Towards Sustainable Urban Mobility: Voronoi-Based Spatial Analysis of EV Charging Stations in Bangkok (2024)

1. Introduction

Petroleum, or oil, a vital fossil fuel sourced from decomposed organic matter, serves as a primary energy source for transportation, electricity generation, and the petrochemical industry [1]. The global oil industry has undergone dynamic shifts, with projections indicating a future decline in production and volatile prices influenced by geopolitical tensions and market factors [2]. Despite global efforts to transition to alternative energy, Thailand’s oil consumption remained substantial at 1,277,000 barrels per day in December 2022, contributing to pollution and environmental harm [3,4]. Heavy dependence on oil has geopolitical and economic implications, fueling conflicts over resources and impacting energy costs and economic stability [5,6].

In response to these challenges, countries worldwide are actively transitioning towards alternative energy sources, recognizing the imperative need for a cleaner and more sustainable energy future. Among the most prominent alternatives is the adoption of EVs. The ascendancy of electric vehicles represents a significant milestone in the pursuit of a more sustainable global future. Despite facing challenges, such as a nascent charging infrastructure and higher upfront costs, electric car sales surged to 6.6 million in 2021, more than tripling their market share from two years earlier [7].

The global shift towards electric mobility is a response to growing concerns about environmental sustainability. EVs have become a focal point in addressing these concerns due to their environmental benefits, including a substantial reduction in greenhouse gas emissions and improved air quality [8]. This global perspective aligns with Thailand’s commendable progress in embracing EVs as part of its commitment to establishing a sustainable and eco-friendly transportation system. Thailand’s efforts to promote electric vehicles are driven by a combination of government incentives, technological innovations, and the expansion of charging infrastructure [9]. With a target to phase out internal combustion engine vehicles by 2035, Thailand aims to position itself as a hub for EV production while simultaneously reducing greenhouse gas emissions. The ambitious nature of these goals underscores Thailand’s commitment to contributing to global sustainability efforts. Considering the significant progress made, there is potential for an accelerated transition to 100% electric vehicles in Thailand before the projected 2035 deadline, leading to the phasing out of oil-based vehicles and the transformation of traditional fuel stations into electric charging stations [10]. This rapid transition not only aligns with global sustainability goals but also positions Thailand as a leader in the adoption of electric mobility.

Crucial to the success of the electric vehicle ecosystem are EV Charging Stations. These stations play a vital role in facilitating the widespread adoption of EVs by offering convenient and accessible charging solutions. The growth and expansion of this charging infrastructure are imperative to establishing a sustainable transportation ecosystem, addressing range anxiety, and promoting the continued growth of electric vehicle ownership [11,12]. As the demand for electric vehicles continues to rise globally, the expansion of charging infrastructure becomes increasingly pivotal in promoting sustainable transportation practices. Despite significant progress in EV adoption, there remains a critical need to optimize the placement of EV charging stations to ensure they are conveniently accessible to the population, which in turn supports the broader adoption of EVs and contributes to environmental sustainability. This study aims to address this need by analyzing the geographical distribution of EV charging stations, evaluating service provision, assessing population impact, and pinpointing strategic locations for expanding EV charging infrastructure in metropolitan areas. Specifically, this research seeks to develop a foundational framework for global expansion strategies in EV charging operations.

Geographic Information Systems (GIS) are pivotal in the strategic optimization of EV Stations, contributing to enhanced efficiency and accessibility [13]. The integration of GIS technology into EV infrastructure enables station planners to identify optimal locations by considering factors such as population density, traffic patterns, and proximity to major routes. This geospatial approach not only facilitates the present station placement but also supports future expansion plans, guiding policymakers in establishing a comprehensive charging network [14]. To analyze service accessibility at charging stations, GIS leverages network analysis and Voronoi diagrams. Voronoi diagrams are particularly useful in this context as they partition space into cells or polygons that enclose a set of geometric objects, usually points. These polygons represent regions where points are closer to the surrounding objects than to any other object, offering a powerful tool for spatial analysis and decision-making. This ensures not only adequate population access to charging stations but also effective planning to meet future demand. Beyond their application in GIS [15,16,17], Voronoi diagrams have found utility in various fields, including delineating areas of influence in geographical contexts. Additionally, they have proven instrumental in optimizing flight paths for aircraft missions, showcasing their versatility in navigating hazardous environments with low complexity and adaptability to different scenarios [18]. This multifaceted application underscores the effectiveness of Voronoi diagrams in spatial analysis and planning, making them a valuable tool in the advancement of electric vehicle infrastructure and beyond.

This study represents a comprehensive effort to analyze the geographical distribution of EV charging stations, evaluate service provision, assess population impact, and pinpoint strategic locations for expanding EV charging infrastructure in metropolitan areas. Serving as a foundational framework for global expansion strategies in EV charging operations, this research employs Voronoi theory and spatial distribution analysis techniques. By effectively delineating charging station service area boundaries, this analytical approach allows for the segmentation of service coverage into tiers, facilitating the identification of optimal locations for establishing additional charging stations in areas with insufficient service provision. The integration of these analytical tools not only enhances our understanding of current charging station deployment but also provides valuable insights crucial for planning and optimizing future charging infrastructure on a global scale. Furthermore, the study includes practical recommendations for developing a robust EV charging network, emphasizing the importance of robust network connectivity, smart charging infrastructure, interoperability standards, and advanced data analytics to ensure efficiency, scalability, and user convenience.

The content of this paper is organized as follows. First, Section 2 details the materials and methods used in this study, beginning with the analysis of spatial distribution patterns of EV charging stations in Bangkok, including the Nearest Neighbor Index (NNI) calculation, Kernel Density Estimation, and the categorization of EV charging station locations. This is followed by the generation of Voronoi diagrams and an assessment of population density and charger accessibility, concluding with the methodology for assessing the suitability of new charging stations using overlapping Voronoi circles. Section 3 presents the results, covering the spatial distribution patterns of EV charging stations, their service boundaries, and the service burden relative to population data, as well as identifying suitable areas for new charging stations. Finally, Section 4 provides a comprehensive discussion, addressing the spatial distribution patterns, service areas, and service loads of EV charging stations, the infrastructure requirements for effective EV charging networks, and suitable areas for additional charging stations.

2. Materials and Methods

The initial phase involved acquiring geospatial data through systematic field surveys using Google Street View. This method provided intricate details on EV charging stations across Bangkok, including precise geographical coordinates and specific types of locations. This approach not only facilitated the identification of existing charging infrastructure but also enabled an evaluation of its accessibility and distribution within the urban landscape. Complementing the geospatial data, geographical information was gathered from a diverse array of online sources. This encompassed datasets on road networks, population demographics, administrative boundaries, land use patterns, and city maps. The integration of these datasets allowed for a comprehensive analysis of the urban environment, facilitating a deeper understanding of the factors influencing the placement and utilization of EV charging stations. However, the initial phase of our analysis assumes a hom*ogeneous demand for EV charging stations across Bangkok. This simplification is necessary for creating a baseline model. However, we recognize that variations in EV adoption rates, commuting patterns, and urban demographics can lead to disparities in actual service requirements.

Data sourcing was comprehensive, drawing from a wide array of websites and mobile applications, notably PlugShare, EAAnywhere, MEA-EV, PTT-EV-Station, MG-iSMART, and Tesla, among others. These platforms were instrumental in providing granular details on the features of EV charging stations, including location coordinates, charging capacities, connector types, and the availability of ports. Supplementary geographical data pertaining to the urban infrastructure and demographic patterns were meticulously sourced from authoritative digital portals such as MITEARTH, BangkokGIS, the Department of Provincial Administration (DOPA), and the Bangkok Metropolitan Administration (BMA). The integration of these authoritative sources enriched the dataset with vital insights into urban planning and governance, crucial for the study’s scope. The gathered data underpins the spatial analysis conducted in this study, aiming to map and assess the distribution of EV charging stations in Bangkok. This analysis seeks to identify spatial patterns, distribution inequities, and potential areas for infrastructure development, providing a foundation for recommendations on enhancing the city’s EV charging network. The overall methodology can be found in Figure 1.

2.1. Analyzing Spatial Distribution Patterns of EV Charging Stations in Bangkok

2.1.1. Nearest Neighbor Index (NNI) Calculation

The initial objective involved calculating the NNI to understand the spatial distribution pattern of EV charging stations. Data importation involved electric vehicle charging station coordinates, utilizing the Average Nearest Neighbor tool within a GIS environment. This process necessitated setting the projection to WGS 1984 UTM and defining the analysis area based on the administrative boundaries of Bangkok, calculated as 1,568,950,000 square meters. A “Generate Report” function was enabled to document the findings comprehensively. Equation (1) can be used to calculate NNI directly without using ArcGIS 10.8.1.

$NNI = \frac{L}{\sqrt{A}},$

(1)

where L is the observed mean distance between points (in this case, EV charging stations), and A is the area of the study area (administrative boundaries of Bangkok in this case).

2.1.2. Density Mapping with Kernel Density Estimation (KDE)

Following the NNI calculation, we employed Kernel Density Estimation to visualize the concentration of EV charging stations. This involved importing station coordinates and executing the Kernel Density tool, with the projection set to WGS 1984 UTM. The analysis was refined by specifying Area Units as square kilometers and adjusting Environment Settings to align with Bangkok’s administrative boundaries for accurate representation. The KDE is explained as follows:

$KDE (x) = \frac{1}{n h^{2}} \sum_{i = 1}^{n} K (\frac{x - x_{i}}{h})$

(2)

where x represents the spatial location, n is the number of charging stations, h is the bandwidth parameter, and K is the kernel function.

2.1.3. Categorization of EV Charging Station Locations Using Pivot Table Analysis

Location categories of EV charging stations were identified through Pivot Table analysis in Excel. This step streamlined the classification process by leveraging the detailed coordinate data previously gathered.

2.2. Generation of Voronoi Diagram

The creation of a Voronoi diagram was instrumental in this study for delineating areas closest to each EV charging station, thus establishing a theoretical framework for analyzing spatial exclusivity and service areas of charging infrastructure. Voronoi diagrams partition a plane into regions based on distance to points in a specific subset of the plane, where each point (charging station) has a corresponding region consisting of all points closer to it than to any other. This concept is foundational in spatial analysis, offering insights into area dominance and potential service gaps within urban landscapes [19].

The application of Voronoi diagrams in this context provides a visual and analytical method to assess the distribution and accessibility of EV charging stations across Bangkok. By creating these diagrams using the “Create Thiessen Polygon” tool, informed by the precise coordinates of charging stations, we could identify areas underserved by the current infrastructure and thus potential sites for future development.

2.3. Population Density and Charger Accessibility

To calculate the population density per EV charging station, we employed a comprehensive series of data preparation and spatial analysis steps, integrating population data with the spatial distribution of charging stations to assess accessibility across Bangkok’s districts.

2.3.1. Population Data Preparation

We began by compiling population data for each district in Bangkok into an Excel spreadsheet. This dataset included population figures, essential for assessing the accessibility of charging stations at the district level.

2.3.2. GIS Data Importation and Spatial Join

This population data were then imported into the software, where a spatial join was performed. This operation linked the Excel-based population data with the administrative boundary Shapefile of Bangkok, enabling the analysis of population figures in conjunction with geographical boundaries.

2.3.3. District Size Calculation

After the spatial join, the merged dataset was exported as a new Shapefile. Within this file, we introduced a new field to calculate the area of each district using the “Calculate Geometry” function, ensuring the analysis considered the physical size of each district.

2.3.4. Integration with Voronoi Diagram Data

Subsequent steps involved importing Voronoi diagram data (Thiessen Polygons) into the GIS platform. A “Union” operation overlaid the administrative boundary Shapefile, now containing population data, with the Voronoi diagram, merging the two datasets into a single analytical framework.

2.3.5. Calculation of Voronoi Area and Population per Polygon

In the unified dataset, an “Area” field was recalculated to determine the size of the newly formed areas from the union. We introduced a “diff (Voronoi percentage)” field to calculate the percentage of each Voronoi area within its respective district. The “num_pop” field was calculated by multiplying this percentage by the district’s total population, estimating the population within each Voronoi polygon.

2.3.6. Data Consolidation and Density Calculation

A “Dissolve” operation was applied to consolidate the overlaid data into distinct Voronoi polygons, using the “ID_Polygon” as the key. The population within each polygon (“num_pop”) was summed, and a new calculation for polygon size was conducted. The “density” field was then calculated by dividing the summed population (“num_pop”) by the polygon area, yielding the population density per Voronoi polygon.

2.3.7. Charging Station Association and Final Calculations

To link EV charging stations with their respective Voronoi polygons, we utilized the “Intersect” tool. This enabled us to record the number of charging stations within each polygon. A new field, “PH” (population density per charging head), was calculated by dividing the “density” by the number of charging stations in each polygon.

2.3.8. Visualization

Finally, the GIS data’s symbology was configured based on the “PH” field, setting value ranges to effectively visualize the population density per EV charging station across Greater Bangkok. This visual representation aids in identifying areas with insufficient charging infrastructure relative to population density, guiding strategic decisions for future charging station installations.

These steps offer an analysis of population density in relation to the distribution of EV charging stations, highlighting areas where infrastructure development can significantly improve charger accessibility for the city’s residents.

2.4. Assessing Suitability for New Charging Stations Using Overlapping Voronoi Circles

The process for evaluating the suitability of areas for new EV charging station installations involves a series of precise spatial analysis steps. This method integrates data manipulation and GIS tools to identify optimal locations based on proximity to existing infrastructure and spatial distribution.

2.4.1. Import Voronoi Diagram Data

Begin by importing the Voronoi diagram (Thiessen Polygon) Shapefile into the GIS software. This dataset provides the basis for our analysis, representing the current distribution of charging stations.

2.4.2. Identify Voronoi Vertices

Use the “Feature Vertices To Point” tool with the Voronoi Shapefile as input. Set the point type to “All” to identify all vertices of the Voronoi polygons, which represent potential areas for new stations.

2.4.3. Edit and Refine Vertices

Employ the “Editor” tool to manually remove any vertices that do not contribute to the analysis (e.g., outliers or points in non-relevant areas). This step ensures the focus remains on areas viable for development.

2.4.4. Coordinate Calculation and Export

Add fields “x” and “y” to the dataset, designated as “double” data type, to store the coordinates of each vertex. Calculate these coordinates, then export the data to Excel for further processing.

2.4.5. Duplicate Removal and Re-Importation

In Excel, use the “Remove Duplicates” tool to eliminate any redundant vertices. Save the edited dataset and import it back into the software using the “Display XY Data” functionality.

2.4.6. Distance Calculation to Nearest Charging Station

Apply the “Near” tool to calculate the distance from each Voronoi vertex to the nearest existing EV charging station. Store these distances in a new field, “near_dis”.

2.4.7. Buffer Creation Based on Distance

Utilize the “Buffer” tool on the vertex’s dataset, using the “near_dis” field to set the radius. This creates buffer zones indicating areas within a certain proximity to existing stations.

2.4.8. Buffer Overlay and Separation

Conduct a “Union” operation on the buffer zones to merge overlapping areas. Then, apply the “Multipart to Singlepart” tool to delineate individual buffered areas, facilitating a clear analysis of each unique zone.

2.4.9. Spatial Join for Identical Area Matching

Perform a “Spatial Join” between the separated buffer areas and the original dataset, setting the match option to “ARE_IDENTICAL_TO”. This step identifies buffer zones that precisely match the criteria for potential charging station sites.

2.4.10. Visualization and Analysis

Finally, configure the symbology of the GIS dataset using the “join_count” field. This highlights areas with a higher concentration of suitable locations, guiding the decision-making process for new station installations.

By following these steps, the analysis pinpoints areas within Bangkok that are optimally located relative to existing EV charging infrastructure and thus represent promising candidates for the development of new charging stations. This systematic approach ensures a data-driven strategy for expanding the city’s EV charging network.

3. Results

This research project applies geospatial theory and GIS to analyze the spatial distribution of EV charging stations in Bangkok and identifies suitable locations for additional stations. Our findings are organized into four primary objectives and presented through an integrated framework of content, maps, and tables.

3.1. Spatial Distribution Pattern of EV Charging Stations in Bangkok

The spatial distribution analysis of EV charging stations in Bangkok reveals a distinctly clustered pattern, as evidenced by a NNI of 0.57. This clustering is most pronounced within the Central Business District (CBD), particularly in the Lumphini sub-district, which hosts the highest concentration of charging stations in the city with 21 stations. Following closely are the Wang Mai and Pathum Wan sub-districts with 19 and 14 stations, respectively. Notably, other areas exhibit varied station densities; for example, Khlong Toei Nuea features 18 stations, while both Khlong Toei and Silom each house 11 stations. This distribution pattern is visually represented in Figure 2.

Further analysis into the types of locations hosting charging stations indicates a significant presence within residential areas, which account for approximately 20.31% of all stations. Commercial zones follow at 16.38%, with fuel service stations and car service centers comprising 15.7% and 13.82%, respectively, as detailed in Table 1. The majority of charging stations are strategically located along main roads—503 out of 586 stations, to be exact—with 134 on major roads and 369 on minor roads, enhancing accessibility for EV users. Additionally, 83 stations are situated in alleys, further diversifying their urban distribution (Figure 3).

The study also revealed the relationship between charging station placement and urban planning factors. It was observed that commercial districts, as defined in the city’s urban plan, tend to have a higher density of charging stations (Figure 4). A land use analysis reinforced the urban-centric nature of charging infrastructure, with 97.44% of stations located in urban areas, compared to 1.54% in miscellaneous use areas and 1.02% in agricultural zones, showcasing a significant bias towards urban land use for charging station siting (Figure 5).

3.2. Service Boundaries of EV Charging Stations in Bangkok

The utilization of Voronoi diagrams for the analysis of charging station service boundaries in Bangkok has provided significant insights into the spatial accessibility and potential service load of EV charging stations across the city. The Voronoi cells, which delineate the service areas of individual charging stations, exhibit a considerable variation in size, offering a visual and quantitative measure of service coverage.

Larger Voronoi cells suggest that the encompassed charging station may bear a heavier load, servicing a broader area and potentially a larger number of EV users. While this expansive coverage could indicate a station’s strategic placement to maximize service area, it simultaneously raises concerns about the adequacy of service coverage, particularly in terms of accessibility and charging capacity. Conversely, smaller Voronoi cells are indicative of a higher density of charging stations within a given area, suggesting better service coverage and potentially shorter waiting times for users. This pattern of distribution underscores the varying levels of charging infrastructure density across Bangkok, with certain areas enjoying robust coverage while others may face limitations in accessibility.

This analysis not only complements the findings regarding the spatial distribution patterns of charging stations, as discussed in Section 3.1, but also highlights the disparities in service coverage that may exist across different parts of the city. The size and distribution of Voronoi cells serve as a critical tool for identifying areas with either optimal or insufficient charging station accessibility, thereby guiding future infrastructure development efforts to address these disparities (as illustrated in Figure 6).

3.3. The Service Burden of EV Charging Stations Relative to Population Data in Bangkok

The investigation into the service burden of EV charging stations, conducted through the analysis of population density within service boundaries delineated by Voronoi diagrams, reveals significant variations across Bangkok. The study quantifies service demand by correlating population density with the geographic coverage of each charging station, an essential metric for evaluating the adequacy of existing charging infrastructure. Figure 7 provides a visual representation of this correlation, using gradations of color to denote varying population densities and, by extension, service demands on EV charging stations. The findings indicate the following:

The distribution of population density across service boundaries shows a notable concentration in the range of 1000 to 2000 people per square kilometer per charging station, encompassing 163 service boundaries or 27.82% of the total analyzed. This range indicates a moderate level of service demand.
The delineation of population density into additional ranges—2000 to 3000 (19.45%), 0 to 1000 (19.11%), 3000 to 4000 (14.16%), 4000 to 6000 (11.09%), and 6000 to 8000 (6.14%) people per square kilometer per charging station—allows for a nuanced understanding of service demand across the urban fabric.
The average population density within the service boundaries stands at approximately 2696 people per square kilometer, offering a baseline for the general service demand faced by the city’s EV charging stations.
Extremes in the data reveal areas with the lowest and highest population densities at 65 and 13,889 people per square kilometer per charging station, respectively, highlighting regions under significant service demand pressures.

This comprehensive analysis of the service burden borne by EV charging facilities, in relation to surrounding population densities, illuminates areas of critical need for infrastructure enhancement. The findings advocate for the strategic expansion and equitable distribution of charging stations to meet the diverse demands of Bangkok’s urban landscape, underpinning the significance of geospatial analysis in urban planning and infrastructure development.

3.4. Suitable Areas for Establishing Additional Electric Vehicle Charging Stations in Bangkok

Utilizing Voronoi diagrams, this analysis aimed to delineate areas within Bangkok that are suitable for the introduction of new EV charging stations. The investigation centered on the strategic application of Voronoi Vertices and Circles to assess spatial suitability for infrastructure expansion.

Voronoi Vertices Analysis: These vertices, defined by the intersection of edges equidistant from three neighboring charging stations, pinpoint locations maximally distant from current facilities. The identification of such vertices provides a basis for considering where new stations could most effectively augment the existing network.
Overlap of Voronoi Circles: The degree of overlap among Voronoi circles—generated around vertices to the nearest existing station—serves as an indicator of area suitability. High overlap areas suggest optimal locations where a new station could serve multiple adjacent areas lacking immediate access to current charging infrastructure.
Determination of Suitable Areas: Through the quantitative analysis of Voronoi circle overlaps, regions were categorized based on their suitability for infrastructure development. Areas marked by significant overlap are identified as highly suitable for new charging stations, facilitating strategic expansion planning.

Figure 8 displays the analysis of suitable areas for the deployment of additional EV charging stations in Bangkok, highlighting in green the zones of high suitability determined through Voronoi diagram overlaps.

4. Discussion

4.1. The Spatial Distribution Pattern of EV Charging Stations in Bangkok

The spatial distribution analysis of EV charging stations across Bangkok, as indicated by a NNI of 0.57, reveals a distinct clustering within the city’s CBD, particularly in the Pathum Wan district. This clustering aligns with areas characterized by significant social and economic activities, affirming the strategic placement of charging infrastructure to maximize accessibility and usage. Notably, the analysis underscores a pronounced preference for situating charging stations along major roads, a decision likely driven by the greater traffic flow on these arteries compared to secondary roads and alleys. This pattern of location selection is indicative of an overarching strategy to enhance visibility and convenience for EV users, thereby encouraging the adoption of electric vehicles.

Further observations elucidate that the placement of EV charging stations is not random but is significantly influenced by urban planning considerations and the inherent characteristics of different areas. Commercial zones, in particular, exhibit a higher density of charging stations. This phenomenon can be attributed to the vibrant social and economic activities in these areas, which not only increase the visibility of charging stations but also potentially boost their utilization rates. Such areas are deemed more attractive to investors, considering the direct correlation between high activity levels and increased demand for EV charging services.

Moreover, a comprehensive analysis of land use within Bangkok reveals that an overwhelming majority of charging stations, amounting to 97.44%, are located in urban land-use areas. This distribution is reflective of the strategic approach towards infrastructure investment, where areas with heightened social and economic engagements are prioritized. The urban fabric of Bangkok, characterized by its dynamic commercial centers and residential conglomerates, presents an ideal backdrop for the deployment of EV charging stations. Investors and urban planners recognize these urban areas as having superior potential for high utilization rates, making them prime candidates for the development of charging infrastructure.

In summary, the spatial distribution of EV charging stations in Bangkok is a multifaceted phenomenon, deeply intertwined with the city’s urban planning and socio-economic landscape. The preference for major roads, the influence of commercial activity, and the prioritization of urban land-use areas are all strategic considerations that shape the current and future deployment of EV charging infrastructure. These insights not only highlight the critical role of geographical and urban analysis in infrastructure planning but also underscore the importance of aligning infrastructure development with the broader urban ecosystem to maximize its efficiency and accessibility.

4.2. Service Area Scope of Electric Vehicle Charging Stations in Bangkok

Investigating the service areas of EV charging stations in Bangkok with Voronoi diagrams has highlighted the variability in the scope of these areas. The analysis revealed that the service areas, or “cells,” vary significantly in size, with larger cells indicating a potential for higher service burdens on the encompassed charging stations. This size disparity underscores the logistical challenges of achieving equitable EV charging service coverage citywide. Users’ preference for proximal charging options means that stations covering larger areas might be under more strain to meet demand, leading to possible service gaps.

The use of Voronoi diagrams for this analysis offers an objective means to delineate service areas, based on the premise that geometric simplicity can enhance the understanding of infrastructure distribution without the bias inherent in more subjective analytical methods. This geometric approach aims to provide an unbiased view of how evenly distributed the charging infrastructure is across the urban landscape, potentially highlighting areas underserved by current station placements. Moreover, the study’s methodology allows for an implicit representation of charging station density. Areas with a high concentration of smaller Voronoi cells indicate a robust presence of charging stations, suggesting better service coverage, and reduced individual station load. In contrast, larger cells highlight areas where additional stations could significantly improve service availability and user convenience.

While other analytical tools like Network Analyst offer capabilities for modeling service areas based on actual travel distances and specific criteria, the simplicity and objectivity of Voronoi diagrams make them particularly suited for large-scale, broad-brush analyses. This methodological choice reflects a strategic decision to prioritize spatial equity in service provision, aiming to identify and address coverage disparities across Bangkok. To enhance the accuracy of the infrastructure planning, network analysis, which considers real-world travel distances and urban road networks, may be integrated into our future work. This combined approach could potentially assist in addressing the complexities of urban environments and providing more precise service area delineations.

By employing Voronoi diagrams, the study bypasses the more nuanced, but potentially subjective, inputs required by alternative analytical tools, providing a clear, straightforward picture of service area distribution. This approach underscores the critical role of methodological clarity and objectivity in urban infrastructure planning, especially in the dynamic and complex urban environment of Bangkok.

4.3. Service Load of Electric Vehicle Charging Stations on Population Data in Bangkok

The study’s analysis into the population density within the service areas of EV charging stations across Bangkok reveals a varied spectrum of service demands. Specifically, areas characterized by a population density ranging from 1000 to 2000 people per square kilometer per charging station, which encompass 163 service areas or 27.82% of the total evaluated, exhibit what can be classified as moderate service loads. This categorization is further expanded upon when considering the broader range of densities observed, leading to an average population density of 2696 people per square kilometer per service area. Such findings underscore the existence of high service loads across a significant portion of the charging station network, highlighting the critical need for nuanced and strategic infrastructure planning to effectively cater to areas facing elevated demand.

Integrating the operational characteristics of EV charging further illuminates these findings. Unlike the process of refueling conventional vehicles, charging electric vehicles entails considerably longer durations and offers greater flexibility regarding the location of charging points, including residential settings. This distinction introduces additional layers to the analysis of service loads, as the inherent differences in EV charging behavior—stemming from factors such as charging time requirements and the potential for home charging solutions—complicate direct comparisons with traditional fueling patterns. The flexibility afforded by residential charging options, for instance, might alleviate some of the demand pressures on public charging infrastructure in areas of high population density, yet it also emphasizes the importance of a diversified charging network that accommodates a range of user preferences and needs.

This complex interplay between charging station service loads and the operational dynamics of electric vehicle charging highlights the multifaceted challenges inherent in developing a robust and responsive EV charging infrastructure. Addressing these challenges necessitates a comprehensive approach to planning that not only accounts for spatial and demographic factors but also integrates a deep understanding of EV user behaviors and preferences. Such an approach ensures that infrastructure development is not only strategically aligned with the areas of highest demand but is also responsive to the evolving landscape of electric vehicle utilization.

4.4. Infrastructure Requirements for Effective EV Charging Networks

To ensure the efficiency and effectiveness of EV charging networks, several critical infrastructure requirements must be addressed. Robust network connectivity is paramount for the remote monitoring, management, and payment processing of EV charging stations. This connectivity not only facilitates real-time updates and maintenance but also enhances the overall user experience by enabling seamless communication and control over the infrastructure [20,21]. Detailed discussions on the importance of robust network connectivity highlight how it ensures that station performance can be optimized, and issues can be promptly addressed, thereby increasing the reliability and accessibility of the charging services. Moreover, deploying smart charging infrastructure that includes features such as load management, peak shaving, and demand response is crucial for balancing energy loads and reducing peak demand. These smart features are essential for integrating renewable energy sources into the grid, ensuring that the charging network can operate efficiently without overloading the power system, especially during high-demand periods. For example, Singh et al. (2023) discussed how smart charging can balance energy loads and integrate renewable sources into the grid [22]. This approach promotes a more sustainable and resilient energy system, aligning with global efforts to enhance the sustainability of urban transportation. Adherence to interoperability standards, such as the Open Charge Point Protocol (OCPP), is another vital aspect of EV charging infrastructure. These standards ensure compatibility and interoperability between different charging stations and network operators, allowing EV owners to access a wide network of charging stations using a single account or payment method [23]. This interoperability is critical for providing a seamless and convenient user experience, thereby encouraging the adoption of electric vehicles. Furthermore, implementing advanced data analytics and monitoring systems is essential for tracking charging station performance, energy consumption, and user behavior. These data-driven insights are invaluable for optimizing network operations, understanding usage patterns, and informing future infrastructure development [21]. The use of exploratory data analysis methodologies helps in assessing various indicators such as energy demand, usage intensity, and the environmental impact of the charging infrastructure, thereby supporting informed decision-making for future improvements. By addressing these infrastructure requirements, the effectiveness and efficiency of EV charging networks can be significantly enhanced, contributing to the broader adoption of electric vehicles and the sustainability of urban transportation systems.

4.5. Suitable Areas for Installing Additional Electric Vehicle Charging Stations in Bangkok

The application of Voronoi circles to identify areas suitable for new charging stations highlighted regions of extensive overlap in green, indicating high suitability. This analysis, prioritizing mathematical precision over subjective assessments like the Analytic Hierarchy Process (AHP), aims to offer a more objective basis for identifying expansion areas. The comparison of suitable areas identified in 2020 with the actual locations of charging stations established by 2023 illustrates the practical application of this analysis in guiding infrastructure development. However, the emergence of stations in less suitable areas suggests that additional factors—beyond geometric suitability—play a role in infrastructure deployment decisions.

Figure 9 presents a detailed visual comparison that elucidates the relationship between the geospatial analysis of suitability for EV charging stations and the practical development of charging infrastructure in Bangkok. This analysis, based on the overlay of Voronoi circles generated in 2020 with the actual locations of charging stations recorded in 2023, reveals a complex picture of infrastructure development across the city.

The coordinates at 13°33′ N, 36° E, alongside 13°46′ N, 44° E, are highlighted as key locations where the theoretical model of suitability closely matched real-world development outcomes. These areas, pinpointed as highly suitable in the 2020 study, indeed saw the introduction of new charging stations by 2023, affirming the predictive value of Voronoi-based spatial analysis. This alignment underscores the utility of Voronoi diagrams in pinpointing strategic sites that can significantly enhance service coverage and accessibility due to their advantageous positioning.

However, the analysis also identifies a notable trend of infrastructure development within areas classified as less suitable in the 2020 study. About 46.8% of the new stations were set up in areas deemed less suitable, while moderately suitable locations accounted for 24.6% of new installations. The emergence of stations in these less and moderately suitable areas, especially visible in the bottom right corner of the 2023 map, indicates a departure from theoretical predictions, driven by the reality of multiple factors influencing site selection for new charging stations. This deviation calls for a reassessment of what constitutes a ‘suitable’ area, highlighting that considerations extend beyond mere geometric optimization to include land availability, economic viability, regulatory environments, and localized demand patterns. Furthermore, as we stated in the materials and methods section, the analysis of service areas and population density per charging station using Voronoi theory independently assumes a uniform demand distribution.

To address these discrepancies, the findings advocate for a more integrated approach to suitability analysis. By adding layers of data, such as distinguishing between residential and commercial station locations, refining population density figures, and incorporating detailed urban planning insights, the predictive model’s accuracy for identifying prime locations for charging infrastructure can be significantly improved. Moreover, future work should incorporate demographic data to adjust demand estimates for different areas. Analyzing EV adoption rates in different parts of the city and considering traffic patterns and the locations of major employment centers will provide a more precise understanding of service needs. By integrating these factors, the analysis can better account for the complex dynamics of urban environments and ensure more equitable and effective placement of EV charging stations. Such an enriched methodology not only sharpens the forecast for optimal sites for new installations but also ensures a closer alignment between theoretical planning and the operational realities dictating infrastructure deployment.

Additionally, the accuracy of Voronoi diagrams in representing service boundaries and population density per charging station is inherently linked to the resolution of the underlying geographic data. High-resolution data, which includes detailed information on the precise locations of charging stations and accurate population distributions, ensures that the resulting Voronoi diagrams are more precise and reliable. Conversely, low-resolution data can lead to inaccuracies in delineating service areas and estimating population densities, potentially resulting in erroneous conclusions about service coverage and demand. It is important to recognize that the resolution of spatial data can vary significantly depending on the source and method of data collection. For example, detailed demographic surveys and high-precision GPS coordinates offer high-resolution data, while broader census data and less precise location information provide lower resolution. The use of high-quality, high-resolution data is essential for minimizing errors and enhancing the reliability of the analysis.

Author Contributions

Conceptualization, S.B. and N.P.; Methodology, S.B.; Formal analysis, N.P.; Resources, S.B., P.V. and W.P.; Data curation, N.P.; Writing—original draft, N.P.; Writing—review & editing, S.B. and N.P.; Visualization, N.P.; Funding acquisition, P.V. and W.P. All authors have read and agreed to the published version of the manuscript.

Funding

The research was funded by Thailand Science Research and Innovation: Basic Research Fund, Fiscal Year 2023, under project number FRB660073/0164. Additionally, partial APC funding was provided by the Kasetsart University Research and Development Institute.

Data Availability Statement

The data presented in this study are available on request from the corresponding author.

Conflicts of Interest

The authors declare no conflict of interest.

References

Ferkingstad, E.; Løland, A.; Wilhelmsen, M. Causal Modeling and Inference for Electricity Markets. Energy Econ. 2011, 33, 404–412. [Google Scholar] [CrossRef]
Dai, Y.-H.; Xie, W.-J.; Jiang, Z.-Q.; Jiang, G.J.; Zhou, W.-X. Correlation Structure and Principal Components in Global Crude Oil Market. Empir. Econ. 2016, 51, 1501–1519. [Google Scholar] [CrossRef]
CEIC Data. Thailand Oil Consumption; CEIC: Kuala Lumpur, Malaysia, 2023; pp. 1965–2023. [Google Scholar]
Antonio, F.J.; Mendes, R.S.; Thomaz, S.M. Identifying and Modeling Patterns of Tetrapod Vertebrate Mortality Rates in the Gulf of Mexico Oil Spill. Aquat. Toxicol. 2011, 105, 177–179. [Google Scholar] [CrossRef]
Kronick, D.; Rodríguez, F. Political Conflict and Economic Growth in Post-Independence Venezuela. In Roots of Underdevelopment: A New Economic and Political History of Latin America and the Caribbean; Springer: Berlin/Heidelberg, Germany, 2023. [Google Scholar]
Zeynalov, A.; Tiron, K. Macroeconomic Performance of Oil Price Shocks in Russia. arXiv 2022, arXiv:2211.04954. [Google Scholar]
Venter, I. Electric Car Sales Hit 6.6-Million in 2021, Tripling Their Market Share from 2019–IEA. Engineering News. 2022. Available online: https://www.engineeringnews.co.za/article/electric-car-sales-hit-66-million-in-2021-tripling-their-market-share-from-2019-iea-2022-02-07 (accessed on 8 April 2024).
Vitta, S. Electric Cars–Assessment of ‘Green’ Nature Vis-à-Vis Conventional Fuel Driven Cars. Heliyon 2021, 7, e00339. [Google Scholar]
Khandakar, A.; Rizqullah, A.; Berbar, A.A.A.; Ahmed, M.R.; Iqbal, A.; Chowdhury, M.E.H.; Zaman, S.M.A.U. A Case Study to Identify the Hindrances to Widespread Adoption of Electric Vehicles in Qatar. Energies 2020, 13, 3994. [Google Scholar] [CrossRef]
Money Buffalo. Why “100% Electric Cars” in Thailand Must Wait for Another 14 Years? Money Buffalo: Bangkok, Thailand, 2021. [Google Scholar]
Tiu, J.; Ramharack, S.; Hosein, P. On the Optimal Placement of Electric Vehicle Charging Stations. arXiv 2023, arXiv:2306.11069. [Google Scholar] [CrossRef]
Kong, F.; Liu, X. Smart Charging for Electric Vehicles: A Survey From the Algorithmic Perspectives. arXiv 2016, arXiv:1607.07298. [Google Scholar]
Arief, M.; Akhra, Y.; Vanany, I. A Robust and Efficient Optimization Model for Electric Vehicle Charging Stations in Developing Countries Under Electricity Uncertainty. arXiv 2023, arXiv:2307.05470v1. [Google Scholar] [CrossRef]
Wang, L.; Yang, C.; Zhang, Y.; Bu, F. Research on Multi-Objective Planning of Electric Vehicle Charging Stations Considering the Condition of Urban Traffic Network. arXiv 2022, arXiv:2208.12921v1. [Google Scholar] [CrossRef]
Yaagoubi, R.; Yarmani, M.E.; Kamel, A.; Khemiri, W. HybVOR: A Voronoi-Based 3D GIS Approach for Camera Surveillance Network Placement. ISPRS Int. J. Geo-Inf. 2015, 4, 754–782. [Google Scholar] [CrossRef]
Wang, J.; Kwan, M.-P. Hexagon-Based Adaptive Crystal Growth Voronoi Diagrams Based on Weighted Planes for Service Area Delimitation. ISPRS Int. J. Geo-Inf. 2018, 7, 257. [Google Scholar] [CrossRef]
Figurska, M.; Dawidowicz, A.; Zysk, E. Voronoi Diagrams for Senior-Friendly Cities. Int. J. Environ. Res. Public Health 2022, 19, 7447. [Google Scholar] [CrossRef] [PubMed]
Luangwilai, T.; Welamas, W.; Leelayuth, S.; Meechowna, S. Optimal Path in a Hostile Environment by Using Voronoi Diagram. J. Def. Technol. Inst. 2020, 2, 93. [Google Scholar]
Okabe, A.; Boots, B.; Sugihara, K.; Chiu, S.N. Spatial Tessellations: Concepts and Applications of Voronoi Diagrams, 2nd ed.; Wiley: Hoboken, NJ, USA, 2000. [Google Scholar]
Dimitriadou, K.; Rigogiannis, N.; Fountoukidis, S.; Kotarela, F.; Kyritsis, A.; Papanikolaou, N. Current Trends in Electric Vehicle Charging Infrastructure; Opportunities and Challenges in Wireless Charging Integration. Energies 2023, 16, 2057. [Google Scholar] [CrossRef]
Lucas, A.; Prettico, G.; Flammini, M.G.; Kotsakis, E.; Fulli, G.; Masera, M. Indicator-Based Methodology for Assessing EV Charging Infrastructure Using Exploratory Data Analysis. Energies 2018, 11, 1869. [Google Scholar] [CrossRef]
Singh, P.P.; Wen, F.; Palu, I.; Sachan, S.; Deb, S. Electric Vehicles Charging Infrastructure Demand and Deployment: Challenges and Solutions. Energies 2023, 16, 7. [Google Scholar] [CrossRef]
Anadón Martínez, V.; Sumper, A. Planning and Operation Objectives of Public Electric Vehicle Charging Infrastructures: A Review. Energies 2023, 16, 5431. [Google Scholar] [CrossRef]

Figure 1.Research methodology overview.

Figure 2.The NNI value and charging station density across Bangkok.

Figure 3.Spatial distribution of charging stations along the road network in Bangkok.

Figure 4.Charging stations distribution in relation to city planning zones.

Figure 5.Distribution of charging stations by land use categories in Bangkok.

Figure 6.Service area boundaries of EV charging stations in Bangkok.

Figure 7.Population density and service demand for EV charging stations across Bangkok.

Figure 8.Optimal locations for new EV charging stations based on Voronoi diagram overlaps in Bangkok.

Figure 9.Analysis of EV charging station suitability vs. actual installations (2020, 2023).

Table 1.Types and frequencies of charging station locations.

EV Public Charging Station
Type Place	Count	Percent
Village/Condo/Residence/Hotel	119	20.31%
Shopping Complex	96	16.38%
Gas Station	92	15.7%
Car Service Center	81	13.82%
Office Building	47	8.02%
Public Agency	25	4.27%
Parking Area	21	3.58%
Department Store	17	2.9%
Convenience Store	16	2.73%
Restaurant/Coffee Shop	16	2.73%
School/University	13	2.22%
Bank	11	1.88%
Hospital	9	1.54%
Home Improvement Store	7	1.19%
Sports Complex	6	1.02%
Market	5	0.85%
Foundation Place	2	0.34%
Amusem*nt Park	1	0.17%
Park	1	0.17%
Shopping Complex/Hospital	1	0.17%
Total	586	100%

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).