Development of fire risk visualization tool based on heat map

https://doi.org/10.1016/j.jlp.2021.104505Get rights and content

Highlights

  • Proposing a transformed sigmoid risk model (TSRM).

  • Developing a fire risk visualization tool (FRVT) based on TSRM.

  • Using FRVT to visualize 16,448 fire incidents from a regional perspective.

  • FRVT can be applied to visualize regional process fire risks.

Abstract

At present, relatively few tools are able to visualize regional fire risks both dynamically and efficiently. As a general method of density clustering and visualization, heat maps can be used to analyze the spatial distribution characteristics of fire risks. In this study, based on combining a transformed sigmoid function with a heat map method, a transformed sigmoid risk model (TSRM) was proposed. A numerical mapping method between fire risk values and colors was established based on an RGB color model. By using HTML 5 and JavaScript, a fire risk visualization tool (FRVT) was developed based on the TSRM. Baidu Map was used as the geographic information system engine for the FRVT by calling its JavaScript application program interface. To validate the TSRM and FRVT with case studies, 16,448 historical fires from 2013 to 2017 in Changsha, China, were gathered and visualized. The visualization results indicated that the high-risk areas were mainly distributed in urban areas, and that the medium-risk and low-risk areas were primarily distributed in rural areas. The FRVT can also be applied to visualize and analyze process-related fire risks from a regional perspective. In further research, it will be important to update the FRVT, e.g., to visualize the fire risks posed by the transportation of hazardous materials.

Introduction

Fires have affected the sustainable development of society (Yue and Unger, 2018; McNamee et al., 2019; Zong et al., 2020). Fires in the process industries have also posed critical threats to process safety (Willey, 2017; Ding et al., 2020; Zong et al., 2020). From 2014 to 2017, there were 530 fires at filling stations and 1511 fires at petrochemical enterprises, resulting in several million yuan of property losses and dozens of deaths and injuries, as summarized in Table 1 (Fire Department of Ministry of Public Security of the People's Republic of China (FDMPS), 2015; FDMPS, 2016; FDMPS, 2017; FDMPS, 2018). Although the technologies for fire prevention and extinguishing are continuously improving, fires remain commonplace (Brushlinsky et al., 2018; FDMPS, 2018). To manage frequent fires, fire stations with sufficient emergency rescue capabilities are required in a region, and regional fire risks should be identified effectively, so as to determine the appropriate number of (and locations for) fire stations (Chevalier et al., 2012; Yao et al., 2019; Chen et al., 2020; Ding et al., 2020; Zhou et al., 2020). A regular assessment of regional fire risks can assist in identifying regional fire risks effectively and efficiently (Yu et al., 2020; Liu et al., 2020a, 2020b).

In the past few decades, many methods or models have emerged for fire risk assessment, including hazard and operability analysis, failure mode and effects analysis, event tree analysis, fault tree analysis, analytic hierarchy process, Bayesian networks, Petri-nets, and cluster analysis (Shi et al., 2014; Chaudhary et al., 2016; Amin et al., 2019; Liu et al., 2019a; Ancione et al., 2020; Fang et al., 2020; Zhou and Reniers, 2020). The spatial and temporal distribution characteristics of regional fire risk can be obtained based on visual analysis. Owing to developments in computer technology, geographic information systems (GISs) are commonly used to visualize and analyze regional risks (Zhao and Liu, 2016). Many studies have focused on visualizing the distribution characteristics of fire risks based on GISs. By combining a GIS with fire risk assessment techniques, a visualization of fire risks can be created for identifying the spatial distribution characteristics. Furthermore, temporal and spatial distribution prediction models for fire risks can be built based on GISs (Stojanova et al., 2012; Song et al., 2017). Regional risk visualizations can assist in fire management, and in the location optimization for urban emergency rescue facilities (Chaudhary et al., 2016; Sakellariou et al., 2019). Most regional fire risk visualizations have been extracted from GISs based on a graphical user interface, especially ArcGIS, a famous and commonly used graphical user interface (GUI)-based GIS platform (Stojanova et al., 2012; Chaudhary et al., 2016; Song et al., 2017; Sakellariou et al., 2019; Vallejo-Villalta et al., 2019). However, in the GUI-based GIS, the geographic information is less frequently updated. A web-based GIS may overcome this problem.

With the development of the Internet, web-based GIS tools and online maps (Google Maps, Baidu Map, etc.) have been widely used for route planning, location searches, driving navigation, and so on (Veenendaal et al., 2017). Compared with a GUI-based GIS, online maps can be used across platforms, and can contain user-provided crowdsourcing data (including location and traffic data). In addition, most of these online maps provide a web services application programming interface (API) for displaying maps, geocoding, and visual data (Veenendaal et al., 2017; Xiang et al., 2017; Baidu, 2019; Google, 2019). Online maps have been applied for data visualization in various fields of scientific research (Pan et al., 2017; Chmielewski et al., 2018).

By combining the heat map method with online maps, scientific data can be visualized and analyzed (Zhang et al., 2012; Škuta et al., 2014). Different colors can be used to represent different densities in the heat map, essentially providing a density clustering method for machine learning (Liu et al., 2019a). Through a cluster analysis of several fire risk data, a heat map of the fire risk data can be obtained. Then, the spatial distribution characteristics of the regional fire risk can be obtained. Historical fire data is an ideal source of fire risk data, and previous works have used historical fire data for regional fire risk assessments (Ceyhan et al., 2013; Liu et al., 2019b).

Owing to urbanization and rapid socio-economic development (especially in China), regional fire risks have changed rapidly (Price and Bradstock, 2014; Gu, 2019). The expansions of urban scale, increases in urban population, and their flows have led to severe and complicated situations regarding urban fires (Sorrensen, 2012; Zhang et al., 2019). To prevent and control fires in a timely manner, regional fire risks should be regularly assessed and visualized. Providing high-frequency assessment requires improving the efficiency of assessment with specific tools. However, at present, few tools can visualize fire risks dynamically and efficiently.

To fill this gap in fire risk visualization, in this study, a fire risk visualization tool (FRVT) was developed based on a heat map, in a direct outgrowth of our previous work (Liu et al., 2019a). We also proposed a new method for determining an initial fire risk value from historical fire data, according to the severity. Using the FRVT, regional fire risks were visualized dynamically and efficiently. To validate the FRVT, 16,488 fire incidents from 2013 to 2017 in Changsha, China were analyzed.

Section snippets

Transformed sigmoid function

The sigmoid function is named for its sigmoid shape (“S” curve). Owing to its monotonically increasing nature, it is often used as an activation function for logistic regression, one of the classification algorithms in machine learning (Han and Moraga, 1995; Raschka and Mirjalili, 2017). The sigmoid function can be expressed by Eq. (1) and can be graphed as shown by the red curve in Fig. 1 (Liu et al., 2019a). The sigmoid function has a domain of all real numbers, with the return value

Programming with JavaScript

As a high-level and interpreted programming language, JavaScript has been an essential part of web applications. JavaScript is a scripting language whose source code does not need to be compiled before it runs. Applications programmed with JavaScript can run directly on various browsers, which has allowed JavaScript to become widely used. Hypertext marked language (HTML) is used to define the texts, pictures, sounds, and videos that are displayed on browsers. HTML5 is the fifth and current

Case study

To validate the TSRM and FRVT with case studies, we collected and visualized 16,448 fire incidents from 2013 to 2017 in Changsha, China.

Further research

The FRVT can run on a computer with a high-resolution monitor to produce high-definition visualizations. The appropriate values of Rf should be determined according to the computer monitor resolution and amount of fire risk data. In our next work, we will study a method for determining the optimal value of Rf.

With the rapid development of the social economy, there are generally many petrol stations, storage sites, and process plants in a geographic area; these places have higher fire risks than

Conclusions

The main objective of this study was to develop a tool for dynamically and efficiently visualizing regional fire risks. First, the TSRM was proposed, based on the heat map method. To visualize the fire risk, a numerical mapping method between fire risk values and colors was established, based on the RGB color model. Using HTML5, JavaScript, and SQLite, the FRVT was developed, based on the TSRM. By calling the Baidu Map JavaScript API, the Baidu Map was embedded in the FRVT as a GIS engine.

To

Author statement

Dingli Liu: Conceptualization, Methodology, Software, Validation, Writing – original draft. Zhisheng Xu: Conceptualization, Methodology. Chuangang Fan: Validation, Writing – original draft, Writing - review & editing. Yang Zhou: Validation, Writing - review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This study was supported by the Key Research and Development Program of Hunan Province (No. 2020SK2057), the National Natural Science Foundation of China (Nos. 51804338 and 51974361), and the Fundamental Research Funds for the Central Universities (Nos. 502501004 and 502045009). The authors would like to thank Wencai Li and Fengcai Yan from the fire department of Changsha for providing data and good advice.

References (52)

  • C. Gu

    Urbanization: positive and negative effects

    Sci. Bull.

    (2019)
  • F.H. Hedlund et al.

    Puncture of an import gasoline pipeline—spray effects may evaporate more fuel than a Buncefield-type tank overfill event

    Process Saf. Environ. Protect.

    (2019)
  • D. Liu et al.

    Regional evaluation of fire apparatus requirements for petrol stations based on travel times

    Process Saf. Environ. Protect.

    (2020)
  • M. McNamee et al.

    IAFSS agenda 2030 for a fire safe world

    Fire Saf. J.

    (2019)
  • S.R. Nayak et al.

    Fractal dimension of RGB color images

    Optik

    (2018)
  • O. Price et al.

    Countervailing effects of urbanization and vegetation extent on fire frequency on the Wildland Urban Interface: disentangling fuel and ignition effects

    Landsc. Urban Plann.

    (2014)
  • O.M. Qureshi et al.

    Ahmedpur Sharqia oil tanker tragedy: lessons learnt from one of the biggest road accidents in history

    J. Loss Prev. Process. Ind.

    (2020)
  • M.Y. Shamim et al.

    Development and quantitative evaluation of leading and lagging metrics of emergency planning and response element for sustainable process safety performance

    J. Loss Prev. Process. Ind.

    (2019)
  • L. Shi et al.

    Fuzzy fault tree assessment based on improved AHP for fire and explosion accidents for steel oil storage tanks

    J. Hazard Mater.

    (2014)
  • T. Vairo et al.

    Critical aspects of natural gas pipelines risk assessments. A case-study application on buried layout

    Process Saf. Environ. Protect.

    (2021)
  • R.J. Willey

    West fertilizer company fire and explosion: a summary of the U.S. Chemical safety and hazard investigation board report

    J. Loss Prev. Process. Ind.

    (2017)
  • J. Yao et al.

    Location optimization of urban fire stations: access and service coverage

    Comput. Environ. Urban Syst.

    (2019)
  • Y. Zhang et al.

    How fire safety management attended during the urbanization process in China?

    J. Clean. Prod.

    (2019)
  • M. Zhao et al.

    Regional risk assessment for urban major hazards based on GIS geoprocessing to improve public safety

    Saf. Sci.

    (2016)
  • J. Zhou et al.

    Modeling and application of risk assessment considering veto factors using fuzzy Petri nets

    J. Loss Prev. Process. Ind.

    (2020)
  • J. Zhou et al.

    Simulation analysis of fire truck scheduling strategies for fighting oil fires

    J. Loss Prev. Process. Ind.

    (2020)
  • Cited by (3)

    • A POIs based method for determining spatial distribution of urban fire risk

      2021, Process Safety and Environmental Protection
      Citation Excerpt :

      Moreover, the hot spots of fire accidents were identified based on kernel density analysis. Liu et al. (2021) developed a fire risk visualization tool based on heat map, which can be used to visualize regional process fire risks. In order to improve urban safety, points of interest (POIs) based method is developed for fire risk assessment and visualization in this study.

    View full text