parameters (n). CR ratio is designed somehow that if CR < o.o1, it will give the agreeable area in binary comparisons.

15	14	13	12	11	10	9	8	7	6	5	4	3	2	n
1.59	1.57	1.56	1.48	1.51	1.49	1.45	1.41	1.32	1.24	1.12	0.9	0.58	0	RI

Table 3: random index

The agreement ratio of 0.09 is smaller than 0.1 in the considered problem which shows an acceptable level of agreement in binary comparisons.

Binary comparison method were used to analysis different layers in GIS environment resulting a zoning map which is classified in to 4 different classes as : very good, good, moderate and poor capability for fire extinguishing stations (map 3). Suitable places (very good and good suitability) for sits were selected in this step.

Map 3: suitable sites for establishing fire extinguishing stations

Map 4 shows selected sites for 5 new stations which may cover the whole city together with 4 other existing stations. Network analysis was employed to verify the accuracy of analysis in newly proposed stations. The results showed that areas not covered by the former existing stations will be covered by the newly proposed stations.

Map 5 shows the area covered by all stations including newly proposed ones

Map 4: site selection and location of the proposed stations

Map 5: area covered by all stations including existing and newly proposed ones

Conclusion
Network analysis as well as land use map of the city and the other investigations reveal that major part of Zanjan is not being covered by the pre- existing stations so there is a drastic need to establish new stations. Also this study disclosed that GIS and AHP techniques employing different qualitative and quantitative parameters as well as the compatibility for layer compilation are among those techniques which may be used for urban planning.

References

Baird, Donal. Conmputer programs aid in fire station location. Fire Fighting in Canada. 1990 Dec; Vol.34 (no.9):PP.4-5
Chaiken, Jan M. Methods for allocating urban emergency units: A survey Mangement Science. 1972:Vol. 19:PP.110-13..
Graeme, F., 1994, Geographic Information System for Geoscientists, Modelling With GIS, New York, Carter Publisher.
Toregas, C.The location of emergency services facilities. Operations research. 1971: Vol 19:pp.1363-1373.
Revelle, C. S. Central facilities location. Geographichal Analysis. 1970:Vol. 2: pp 30-42
Gay, William. Fire station location: Analysis and technology. Washington, D.C: International City Management Association: 1987 Jul.19p (MIS Report: Vol.19/no.7).
Bahreini, H. Urban design process, Tehran University publication, 1998.
Zebardast, E. Analytic Hierarchy Process in Urban planning, Art journal, 2001, vol10, pp13-22.