UNIVERSIDADE ESTADUAL PAULISTA “JÚLIO DE MESQUITA FILHO” 

FACULDADE DE ENGENHARIA DE ILHA SOLTEIRA 

PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA ELÉTRICA 

 

 

 

IGOOR MORRO MELLO 

 

 

 

Spatial-temporal approach of electric vehicles in urban zones: traffic, 

infrastructure, decrease in local pollution, recharge, and impact on the electric 

network. 

 

 

 

 

 

 

 

 

 

 

Ilha Solteira 
2022 

 



 
 

 
 

UNIVERSIDADE ESTADUAL PAULISTA “JÚLIO DE MESQUITA FILHO” 

FACULDADE DE ENGENHARIA DE ILHA SOLTEIRA 

PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA ELÉTRICA 

 

 

IGOOR MORRO MELLO 

 

 

Spatial-temporal approach of electric vehicles in urban zones: traffic, 

infrastructure, decrease in local pollution, recharge, and impact on the electric 

network. 

 

 
 
 
Thesis presented to the Faculty of 
Engineering of the Campus of Ilha Solteira 
UNESP as part of the requirements for 
obtaining the title of Doctor in Electrical 
Engineering. 
Area of concentration: Automation. 

 

Prof. Dr. Antonio Padilha Feltrin 
Orientador 

 

 

 

 

 

     

                                                           

Ilha Solteira 
2022 



 
 

 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 



 
 

 
 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 



 
 

 
 

AGRADECIMENTOS 

 

À Deus, criador inefável, que é fonte da verdadeira luz e da ciência. 

Agradeço aos meus pais Luciene Morro Mello e Márcio Orélio de Mello que rezaram 

por mim diariamente durante todos estes anos e estiveram em toda a minha jornada. 

Agradeço à minha noiva Lílian de Fátima e Silva pela busca da santidade ao meu 

lado, com um sério caminho baseado na verdade, fé e amor para juntos chegarmos 

ao céu. 

Agradeço ao meu orientador Dr. Antonio Padilha Feltrin pelo auxílio durante todo o 

trabalho de doutorado e ao Dr. Joel David Melo Trujillo pelas contribuições que 

enriqueceram a tese. 

Agradeço ao meu diretor espiritual Dr. Luiz Henrique Brandão de Figueiredo, padre 

diocesano que me auxilia no caminho de santidade, com uma vida dedicada as 

coisas do alto, baseada na verdade e na fé. 

Agradeço aos professores de Engenharia Elétrica da UNESP-FEIS e ao Laboratório 

de Planejamento de Sistemas de Energia Elétrica – LaPSEE e seus integrantes, em 

especial meu amigo Henrique Molina Barradas. Agradeço ao Instituto SENAI de 

Tecnologia em Automação (IST Automação) e colegas de trabalho. 

O presente trabalho foi realizado com apoio da Coordenação de Aperfeiçoamento de 

Pessoal de Nível Superior – Brasil (CAPES) – Código de Financiamento 001; 

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) (Processos 

2015/21972-6, 2017/01909-3, 2017/22577-9, 2019/00466-6); Conselho Nacional de 

Desenvolvimento Científico e Tecnológico-Brazil (CNPq) (processos: 307281/2016-7, 

422044/2018-0); pelo Instituto Nacional de Ciência e Tecnologia de Energia Elétrica 

(INCT-INERGE). Agradecimentos à empresa AIMSUN pela licença concedida do 

Aimsun Next Postgraduate License para o desenvolvimento da tese de doutorado. 

 

 



 
 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

“Cumpre o pequeno dever de cada momento; faz o que deves e está no que fazes.” 

São Josemaría Escrivá 



 
 

 
 

ABSTRACT 

Electric vehicles have been encouraged in recent years as they have become 

widespread in society in terms of new technology, more efficient than combustion 

vehicles, and can reduce local pollution. However, due to the high price for purchase 

an electric vehicle, it is expected that their acquisition will occur gradually through 

financial incentives, price reduction, and diffusion of technology in society. With the 

increase in the purchase of these vehicles over the years, network infrastructure will 

meet the demand for recharging these vehicles in the urban zone and planning the 

distribution networks to meet the demand for recharging. Thus, this work proposes a 

complete methodology for the introduction of electric vehicles in an urban zone. First, 

a method is proposed for the acquisition of electric vehicles over time. With the 

introduction, a vehicle traffic methodology is proposed to find variables such as state 

of charge and driving patterns in the urban zone. With the results of vehicular traffic, 

charging stations are presented in the most critical locations. A module is created to 

find the local pollution reduction with the introduction of electric vehicles over the 

period. The increase in residential load from residential night charging is calculated, 

and finally, a methodology for connecting the charging stations to the distribution 

network is proposed. The method is applied in an urban zone of about 200,000 

inhabitants using and integrating computational programming, traffic, geographic 

information systems, and energy distribution systems tools. The results are shown 

through spatial thematic maps over the study period. Helps to assist distribution 

companies in decision-making and planning, owners of charging stations, owners of 

electric vehicle dealerships, urban planners, and other public and private agents who 

will be involved in introducing electric vehicles in some stage urban zones. 

Keywords: electric vehicles; geospatial analysis; power distribution system planning; 

spatial load forecasting; traffic and transport network; urban infrastructures and 

mobility. 

 

 

 

 



 
 

 
 

RESUMO 

Veículos elétricos têm sido incentivados nos últimos anos à medida que se tornam 

difundidos na sociedade em termos de uma nova tecnologia, por serem mais 

eficientes que os veículos a combustão e podendo reduzir a poluição local. No 

entanto, devido aos preços ainda elevados para aquisição dos veículos elétricos, 

espera-se que sua aquisição ocorra de forma gradual, através de incentivos 

financeiros, diminuição do preço, difusão da tecnologia na sociedade. Com o 

aumento, mesmo que gradual, da aquisição destes veículos ao longo dos anos, será 

necessária uma rede de infraestrutura para atender a demanda de recarga destes 

veículos na zona urbana e um planejamento das redes de distribuição para que se 

tenha demanda para a recarga. Assim, este trabalho propõe uma metodologia 

completa para a introdução dos veículos elétricos em uma zona urbana. 

Primeiramente, uma metodologia é proposta para a aquisição de veículos elétricos 

ao longo do tempo. Com a introdução, uma metodologia de tráfego veicular é 

proposta para encontrar as variáveis relacionadas aos veículos elétricos como 

estado da carga e padrões de condução na zona urbana. Com os resultados do 

tráfego veicular, são propostas estações de recarga nos locais mais críticos. Um 

módulo para obter a diminuição da poluição local com a introdução dos veículos 

elétricos ao longo do período é criada. Calcula-se o aumento de carga residencial 

oriundo da recarga residencial noturna e por fim uma metodologia para a conexão 

das estações de recarga na rede de distribuição é proposta. A metodologia é 

aplicada em uma zona urbana de cerca de 200 mil habitantes com o uso e 

integração de ferramentas de programação computacional, de tráfego, sistemas de 

informação geográfica e dos sistemas de distribuição de energia. Os resultados são 

mostrados através de mapas temáticos espaciais ao longo do período de estudo e 

auxiliam as empresas de distribuição na tomada de decisões e planejamento, donos 

de estações de recarga, donos de concessionárias de veículos elétricos, 

planejadores urbanos e demais agentes públicos e privados que estarão envolvidos 

em alguma etapa do processo de introdução dos veículos elétricos nas zonas 

urbanas. 

Palavras Chaves: análise geoespacial; infraestrutura e mobilidade urbana; rede de 

tráfego e transportes; veículos elétricos; previsão espacial de carga; planejamento 

de sistemas de distribuição de energia. 



 
 

 
 

LIST OF FIGURES 

Figure 1 – Examples of the types of DNCEV……...……………………………………20 

Figure 2 – Modeling and simulation levels – City………………...…………………….24 

Figure 3 – Modeling and simulation levels – Vehicles………………………...…….…25 

Figure 4 – Modes of charging a DNCEV………………………………..……………….29 

Figure 5 – Stationary WPT for a private DNCEV……..…………………………….….29 

Figure 6 – Stationary WPT for a BEB and PHEB…………..……………..…………...30 

Figure 7 – Wireless Power Transfer System……………………………………………30 

Figure 8 – LEM micro trips - speed vs. time in an urban drive cycle…………….…32 

Figure 9 – Example of the spatial map with local pollution caused by vehicles…….33 

Figure 10 – Distribution network model in OpenDSS……………………………….....36 

Figure 11 – DNCEV shortest path algorithm……………..…………………………..…58 

Figure 12 – Example of a spatial intersection…………………………………………..62 

Figure 13 – Possible connections for CS………….…………………………………….67 

Figure 14 – Flowchart with the allocation algorithm..…………………………...……..71 

Figure 15 – Flowchart with the methodology steps…………………………………….72 

Figure 16 – Subareas of the urban zone………………………..………………………73 

Figure 17 – Streets of the urban zone…………………………………………….……..74 

Figure 18 – Spatial lines and substation of the urban zone…………………………...74 

Figure 19 – Spatial transformers of the urban zone……………………………………75 

Figure 20 – Percentage of vehicles traveling the city at each hour……………..……76 

Figure 21 – Spatial distribution of DNCEV in year 0…………………….……...…….77 

Figure 22 – Spatial distribution of DNCEV in year 1…………………….……...…….77 

Figure 23 – Spatial distribution of DNCEV in year 2…………………….……...…….78 

Figure 24 – Spatial distribution of DNCEV in year 3…………………….……...…….78 



 
 

 
 

Figure 25 – Spatial distribution of DNCEV in year 4…………………….……...…….79 

Figure 26 – Spatial distribution of DNCEV in year 5…………………….……...…….79 

Figure 27 – Streets visited by DNCEV in year 0…………………...………………….81 

Figure 28 – Streets visited by DNCEV in year 1…………………...………………….82 

Figure 29 – Streets visited by DNCEV in year 2…………………...………………….82 

Figure 30 – Streets visited by DNCEV in year 3…………………...………………….83 

Figure 31 – Streets visited by DNCEV in year 4…………………...………………….83 

Figure 32 – Streets visited by DNCEV in year 5…………………...………………….84 

Figure 33 – SOC in year 0………………………………………………………...………84 

Figure 34 – SOC in year 1………………………………………………………...………85 

Figure 35 – SOC in year 2………………………………………………………...………85 

Figure 36 – SOC in year 3………………………………………………………...………86 

Figure 37 – SOC in year 4………………………………………………………...………86 

Figure 38 – SOC in year 5………………………………………………………...………87 

Figure 39 – CS in year 0……………………………………………….………………….89 

Figure 40 – CS in year 1……………………………………………….………………….89 

Figure 41 – CS in year 2……………………………………………….………………….90 

Figure 42 – CS in year 3……………………………………………….………………….90 

Figure 43 – CS in year 4……………………………………………….………………….91 

Figure 44 – CS in year 5……………………………………………….………………….91 

Figure 45 – ROI of CS…….………………..………………………….………………….92 

Figure 46 – Example of average DNCEV SOC enters and leaves the CS……..…...93 

Figure 47 – Spatial local pollution in year 0 – CO2…………………………………….95 

Figure 48 – Spatial local pollution in year 1 – CO2…………………………………….95 

Figure 49 – Spatial local pollution in year 2 – CO2…………………………………….96 



 
 

 
 

Figure 50 – Spatial local pollution in year 3 – CO2…………………………………….96 

Figure 51 – Spatial local pollution in year 4 – CO2…………………………………….97 

Figure 52 – Spatial local pollution in year 5 – CO2…………………………………….97 

Figure 53 – Spatial local pollution in year 0 – NOx…………………………….………98 

Figure 54 – Spatial local pollution in year 1 – NOx…………………………….………98 

Figure 55 – Spatial local pollution in year 2 – NOx…………………………….………99 

Figure 56 – Spatial local pollution in year 3 – NOx…………………………….………99 

Figure 57 – Spatial local pollution in year 4 – NOx…………………………….……..100 

Figure 58 – Spatial local pollution in year 5 – NOx………………………………...…100 

Figure 59 – Diversity factor for residential Charging Factor…………………………102 

Figure 60 – Maximum diversified demand in kW - year 0……………………..……..102 

Figure 61 – Maximum diversified demand in kW - year 1……………………..……..103 

Figure 62 – Maximum diversified demand in kW - year 2……………………..……103 

Figure 63 – Maximum diversified demand in kW - year 3……………………..……104 

Figure 64 – Maximum diversified demand in kW - year 4……………………..……104 

Figure 65 – Maximum diversified demand in kW - year 5……………………..……105 

Figure 66 – Lower-cost connections – year 0…………………………………..……107 

Figure 67 – Lower-cost connections – year 1…………………………………..……108 

Figure 68 – Lower-cost connections – year 2…………………………………..……108 

Figure 69 – Lower-cost connections – year 3…………………………………..……109 

Figure 70 – Lower-cost connections – year 4…………………………………..……109 

Figure 71 – Lower-cost connections – year 5…………………………………..……110 

Figure 72 – Daily voltage curve in low voltage connection for a blue CS..…..……112 

Figure 73 – Daily voltage curve in low voltage connection for a gray CS.…..……112 

Figure 74 – Daily voltage curve in medium voltage connection for a black CS..…113 



 
 

 
 

Figure 75 – Daily voltage curve in medium voltage connection for a gray CS….…113 

Figure 76 – Daily voltage curve in medium voltage connection for a gray CS….…113 

Figure 77 – Integration of public transport……………………………………………115 

Figure 78 – Lower-cost connections stationary WPT CS..…………………………116 

Figure 79 – Spatial Comparison of allocation methodologies..……………………118 

Figure 80 – Comparison of allocation methodologies..………….…………………119 

Figure 81 – RI of CS in year 0……………………………...…………………………138 

Figure 82 – RI of CS in year 1……………………………...…………………………138 

Figure 83 – RI of CS in year 2……………………………...…………………………139 

Figure 84 – RI of CS in year 3……………………………...…………………………139 

Figure 85 – RI of CS in year 4……………………………...…………………………140 

Figure 86 – RI of CS in year 5……………………………...…………………………140 

 

 

 

 

 

 

 

 

 

 



 
 

 
 

LIST OF TABLES 

TABLE 1 – CHARACTERISTICS OF THE TYPES OF RECHARGE………………...28 

TABLE 2 – CS CHARACTERISTIC………………………………………………...……88 

TABLE 3 – TYPE OF CS…………………………………………………….……………92 

TABLE 4 – GHG REDUCTION OVER THE YEARS…………….….………………..101 

TABLE 5 – FINAL COSTS……………………….…………….….…………………….111 

TABLE 6 – OVERLOAD ELEMENTS……………..………….….…………………….114 

TABLE 7 – VALUES FOUND BY THE METHODOLOGIES.….…………………….118 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 
 

 
 

LIST OF ACRONYMS AND ABBREVIATIONS 

BEB  Battery Electric Bus 

BEV  Battery Electric Vehicle 

CO2  Carbon Dioxide 

CS  Charging Station 

CV  Coefficient of Variation 

DNCEV Distribution Network Connection Electric Vehicle 

EREV  Extended-Range Electric Vehicles 

EGT  Electric Garbage Truck 

ET  Electric Taxi 

EV  Electric Vehicle 

FCEV  Fuel Cell Electric Vehicle 

GHG  Greenhouse Gases 

GIS  Geographic Information System 

GWR  Geographically weighted regression 

HEV  Hybrid Electric Vehicle 

HSAR  Hierarchical spatial autoregressive 

ICEV  Internal Combustion Engine Vehicle 

LEM  London Emission Model 

MCS  Monte Carlo Simulation 

MHV  Mild hybrid vehicle 

MIHV  Micro-Hybrid Vehicle 

NOx  Nitrogen Oxides 



 
 

 
 

O-D  Origins-destinations 

PHEB  Plug-in Hybrid Electric Bus 

PHEV  Plug-in Hybrid Electric Vehicle 

PHGT  Plug-in Hybrid Garbage Truck 

RI  Region of Influence 

SOC  State of Charge 

TSS  Transport Simulation Systems 

V2G  Vehicle to Grid 

WPT  Wireless Power Transfer 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 
 

 
 

SUMMARY 

1. INTRODUCTION .................................................................................................................. 17 

1.1 CONTEXT ............................................................................................................................. 17 

1.2 INTRODUCTION OF EVS IN URBAN ZONES .................................................................... 18 

1.3 URBAN TRAFFIC MODEL ................................................................................................... 23 

1.4 INFRASTRUCTURE FOR CHARGING DNCEV .................................................................. 27 

1.5 DECREASE IN LOCAL POLLUTION ................................................................................... 31 

1.6 INFRASTRUCTURE AND DNCEV ELECTRIC NETWORK CONNECTION ...................... 33 

1.7 CONTRIBUTIONS TO RESEARCH ..................................................................................... 37 

2. LITERATURE REVIEW ........................................................................................................ 40 

2.1 DNCEV ACQUISITION REVIEW ......................................................................................... 40 

2.2 TRAFFIC MODEL REVIEW ................................................................................................. 43 

2.3 INFRASTRUCTURE FOR CHARGING REVIEW ................................................................ 45 

2.4 GHG REDUCTION REVIEW ................................................................................................ 47 

2.5 IMPACT ON THE DISTRIBUTION NETWORK REVIEW .................................................... 49 

3. PROPOSED METHODOLOGY............................................................................................ 51 

3.1 MODULE 1: ACQUISITION OF DNCEV .............................................................................. 51 

3.2 MODULE 2: TRAFFIC MODEL ............................................................................................ 54 

3.3 MODULE 3: INFRASTRUCTURE FOR CHARGING DNCEV ............................................. 59 

3.4  MODULE 4: DECREASE IN LOCAL POLLUTION .............................................................. 61 

3.5  MODULE 5: IMPACT ON THE DISTRIBUTION NETWORK.............................................. 63 

3.5.1 Residential impact ................................................................................................................. 63 

3.5.2 CS distribution network connection ....................................................................................... 64 

3.6   ALGORITHM FOR ALL MODULES ................................................................................... 71 

4. RESULTS AND DISCUSSION............................................................................................. 72 

4.1 RESULTS OF INTRODUCTION OF DNCEV ...................................................................... 76 

4.2 RESULTS OF TRAFFIC SIMULATION ................................................................................ 80 

4.3 RESULTS OF INFRASTRUCTURE FOR CHARGING DNCEV .......................................... 88 

4.4 RESULTS OF DECREASE IN LOCAL POLLUTION ........................................................... 94 

4.5 RESULTS OF RESIDENTIAL IMPACT .............................................................................. 101 

4.6 RESULTS OF CS CONNECTION ...................................................................................... 105 

4.7 RESULTS OF STATIONARY WPT CS FOR BEB ............................................................. 115 

4.8 COMPARISON OF RESULTS ........................................................................................... 116 

4.9 FUTURE WORKS ............................................................................................................... 120 

5. CONCLUSION .................................................................................................................... 122 

          REFERENCES ................................................................................................................... 124 

          APPENDIX 1 ...................................................................................................................... 136 



17 
 

 
 

1 INTRODUCTION 

1.1 CONTEXT 

The use of Electric vehicles (EVs) in urban zones has intensified in the last 

years in several countries as a part of new trends in the automotive sector, helping to 

reduce the emission of greenhouse gases (GHG) (ZIMM, 2020). Furthermore, 

besides helping reduce local pollution, the electrification of transportation contributes 

to reducing other environmental, social, and economic problems (PARDO-BOSCH et 

al., 2021). 

The high-income public initially acquired the high price to purchase an EV in 

developed countries. However, with the increase in multi-level recharge 

infrastructure, maintenance, regulations, tax incentives, advertising, public 

awareness, a decrease in price, and the introduction of different EVs, the number of 

them has expanded in these countries. 

Over these years of changes in the transportation system in developed 

countries, with the introduction of new technologies and challenges, developing 

countries have become almost passive observers of transportation electrification. 

Moreover, the high price for EVs acquisition, lack of infrastructure, and other varied 

problems have made their introduction difficult. However, with the recent increase in 

environmental concern, the need to reduce local pollution in urban zones, combined 

with the search for new technologies, is becoming a reality (BAMISILE et al., 2021). 

These countries will detect several challenges in the coming years to electrify 

transportation. They will have technological, economic, social, environmental, cultural 

challenges that can make it difficult for EVs to enter their markets (GOEL; SHARMA; 

RATHORE, 2021). To mitigate such problems, the present work proposes introducing 

EVs in urban zones considering several aspects that will gradually help use EVs. 

Moreover, this work tries to cause a minor problem on the systems. As an example, 

this work will help distribution companies to find connection points in the network, 

prioritizing the connection in the current elements and consequently reducing 

expansion costs. The systems that will be impacted are considered in this work, to 

help the most significant number of agents participating in some stage. 



18 
 

 
 

This work presents a different model that will be integrated to consider the 

dynamic of EVs in the urban zones. The proposed methodology is composed by five 

modules. First, an approach proposes the gradual introduction of several EVs in the 

urban zones over the years. Second, an urban traffic model considering the EVs and 

their characteristics is presented. Third, this work creates criteria to introduce 

different infrastructures for EVs. Fourth, a model that locates and quantifies the 

decrease in local pollution with the introduction of EVs is presented. And finally, the 

fifth, the impact in the distribution network is studied and mitigated, with a model that 

explores the possibilities of connection of EVs in the network using home and 

charging stations (CS) modes of recharging. 

In the following, each of these modules will be covered in its most varied 

aspects. A theoretical and state-of-the-art approach will be shown with the main 

features of each of the modules. Finally, it is essential to emphasize the 

multidisciplinary of this work, contemplating several aspects related to EVs and their 

infrastructure. 

1.2 INTRODUCTION OF EVS IN URBAN ZONES 

The introduction of EVs in urban zones must consider financial, economic, 

social, infrastructure, incentive advertisements, and exemptions. For this penetration, 

it is necessary to identify acquisition patterns for the various types of existing EVs. 

They can be divided into the following categories (DENTON, 2018): 

i) Battery electric vehicle (BEV): EV whose power is provided only by the 

battery as a source of electrical energy. 

ii) Plug-in hybrid electric vehicle (PHEV): A vehicle with a battery 

connected to the electrical network and an internal combustion engine 

(ICE). It generally changes its operation after the electrical autonomy is 

exhausted. 

iii) Extended-range electric vehicle (EREV): A vehicle powered by a battery 

that receives power from a generator coupled to an internal combustion 

engine. Similar to BEV but with low autonomy. The autonomy increases 

through a generator. The electric motor always provides movement, and 

this system can be called a series hybrid. 



19 
 

 
 

iv) Hybrid electric vehicle (HEV): Execute traction by a battery and/or by 

ICE. The vehicle selects the power source automatically, depending on 

speed, battery state of charge (SOC), and engine load. The battery is 

recharged using a regenerative braking system without using the power 

network (sometimes called parallel EV). 

v) Mild hybrid vehicle (MHV): Similar to HEV but do not operate only in 

electric mode. Generally, it uses the electric motor in acceleration 

moments. 

vi) Micro hybrid vehicle (MIHV): Has a start-stop system (the system that 

turns off the vehicles at short stops and automatically turns on when 

accelerating the vehicle), and the regenerative braking system is used to 

recharge the 12V battery. 

vii) Motorcycle and electric quad: powered by batteries that can connect to 

the distribution network. 

viii) Fuel cell electric vehicle (FCEV): EV which uses a fuel cell, instead of a 

battery, or in combination with a battery or supercapacitor, to power its 

onboard electric motor. Fuel cells in vehicles generate electricity to power 

the engine, mostly oxygen from the air and compressed hydrogen. 

ix) Vehicle to grid (V2G): Enables energy stored in EVs to be fed back into 

the electricity network to help supply energy at times of peak demand. 

x) Electric Taxi (ET): EV that can be subcategorized in i) to viii) but have 

different driving patterns because of its almost constant operation. 

xi) Trolleybus: A vehicle powered by electricity drawn from two overhead 

wires by trolley poles. Currently, trolleybuses can have batteries if the 

coupling system fails. 

xii) Battery Electric Bus (BEB): whose power is provided only by the battery 

as a source of electrical energy. 

xiii) Plug-in Hybrid Electric Bus (PHEB): a bus similar to PHEV. 

xiv) Electric Garbage Truck (EGT): Electric truck that travels the urban zone 

collecting the garbage. 

xv) Plug-in Hybrid Garbage Truck (PHGT): a garbage truck similar to PHEV. 

The focus of the present work is to analyze the dynamics of EVs that can 

connect to the distribution network and will be called hereafter by distribution network 



20 
 

 
 

connection electric vehicle (DNCEV). Therefore, DNCEV means that the EV can 

connect to the distribution network to recharge their battery (BEVs, PHEVs, EREVs, 

ETs, motorcycles, electric quads, BEBs, PHEB, EGT, PHGT). This work excludes 

V2Gs, as the state of the art of distribution network precludes a more detailed 

analysis of the behavior of these vehicles. Then, in Fig.1 are shown the examples of 

the types of DNCEV covered in this work. 

Figure 1 – Examples of the types of DNCEV 

Source: Author’s own elaboration. 

In Fig. 1, a) corresponds to BEV; b) to PHEV; c) to EREVs (where a to c can 

be ETs); d) represents motorcycles (electric squads are similar but with four wheels); 

e) represents BEBs, which can also be PHEB if they have a combustion motor and f) 

represents EGT which can also be PHGT if they have a combustion motor.  

Some developing countries such as Brazil have created laws and regulations 

to encourage their acquisition (MORRO-MELLO; PADILHA-FELTRIN; MELO, 2016), 

and further acquisition of DNCEV is expected for the next few years. However, the 

DNCEV price, lack of knowledge of technology, lack of infrastructure for 



21 
 

 
 

maintenance, few CS, low autonomy of batteries, lack of other incentives (such as 

exclusive corridors or use of bus lanes) are still limiting factors for its purchase 

(FAPESP RESEARCH, 2017). Nevertheless, although still slow, the insertion of EVs 

in Brazilian cities should occur in the coming decades. The insertion depends on 

incentive policies such as free CS, licenses for installing residential charging 

infrastructure, priority access (access to restricted corridors), parking, free access to 

high-occupancy vehicles or toll roads, and permission vehicles on the rotations are 

implemented (MACHADO et al., 2020).  

The high price for DNCEV in some developing countries will make that their 

purchase follows a heterogeneous pattern (HEYMANN et al., 2017; HEYMANN et al., 

2019). This pattern can be influenced by the socioeconomic dynamics of the city 

(ELNOZAHY; SALAMA, 2014). Higher-income regions tend to acquire DNCEV first 

and affect other areas over time (MORRO-MELLO et al., 2018). Thus, it is expected 

that there is a heterogeneous spatial distribution of DNCEV in the cities, with clusters 

in some regions with more DNCEV penetration. 

One difficulty purchasing DNCEV is that consumers may have a relatively poor 

understanding of fuel economy and the advantages of DNCEV operating costs. So, 

socioeconomic factors also influence the type of DNCEV to be acquired. As seen in 

(LANE et al., 2018), PHEVs have greater adherence to younger, wealthier, more 

studied, perception of ease of use and disadvantages of the technology and available 

charging infrastructure, compatible lifestyle, previous experience with a PHEV. The 

social interaction (hear about trusted colleagues' experience with vehicle 

technologies) can also be attracted to alternative vehicles. Their use can serve as a 

means by which to transmit attitudes of greenhouse gas (GHG) reduction or 

enthusiasm for new technologies, lifestyle choices, including financial management, 

activity, and fitness, and have a strong sense of community. Among the individuals 

who prefer a BEV are those who prioritize high fuel economy and environmental 

performance. The preference for purchasing a model for another of the BEVs 

considers size, price, space, and performance (REZVANI; JANSSON; BODIN, 2014). 

BEVs and PHEVs attract different subsets of Norway: BEV owners are more 

likely to live in urban centers and own multiple vehicles. In contrast, PHEV owners 

are geographically dispersed and more likely to live in single-vehicle households 



22 
 

 
 

(ZIMM, 2020). Moreover, in developing countries like Brazil, the lack of information 

between BEV, PHEV, and other DNCEV makes studies necessary to differentiate 

their acquisition, driving, flow, SOC, and charging patterns. 

The spatial estimation of DNCEV buyers is essential to assist in the electrical 

network planning studies with the increasing load due to the EVs' recharge. In 

countries with many DNCEV, it is observed that most private DNCEV owners perform 

recharges at night. In this way, the demand of the distribution network can be 

increased and cause problems, in the case of simultaneous charging and coinciding 

with the maximum peak of the system. In the transport system studies, the planning 

of CS, the driving pattern of DNCEV are essential to the transportation system, 

operation, and planning.  

In general, the influence of DNCEV buyers on their neighbors is not 

considered but important, as shown in (MORRO-MELLO; PADILHA-FELTRIN; 

MELO, 2017). The methodologies that model such impact use socioeconomic 

variables without modeling spatial interaction among inhabitants to obtain EVs. 

Spatial regressions can estimate the region’s most likely to acquire EVs per subarea 

considering the interaction among the inhabitants, modeling how initial acquisition in 

a subarea may influence other inhabitants.  

The results from these regressions can be spatial databases inserted in 

geographic information systems (GIS) to build spatial heat maps with the rate of EV 

buyers per subarea. These maps allow us to visualize the regions that may have a 

load growth because of the recharging of EVs. In this way, to show the usefulness of 

spatial regressions in load forecasting studies, this work presents in the first module 

the application of spatial regressions in a city, showing the different scenarios of load 

growth that can be obtained from such regressions. 

This work compares each of the stages of acquisition of DNCEV over five 

years, providing spatial heat maps with the rate of EVs per subarea. This work differs 

from others found in the literature because, models the spatial interactions among the 

inhabitants, creating a spatial database with the number of DNCEV acquired and the 

influence of the neighborhood in the acquisition of EVs. Such a database can be 

incorporated in any GIS available in the electric sector companies, helping to plan 



23 
 

 
 

residential load increase; the transport sector, with their preferences; in car 

dealerships, sales studies; in public companies and planning studies. 

1.3 URBAN TRAFFIC MODEL 

The use of DNCEV in developed countries is already a widespread 

technology. Therefore, it is possible to find a database with the circulation history of 

these vehicles, real-time SOC, residential and CS electrical consumption actual data, 

reduction of gases that intensify local pollution, operation of distribution networks with 

recharges, among other aspects. 

In developing countries like Brazil, the introduction of DNCEV is still slow, and 

the database with historical data does not exist. Therefore, to assist public and 

private planners involved in some stage of the introduction, maintenance, 

infrastructure, and production chain related to DNCEV, scenarios for all agents 

involved in this process are necessary. After an analysis of the introduction of these 

vehicles in the urban zone, the behavior along the streets and avenues of the cities 

(modeling of variables associated with DNCEV) should be performed to assist in the 

decision-making of the agents involved. 

Several analyzes can be performed with the help of transport simulation 

systems (TSS), more simplified, or with a higher degree of complexity. However, due 

to the characteristics of DNCEV and their need to connect to the network, it is 

necessary to acquire information on residential and CS charging, real-time SOC 

consumption, driving patterns, among others. To obtain the data, it is required to 

model and simulate characteristics of origins and destinations (O-D) in an urban 

zone. Also, it is essential to model the influence of other vehicles such as congestion 

points, stops, public transport, traffic control, CS location, recharge time in CS 

(MORRO-MELLO et al., 2019). 

For more simplified analyzes, the use of computational tools for modeling 

transport systems that allow macroscopic simulations (fewer variables and details, 

simulating the flow as a whole) is sufficient. As the degree of complexity and the 

number of variables to be modeled and information to be acquired increases, 

computational tools that allow mesoscopic simulations (modeling parameters for 

groups of vehicles) or microscopic simulations (modeling parameters of each vehicle, 



24 
 

 
 

with more detail) are necessary (PUPPO; TOSIN, 2021). In Fig. 2, an example of the 

modeling and simulation for a generic city is shown, and in Fig. 3, an example of the 

vehicle modeling at each simulation level. 

Figure 2 – Modeling and simulation levels - City 

 

Source: Author’s own elaboration. 



25 
 

 
 

Figure 3 – Modeling and simulation levels - Vehicles 

Source: AIMSUN, 2021. 

In Fig. 2, at the macroscopic, mesoscopic, and microscopic level, the 

information modeled and acquired after the simulations is, for the city, for groups 

divided spatially and for each street and avenue, respectively. In Fig. 3, the vehicles 

modeled and simulated at a macroscopic level have similar characteristics; at 

mesoscopic, vehicle groups have identical features, and each vehicle has its 

attributes at a microscopic level. 

Due to the large volume of information required for DNCEV, microscopic 

simulations (AIMSUN, 2021) should be performed for this vehicle category. These 

simulations consider the individual characteristics of the DNCEV and are possible to 

obtain unique data after the simulations. However, for the other internal combustion 

engine vehicle (ICEV), a mesoscopic simulation with similar characteristics can be 

performed since individual information for ICEV is not required. Thus, the model 

becomes hybrid by merging mesoscopic and microscopic simulations depending on 

the type of vehicle to be simulated. 

Few computational tools allow hybrid simulations of the transport system. One 

of these tools is AIMSUN (AIMSUN, 2021). The AIMSUN tool, a traffic simulator 

software, has a complete package for studying transport and traffic with simulations 

that allow: 

A. Transport and traffic studies: traffic light optimization, impact reports, traffic 

demand, public transport demand studies, mobility plan, capacity study, 

road concessions study, pollutant emissions analysis, among others. 

B. Traffic surveys: volume, demand, congestion, traffic conditions. 



26 
 

 
 

C. Logistics consulting: dimensioning and optimization of resources, process 

simulation, supply chain, logistics terminals. 

D. People flow, pedestrian simulation. 

With several possibilities of input data, data modeling, and programming 

language, AIMSUN allows the simulation of DNCEV in urban zones and between 

cities (highways) to obtain information to assist in planning CS and other planning 

and studies. Among the primary data that can be modeled and obtained with the 

simulations performed, stand out: 

i) Allocation of O-D from DNCEV and ICEV in the urban zone allows the 

planner to use a database already available for O-D or estimate the 

acquisition of DNCEV and their destinations over periods. 

ii) Driving patterns for DNCEV and ICEV - streets and avenues in urban 

zones in which vehicles will travel from their O-D. 

iii) Use of programming language (python) for modeling variables that are 

not included in the tool. AIMSUN users can use such as modeling of 

DNCEV, engine efficiency, modeling of SOC in real-time, regenerative 

brake, hybrid vehicles with mixed consumption (battery, fossil fuels, 

ethanol, gas). 

iv) Identify locations with better conditions to install CS (places with 

available physical space, strategic locations). 

v) New simulations can be performed with CS spatial allocated in the tool. 

Simulations provide vehicle information of vehicles entering, recharging, 

leaving the CS, recharge time, the number of vehicles to enter, waiting 

time, and information distribution network planners increasing load and 

planning. 

vi) The simulations allow an analysis of the emission of pollutants that 

contribute to increasing the local pollution that will no longer be 

generated in the environment and can serve purposes of studies and 

awareness, advertising and incentives for the use of DNCEV. 

vii) Modeling can be carried out at the macroscopic, mesoscopic, 

microscopic, or hybrid levels. 

viii) Modeling of street parameters as exclusive lanes for buses (which can 

simulate the circulation of DNCEV as a form of incentives for the 



27 
 

 
 

circulation of vehicles), traffic light plan, modeling of ICEV stops, speed, 

among others. 

ix) Modeling of pre-programmed bus routes to purchase electric buses, 

electric garbage trucks, and security services using DNCEV. 

x) Model different driving levels: from aggressiveness to smooth, 

depending on the planner's experience from the urban study zone. 

Simulations can be performed similar to applications like WAZE, with O-

D minimum travel times and/or shortest paths. 

Other functionalities in the AIMSUN tool can be used to obtain information 

from DNCEV or modeled after the needs of planners. All this information and others 

that can be modeled will be available in a database, which can be used to build 

spatial maps, in the tool or with the help of GIS tools, with the variables that the 

planner needs. Finally, an urban traffic model, in addition to the features already 

described, assists (with its modeling, simulation, and database generated) in 

obtaining parameters, variables, and decision-making for the various types of 

infrastructure for charging for DNCEV. Thus, this work uses the AIMSUN tool to 

model DNCEV and ICEV to find variables that will assist in obtaining parameters for 

the installation and infrastructure maintenance for DNCEV, the influence of DNCEV 

on the electricity distribution network, decrease in local pollution, among others. 

1.4 INFRASTRUCTURE FOR CHARGING DNCEV 

There are different ways to charge the DNCEV. Several factors must be 

considered to select the type of recharge, such as recharge time, battery life, 

availability of technology, resources, energy, economic factors, and compatibility with 

DNCEV. There are three recharge methods. AC recharge, being implemented in 

public and private facilities, with relatively low investment; DC recharge, requiring 

investment and super-fast recharge and inductive recharge, occurring through 

contactless inductors with high complexity and costs. The ways of AC and DC charge 

are divided into four modes of recharge described below, in Table 1 and Fig. 4. 

Mode 1: Single-phase recharging with a maximum of 32A. The recharging 

device is integrated with the vehicle, with a residual current protection device and a 

standard plug and socket to make the connection. 



28 
 

 
 

Mode 2: Three-phase charging with a maximum of 32A. A recharging device is 

installed in the vehicle, with a protection and control device integrated with the cable 

or installed in the wall. 

Mode 3: Three-phase charging at CS with a maximum of 63A. The charging 

device is fixed to the "electro station." The CS has communication, overcurrent 

protection devices, a disconnection system, payment, and specific outlets for 

recharging. 

Mode 4: DC charge at CS. The charging device is fixed in the CS with 

protection. System of two plugs and plugs, quick recharge. 

TABLE 1 – CHARACTERISTICS OF THE TYPES OF RECHARGE. 

Mode 
Recharge time for 

average range of 100 km 

Power 

source 

Electric 

Power (kW) 

Voltage 

(V) 

Maximum 

current (A) 

1 6 – 8 hours 
Single-

phase 
3.3 230 16 

1 3 – 4 hours 
Single-

phase 
7.4 230 32 

2 2 - 3 hours 
Three-

phase 
10 400 16 

2 1 - 2 hours 
Three-

phase 
22 400 32 

3 20 - 30 minutes 
Three-

phase 
43 400 63 

4 20 - 30 minutes DC 50 400 - 500 100 - 125 

4 10 minutes DC 120 300 - 500 300 - 350 

Source: Adapted from DENTON, 2018. 

This system works in a massive range of recharge time, conditions, and 

environments that can be configured to recharge all DNCEV, especially public 

transport buses, when they stop at the passenger stations. The specifications of 

these systems depend on construction parameters such as size, area of application, 

frequency, efficiency, height between DNCEV and WPT, and they can reach 200 kW 

(DENTON, 2018). Figs. 5 and 6 are shown a stationary WPT for a private DNCEV, 

BEB, and PHEB, respectively. 



29 
 

 
 

Figure 4 – Modes of charging a DNCEV 

Source: Charging Modes - Deltrix charging solutions, 2021. 

Figure 5 – Stationary WPT for a private DNCEV 

Source: (BRECHER; ARTHUR, 2014). 



30 
 

 
 

Figure 6 – Stationary WPT for a BEB and PHEB

  

Source: (BRECHER; ARTHUR, 2014). 

Dynamic WPT is a relatively new technology in which DNCEV recharges when 

traveling on a highway, as seen in Fig. 7. For this type of technology, the main 

challenges are the coil synchronization, acceptable power levels, DNCEV alignment, 

speed profiles, multiple DNCEV at the same time (DENTON, 2018). This technology 

has been applied on some highways to create green corridors to encourage DNCEV. 

In addition, research (JORGETTO, 2018) has been done to improve the efficiency of 

recharging with the dynamic WPT. 

Figure 7 – Wireless Power Transfer System 

Source: (FUJITA; YASUDA; AKAGI, 2017). 

Recharges can be carried out in homes, commonly called residential 

recharging. With less power demand and longer recharging, mode 1 in homes can 

even be mode 2 in residential shared parking lots and static WPT. Mode 2 can also 

be present in residential, commercial areas, and parking lots of shopping malls, 



31 
 

 
 

parks, and the payment for recharge may or may not be necessary. Mode 3 and 4 

are present in specific CS requiring payment for recharge. Dynamic WPT needs a 

green corridor on the highways, which is rare and currently exists mainly for testing 

and research. 

CS should not be located anywhere in an urban zone. Criteria for allocating 

the CS must be adopted to prevent the CS from being located in an unnecessary 

location, with little DNCEV traffic, or that the DNCEV that pass through do not need 

to be recharged, without physical space for allocation, expensive and difficult 

connection to the electrical network, among others. Many techniques to find the 

location of CS seek solutions without considering the stochasticity of the problem 

and/or do not consider the constraints of the various agents involved, which can 

cause problems to some agents participating in the CS allocation process.  

In addition, spatial and technical information should be available to all agents 

to be incorporated in their GIS tools since the agents involved in this process have 

used spatial analysis to allocate CS. To help in sustainable city planning, to use a 

methodology that allows to integrate and attend to the planning requirements of 

interested agents in the installation process is important, as well as considering some 

sources of stochasticity. In addition, to the location of the CS, it is also necessary to 

quantify the SOC of DNCEV when returning to their homes and estimate the energy 

consumption at the end of the day, based on residential recharge. 

Thus, to find the infrastructure for DNCEV, a spatial-temporal methodology 

was developed. Based on the previous modules, criteria such as SOC, driving 

pattern, spatial availability for CS allocation, places of significant influence, the 

concentration of DNCEV are used to allocate CS in the urban zone. Finally, WPT 

stationary is proposed in bus stops helping in the introduction of BEB and PHEB. 

1.5 DECREASE IN LOCAL POLLUTION 

GHG reduction in urban zones to decrease local pollution can be achieved 

through electrification of transportation with DNCEV (APARECIDO et al., 2017; 

KUMAR; REVANKAR, 2017). Vehicles are primarily responsible for local pollution, 

causing several environmental, social, and economic problems (ANDRÉ; 

PASQUIER; CARTERET, 2018; SENGUPTA; COHAN, 2017). The decarbonization 



32 
 

 
 

of transportation has been used in many cities today (SANTOS; ASENSIO, 2019), 

associated with an infrastructure and new technologies available that can contribute 

to increasing the use of DNCEV. 

In (MCLAREN; MILER; O’SHAUGHNESSY, 2016), the reduction in carbon 

dioxide (CO2) that is no longer generated is on average 135.28 g/km for a pure EV 

and 81.73 g/km for a hybrid EV. The DNCEV recharging at CS will leave to produce 

an average of 680.38 g/kWh of CO2. The sustainable planning for the introduction of 

DNCEV in urban zones can be carried out with the help of computational tools of 

TSS and GIS (CIHAT et al., 2019; HUANG et al., 2019). The AIMSUN tool can assist 

in the modeling of a decrease in local pollution through the GHG emissions model 

(MORRO-MELLO et al., 2019). Depending on the location of CS, the reduction in 

pollutant emission values in an urban zone may differ as vehicles may have to travel 

more to recharge. As part of a GHG reduction policy through the decarbonizing the 

transportation system (CAROLINA; TEIXEIRA; RICARDO, 2018; MANJUNATH; 

GROSS, 2017; MUÑOZ-VILLAMIZAR; MONTOYA-TORRES; FAULIN, 2017; HILL et 

al., 2019), information on the reduction of gases (CO2 and nitrogen oxide - NOx) with 

of the use of DNCEV are modeled with the London emission model (LEM) presented 

in the AIMSUN tool (AIMSUN, 2021). 

The LEM is an emission model developed by transport for London (AIMSUN, 

2021) with observations through average speed models. It uses equations to derive 

the emissions for an individual vehicle in a set of micro trips that add up to integrate a 

complete trip. A micro trip is defined as a trip segment where the speed rises from 

stationary to more than 5 km/h and back to static, as illustrated in Fig. 8. 

Figure 8 – LEM micro trips - speed vs. time in an urban drive cycle 

 

Source: (AIMSUN, 2021). 



33 
 

 
 

After driving pattern simulations of vehicles, the LEM shows the emissions of 

CO2 and NOx in g/km. Therefore, to quantify the reduction in local pollution, the LEM 

is included in all DNCEV and ICEV. The quantification of DNCEV pollution is the 

decrease expected with their introduction in urban zones. With the expected 

reduction shown in spatial maps of the city as shown in an example in Fig. 9, further 

studies related to the area of health and environment, with the consequences of this 

decrease in local pollution, can be carried out. 

Figure 9 – Example of the spatial map with local pollution caused by vehicles 

Source: Pun; Manjourides; Suh, (2017). 

This module intends to investigate the sustainability of the introduction of 

DNCEV in the urban zone, in each street, observing the reduction of local pollution 

through the behavior of travel, time, congestion, speed, distance that will provide 

guidelines for urban planners and other companies that need this data for analysis 

and planning. 

1.6 INFRASTRUCTURE AND DNCEV ELECTRIC NETWORK CONNECTION  

The installation of CS has happened in many cities as a part of the necessary 

infrastructure for DNCEV. CS must meet the demand for DNCEV that will be 



34 
 

 
 

recharged throughout the day and located in strategic places (ERBAS et al., 2018). 

Due to the short time required to charge a DNCEV during the day, CS has high 

energy demands and requires a specific kind of installation. 

The price of kWh in residential recharges and semi-fast charging stations are 

lower compared to fuel prices for ICEV considering both costs and efficiency 

(LUTSEY; CUI; YU, 2021) and scenarios for the coming years, even considering 

thermoelectric (more expensive energy generation) still makes the scenario favorable 

for EVs (KEWEN et al, 2021). In the case of fast charging stations, the charging price 

is higher (MURATORI et al, 2019). The introduction of photovoltaic panels and 

energy storage (battery) helps to reduce the price of kWh, making it economically 

viable to mitigate fixed cost and demand charges. Energy storage alone mitigate 

demand charges, reducing costs for peaky or low-utilization loads. Photovoltaic 

panels help mitigate energy charges. Their combination can synergistically deploy to 

provide cost reductions (MURATORI et al, 2019). 

Distribution companies must fulfill the growing connection requests for the CS. 

Due to the low acquisition of DNCEV in most cities, there may be no significant 

impacts on the network. However, due to the goals and incentives for electric 

mobility, it is necessary to analyze the connection point better to mitigate future 

problems (VISWANATHAN et al., 2018). Therefore, the connection of these stations 

in the distribution network should consider studies of how to ensure adequate energy 

levels, since CS can cause voltage deviations, unbalanced loads, and abrupt 

increases in power requests (FLORES; SHAFFER; BROUWER, 2016; ELMA, 2019; 

ZHANG, 2018). These issues need to be considered by planners seeking to identify 

the most favorable conditions for the agents involved in the installation process 

(WARGERS et al., 2018; GLOBAL EV OUTLOOK, 2020; FLAMMINI et al., 2018; 

SBORDONE et al., 2015). 

Some studies in specialized literature focus on expanding the distribution 

network for CS that can increase investments (WANG et al., 2015; HEYMANN et al., 

2018). However, these studies are not articulate how stations can meet this new 

demand without new investments in the electric network. To accomplish this goal, 

this work argues that with the request (several stations that must be connected), the 

best way to meet them must be found without causing problems with the network 



35 
 

 
 

planning, reducing or avoiding investments in new installations and reinforcement in 

the distribution network. This evaluation must allow distributors to take advantage of 

the current network elements' unused capacity (supply capability) (CELG-D, 2016). It 

must also allow them to create a zone for exploring other points within the electrical 

network instead of connecting CS to the nearest network point. This strategy restricts 

the number of potential solutions to this power distribution problem. Exploring these 

zones increases the possibilities for maximizing the network elements’ supply 

capability. 

Distribution companies have technical criteria stipulating the connection of 

new loads to their less overloaded and/or less distant elements. This strategy is 

effective in dealing with low-usage new loads. However, for high-usage new loads 

like CS, a zone to explore other points of the electrical network can be advantageous 

for distribution companies, allowing them to reduce investments, avoid overloads, 

and prevent inadequate voltage drops and losses. Also, looking for the unused 

capacity of the current elements to connect the CS is a strategy for minimizing 

network reinforcement and expansion investments. Finally, spatial tools (MELO; 

CARRENO; PADILHA-FELTRIN, 2012) can help choose connection strategies for 

different locations since CS can be connected at low or medium voltage depending 

on installed power. 

In contrast to recharging in CS, many vehicles are recharged in their homes, 

mostly at night. The DNCEV owner uses the vehicle throughout the day and at night, 

arriving at his home, recharges it. With few vehicles connecting at night, there should 

be no significant impact on the distribution network. As the introduction of DNCEV 

increases, many vehicles will begin to connect to the distribution network and may 

overload it. The overload may occur due to the disproportionate increase in 

recharging simultaneously, and/or many DNCEV owners will make the connection to 

the network in the peak period in the early evening. Therefore, it is necessary to find 

trends in the charging profile by subareas of the urban zone with the increase of 

DNCEV and create strategies to mitigate such problems with incentives for charging. 

In addition, the regions will have different charging profiles, and depending on the 

profile, it is possible to create stratified charging strategies per region, to avoid 

overload. 



36 
 

 
 

Tools that model the distribution network characteristics (MONTEIRO et al., 

2005) are necessary to obtain voltage and current conditions, plan for a safe 

connection of services, and deliver reliable energy (HEYMANN et al., 2017; XIAO et 

al., 2014). These tools provide a large amount of data about distribution network 

elements, supply capability, spatial location, and other aspects of the system. The 

OpenDSS tool (ELECTRIC POWER RESEARCH INSTITUTE, 2018) allows the 

modeling and simulation of the distribution network with a high level of detail of the 

components and therefore was used in this work. Also, distribution companies use 

GIS tools for analyzing, forecasting, and planning the connection of new loads in the 

distribution system (MELO; CARRENO; PADILHA-FELTRIN, 2012). These GIS tools 

allow for the coordination of urban planning information with the data obtained via 

network planning tools; in particular, CS information can be placed in a spatial 

database for electrical utilities. Fig. 10 shows an example of using OpenDSS to 

model and simulate the distribution network's power flow, with spatial characteristics 

model and communication with other GIS tools. 

Figure 10 – Distribution network model in OpenDSS 

Source: Author’s own elaboration. 

In Fig. 10, it is shown in a) an example of modeling a distribution network in 

OpenDSS, in b) an example of a spatial distribution network in OpenDSS, and in c) 

an example of a spatial distribution network a GIS tool. Thus, the results obtained in 

OpenDSS analyzes can be used in GIS tools, present in distribution utilities. 

Thus, this work proposes a methodology to find the connection points between 

the CS and the distribution network through cost functions, finding the least 

expensive connection for the distribution companies, prioritizing the connection of 

elements with available load (elements that have supply capacity to meet the CS 



37 
 

 
 

demand) and the introduction of photovoltaic panels. In addition, an analysis of the 

DNCEV residential recharge is carried out to find the demand for recharge. 

1.7 CONTRIBUTIONS TO RESEARCH 

Each of the five modules produces a spatial database with the introduction of 

DNCEV over a planning horizon, with the characteristic of urban traffic, the 

introduction of infrastructure, decrease in local pollution, and connection of DNCEV 

and CS in the distribution network. The database produced in each module serves as 

a basis for the others. Once all five modules for a period of time are finished, and a 

database with all the results is generated, this database will serve as a starting point 

for planning a new time horizon. A new time horizon, therefore, considers the results 

of the previous planning. Therefore, for this work, a time horizon of five years to 

introduce DNCEV will be carried out from the initial state (year 0) until the fifth year. 

The results of each of the five study modules produced for each planning year 

will be presented in spatial thematic maps with the help of GIS tools. Due to the 

multidisciplinary of this study, the presentation of these data in spatial maps helps 

interpret results. Furthermore, it allows communication between the various agents 

who will introduce DNCEV and their infrastructure, such as research studies, 

engineering, environment, sustainability, architecture, social policies, public policies.  

The methodology proposed in this work was applied to a specific city. The 

reader will understand throughout the work that due to the amount of data processed 

and the integration of computational tools, it was not possible to apply the proposed 

methodology in other databases, for other study areas in the doctorate time. 

However, it now engages, from the proposed methodology and path to such 

methodology that will be exposed below, the readers the replication of this work in 

other cities. This work should not be understood as applicable only to the urban zone 

used. It is expected that it will be replicated in other cities and that the results will help 

the various agents that will participate in the process of introducing the DNCEV. To 

assist in understanding, the reader can also access the papers that were produced 

until the date of completion of this work and are described in appendix 1. 

In this work of introducing DNCEV in urban zones, a purchase intention survey 

was not carried out. The methodology covered in this work can be applied to any 



38 
 

 
 

urban study zone that wants to introduce DNCEV and consider several public and 

private agents that will participate in the introduction process considering the 

acquisition of DNCEV, traffic model, infrastructure for charging, decrease in local 

pollution and impact on the distribution network. The urban zone which the 

methodology was applied is located in a medium-sized Brazilian city. With the 

Brazilian sociocultural reality, it was observed that a DNCEV purchase intention 

methodology would not reach its purpose to be applied, since the inhabitants do not 

have sufficient knowledge of the DNCEV advantages and technology. Therefore, 

DNCEV buyers will acquire knowledge from an utility marketing policy, government 

agencies, tax incentives, among others, which are still being built for the urban zone 

of application as well as for the entire Brazilian scenario. Finally, the main 

contributions of this work are described below: 

i) The proposal uses a spatial model with interactions among the inhabitants to 

find the temporal rate of acquisition of DNCEV per subarea. 

ii) The proposal model simulates urban traffic dynamics with ICEV and 

DNCEV, using a hybrid simulation (mesoscopic and microscopic), considering 

stochastic factors, such as the SOC at the beginning of the traffic simulations. 

iii) The proposal helps quantify the decrease in local pollution by introducing 

DNCEV, important information to various sectors involved such as selling, 

developments, planning horizon, and new engagement policies. 

iv) The methodology assists with the process of choosing connections for CS 

in the distribution network, proposing a local search algorithm based on graph theory, 

calculating the total cost of connection related to the connection branch. The 

methodology uses the supply capability of the current network elements to reduce 

investments, ensures a safe connection, and considers whether the type of 

connection will be at low or medium voltage. The model can help distribution 

planners to reduce new investments of expansion. 

v) The proposal characterizes the requirements of the planners. As a result, a 

spatial-temporal database can be processed by any GIS of the public or private 

agent involved in the introduction and infrastructure process, as seen with the 

creation of spatial maps of all module results. 



39 
 

 
 

vi) Using a GIS tool to link spatial databases from other urban planning areas 

with distribution network planning. The tool helps decide how and where to connect 

CS, identifying the places that need more attention in planning power-distribution 

systems. 



40 
 

 
 

2 LITERATURE REVIEW 

Many works are found in the literature, with different approaches to DNCEV, 

such as their acquisition, infrastructure, electric network operation, and planning. On 

the other hand, there are few works found for traffic flow and GHG reduction with the 

introduction of DNCEV. Moreover, a few works spatially treat information, presenting 

a methodology and results that can be used by several agents who will participate in 

the process of introducing DNCEV. Generally, such works are restricted to one area 

of knowledge, solving a specific problem for an agent. Below, a summary of some 

works in the specialized literature for each of the modules will be presented. 

2.1 DNCEV ACQUISITION REVIEW 

Some works found in the specialized literature deal with the acquisition of 

DNCEV. However, these works have not considered the spatial interaction among 

the potential buyers (BERNARDS; MORREN; SLOOTWEG, 2018). This interaction 

also has been observed for other technologies (BOLLINGER; GILLINGHAM, 2012) 

and results from the influence that a DNCEV buyer can accomplish in the vicinity. 

Furthermore, this interaction has been studied and characterized to quantify the 

potential market (LI et al., 2017; CHEN; JI, 2016). 

A methodology to estimate the DNCEV penetration based on a Monte Carlo 

Simulation (MCS) technique was proposed (MU et al., 2014). The methodology used 

the yearly adoption rate, the SOC, and the owners driving behavior as model 

parameters. The results of such estimation were inserted in GIS to analyze the 

charging impact on a distribution network. In this class of simulations, it is used 

stochastic parameters that do not have their interdependence characterized. 

In (HEYMANN et al., 2017), the authors presented a spatial approach to 

estimate DNCEV charging events and their impact on the electrical network. While 

this work uses a diffusion theoretic approach that predicts spatial charging pattern 

discrepancies for midday and overnight charging, the influence of the neighborhood 

is not considered. 

A conditional autoregressive regression model was used to estimate the 

number of conventional vehicle and DNCEV buyers using census data of 

Pennsylvania’s region (CHEN; WANG; KOCKELMAN, 2015). The results show an 



41 
 

 
 

interaction between the residents with their neighbors in decision-making to buy a 

DNCEV. However, the methodology determines the spatial relationships between the 

inhabitants in a global form for the whole city without showing the local spatial 

influence in small areas (THORSON et al., 2015). 

In the contributions above, the spatial interaction among the residents in local 

form was not modeled. This interaction may influence DNCEV purchases in small 

areas, decreasing the distrust to purchase in the first years. To model this interaction, 

spatial regressions models were used to estimate the number of DNCEV adopters. 

DNCEV in several countries are still a new technology and unknown to many 

consumers. Thus, intention surveys can be developed by specialists in market 

behavior and the acceptance of new technologies are important to estimate the 

acquisition of DNCEV. Research can be published under different aspects such as 

regional, socioeconomic, influence, among others. To assist in the development of 

technology acceptance surveys, it is possible to refer to the specialized literature to 

understand how the introduction of DNCEV and intent surveys are carried out. 

In (ROEMER; HENSELER, 2022), the authors used a grounded theory 

approach to explain the drivers and barriers of acceptance of EV on the employee’s 

level. The authors extracted five main determinants from interviews and the results 

show that environmental concern determinant triggering the first EV usage. Later, the 

acquisition occurs after the comprehension of ease of use, perceived risks, and 

relative advantage increases. 

In (JAISWAL; DESHMUKH; THAICHON, 2022), the study applied analytical 

procedures with cluster analysis, multiple discriminant analysis and Chi-square test to 

explore and identify distinct sets of potential buyer segments for EVs based on 

psychographic, behavioral, and socio-economic characterization by employing an 

integrated research framework of ‘perceived benefits-attitude-intention’. The results 

shown that younger consumers are a greater degree of proclivity to espouse internet 

technology and innovative products or EV rather than their older counterparts. 

In (ZHANG et al, 2021) the work explores the relationships between EV 

related information and consumers’ perceptions and adoption intentions. Based on 

survey data from 343 respondents, the results indicated that environmentally friendly 



42 
 

 
 

information, performance, and attribute information regarding EVs are positively 

associated with consumers’ perceived value. The research highlights the importance 

of EV-related information and information quality on consumers’ perceptions and 

adoption intentions of EVs, highlighting that clear information about the technology 

will help in its adoption. 

In (VAFAEI-ZADEH et al., 2022), the authors investigate 213 usable data 

(using structured questionnaires) of EV purchase intention among Generation Y 

consumers considering the impact of perceived usefulness, attitude, subjective 

norms, perceived behavioral control, price value, perceived risk, environmental self-

image, and infrastructure barrier. The study shows that only perceived risk had a 

negative impact on intention and give instructions to better explain EV purchase 

intention among the Generation Y consumers. 

In (SONG; CHU; IM, 2022), the authors study the effect of cultural and 

psychological characteristics on the purchase behavior and satisfaction of EV. The 

study shows that motivation for EV usage, reasons for purchase, and satisfaction are 

different depending on the country. US owners showed a higher propensity to 

innovate and greater knowledge of EVs than Chinese owners for example. Chinese 

owners cited reputation and signaling social responsibility as being important 

considerations in their purchase. The results show how culture and stage of 

motorization are reflected in consumer behavior and satisfaction regarding EVs. 

In (GUNAWAN et al. 2022), the authors perform intention research with 526 

respondents and discuss a theoretical model and risk perception, using the structural 

equation modeling method. The aim of the study is to provide an overview of the 

factors that drive interest in adopting EV. The results showed that the model can 

estimate the influence by performance and effort expectancies, hedonic motivation, 

price value, as well as functional, financial, and social risks, giving directions in the 

adoption of EV. 

Five factors (financial incentives, charging infrastructure, social reinforcement, 

environmental concern, and price) that influence the adoption of EV are studied in 

(ALI; NAUSHAD, 2022) and serve as a basis for data collected study from 366 

randomly selected respondents. Structural Equation Modeling and confirmatory 



43 
 

 
 

Factor Analysis were used to analyze the data and conclude that pricing has a 

substantial impact on the adoption of EV. 

The specialized literature shows that the environmental concern is a positive 

factor when raised in the research while the price, lack of infrastructure are concerns 

of the interviewees that can influence the acquisition. 

All these studies and others found in the specialized literature show the 

variables that should be considered to assess the potential for EV acquisition. These 

variables can be conveyed to purchase intent surveys and can be modeled through 

equations and methodologies to create EV acquisition scenarios in urban zones. 

Such research can be carried out spatially, considering different aspects in the 

regions of the urban zone and temporal as the introduction of EV occurs, like 

financial incentives, charging infrastructure, environmental concern, decrease in EV 

price, and others. 

These EV purchase intention surveys can be incorporated into the 

methodology presented in this work. The DNCEV acquisition module can be spatially 

modified to incorporate such surveys. The intention survey can be carried out by sub-

areas, considering socioeconomic aspects, EV price, environmental concern, age, 

among other aspects. With the surveys carried out, each of the variables and aspects 

conveyed in the questionnaires and their results can be incorporated into GIS tools. 

Thus, mathematical models can be created for each of the variables to find the 

purchase rate. 

It can also be incorporated into the methodology in a temporal manner. As EV 

are acquired, after a certain period (five modules of the methodology applied for a 

period), intention surveys can be made to understand the difference in behavior and 

acquisition expectation for the next period. Thus, both spatially and temporally, 

intention surveys can be applied and incorporated into the DNCEV acquisition 

methodology. 

2.2 TRAFFIC MODEL REVIEW 

Related to traffic, DNCEV modeling and simulation are still scarce in the 

specialized literature. Some papers that deal with the traffic of DNCEV are presented. 



44 
 

 
 

In (QU et al., 2020), a development of a DNCEV following model based on 

reinforcement learning to dampen traffic oscillations (stop-and-go traffic waves) 

caused by human drivers and improve electric energy consumption is proposed with 

the capability of self-learning and self-correction. The results show that the proposed 

model improves travel efficiency by reducing the negative impact of traffic 

oscillations, and it can also reduce SOC consumption. However, the work does not 

evaluate the dynamics of DNCEV spatially and temporally, and the focus is not on 

various agents that will participate in the introduction and maintenance of them. 

In (CHEN; WU; ZHANG, 2020), the impact of traffic network topological 

characteristics on the charging characteristics of large-scale DNCEV is presented 

with a model of the complex adaptive system, using a multi-agent technique in a 

spatial-temporal manner. The traffic networks of various cities in comparison testify 

that improving the connectivity of the traffic network can effectively reduce the effect 

on the power network brought by the charging load of large-scale DNCEV. However, 

in this work, the method and simulation model considering the DNCEV in a 

macroscopic way, as a large and unique group, disregarding the difference in the 

DNCEV group like types, location in the urban zone, and different behavior patterns. 

In (RODRIGUEZ, 2020), a study is conducted to evaluate the impact of the 

SOC driving behavior under various traffic scenarios. To simulate the driver behavior, 

a widely used intelligent driver model is chosen to characterize the different levels of 

driver aggressiveness. A microscopic traffic simulator, PTV Vissim, is used to 

simulate various realistic traffic environments. The co-simulation of the PTV Vissim 

Component Object Model interface in conjunction with MATLAB allows the energy 

consumption performance on a DNCEV to be determined for various levels of driving 

aggressiveness. The results of these driving tests demonstrate that the level of 

driving aggressiveness cannot be fixed and should instead adapt to the traffic 

environment to maximize the battery life and range of an EV. However, in this work, 

an approach to charging locations is not described and/or model. These locations 

influence the SOC, contributing to driving aggressiveness and instant level of charge 

along the streets. 

In (LEI, 2019), an adaptive equivalent consumption minimization strategy 

considering traffic information is proposed using traffic simulator Vissim to facilitate 



45 
 

 
 

the effective energy management of DNCEV. First, using a genetic algorithm, factors 

in determining initial SOC and driving distance are found. Then, dynamic 

programming is used to determine the optimal SOC trajectory according to the traffic 

information, and finally, a fuzzy controller is employed to regulate the equivalent 

factor dynamic. Simulation in the Vissim and experimental results shows that the 

proposed strategy can lead to less fuel consumption than the traditional equivalent 

consumption minimization strategy. The authors do not use the SOC information and 

driving patterns for other applications like resource allocation for DNCEV users. 

Thus, the focus of the paper assists only in one of the topics of analysis of DNCEV. 

In (BI; TANG, 2019), a dynamic DNCEV routing problem model is designed to 

plan the logistics industry's itinerary. To reflect the real situation, the model considers 

a time-dependent stochastic traffic condition and captures the discharging/charging 

pattern to minimize the overall service duration and planning strategy. The results 

show benefits for optimal policies and the benefits from accurate modeling of battery 

dynamics. The traffic characteristics adopted in the paper are based on estimates 

given by traffic companies without characterizing the real-time driving pattern with a 

traffic simulation that considers the dynamic of DNCEV applied in a real traffic flow. 

2.3 INFRASTRUCTURE FOR CHARGING REVIEW 

Others works studied the infrastructure for DNCEV in urban zones. In 

(ZHANG; KANG; KWON, 2020), the authors using a multi-day data sampling method 

and analysis of scenarios to provide insights on the bounds of potential DNCEV 

market under different charging opportunities, including CS charging divided in AC 

and DC charging. The results showed that facility utilization could be improved 

without affecting travelers’ charging demand. Smart grid charging strategy can 

significantly reduce the total number of operating chargers during the same time in a 

day, and DNCEV users’ charging behaviors have a minor impact on this 

improvement. The numerical results indicate that an appropriate number of chargers 

installed in leisure locations should be more profitable and have a higher charger 

utilization rate since those chargers help cover DNCEV users’ trips. The work uses 

DNCEV real data to analyze the scenarios. Still, for cities that do not yet have 

DNCEV, such data do not exist, and simulations to find their driving patterns, among 

other aspects, are necessary to create and analyze such scenarios. 



46 
 

 
 

In the same line as the previous one, the authors in (LEE et al., 2020) examine 

the charging behavior of thousands of DNCEV in California. The study investigates 

the charging locations (at work or public location) and the level of charging AC slow 

charging, AC fast charging, or DC fast charging. The charging behavior differs among 

DNCEV owners based on driving patterns, preferences, and infrastructure access. 

The authors identify factors associated with the DNCEV owner’s choice of charging 

location and charging level based on socio-demographic, DNCEV characteristics, 

commute behavior, and workplace charging availability. 

In (FANG et al., 2020), the authors studied the impacts of policy incentives 

and consumer preferences to promote CS construction. To improve economic 

efficiency, taxation and subsidy policies are employed. Consumers can choose an 

ICEV or DNCEV in response to the policy incentive. The time-varying needs of CS 

and fuel stations in each area are characterized by an evolutionary game model 

considering the competitive connections between the stations. The results prove the 

advantages of the balanced dynamic subsidy and taxation policies on the promotion 

of CS. The penetration level of DNCEV and charging prices are the main driving 

forces. However, in some countries, the price of DNCEV is still very high, and the 

study of choice ICEV will always have an advantage. Thus, applying this study to 

such countries may not be advantageous, requiring studies on the introduction of 

DNCEV based on other criteria. 

In (LAM; LEUNG; CHU, 2014), the authors studied DNCEV CS placement 

considering human aspects in the long term with many social constraints like 

satisfaction, drivers' convenience, environmental impact, DNCEV accessibility. The 

authors propose four CS placement methods, showing their solutions and their pros 

and cons for quality, algorithmic efficiency, problem size, and nature of the algorithm. 

However, the available physical space for CS installation is not considered and can 

cause location and installation problems. 

In (SACHAN; DEB; SINGH, 2020), the paper comprehensively discusses 

basic infrastructures for charging DNCEV such as distributed infrastructure (AC 

charging) and fast charging infrastructure (DC charging). It has been found that 

distributed infrastructure shows the best results for the charging of DNCEV. The 

other infrastructure prove costlier and increase power demand. However, the study 



47 
 

 
 

does not consider that there are different patterns of conduction of DNCEV. For 

example, ETs, modeled in (MORRO-MELLO et al., 2019), need recharging in a short 

period of time, requiring fast charging. In addition, an uncoordinated distributed 

infrastructure can also cause problems. 

2.4 GHG REDUCTION REVIEW 

There are works in the specialized literature that deal with GHG reduction 

through the introduction of DNCEV. In (XIONG; JI; MA, 2020), the authors quantify 

the environmental impacts and costs of remanufacturing a battery cell. The results 

indicate that the reductions in energy consumption and GHG emissions by battery 

remanufacturing are 8.55% and 6.62%, respectively. In (MICHAELIDES, 2020), the 

authors study the regional mix of electricity generation, and the effect of the shift to 

DNCEV on GHG emissions is determined. The authors concluded that the 

simultaneous charging of DNCEV would strain the electricity network's capacity and 

examine the effects of DNCEV on the further utilization of renewable energy sources.  

In (CHOI et al., 2020), the work compares the well-to-wheel GHG emissions of 

different types of DNCEV and ICEV of various energy policies that could affect future 

emissions. A framework was proposed to evaluate the impacts of energy policy 

regarding electricity and hydrogen production on the benefits of using the different 

types of DNCEV. In (WANG et al., 2020), the study presents a GHG emissions and 

energy consumption accounting approach for on-road transportation, developed to 

estimate well-to-wheel emission distributions for household ICEV and DNCEV. The 

study combines real data with the effects of vehicle electrification and well-to-wheel 

emission. Mean values for daily regional GHG emissions from household 

transportation were estimated. The results of the policy scenarios present insight into 

the effectiveness of DNCEV at reducing emissions. 

In (BENAJES et al., 2020), the paper assessing the potential of different types 

of DNCEV when used together with a low-temperature combustion mode. A deep 

analysis is performed in terms of emissions. The results show that the PHEV has the 

highest benefits in terms of fuel consumption. With this technology, it is possible to 

achieve the 50 g/km CO2 target for the PHEVs while NOx is under the limits. In (KIM; 

KIM; LEE, 2020), the paper approaches GHG emissions caused by DNCEV by 

examining the power mix and resulting change of market share. In numerical terms, 



48 
 

 
 

the estimated GHG reduction is 122,441 tCO2, which constitutes only 4.7%. 

Therefore, the authors concluded that it is necessary to clarify the segmentation of 

DNCEV to optimize efficient infrastructure investment and a stable diffusion of 

environment-friendly vehicles. 

The authors in (DOLUWEERA et al., 2020) examine the energy and GHG 

emissions impacts of DNCEV on electricity and transport systems, using a hybrid 

simulation model and develop a new component to model DNCEV fleets. The 

analysis shows that the adaptation of DNCEV can contribute to the 2030 emissions 

reduction target of 30% below 2005 levels. In (ALIMUJIANG; JIANG, 2020), the 

paper shows the co-benefits generated by using three types of DNCEV by combining 

environmental benefit and cost-effectiveness analyses. The results show that BEBs 

provide the highest co-benefits. Thus, replacing traditional fuel buses with BEBS can 

reduce air pollution and CO2 emissions. In addition, private DNCEV and ETs also 

provide the co-benefits of reducing CO, NOx, NMHC, and PM10 emissions. 

In (VILCHEZ; JOCHEM, 2019), the paper investigates the possible impacts of 

DNCEV on future energy demand and GHG emissions until 2030 by using a system 

dynamics model covering nine-car technologies in different countries. GHG 

emissions from cars in the countries are simulated to reach up to 2.6 gigatons in 

2030. The main conclusion from model-based policy analysis is that DNCEV may 

positively contribute to emissions mitigation in the passenger road transport system. 

All works cited and others found in the literature use models, simulations, and 

perspectives to quantify GHG reduction with the introduction of DNCEV. However, 

the papers found do not conduct a study to find the reduction in local pollution with 

the introduction of DNCEV. As renewable sources of energy generation will be 

necessary to quantify GHG emissions, and this may vary depending on the study 

regions and different characteristics and perspectives, the introduction of DNCEV, in 

terms of pollution, has primary consequences in the cities its decrease. Therefore, 

quantifying local pollution reduction in urban zones, subareas, streets, and avenues 

is necessary and extremely important. However, the quantification is absent in the 

works cited and serves to raise the awareness of the local population, improve the 

quality of life, reduce respiratory diseases, and only ultimately (also important) the 

prospects for GHG reduction. 



49 
 

 
 

2.5 IMPACT ON THE DISTRIBUTION NETWORK REVIEW 

To connect DNCEV in the distribution network, many papers are found in the 

literature that deal with the residential and CS connection. In (ZHANG; YAN; DU, 

2015), the authors show that the DNCEV demand has a significant impact on the 

location of CS in the distribution network, comparing the total electricity required and 

the charging profile. In (MOHAMED et al., 2016), a study is presented with the impact 

of the CS on the capacity of distribution feeders. The methodology examines the 

different factors that affect the impact of the DNCEV charger on the network. The 

results show overload in the distribution networks under a certain system of charging 

with high levels as the use of DNCEV grows, impacting the planning and operation. 

Other works in the literature deal with factors that affect the way of charging in 

the distribution network (PARKS; DENHOLM; MARKEL, 2007; CAROLINA; 

TEIXEIRA; RICARDO, 2018) such as impacts on the voltage levels, different 

charging variables, time, and frequency of charge. To avoid these important 

problems in the distribution network, the use of a coordinated charging methodology 

is recommended. EV charging is done by smart charging at a specific time in a 

specific period when the load is low. As a result, there is a decrease in the daily cost 

of electricity, voltage deviations, peak loads in transformers and current lines, 

congestion in the distribution network, waste of renewable energy resources, energy 

losses, and incremental investments. Coordinated charging is an efficient and 

valuable strategy for DNCEV owners and network operators. 

Of all five modules, the charging and influence of DNCEV in the distribution 

network is the most discussed topic in the specialized literature. Thus, the present 

work proposes a different approach from those found to assist distribution companies 

in decision-making for CS connection. Using spatial modeling, this work finds the 

least-cost expansion for the distribution company to connect the CS, prioritizing the 

current elements of the network, decreasing the cost of new investments. Finally, it 

finds the increase in residential load with night charging exposes such increase in 

spatial maps by zone. 

Several works are found in the specialized literature related to DNCEV, as 

demonstrated in this literature review with the introduction, traffic, infrastructure, GHG 

emission, and distribution network. The difference between this work to the others is 



50 
 

 
 

integrating all these modules in a spatial-temporal manner. Thus, each result of a 

module serves as a basis for the others and the result for a certain period of time, 

serving as a start for the subsequent period. In addition, each of the modules forms a 

spatial database that is used in other modules to create planning criteria. Finally, all 

the results are shown on spatial maps of a study zone, with the help of GIS tools, 

providing the visualization of the results for several agents that will participate in 

some stage of introducing DNCEV. 



51 
 

 
 

3 PROPOSED METHODOLOGY 

The methodology proposed in this work is divided into five modules. Each 

module result will be the input database for the next module. The result of the first 

five modules for a given period will serve as a basis for the next period. The input 

data for the proposed methodology are: 

a) Urban zone, with streets and avenues, sectors, and socioeconomic 

information characterized spatially divided into subarea. 

b) Number of ICEV and DNCEV in each subarea. 

c) O-D of vehicles divided by periods of the day (i.e. residential to commercial 

and industrial in the morning and the opposite in the end of the afternoon). 

d) Electrical network spatial modeled with technical information on network 

elements, consumption, and current supply capacity. 

Such input data will be used during the five modules, being available in a 

spatial database and used with the help of GIS tools. 

3.1 MODULE 1: ACQUISITION OF DNCEV 

The acquisition of DNCEV is developed using spatial regression, with a 

combination of three methods: economic analysis (MORRO-MELLO, 2016), 

Geographically weighted regression (GWR) (MORRO-MELLO et al., 2018), and 

hierarchical spatial autoregressive (HSAR) (RODRIGUES et al., 2019). 

Spatial regression is a statistical analysis that relates two or more spatial 

variables and that one of them can be explained by the others. Serve to determine 

and describe how the variables are related and predict dependent variables' future 

values. If there is a strong tendency or spatial correlation, the results will be 

influenced, presenting statistical association in the locations where it did not exist. 

Different techniques to find the rate of EV buyers per subarea in an urban area are 

presented to understand the spatial regression better. After, in the case study, each 

technique is used to find the most viable method that estimates the number of EVs in 

a subarea. 



52 
 

 
 

The economic analysis described in (MORRO-MELLO, 2016) considers cost-

benefit analysis to find regions with sufficient income to acquire a DNCEV. In the 

cost-benefit analysis, the price of a DNCEV, the average income in the subarea, 

number of residents, individuals, education, among other socioeconomic 

characteristics, were considered. 

A study with the possibility of purchasing a DNCEV by category (LANE et al., 

2018) was also used. In this study, it is established that the purchase of a PHEV has 

a greater acquisition in young groups, wealthy, more studied, with a greater 

perception of ease of use and greater social interaction (individuals that hear about 

the experience of trusted colleagues with PHEV technologies, shaping purchase 

preferences), and prioritizing autonomy. 

People attracted to BEVs tend to transmit pro-environmental attitudes or 

enthusiasm for new technologies, with other lifestyle choices, including financial 

management, activity, fitness, and a strong sense of community, being individuals 

who prioritize high fuel economy and environmental performance. Important issues 

for the purchase of DNCEV: the importance of the policy, such as discounts or tax 

incentives that minimize the acquisition cost, free CS, free installation of residential 

charging infrastructure, access priority (access to restricted corridors), parking lots in 

urban zones, free access to high-occupancy vehicles or toll roads, permission for 

vehicles on rotations. 

Such information can be found in the urban study zone and socioeconomic 

reports database, with preference patterns of the region's inhabitants. Thus, from this 

information, the rate of DNCEV per subarea is found (¥̂𝑖), through an analysis 

described in (MORRO-MELLO, 2016; MORRO-MELLO; PADILHA-FELTRIN; MELO, 

2016). 

The GWR model proposed in (MORRO-MELLO et al., 2018) can be used to 

find the initial acquisition rate of EVs. The model relates the input data to output data 

by creating groups of explanatory and response variables to obtain the proportion of 

EVs acquired in a given region. For example, the GWR model uses as input data the 

socioeconomic data for each subarea, the number of houses with favorable 

conditions to acquire an EV, average subarea income, and sensitivity measurements. 



53 
 

 
 

All this information is classified as explanatory variables. As a response, the model 

provides the rate of EVs acquired in each region determined by: 

�̂�𝑖 =  𝛽0(𝑢𝑖, 𝑣𝑖) + ∑𝑘 𝛽𝑘(𝑢𝑖, 𝑣𝑖)𝑥𝑖𝑘 +  𝜖𝑖                                                                               (1) 

In which 𝒙𝒊𝒌 is the k explanatory variable considered in region i, �̂�𝑖 is the 

response variable in region i. The explanatory variables in the proposed method are 

the socioeconomic variables that correlate with the purchase of EVs. The variable �̂�𝑖 

represents the estimated proportion of households in region i that has the financial 

condition to acquire EVs. This method adjusts a spatial function βk for each k 

explanatory variable using the geographic 𝒖𝒊 and 𝒗𝒊. The values of βk are: 

β̂(𝑖) = (𝑋′𝑊 𝑋)−1 X′W𝑦𝑖                                                                                                                                (2) 

 

In which X is the matrix of the explanatory variables, W is the weight matrix 

that considers the influence of neighboring regions, 𝑦𝑖 is the number of houses with 

economic conditions.  

The HSAR model allows characterizing heterogeneous distributions 

considering the local influence of several variables. The input data for the HSAR 

model are the socioeconomic variables by subareas of the urban zone and the initial 

acquisition rate of EVs. This information can be processed by a GIS tool (ESRI, 

2017). To characterize the estimated heterogeneous spatial rate of EV in each 

subarea (𝑌𝑒𝑠𝑡(𝑖,𝑗)) Eq. (3) is used. 

𝑌𝑒𝑠𝑡(𝑖,𝑗) = 𝜌𝑊𝑌𝑜𝑏𝑠(𝑖,𝑗) + 𝛽𝑋(𝑖,𝑗) + 𝛾𝑧(𝑖,𝑗) + Δ𝜃(𝑖,𝑗) + 𝜀(𝑖,𝑗)                                                  (3) 

In which 𝜌 is the coefficient that measures the spatial interaction globally, 𝑊 is 

the spatial weights matrix that measures the influence of a determined subarea, 

𝑌𝑜𝑏𝑠(𝑖,𝑗) is the initial acquisition rate of EVs, 𝜀(𝑖,𝑗) is the difference between the EV 

acquisition share within and outside the spatial influence of each subarea, 𝜃(𝑖,𝑗) and 

the Δ weights matrix characterize the stochastic nature in the EV acquisition, 𝑋(𝑖,𝑗) 

and 𝑧(𝑖,𝑗) characterize the socioeconomic variables in a subarea, 𝛽 and 𝛾 measure 

the influence of 𝑋(𝑖,𝑗) and 𝑧(𝑖,𝑗) respectively. The 𝜃(𝑖,𝑗) values are calculated for 

different levels. Eq. (4) represent the 𝜃(𝑖,𝑗) for low level. 



54 
 

 
 

𝜃(𝑖,𝑗) = 𝜆𝑀𝜃𝑖𝑛𝑖𝑡𝑖𝑎𝑙(𝑖,𝑗) +  𝜇(𝑖,𝑗)                                                                                       (4) 

In which 𝜆 is a coefficient that measures the intensity of the spatial 

interactions, 𝑀 is the spatial weights matrix, 𝜃𝑖𝑛𝑖𝑡𝑖𝑎𝑙(𝑖,𝑗) is a random number between 

0 and 1 that characterizes the stochastic nature of EV acquisition, 𝜇(𝑖,𝑗) is the 

parameter that accounts for the deviation between the 𝜃𝑖𝑛𝑖𝑡𝑖𝑎𝑙(𝑖,𝑗) and 𝜃(𝑖,𝑗). The 

parameters 𝜀(𝑖,𝑗) and 𝜇(𝑖,𝑗) are calculated through normal probability functions 

determined by the planner as shown in Eq. (5) and (6), respectively. 

𝜀(𝑖,𝑗) = 𝑁(0, 𝜎1)                                                                                                           (5) 

𝜇(𝑖,𝑗) = 𝑁(0, 𝜎2)                                                                                                           (6) 

In which 𝜎1 and 𝜎2 can be adjusted by expected or historical values of EV 

rates. Finally, GIS and statistical software can be used to model the HSAR and find 

the acquisition rate of EVs in each subarea. 

To find DNCEV in each subarea, a combination of three methodologies 

described above is proposed, named three-stage DNCEV acquisition. Eq. (7) 

proposed in this work, assigns weights to each of the methodologies to find the rate 

of DNCEV per subarea (𝐴𝑐(𝑖,𝑗)). 

 𝐴𝑐(𝑖,𝑗) = (𝑅 ×  ¥̂𝑖) + (𝑆 ×  �̂�𝑖) + (𝑇 × 𝑌𝑒𝑠𝑡(𝑖,𝑗))                                                          (7) 

Where R, S, and T are weighted with a sum of 1. If the urban zone does not have 

DNCEV, the DNCEV rate should be found only with economic analysis for the first 

period. 

3.2 MODULE 2: TRAFFIC MODEL 

Transportation tools that model the DNCEV characteristics are important to 

help plan, construct, and maintain their infrastructure. The simulation of the DNCEV, 

modeling their characteristics and produce information of real-time SOC, driving 

patterns along the streets, streets most visited, speed, congestion points, among 

others. This information can be saved in a spatial database for planner users. For the 

traffic model, some input data are necessary and summarized as follow: 



55 
 

 
 

- Transportation network map: georeferenced map with information of urban 

zone with their main and secondary roads, traffic lights, allowed speed 

limits, different sectors of the city (residential, industrial, and/or 

commercial), among other characteristics. The map and information can be 

found in the transportation department files or georeferenced maps with 

streets in specialized sites. 

- Number of DNCEV obtained in module 1 and/or already allocated in the 

city. 

- Number of ICEV: The number of ICEV that cross the urban zone can be 

obtained through reports in the transportation department of the analyzed 

city. 

- ICEV and DNCEV O-D: For the simulation of real traffic during the day, 

modeling ICEV and DNCEV circulation in urban zones provide congestion 

points, speeds, and travel times. Such information is necessary for the 

SOC. Transportation departments often take field measurements and 

divide them into average estimates by hour. O-D characteristics are 

described in the transport department database used on transport 

modeling tools at each hour of the day. The O-D information varies from 

home and/or work to places of leisure, education, health, etc. If the planner 

does not have complete information, the database of cities with similar 

characteristics can assist in the decision-making to allocate O-D data. 

The map of the urban zone is plotted on a TSS capable of simulating the 

driving patterns of vehicles over periods of time. ICEV and private DNCEV are 

allocated to their origins and their multiple destinations determined throughout the 

day following reports from the urban zone containing such information and returning 

to their origins at the end of the day. The routes of the BEBs and EGTs are found in 

the reports and trace previously. 

DNCEV's SOC in the city streets is an important piece of information to 

evaluate CS building. Indicating locations where DNCEV will usually require 

recharging avoids CS allocation in regions with high SOC in DNCEV batteries. 

Furthermore, the vehicles always look for the shortest route and/or shortest travel 

time, and the shortest path algorithm is required for this purpose. 



56 
 

 
 

The SOC consumption model in each street is modeled and simulated with a 

TSS tool such as AIMSUN (AIMSUN, 2021; SUMO; GE; CIUFFO; MENENDEZ, 

2014), or VISSIM (GRIGGS et al., 2015). Simulations must be performed for all hours 

of the day, and a hybrid (microscopic and mesoscopic) traffic simulation must be 

used. To characterize each DNCEV and SOC consumption behavior at each instant, 

a microscopic simulation is used.  

To characterize O-D for ICEV at each hour and zone, groups of vehicles with 

similar characteristics of O-D in each zone are placed together in clusters, and the 

mesoscopic simulation is used. Then, during each hour, a percentage of vehicles in 

each cluster is assumed to leave and/or arrive at the center of the different zones, 

such as industrial, educational, health, etc. (this information and the corresponding 

percentages should be available in the transport department files). 

To determine DNCEV's SOC along the city's roads, the model of SOC 

(GOEKE; SCHNEIDER, 2015) is chosen becaus