UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Optimal placement and operation of hydrogen fueling stations in Metro Vancouver and Victoria Azizkhani, Sara 2019

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2019_may_azizkhani_sara.pdf [ 1.37MB ]
Metadata
JSON: 24-1.0378353.json
JSON-LD: 24-1.0378353-ld.json
RDF/XML (Pretty): 24-1.0378353-rdf.xml
RDF/JSON: 24-1.0378353-rdf.json
Turtle: 24-1.0378353-turtle.txt
N-Triples: 24-1.0378353-rdf-ntriples.txt
Original Record: 24-1.0378353-source.json
Full Text
24-1.0378353-fulltext.txt
Citation
24-1.0378353.ris

Full Text

  Optimal Placement and Operation of Hydrogen Fueling Stations in Metro Vancouver and Victoria  by Sara Azizkhani B.Sc., Iran University of Science and Technology, 2014  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Mechanical Engineering) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2019 © Sara Azizkhani, 2019ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, a thesis/dissertation entitled:    Optimal Placement and Operation of Hydrogen Fueling Stations in Metro Vancouver and Victoria  submitted by Sara Azizkhani in partial fulfillment of the requirements for the degree of Master of Applied Science in Mechanical Engineering  Examining Committee: Farrokh Sassani, Mechanical Engineering Supervisor  Shahabaddin Sokhansanj, Chemical and Biological Engineering Supervisory Committee Member  Ali G. Madiseh, Mining Engineering Supervisory Committee Member  Additional Examiner   Additional Supervisory Committee Members:  Supervisory Committee Member  Supervisory Committee Member  iii  Abstract Nowadays, road transportation is responsible for a substantial and growing share of GHG emissions. Canada is ranked among the top 10 worst offenders in terms of GHG emissions. Specifically, in the Province of British Columbia (BC), 25% of GHG is from the road transportation sector. BC is a test bed for fuel cell products for transportation and possesses outstanding potential for hydrogen production. Recently, hydrogen fuel cell vehicles have attracted more customers, therefore an estimation of the total hydrogen demand is required to satisfy the hydrogen demand for fuel cell vehicles. Hydrogen has more advantages than the existing fuels because it is high-quality carbon-free energy carrier with lower or zero GHG emissions. In this study, a simulation analysis in Metro Vancouver and Victoria which are located in the South West corner of BC is conducted to determine the appropriate arrangements of hydrogen stations in various time horizons. Also, the essential stages of building new stations during demand growth are discussed. To this aim, after reviewing the literature on optimisation and simulation methods, a new model to evaluate the demand of hydrogen in each municipality in Metro Vancouver and Victoria is proposed. This model classifies the parameters that are crucial in ranking potential locations and finds daily hydrogen demand. Furthermore, various scenarios are identified to find the best arrangement for hydrogen fueling infrastructure with the lowest average queue length and average waiting time for different time periods. It is shown that the total number of hydrogen stations that are essential to be built in the 30-year time horizon to cover the demand of all fuel cell vehicles in Metro Vancouver and Victoria is 52. 12 stations in each of Surrey and Vancouver; 4 in Burnaby; 3 in each of iv  Richmond, Coquitlam, Langley, and Victoria; and 2 stations in each of North Vancouver, West Vancouver, Delta, Maple Ridge, and University of British Columbia (UBC).                v  Lay Summary Fossil fuels are a significant source of air pollution and global warming. Apart from being unsustainable, these fuels are responsible for global warming. On the other hand, hydrogen is the cleanest and the most effective fuel which allows more sustainable energy systems. Since current natural resources are finite and hydrogen is infinitely renewable, hydrogen becomes the best resource for sustainable energy and consequently is a must for our future. Therefore, hydrogen-powered vehicles may be the best alternative to increase air quality and overcome pollution-related issues. However, without an adequate number of refueling stations, vehicle manufacturers may not take the risk to produce these kinds of vehicles. On the other side, energy companies prefer not to install the refueling stations unless there are some hydrogen-powered vehicles available. Therefore, an estimation of the hydrogen-powered vehicles and also the sufficient number of stations is required for both energy companies and vehicle manufacturer to have a better overview of the hydrogen market.   vi  Preface This thesis entitled “Optimal Placement and Operation of Hydrogen Fueling Stations in Metro Vancouver and Victoria” presents the research conducted by Sara Azizkhani, based on the initial research question by the supervisor, Prof. Farrokh Sassani. The proposed methodology in this manuscript is original, unpublished, independent work by the author.     vii  Table of Contents Abstract ............................................................................................................................ iii Lay Summary .................................................................................................................... v Preface .............................................................................................................................. vi Table of Contents ............................................................................................................ vii List of Tables ..................................................................................................................... x List of Figures .................................................................................................................. xi List of Symbols ............................................................................................................... xiii Glossary .......................................................................................................................... xiv Acknowledgement ........................................................................................................... xv Dedication ....................................................................................................................... xvi Chapter 1 ........................................................................................................................... 1 Introduction..................................................................................................................... 1 1.1 Background ............................................................................................................... 1 1.1.1 Hydrogen Infrastructure Pathway Options ..................................................... 2 1.1.2 Supply Chain Management ............................................................................ 4 1.2 Motivation................................................................................................................. 5 1.3 Objectives ................................................................................................................. 7 1.4 Thesis Layout............................................................................................................ 8 Chapter 2 ........................................................................................................................... 9 Literature Review ........................................................................................................... 9 2.1 Introduction............................................................................................................... 9 viii  2.2 Optimisation ........................................................................................................... 10 2.3 Simulation ............................................................................................................... 13 2.3.1 Discrete-Event Simulation ............................................................................... 14 Chapter 3 ......................................................................................................................... 17 Demand Evaluation ...................................................................................................... 17 3.1 Introduction............................................................................................................. 17 3.2 The Spatially Aggregated Demand Model ............................................................. 18 3.3 Hydrogen Demand .................................................................................................. 27 Chapter 4 ......................................................................................................................... 32 Method and Results ...................................................................................................... 32 4.1 Introduction............................................................................................................. 32 4.2 Modeling ................................................................................................................. 36 4.3 Scenario Definitions ............................................................................................... 42 4.3.1 Scenario One .................................................................................................... 44 4.3.2 Scenario Two ................................................................................................... 46 4.3.3 Scenario Three ................................................................................................. 48 4.3.4 Scenario Four ................................................................................................... 49 4.4 Data Collection ....................................................................................................... 50 4.5 Summary and Remarks ........................................................................................... 52 Chapter 5 ......................................................................................................................... 55 Conclusions .................................................................................................................. 55 5.1 Summary ................................................................................................................. 55 5.2 Conclusions ............................................................................................................ 57 5.3 Contributions .......................................................................................................... 58 5.4 Limitations .............................................................................................................. 59 ix  5.5 Future Research Direction ...................................................................................... 59 Bibliography .................................................................................................................... 61    x  List of Tables Table 3-1 Attributes affecting hydrogen vehicle adoption by consumers ........................ 19 Table 3-2 Household income ............................................................................................ 22 Table 3-3 Number of registered car per capita ................................................................. 23 Table 3-4 Number of people with 15 minutes or more commute duration ...................... 24 Table 3-5 Number of people with at least a bachelor’s degree ........................................ 25 Table 3-6 Ranking of the municipalities in Metro Vancouver and Victoria to build refueling stations ............................................................................................................... 26 Table 3-7 Prediction of the Number of Vehicles and the Total Demand ......................... 30 Table 4-1 Surrey station schedule..................................................................................... 38 Table 4-2 Station schedule in the shift block for the other municipalities ....................... 38 Table 4-3 Covered municipalities in each scenario .......................................................... 43 Table 4-4 Location and capacity of the available or planned stations .............................. 51    xi  List of Figures Figure 1-1 HSC network and components ......................................................................... 3 Figure 3-1 Hydrogen market penetration ......................................................................... 27 Figure 3-2 Demand estimation validation ........................................................................ 31 Figure 4-1 The major cities in BC .................................................................................... 33 Figure 4-2 Schematic of the simulation model ................................................................. 34 Figure 4-3 Modeling flow chart ........................................................................................ 35 Figure 4-4 Hydrogen station structure in Surrey .............................................................. 37 Figure 4-5 Multiple stations in Burnaby ........................................................................... 39 Figure 4-6 Simplified simulation model ........................................................................... 41 Figure 4-7 Metro Vancouver municipalities..................................................................... 44 Figure 4-8 Average queue length, waiting time in each district in scenario one for year 2020 .................................................................................................................................. 45 Figure 4-9 Average queue length, waiting time, and number of balks in each municipality in scenario one for the year 2030 ...................................................................................... 45 Figure 4-10 Average queue length, waiting time in each district in scenario two for the year 2030 .......................................................................................................................... 47 Figure 4-11 Average queue length, waiting time, and number of balks in each district in scenario two for the year 2040.......................................................................................... 47 Figure 4-12 Average queue length, waiting time in each municipality in scenario three for the year 2040 ............................................................................................................... 48 Figure 4-13 Average queue length, waiting time, and number of balks in each district in scenario three for the year 2050........................................................................................ 49 xii  Figure 4-14 Average queue length, waiting time, and number of balks in each district in scenario four for the year 2050 ......................................................................................... 50 Figure 4-15  Number of stations in each grid for 30-year time horizon ........................... 52 Figure 4-16 Total number of Station from 2020 to 2050 ................................................. 53 Figure 4-17 - Summary of the results ............................................................................... 54    xiii  List of Symbols  𝐴𝑊𝑖 Attribute weight HC Hydrogen consumption HD Hydrogen demand IF Impact factor 𝐿𝑞 Average queue length 𝐿𝑠 Average number of customers in the system MP Market penetration NV Number of vehicles PGR Population growth rate R Ranking 𝑆𝑊𝑖𝑗 Scoring weight 𝑊𝑞 Average waiting time in the queue 𝑊𝑠 Average waiting time in the system 𝜆 Mean arrival rate µ Mean service rate 𝜌 Utilisation    xiv  Glossary BC British Columbia GHG Green House Gasses HSC Hydrogen Supply Chain HTEC Hydrogen Technology and Energy Corporation LP Linear Programming MILP Mixed Integer Linear Programming MINLP Mixed Integer Non-Linear Programming NLP Non-Linear Programming SCM Supply Chain Management UBC University of British Columbia ZEV Zero Emission Vehicle      xv  Acknowledgement First of all, I would like to express my sincere appreciation to my supervisor, Prof. Farrokh Sassani, for his distinguished supervision, kindness, and continuous encouragement. I am very grateful to Prof. Sassani for his patience and have deeply enjoyed the pleasure of working under his supervision.  I am strongly thankful to Mr. Colin Armstrong, the president and CEO of Hydrogen Technology and Energy Corporation (HTEC) who provided me with the required data. My special thanks and gratitude go to my parents, who devoted their lives to me. Thank you for filling me with your unfailing love and kindness and always being there for me.  Last but not least, I would like to thank the person who always supports me with his endless love and patience, my beloved husband, Arash.    xvi  Dedication   To my beloved parents, Pari and Mahmoud To the love of my life, Arash               Chapter 1   Introduction 1.1 Background Non-renewable fossil fuel resources, such as natural gas, petroleum, and coal are the primary resources to supply most of the global energy demand. However, the growing demand for energy causes a fast depletion of finite fossil fuels which is not sustainable. Besides, using fossil fuel is considered a major contributor to the Green House Gas (GHG) emissions, which is responsible for aggravating global warming. Therefore, it is widely known that clean and renewable fuel alternatives should replace fossil fuels for environmental and sustainable concerns [1]. Hydrogen can be regarded as the cleanest and the most efficient fuel as it offers the highest amount of energy per unit weight without emitting much pollutants and GHGs. Furthermore, it could serve as an attractive and efficient energy carrier for storing and delivering sustainable and renewable energies such as wind, solar, and bioenergy [2]. Hence, hydrogen plays a significant role in developing an environmentally-friendly and sustainable energy system in the future for the world. 2  Hydrogen is the lightest, simplest, and most abundant chemical element in the world and readily interacts with other elements, e.g., free oxygen, and carbon, nitrogen, and oxygen in organic compounds, to form chemical compounds [3]. Its interaction with oxygen in the air produces heat energy and water. Hydrogen is widely considered as sustainable energy for decarbonising road transport, decreasing the emission of harmful gases, and enhancing the safety of energy supply. Not only it decreases the GHG emissions but also enhances the security of energy supply. Accordingly, the hydrogen economy has attracted more and more attention around the world recently [3], [4]. 1.1.1 Hydrogen Infrastructure Pathway Options A hydrogen infrastructure is defined as the supply chain required to produce, store and deliver hydrogen to the consumer. Like the current petroleum supply chain, Hydrogen Supply Chain (HSC) consists of several distinct components such as energy resources, production facilities, conditioning and storage, transportation, and refueling stations. Production processes are required to convert primary energy resources into hydrogen. Storage units are needed to compensate for the fluctuations in demand.  However, unlike most other fuel infrastructures, an additional dimension exists when defining the location of production within the supply chain. Hydrogen can be produced either centrally or distributed. Centralised productions are similar to the current gasoline supply chain. In the centralised production, large quantities are produced at a central site and then distributed, while, small-scale production facilities like reformers and electrolysers are closer to the point of use (onsite). In this scenario, the natural gas and electricity grid is used to produce 3  hydrogen at the forecourt refueling stations, which decreases the cost of distribution. Hydrogen is transported from the production facilities to the point of sale using different transportation systems. Finally, refueling technologies allow the transfer of hydrogen to users at retail stations [5]. Figure 1-1 shows that there is a variety of potential technological options at each node of the HSC. Hydrogen can be generated from a variety of sources and distributed through different forms of transportation. Gaseous hydrogen can be transported using either tube trailer or tanker truck, while a tanker truck is also used for liquid delivery [3].  Although production of hydrogen requires energy that may produce pollution itself, this will be at a fixed physical facility where the pollution can be contained and controlled.  Electrolysis Steam Methane Reforming Gasification Liquefaction Compression Tanker Truck Pipeline Tube Trailer RES (Solar, Wind, Hydro) Biomass Coal Natural Gas Refuelling Station m Refuelling Station n Refuelling Station 1 Compressed H2 Liquid H2 Energy Sources Compression Onsite Production I. Energy Source II. Production Technology III. Conditioning and Storage IV. Transport IV. Refuelling Station Figure 1-1 HSC network and components 4  A considerable number of studies have been carried out to investigate various pathway options. To optimise the cost, GHG emissions or energy efficiency of the various pathways, different assumptions on demand, size of production units and prices on the primary energy feedstocks are made. Also, a few studies examined the dynamic changes and transition from pathways, while the emphasis of most of the works was on individual pathway “steady-state” optimisation [6].  1.1.2 Supply Chain Management The hydrogen economy or more specifically the use of hydrogen as an energy resource includes a variety of aspects such as economic, environmental, and social impacts. Supply Chain Management (SCM) method can be employed to minimise the total cost of the HSC network while paying attention to environmental concerns and satisfying the demand requirements. This method applies a set of approaches to produce and distribute hydrogen more efficiently. SCM reduces the environmental impact by considering the local and global regulations for GHG emissions while being economically vital. The large size of the physical supply network and the variables that involve the decisions make the supply chain management very complex. The variables can be from details of production sites and distribution centers to the planning and scheduling of plant production and also network connectivity.  Supply chain network takes production and necessary processes, storage and transportation goods into consideration. To determine an acceptable configuration of the supply chain, the supply chain network nodes and the way of accomplishing its functionality and also the constraints between nodes should be taken into account. Based 5  on these considerations, the potential options at each node and the location of inventory are finalised [7]. Constraints and factors such as lead time and cost relationships affect the decision-making process. To assess various pathways as a potential long-term alternative, the parameters that affect performance should be taken into consideration. Each of these factors has its attributes, and the trade-offs between these factors and comparison of the different technological options lead to the best pathway. Therefore, the study of supply chains related to sustainability and resource efficiency is still needed.  1.2 Motivation Urban air quality and global warming caused by GHG emissions and energy security are currently known as the main concerns in the world. Therefore, a transition from the current global energy system to a sustainable one is attracting serious attention. While many alternatives are being suggested, the growing energy demand of the future is met with higher efficiency and more renewable energy resources such as wind, solar and biomass [5]. As our natural resources will not last forever, finding a new fuel that is clean and able to provide the power that we need is a must for our future. Hydrogen is infinitely renewable, readily available, and most importantly, it is clean. Countries like the USA, Germany, China, South Korea, and Japan are all accelerating their development of hydrogen fuel infrastructures. Japanese efforts are ahead of all other countries. 6  Using hydrogen in the transportation sector offers some advantages over existing fuels and other emerging competitors. Additionally, it is a high-quality carbon-free energy carrier, which can achieve improved efficiencies at the point of use with reduced or zero GHG emissions. To support these benefits, it should be mentioned that hydrogen can be produced from some primary energy sources, such as natural gas, coal, biomass, wind, hydro, and solar energy, contributing towards greater energy security and flexibility [8]. Nowadays, road transportation is responsible for a substantial and growing share of global GHG emissions. When it comes to global GHG emissions, Canada is ranked among the top 10 worst offenders. In the meanwhile, 25% of the total GHG emissions are from the road transportation sector in the province of British Columbia (BC). Also, 85.8% of the total GHG emission in the transportation system in BC comes from light and medium duty vehicles [9]. Hydrogen, as a low-carbon fuel, can reduce GHG emissions, particularly if produced by renewable energy sources such as solar, wind, and hydro. However, cost and efficiency are still a concern. To propose hydrogen as the fuel of the future, some technical and commercial innovations not only in vehicle technology but also for the creation of an entirely new refueling infrastructure are required [5]. New technology penetration into the transportation sector requires significant capital investment and long-term commitment. To this aim, immediate delivery of the new fuel at the refueling stations after the introduction of new vehicles is necessary. An adequate number of refueling stations is crucial for a vehicle manufacturer to convince itself before a considerable investment in the mass production of fuel cell vehicles. On the 7  other hand, energy companies’ concerns about the installation of hydrogen production, distribution and refueling infrastructures with no estimate of a profitable demand level cannot be ignored.  The balance between vehicle growth and supply chain development is a necessary pre-requirement for new technologies penetration into the transportation system. However, adequate initial fueling network coverage is required in advance of vehicle rollout. Therefore, the transition to a sustainable hydrogen economy is considered a complex strategic planning problem with significant economic consequences [10]. Generally, the network design problem can be categorised to location, allocation, routing, location allocation, and location routing by considering different levels of interest. In these problems, planning levels are strategic, tactical, or operational aspects. To study the network design problem, one can either apply optimisation or simulation methods with deterministic or stochastic data for long-, intermediate-, or short-term period. The problem that we considered in this study is a simulation of a large-scale location-allocation problem for a long-term period. Some of the required data for this problem is provided by Hydrogen Technology and Energy Corporation (HTEC) which is located in North Vancouver. HTEC works in the area of hydrogen production, distribution and fueling stations in North America. 1.3 Objectives The proposed simulation model of hydrogen infrastructure has been adapted to Metro Vancouver and Victoria to answer the following questions: 8  • What is the best arrangement for hydrogen stations regarding number and location in the given time horizon in Metro Vancouver and Victoria? • When should the new stations be built as the demand grows during the time horizon? • How do different arrangement scenarios work? 1.4 Thesis Layout The remainder of the thesis is organised as follows: In Chapter 2 we review and discuss the papers that applied optimisation and simulation methods. Chapter 3 presents a demand evaluation model to find hydrogen demands in each municipality. Methods, scenarios and the results are presented in Chapter 4; and Chapter 5 concludes the thesis based on the results.   9     Chapter 2  Literature Review 2.1 Introduction The research activities about hydrogen supply networks can be classified in terms of the implemented approach, namely optimisation and simulation. In the optimisation method, different objective functions are considered to determine the best solution. While the simulation approach, which is a collection of entities and uses logic input data to determine the performance of a system. From the mathematical modeling point of view, optimisation method benefits from a wide range of tools and techniques including Linear Programming (LP), Non-Linear Programming (NLP), Mixed Integer Linear Programming (MILP), Mixed Integer Non-Linear Programming (MINLP), and Dynamic programming. There are various commercially available software packages to solve a full range of models for HSC such as GAMS, AMPL, and MATLAB. A comprehensive review of the studies which employed optimisation method is presented in Section 2.1. Subsequently, Section 2.2 reviews the simulation-based articles, which are limited compared to optimisation-10  based studies. ExtendSim, Any Logic, and Arena are regarded as the most popular software programs for simulation purposes. The availability of ExtendSim as well as its distinguishing features, flexibility, and also the author’s previous knowledge lead to use this software in this study. 2.2 Optimisation To begin with, Han et al. [11] maximised the total net profit of Korean hydrogen supply network considering the amount of hydrogen production and the location of its storage. Almansori and Shah [12] developed a model to take the availability of energy sources (i.e., raw materials) and their logistics, as well as the variation of hydrogen demand over a long-term planning horizon, into account for Great Britain. The results of [11][12] showed that as demand grows, more production plants of different sizes should be built to meet the demand. Regarding the constraints, both papers assumed demand, production, storage, transportation, and non-negativity constraints. Despite the simplicity of this method, the main drawback is that in the single objective optimisation, where only the economic criterion is considered. While environmental impact is one of the main concerns in the HSC. Therefore, the most economical solution is not always the best alternative. Almaraz et al. [3], Kim et al. [13], and Hugo et al. [5] used multi-objective optimisation to incorporate the environmental concerns in conjunction with the traditional economic-based criteria. Almaraz et al. [3] implemented a multi-objective optimisation to design the hydrogen supply chain in the Midi-Pyrenees region in France. They also examined energy source availability on a multi-period long-term problem from 2020 to 2050, and finally showed how different geographical scales could affect the development 11  of a sustainable supply chain. In another study by Hugo [5], the MILP technique was used to analyse the long-range investment planning and design of HSC in Great Britain. The model considered British Petroleum’s strategic hydrogen infrastructure planning using high-level optimisation programming. Furthermore, Kim [13] formulated a network design problem as a MILP problem to find the optimal supply chain configurations from various alternatives in Korea. All three papers solved their multi-objective problems through a constraint method. The difference between them is the objective functions. Almaraz et al. [3] developed a sophisticated model to optimise the hydrogen network. They considered the total daily cost which is the sum of facility capital cost, energy sources, and transportation cost. While Hugo et al. [5] and Kim et al. [13] ignored one of the objective functions. Hugo [5] considered only economic and environmental criteria to set an optimal trade-off solution and Kim et al. [13] focused on cost, efficiency, and safety. For decision-making purposes, the optimal Pareto solution is needed which can be found by a trade-off between the objective functions. These three studies provided a set of Pareto solutions to apply the concept of dominance directly. Almaraz et al. [3] found the optimal solution from low to high-risk problems. Hugo et al. [5] found that as GHG emission increases, the total cost decreases, which means the net present value also increases. The total network cost based on various safety level was plotted by Kim et al. [13]. It shows that at each safety level area there is a corresponding cost which represents a specific configuration of the supply chains. By considering the risk level and the related costs, the decision makers can choose the best configuration for a supply chain. The main advantage of the multi-objective approach is that since this method generates a full set of trade-off solutions and not just 12  one single “best” alternative, the decision maker can decide on a particular design that satisfies his/her willingness to compromise by assessing interesting trade-offs between solutions. However, the drawback of this method is that, in the real world, when the decision making is about the long-term planning or even the short-term planning, there exist some source of uncertainty which may considerably affect the decision over the time horizon. To deal with this issue stochastic programming can be considered. Kim et al. [14] and Dayhim et al. [15] considered a set of uncertain parameters in their work. Kim et al. [14] investigated the design of HSC with different components such as production, storage, and transportation. They examined the total network cost for various configurations of an HSC in Korea in an uncertain environment for hydrogen demand. To introduce uncertainty to the system, a stochastic model based on the two-stage programming approach was developed. In a similar work, Dayhim et al. [15] considered the influence of uncertainty on the hydrogen production, storage, and usage in macro view (e.g., county level) to minimise the total daily social cost of the HSC network for State of New Jersey. They [15] proposed a spatially aggregated demand model to estimate the potential demand based on different household attributes such as income and education. To justify the source of uncertainty in demand, Dayhim et al. [15] stated that there are several factors which can affect consumer choice in purchasing a fuel cell vehicle. Concurrently, consumer preference on the demand side is the most important factor in predicting changes in the auto market. Thus far, we reviewed various optimisation studies and outlined the main focus and outcomes of related papers. Regarding the disadvantages of the optimisation method, we can say that the main drawback is that it is limited to moderate-sized problems. If we want 13  to apply optimisation for a large-scale problem, we should consider many simplifications which may make the problem too unrealistic. Moreover, despite offering the optimal solution in a wide range of problems, this method is not able to assess the interaction between the entities which can affect the consumers’ satisfaction, especially in the long-term governmental projects. Developing an HSC in Canada and specifically in BC, is more challenging than HSC in pioneering countries like USA and Germany due to the shortage of hydrogen roadmap at the regional and national level. This drawback leads to problems in collecting data and planning a strategy. 2.3 Simulation Apart from optimisation, simulation approach has been implemented in some studies to investigate the fuel supply chain  [19]. Simulation is considered a mathematical tool to assess the performance and efficiency of a supply chain network. The competitive market encourages manufacturing and service industries to improve their services. The population growth forces companies to enlarge their capacity or introduce new products. To predict and analyse the flexibility, performance and costumer’s satisfaction, computer simulation can be very helpful.  The goal of the simulation is to imitate reality which comes from considering all processes in an operating system [18]. Applying the simulation approach opens decision-makers’ eyes to the pros and cons of a system where an analytical solution is unavailable, 14  difficult to develop or more expensive compared to numerically solving a mathematical model. Simulation allows any details, exception, and condition to be included in the model. Simulation of the processes of a system, which is a set of interactions between receiving input and producing output, is beneficial to enhance the system performance and make better decisions based on the results [16]. A hydrogen fuel supply network is a complex combination of various factors such as different resources and machinery, facility location, and location population before implementation. Simulation of such a model would offer a comprehensive overview of the system performance [17]. Various simulation software packages can be employed to model and study complex systems. In this study, ExtendSim simulation software is selected because this program possesses both the power and flexibility simultaneously. Subsequently, based on the results of simulating different scenarios, the best option could be determined before starting the project.  The particulars of discrete event modeling and simulation are described in the next section. 2.3.1 Discrete-Event Simulation A collection of interacting objects forming an integral or complex structure with a proposed function is defined as a System [19]. To analyse a system, either the actual system or a model of the real system can be examined.  A model represents a system by general rules and concepts. It can be a simplified abstract of the real world with a sufficient level of details for the projected purpose. A model can be categorised as a mathematical, 15  physical, or logical model. To solve a mathematical model, either analytical methods or simulation can be used.  Simulation is the process of performing experiments with a simulation model to understand the behaviour of the system over time and evaluate scenarios for the operation of the system while the real system is replaced with a model. There exist two different simulation approaches; continuous and discrete event simulation. A model can also be static or dynamic and stochastic or deterministic [20], [21]. It may be challenging to derive the analytical equations that describe a complex system state [22]. In numerical modeling, it is possible to work with either simulation or optimisation. Several scenarios with different inputs and assumptions are often examined by running the simulation models, and the system behaviour is monitored. Therefore, the best scenario is selected. On the other hand, optimisation can find the optimal solution subject to the specified conditions and constraints. However, computing developments have made it possible to run multiple sequences of simulation and system configurations [23]. The best configuration among the many possible is selected through multiple scenario evaluations [22].  There exist different standards in the discrete-event simulation methodology. However, a common structure exists among different simulation software. There are different ways to introduce Input data to the system such as probability distributions with a random number generator which produces pseudo-random-numbers. To express the discrete-event system at a specific point, system state variables are needed to transfer all the information to the system. The system is comprised of Entities or Items which possess 16  attributes and information and are in a pool of interlinked blocks. The system changes when an event occurs, where an item interacts with an Activity (a process within the simulation). Delay, Queue or Logic are various types of Activities. Simulation acts as a support in the identification of the most efficient operating plans and or bottlenecks without interfering with the real system. In large physical systems, finding the results based on the simulation is considerably more cost-effective and faster than field studies, because no material or physical structure is needed. Moreover, all the potential questions, specifically “what if” questions, can be answered in the simulation without building the real system [20], [21]. Examples of real-world systems that can be modeled by discrete-event simulation are manufacturing systems, business processes and supply chains [24]. This method is a frequent decision support tool in that this technique is more flexible, versatile and has a high analysis potential [20], [21]. Based on the simulation approach, bottlenecks and influential factors in the system can be revealed. This method not only considers the interaction between processes and random events but also provide the flexibility to use various distributions for events. These make this method a powerful tool to simulate real-life processes. Sensitivity analysis and assessing the performance of various system configurations could be easily carried out by discrete-event simulation [25]. The simulation method is also popular in logistics and transportation systems [21]. Undoubtedly, supply chain management is the most common and suitable area for discrete-event simulation [21]. As a result, this methodology has been successfully applied to various supply chain and logistic studies [26]–[29]. 17     Chapter 3  Demand Evaluation 3.1 Introduction The total demand for hydrogen plays a key role in designing a sustainable energy infrastructure for hydrogen distribution. The goal of this chapter is to estimate the required hydrogen amount for each municipality in Metro Vancouver and Victoria and subsequently determine the priority of locations to build the hydrogen refueling stations.  In the last few years, there has been a growing interest in fuel cell vehicles. To satisfy the total hydrogen demand for the fuel cell vehicle market, an estimate of the hydrogen demand in each municipality is necessary. The literature on the total hydrogen demand equation expresses various approaches and parameters that affect the amount of required hydrogen. Dyckman [30], Cao and Mokhtarian [31] discuss the effect of income, the price of vehicles, and vehicle stocks in the United States on the total vehicle demand. Various studies [32], [33], [34] discuss that household income plays a key role in automobile ownership. Other important parameters to measure the total demand include 18  automobile stocks, population, and the registered automobiles [30], [35] and driving time  [32], [34]. In the studies above, it was observed that the total hydrogen demand is a function of parameters such as income, registered vehicles per household, education, and commute duration. In the present study, we attempt to consider these issues for the cities of interest. In the first step, the classification of the influential factors on the ranking of the potential locations is considered. After deciding on the top ten important municipalities, we define the total hydrogen demand equation, based on which the estimation of the number of hydrogen vehicles and hydrogen demand up to the year 2050 is calculated. 3.2 The Spatially Aggregated Demand Model To rank the first ten municipalities, the key factors affecting consumers’ choice are identified. The main steps in this model are:   Step 1: Identify the primary attributes that affect consumer acceptance of fuel cell vehicles  There are many parameters which can directly or indirectly influence the consumer vehicle choice. Table 3-1 shows the impact and rationale of the attributes which are considered in this study [30], [31], [33], [35], [36]. The considered issues are as follows: 1) Household income: Households with higher income are more likely to purchase fuel cell vehicles. 19  2) Registered car per capita: A higher number of registered car per capita implies the potential of households in buying fuel cell vehicles. 3) Commute duration: The more time using a vehicle for transportation, the more eagerness to purchase a new and more efficient vehicle (especially a fuel cell vehicle). 4) Education: A higher education level potentially motivates an individual to be an initial customer. The required data to support the analysis of these attributes are obtained from the Canadian Census Bureau [37]. Table 3-1 Attributes affecting hydrogen vehicle adoption by consumers Attribute Impact Rationale Household Income High Higher incomes lead to earlier adoption Registered Car per Capita High Households with multiple vehicles more likely to adopt hydrogen vehicles Commute Duration Medium More time spent commuting in a vehicle interests consumer in newer and more efficient vehicles Education Medium Higher education leads to earlier adoption 20  There are other attributes such as Air Quality, Clean Cities Coalitions, Hybrid Vehicle Registrations, Provincial and Federal Government Incentives, and ZEV Sales Mandate which affect the hydrogen vehicle adoption by consumers. Yet, they are not very heterogeneous within a given province.  Step 2: Classify and score the attributes We can classify and define scores for each of the attributes above (1-4).  For example, classification of household income is based on the annual income of households in Canadian dollar. The classified groups are scored from 1, for the lowest income group, to 5, for the highest income group.  The criterion of the classification is in a way that the percentage of each attribute in each region is found, and then scores from 1 to 5 are assigned to each municipality in Metro Vancouver and Victoria. The lowest score is assigned to the municipalities which are between 0% and 20%, and the highest score is given to the municipalities from 80% to 100%.  - Step 3: Weigh each attribute based on its impact level. To consider the importance of each attribute, they can be given weights as below [15]: • Household income (𝐴𝑊1  =  30%)  • Registered Cars per capita (𝐴𝑊2  =  30%) 21  • Commute Duration (𝐴𝑊3  =  20%) • Education (𝐴𝑊4  =  20%) Finally, to determine the ranking of the municipalities in Metro Vancouver and Victoria, hydrogen demand equation which was developed by Johnson and Ogden research report [38] can be modified. Based on Eq. 3-1 we can recognise the municipalities which possess the highest priority to build the hydrogen station. To this end, the 100% market penetration is considered because only the ranking of the municipalities matters at this stage of our analysis. Later, the equation with which the hydrogen demand can be calculated will be defined. 𝑅 = ∑ 𝐴𝑊𝑖 × 𝑆𝑊𝑖𝑗 4𝑖=1  3-1 Where 𝑅 in Eq. 3-1 shows the ranking, attribute weights are denoted by 𝐴𝑊𝑖, and 𝑆𝑊𝑖𝑗 represents the scoring weight. Table 3-2 and Table 3-3 show the classification of household income and registered cars as five different groups. The categorisation of commuting duration is based on the number of households who travel to work more than 15 minutes every day (Table 3-4). The last influential parameter is the education level, which is based on the number of people who have at least a bachelor’s degree (Table 3-5). It is worth adding that scores are normalised so that the attributes for each group are equal to 100%. Melendez and Milbrandt [39] defined a scoring classification approach for the State of New Jersey. Then, we modified this method for Metro Vancouver and Victoria. 22  The required data to classify attributes were obtained from 2016 Canada Census and shown in Table 3-2, 3-3, 3-4, and 3-5 [37]. Household Income o Data representation: income  o Rationale: Higher income motivates customers to experience hydrogen vehicles.  Table 3-2 Household income Classification of Income ($CAD) Score 0–26,999 1 27,000–54,999 2 55,000–79,999 3 80,000–104,999 4 Over 105,000 5  Registered Car per Capita o Data representation: number of registered cars per population of each municipality. 23  o Rationale: Due to limited hydrogen range and refueling stations, those municipalities who have a higher number of registered cars will be the first customers.  Table 3-3 Number of registered car per capita Classification of Number of Registered Cars per Capita Score 0 – 0.2 1 0.2 – 0.4 2 0.4 – 0.6 3 0.6 – 0.8 4 Over 0.8 5  Commute Duration o Data representation: Workers who are older than 15 years and commute more than 15 minutes per day. o Rationale: Those who spend more time commuting in a vehicle every day might be more interested in the more efficient and newer vehicles.  24  Table 3-4 Number of people with 15 minutes or more commute duration Classification of Number of Commuters Travelling Longer than 15 Minutes Scores 0–52,999 1 53,000–104,999 2 105,000–159,999 3 160,000–209,999 4 Over 210,000 5  Education o Data representation: Number of people who are older than 25 with at least a bachelor’s degree o Rationale: Higher education level is another motivating factor for considering the fuel cell vehicles as a premier vehicle     25  Table 3-5 Number of people with at least a bachelor’s degree Classification of Number of Educated People Scores 0–43,999 1 44,000–87,999 2 88,000–131,999 3 132,000–175,999 4 Over 176,000 5  Implementing the results of considered categories in the proposed equation (Eq. 3-1) leads to the ranking of municipalities, as presented in Table 3-6.     26  Table 3-6 Ranking of the municipalities in Metro Vancouver and Victoria to build refueling stations Priority Municipality R (Assuming 100% market penetration in each municipality) 1 Vancouver 4.00 2 Surrey 3.95 3 Langley 3.10 4 North Vancouver 3.00 5 Burnaby 2.90 6 Delta 2.75 7 Richmond 2.75 8 Coquitlam 2.60 9 Victoria 2.50 10 New Westminster 2.50 This table represents the first ten municipalities that have the highest value in ranking 𝑅, Eq. 3-1. 27  3.3 Hydrogen Demand One of the essential parameters that can directly affect the estimation of hydrogen demand is market penetration. This parameter indicates the adoption of a hydrogen production compared to the total theoretically expected market. In several studies [3], [40]–[42], the analysis of the market penetration factor was considered to observe the variation of demand over time. The modified market penetration variation for Metro Vancouver and Victoria is illustrated in Figure 3-1. This figure shows that market penetration factor varies as an S-shape trajectory. This S-shape indicates that at the first stage which is the introductory stage, the penetration is less than 10%. This stage is mainly limited to vehicles with the daily route and the regular refueling intervals probably in dense urban areas such as Vancouver. At the second stage, the cost of hydrogen production and manufacturing of hydrogen fuel cell vehicles is more affordable. On the other hand, obstacles are overcome thanks to the experiences gained at the first stage. Therefore, the demand or market penetration percentage trajectory might rise sharply [43].       Introductory Stage Growth Stage  Saturated Stage  Figure 3-1 Hydrogen market penetration 28  Ultimately, in the saturated market where the market penetration is more than 90%, the demand trajectory will reach a plateau of approximately a hundred percent.  After determining the market penetration variation, the total hydrogen demand can be calculated. In the previous works [3], [44], the total hydrogen demand was predicted by the total number of vehicles, the average distance traveled and also population density in each grid. As an improvement, in some studies such as conducted studies in [3], [8], only the saturated market (penetration of hundred percent) was considered which is despite real hydrogen economy. Hydrogen demand will increase gradually and ultimately reaches the ideal situation of hundred percent market penetration [43]. To further improve this approach, the daily hydrogen demand (HD) formula (Eq. 3-2) is modified in the current study based on market penetration (MP) - which follows an S-shaped trajectory - population growth rate (PGR), number of vehicles (NV) and hydrogen consumption (HC), and a new parameter impact factor (IF): 𝐻𝐷 = 𝑁𝑉 × 𝑃𝐺𝑅 × 𝑀𝑃 × 𝐻𝐶 × 𝐼𝐹 3-2 In this study, HC is considered as 0.5 kg of H2/day/vehicle, and PGR is calculated with respect to the growth rate from the year 2006 to the year 2016 collected from Canada Census data. The estimation of HC is based on the assumption that the average vehicle travels is 20,000 km/year and the fuel economy of 65 miles/kg (roughly equivalent to a gasoline fuel economy of 65 miles per gallon). Note that the potential energy of 2.2 pounds (1 kg) of hydrogen gas is nearly the same as that of 1 gallon of gasoline [45].  29  Consequently, the total hydrogen demand per day and the estimation of the number of vehicles per year can be determined. In our analysis, the derived equation is a function of not only income and registered car per capita, but also commuting duration and educated people for each municipality in Metro Vancouver and Victoria. We assume that normalised value of income and registered car per capita affect IF calculation with the weight of 0.3 while the number of educated people and commuting duration have the equal weights of 0.2 (Eq. 3-3). Hence, IF formula is: 𝐼𝐹 = 0.3 × (𝐼𝑛𝑐𝑜𝑚𝑒 + 𝑅𝑒𝑔𝑖𝑠𝑡𝑒𝑟𝑒𝑑 𝑐𝑎𝑟𝑠 𝑝𝑒𝑟 𝑐𝑎𝑝𝑖𝑡𝑎)+ 0.2 × (𝑐𝑜𝑚𝑚𝑢𝑡𝑖𝑛𝑔 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛 + 𝑒𝑑𝑢𝑐𝑎𝑡𝑒𝑑 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛) 3-3 Therefore, by finding the total hydrogen demand for Metro Vancouver and Victoria, the total number of hydrogen fuel cell vehicle is also predicted. The results which were calculated by Excel can be seen in Table 3-7:   30  Table 3-7 Prediction of the Number of Vehicles and the Total Demand Prediction Year MP% Number of Vehicles Demand kgH2/Day 2015 0 0 0 2018 0.4 234 160.3 2020 1 585 400.7 2025 3 1755 1202.1 2030 9 5265 3606.2 2035 23 13456 9216.4 2040 46 26912 18432.9 2045 79 46219 31656.8 2050 100 57920 39671.2   To validate our anticipations, the estimated number of vehicles is compared with the prediction made by the government [46] from 2018 to 2030. Figure 3-2 substantiates that our results are in close agreement with those from the government for 2018 to 2030. 31  The number of vehicles from 2030 onwards is estimated based on Eq. 3-2 which is required for our simulation to find the number of refueling station in the future.  As shown in Figure 3-2, the number of vehicles reaches 5,265 for the market penetration of 9% by 2030 and falls in the introductory phase which leads to a gradual increase in the number of vehicles from 234 in 2018 to 5,265 in 2030. From 2030 to nearly 2045, the second phase of the hydrogen fuel cell vehicle market occurs where the number of vehicles rises more than seven times over 15 years and reaches 46,219. Eventually, by 2050 when the market penetration is 100%, the total number of vehicles based on our prediction will be 57,920 which means hydrogen demand is slightly below 40,000 kg of hydrogen per day. In our simulation we considered a 30-year time horizon starting from 2020 to find the required hydrogen stations. However, to compare the results with the Government data, we find the demand from 2015 when market penetration is zero.    Figure 3-2 Demand estimation validation 32     Chapter 4  Method and Results  4.1 Introduction This chapter is devoted to the hydrogen fueling infrastructure design. Here, we study the hydrogen fueling infrastructure for the case study of Metro Vancouver and Victoria: a multi-period problem is considered (2020-2050). Furthermore, a geographic division, based on municipalities instead of grid squares is used. This approach is more consistent with administrative data which facilitates data collection [3]. Figure 4-1 shows the major cities in BC [47]. Metro Vancouver, the largest region of Lower Mainland, is located in the South West corner of the province of BC in Canada.       33             BC is considered a leading developer and a test bed for fuel cell products for transportation and has excellent potential for producing hydrogen based on renewable sources [48]. The BC government has legislated substantial pieces of climate actions (Greenhouse Gas Reduction Targets Act) to reduce emissions and transition to a low-carbon economy. Under the Act, BC's GHG emissions should be lowered by at least 80 percent below 2007 levels by 2050. In December 2008, the Act provided authority for the Greenhouse Gas Emission Control Regulation [49] and the Carbon Neutral Government Regulation [50]. These objectives lead to study new scenarios related to the transportation system in this region. The hydrogen fueling infrastructure program is managed by the Canadian Hydrogen and Fuel Cell Association and was initiated in 2015 [51]. This study emerged as an initiative to evaluate the hydrogen infrastructure based on Figure 4-1 The major cities in BC  34  the estimated demand in Metro Vancouver and Victoria. More specifically, the objectives of the project are based on the following items: • Demand estimation • Scenario definitions • Data collection and assumptions • Modeling • Results analysis The methodological framework of the study is proposed in Figure 4-2. The input block corresponds to all the databases, assumptions, and scenarios. The integration of the model and the scenario-based approach constitute the core of the approach. The result concerning the fueling station arrangement is the main output.         Input   -Potential location of hydrogen stations        -Demand  -Demand ratio -Dispenser capacity -Technical data   Simulation Framework   Applying different scenarios to determine “the optimal arrangement of hydrogen stations in Metro Vancouver and Victoria" using discrete event simulation           Output  -Find the best scenario (number of hydrogen stations) for each time period in each municipality -Average waiting time -Average queue Length        Figure 4-2 Schematic of the simulation model 35  Based on the model flowchart provided in Figure 4-3, hydrogen is produced and transported to different locations based on the defined scenarios in section 4.3. At this point, based on the distribution of cars, cars go to different refueling stations and refill their tank. The probability of the hydrogen distribution to each location is the same as the demand ratio which was calculated in Chapter 3. After car arrival at the stations, car tank is filled and after a while car needs to come back to the refueling stations. In case that the queue length is more than five, cars will exit the model without refueling. It means that the scenario is no longer suitable due to the demand increase. This is the time to stick to another scenario.   Legend: Hydrogen production  Hydrogen delivery  Refueling station  Car distribution (demand ratio)  Exit fueling station and come back after awhile  Car arrival (demand)  Surrey  Burnaby  Vancouver  Victoria  Refueling  Hydrogen flow  Queue>5  Yes  No  Balk  Car flow  … Figure 4-3 Modeling flow chart  Car and hydrogen flow  Car flow  Car flow  Car flow  Car flow  Car flow  Car flow  Car flow …  … 36  4.2 Modeling Thus far we have estimated the total hydrogen demand in Metro Vancouver and Victoria which leads to the number of vehicles in each municipality. Figure 4-4 shows the structure of Surrey hydrogen station which is a part of the model that we developed in this study. This figure indicates the details of the station located in the municipality of Surrey in the simulation model. Station capacity is set using a Tank block from the Rate library. The Valve block next to the Tank block shows the station nozzle. This block calculates station utilisation, busy time, and idle time. To define the car fuel tank in the model, an Interchange block is used. Hydrogen flow and car flow connect at this block. The assumption of refueling duration is considered in the Valve blocks.            37  Figure 4-4 Hydrogen station structure in Surrey   38  As all the stations are not available 24 hours, each station should be scheduled in the model. To control the station schedule, shift block can be used. Table 4-1 shows the working hour of hydrogen station located in Surrey (8:30 – 16:30) and Table 4.2 shows the other stations schedule (6 - 23) in the shift block, as suggested by HTEC. Service time is defined for one day and repeated every 24 hours. The start point of the shift block should be 0 which represents 24:00. Table 4-1 Surrey station schedule Time On/Off 0 Off 8.5 On 16.5 Off  Table 4-2 Station schedule in the shift block for the other municipalities Time On/Off 0 Off 6 On 23 Off To define the car distribution using the demand ratio, which was calculated in Chapter 3, a “select item out” block was used. In the next time stages, when apart from the existing station new stations are needed in each municipality, we should use this block to direct the cars into different stations (Figure 4-5)39  Figure 4-5 Multiple stations in Burnaby 40  In the original model, the items are circulated in the system by the activity block to refill their tank over time. In this model, if the tank of a car cannot be refilled due to the long queue ahead, the car will be balked. In the very simplified model (system 1 in Figure 4-6), we assume that each car exits the model after the first refueling. To compare the model result with queuing theory, we considered the entire refueling station (station #1 in the system 1 shown in Figure 4-6) as an activity block (station #2 in the system 2 shown in Figure 4-6). We replaced the refueling station in the system 1 by the refueling station in system 2 in a way that the average length in both systems is equal. Thus, it can be concluded that both systems are working in the same way. The system 2 in Figure 4-6 shows a single channel queue model that the items enter the model exponentially with exponential inter-arrival time, and the service time is constant. Eq. 4-1 to 4-5 show the queuing formulas for an M/D/1 queue model. 𝜌 =𝜆µ 4-1 𝐿𝑞 =𝜆22(µ − 𝜆)=𝜌22(1 − 𝜌) 4-2 𝑊𝑞 =𝜆2(µ − 𝜆) 4-3 𝐿𝑠 = 𝐿𝑞 +𝜆µ 4-4 𝑊𝑠 = 𝑊𝑞 +1µ 4-5  41  where 𝜌 is the utilisation, 𝜆 is the mean arrival rate, and µ is the service rate, 𝐿𝑞 is the average queue length, 𝑊𝑞 is the average waiting time in queue, 𝐿𝑠 is the average number of customers in the system, and 𝑊𝑠 is the average waiting time in system. The mean arrival rate can be defined by dividing one over the mean arrival time where in this case is (Eq. 4-6): 𝜆 =1𝑚𝑒𝑎𝑛 𝑎𝑟𝑟𝑖𝑣𝑎𝑙 𝑡𝑖𝑚𝑒=11.2= 0.83 4-6 The service rate (µ) is one over the service time which is one hour. The value of the utilisation can be calculated by Eq. 4-1 which is 0.83 and is very close to the simulation result, which is 0.84. Additionally, the average queue length calculated by the queuing theory (Eq. 4-2 and 4-3) is 2.04 which is close to the one determined by the simplified simulation model, which is 2.06. Also, the waiting time computed by the model is 2.48 and is very close to the one calculated by queuing theory, which is 2.44. Similarly, we can show that 𝑊𝑠 and 𝐿𝑠 possess close values for both the queuing theory and the model results. The close agreement between the queuing theory and the model results validates our simulation.       Figure 4-6 Simplified simulation model 42  4.3 Scenario Definitions Among the many tested scenarios, the best options are chosen to expand regarding the number and location of stations, and results. The first scenario is based on the available and already planned stations by HTEC. As the demand increases, the average queue length and waiting time are also increases, specifically in Vancouver and Surrey. One solution may be going ahead with the available stations and cars change their route to another station when they encounter a long queue ahead. This approach seems very cost effective, but it does not take the driver convenience factor into account as the queue length and waiting time may be very long. The more logical approach is to increase the number of stations either in the same place that already exists or in other places to provide more coverage all over the area.  There are four different scenarios to overcome this issue (Table 4-3). In the first one, we start with the available stations (six stations) and then looking for the time that we need to increase the number of stations to remove the bottlenecks in the system.  At the point that we need to launch new stations, two approaches will be considered. One is to add more stations in the places that are more crowded, and the other is to build new stations in the new places based on the ranking which was shown in Chapter 3.   43  Table 4-3 Covered municipalities in each scenario 2020 2030 2040 2050 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Surrey Surrey Surrey Surrey Vancouver Vancouver Vancouver Vancouver Burnaby Burnaby Burnaby Burnaby Victoria Victoria Victoria Victoria UBC UBC UBC UBC North Vancouver North Vancouver North Vancouver North Vancouver  Delta Delta Delta  Richmond Richmond Coquitlam Coquitlam New Westminster New Westminster Langley Langley West Vancouver West Vancouver Maple Ridge Maple Ridge 44  The geographic breakdown of Metro Vancouver municipalities can be seen in Figure 4-7 [52]. In this study, Langley denotes the city and the township of Langley; and North Vancouver denotes both the city and the district of North Vancouver. 4.3.1 Scenario One  In the first scenario, six different municipalities are considered as the potential hydrogen stations in 2020 locations including Surrey, Vancouver, Burnaby, North Vancouver, UBC, and Victoria. Based on our simulation we can find the queue length and waiting time on average. Therefore, we can show the performance of this scenario (Figure 4-8).  Figure 4-7 Metro Vancouver municipalities  45  05101520LocationScenario One in 2030Average queue length Average waiting time (minute) Balk (/100)If we continue with this scenario for the year 2030, the queue length will be higher than the reasonable value, five in this study, which is not convenient for the customers. Apart from queue length, there are some customers in some municipalities that leave the fueling station without filling their tank due to the long queue.     Figure 4-9 Average queue length, waiting time, and number of balks in each municipality in scenario one for the year 2030  00.10.20.30.40.5LocationScenario One in 2020Average queue length Average waiting time (minute)Figure 4-8 Average queue length, waiting time in each district in scenario one for year 2020 46  Although the average queue length in Figure 4-9 is less than five, there are balks in Surrey, Vancouver, and Burnaby which means that this scenario is no longer efficient. Balks in Surrey, Vancouver, and Burnaby is approximately 300, 850, 100, respectively. Therefore, we have to move to the second scenario to consider higher customer satisfaction for the year 2030. 4.3.2 Scenario Two In the second scenario, nine potential hydrogen stations are considered to be located in seven different municipalities. Two stations in Surrey, two in Vancouver, and one in each of Burnaby, North Vancouver, UBC, Delta, and Victoria municipalities. These stations are for 2030, and the performance of the scenario can be shown based on the average queue length and the waiting time from the simulation model. Figure 4-10 shows the average queue length and the waiting time in this scenario. As it can be seen, scenario two appropriately works in the year 2030, however, it cannot be used in the year 2040 due to a large number of balks in all the municipalities (Figure 4-11). Therefore, it is time to resort to a new scenario.  47   Figure 4-10 Average queue length, waiting time in each district in scenario two for the year 2030  Figure 4-11 Average queue length, waiting time, and number of balks in each district in scenario two for the year 2040 In Figure 4-10 we see that the average waiting time can be as high as 11 minutes in Vancouver. This indicates that waiting time can be at times much longer than 11. The 02468101214Surrey Vancouver Burnaby Victoria UBC NorthVancouverDeltaLocationScenario Two in 2030Average queue length Average waiting time (minue)020406080100120Surrey Vancouver Burnaby Victoria UBC NorthVancouverDeltaLocationScenario Two in 2040Average queue length Average waiting time (minute) Balk (/100)48  scenarios examined in this thesis were typical and used to show the power of discrete-event simulation. Indeed, other scenarios can be examined with a view of reducing the average waiting time. 4.3.3 Scenario Three In the third scenario, 32 new hydrogen stations should be established in 13 different municipalities. Eight stations in Surrey, 10 in Vancouver, three in Burnaby, and two in each of North Vancouver, UBC, Coquitlam, Maple Ridge, Delta, West Vancouver, Richmond, New Westminster, Langley and Victoria municipalities. The performance of this scenario in the year 2040 can be seen in Figure 4-12.         Figure 4-12 Average queue length, waiting time in each municipality in scenario three for the year 2040  024681012LocationScenario Three in 2040Average queue length Average waiting time (minute)49  If we apply this scenario in the next period (2050), the convenience factors including waiting time, queue length, and the number of balks will increase significantly and affect the consumers’ satisfaction (Figure 4-13).  Figure 4-13 Average queue length, waiting time, and number of balks in each district in scenario three for the year 2050  At this point, we need to stick to a new scenario. 4.3.4 Scenario Four In the last period, at least 11 new hydrogen stations need to be considered to cover the demand all over Metro Vancouver and Victoria. New stations are four more in Surrey, two in Vancouver, and one in each of Burnaby, Coquitlam, Langley, Richmond and Victoria municipalities. The results of this scenario are shown in Figure 4-14. 051015202530LocationScenario Three in 2050Average queue length Average waiting time (minute) Balk (/100)50   Figure 4-14 Average queue length, waiting time, and number of balks in each district in scenario four for the year 2050 4.4 Data Collection  Demand: Total number of cars that go to the refueling stations are different due to the different demand each year.   Car tank capacity: Drivers always show up with different amount of fuel in the tank. Average in California is 3.5 kg, but it is increasing as more stations become operational, and confidence builds.  Car consumption: Each car consumes about 1 kg hydrogen per 100 km. Therefore, car consumption is about 0.5 kg/day on average.  The time between refueling: As the car tank should not run empty, we assumed refueling one day early, with about one day’s fuel in the tank, which is 10 %. 024681012LocationScenario Four in 2050Average queue length Average waiting time (minute)51   Station schedule: Table 4-4 show the planned station schedule. Service times in each station is assumed not to change during the time horizon.  Refueling duration: Refueling takes about 8 minutes long on average.  Station capacity: Table 4-4 shows the station capacity. Table 4-4 Location and capacity of the available or planned stations Station Location Capacity Schedule Powertech Surrey 500 kg/day 8:30 - 16:30 HTEC Vancouver 150 kg/day 6:00 – 23:00 HTEC Burnaby 150 kg/day 6:00 – 23:00 HTEC UBC 150 kg/day 6:00 – 23:00 HTEC North Vancouver 150 kg/day 6:00 – 23:00 HTEC Victoria 150 kg/day 6:00 – 23:00 In the first and second stage (early stages of the market) the stations are small (150 kg/day) and medium (500 kg/day), while the stations that will be established after 2030 are assumed to be a large size (1000 kg/day). The reason that larger stations such as the reformer station and liquid hydrogen station should be used in the saturated state of the market is that they exhibit the lowest costs since they can spread their installation and capital costs over a large volume of hydrogen sale [53].  52  4.5 Summary and Remarks The results show that 52 hydrogen stations need to be established in the 30-year time horizon to satisfy the demand all over Metro Vancouver and Victoria. Figure 4-15 shows the number of stations in each grid for a 30-year time horizon.       As the demand increases substantially at the end of the second time period, the majority of the stations (more than 50%) will be built in the year 2040 which is shown in Figure 4-16. Figure 4-15  Number of stations in each grid for 30-year time horizon 02468101214Number of hydrogen stationsLocationNumber of stations in each grid2020 2030 2040 205053  Figure 4-16 Total number of Station from 2020 to 2050    Figure 4-17 shows the summary of the results at each year. Based on this figure, it can be observed that which scenario works better in each period. The average length of five is considered as the maximum average length in this study and is the criterion to move from one scenario to a new one. The results reported in this research will undoubtedly be affected by change in the hydrogen technology, competing technologies, such as electric vehicles, and the nature of data in the years ahead. Therefore, predictions for the decades of 2040, for example, are just that: “predictions”. The correct and logical course of action will be to accept and use the results for say, a period of 5 years, which can be considered as “more probable predictions” and after 5 years re-run simulations, perhaps with a revised model, using new data that changes in technology will bring about.54       Figure 4-17 - Summary of the results 55     Chapter 5 Conclusions 5.1 Summary  In this section, a summary of this study for the optimal placement of hydrogen stations in Metro Vancouver and Victoria is presented As discussed, fossil fuels have a major contribution in the GHG emission and are unsustainable, while hydrogen is infinitely renewable and is the best resource for sustainable energy. Consequently, hydrogen is a must for our future. Since 25% of the total GHG emissions in BC is from the road transportation sector, we decided to consider a simulation study to find the best potential locations for hydrogen fueling stations in Metro Vancouver and Victoria in British Columbia. In this study, we considered the simulation of a large-scale location-allocation problem for a long-term period. Through studies that apply optimisation or simulation methods, the modified demand evaluation model was presented to determine hydrogen demand in each municipality. The hydrogen demand model considered all essential factors 56  such as income, number of registered cars per capita, education, commute duration to estimate the demand. Then, these parameters were classified to evaluate the total hydrogen demand to satisfy the hydrogen needed in the market of fuel cell vehicles. The spatially aggregated demand model was applied to find both the number of hydrogen vehicles and the hydrogen demand up to the year 2050.  It was shown that the daily hydrogen demand depends on market penetration, population growth rate, number of vehicles, hydrogen consumption, and the impact factor. The estimated number of vehicles from 2018 to 2030 were compared with the government data [46] to validate the results. After the validation, the number of vehicles and the total demand from 2030 to 2050 were predicted. Another objective of this study was to determine the best arrangement of hydrogen stations, the number of stations and their locations, in the given time horizon in Metro Vancouver and Victoria. The years and the number of allocated hydrogen stations as the demand increases were also determined.  Moreover, by assessing the various scenarios using ExtendSim, the best scenarios with the required number of stations were estimated. HTEC suggested the first scenario, but we still need to define new scenarios, because as the demand increases the average queue length and the average waiting time rise, as expected. Therefore, we considered the second scenario by increasing the number of stations to remove the bottlenecks in the system. The criterion for approaching a new scenario in this study is the average queue length. When this average is more than five which shows high traffic in the fueling stations, a new scenario should be defined. In the first scenario, six different municipalities were considered as the potential hydrogen stations including Surrey, Vancouver, Burnaby, 57  North Vancouver, UBC, and Victoria starting from 2020, while in the second scenario, nine potential hydrogen stations were considered to be located in seven different municipalities from 2030.  Based on the results, in the third scenario, 32 new hydrogen stations are needed in 13 different municipalities from 2040. Finally, it was shown that 11 new hydrogen stations should be established to cover the demand for Metro Vancouver and Victoria in 2050. To sum up, it was computed that 52 hydrogen stations are needed in the 30-year time horizon to cover the demand of all fuel cell vehicles in Metro Vancouver and Victoria. 5.2 Conclusions This study showed that simulation could be used as a powerful decision-making tool and provide the decision makers with the flexibility to model various levels of detail and complexity. Especially that it allowed a high degree of system details and particulars to be modeled. It produced tangible and useable results, verified through the queueing theory, whereby decisions can be made with confidence.  Through the research conducted it became evident that successful simulation requires careful modeling, and reliable data which is one of the important aspects of any simulation study. It also allowed seeing how the system works by changing input data, adding or removing any recourses. The simulation-based framework can be easily scaled for additional customers, resources, and different stations configurations. Energy corporations face challenges in efficiently satisfying hydrogen demand with limited resources and capacity. As demand increases, it becomes a challenging issue 58  for the stakeholders to make operational decisions. This study provides an overview of demand growth in Metro-Vancouver and Victoria to help not only decision makers to prepare the required infrastructure but also fuel cell vehicles manufacturers in the 30-year time horizon. 5.3 Contributions This study is a preliminary effort to evaluate the hydrogen infrastructure based on the estimated demand. The applied simulation in Metro Vancouver and Victoria helps to predict the number of hydrogen station base on the demand input. More specifically, the contributions of this study are as follows: - Estimating the total demand and the total number of vehicles in the time period from 2020 to 2050 which aids the manufactures to produce sufficient vehicles that meet the expected market demand. - Determining the number of stations in each time horizon which assist in establishing the required infrastructure. - Estimating the average queue length and the average waiting time which are representatives of customer convenience in the refueling, and which aid when new stations must be built. 59  5.4 Limitations One of the assumptions that can improve the model is refueling behavior which was not considered in this study due to the lack of data. Refueling behavior refers to a customer’s fueling preference or habits.  In this study, we assumed that drivers refuel their car near where they live. However, suppose that one who lives in Vancouver but works in Coquitlam may prefer to use the Coquitlam station or Burnaby station during his/her trip rather than using the Vancouver station. This data can be obtained in different ways such as surveys or interviews with drivers. 5.5 Future Research Direction Determining more feasible locations of the fueling stations based on land availability, land price, production location, and delivery issues is suggested as future research.  Furthermore, studying the effect of the related costs of hydrogen production methods and location, hydrogen transportation, establishing the refueling stations, and employing the available resources can be considered as future research. The combination of the cost function and the consumer convenience factor (which was considered in this study), would provide a detailed infrastructure proposal for the future of the hydrogen market. 60   This research focused on light and medium vehicles. However, in the future, the provincial and federal government can legislate that transportation in specific sectors, such as Canada Post or public transit must adopt hydrogen-powered vehicles.    61  Bibliography [1] A. Kontogianni, C. Tourkolias, and E. I. Papageorgiou, “Revealing market adaptation to a low carbon transport economy: Tales of hydrogen futures as perceived by fuzzy cognitive mapping,” Int. J. Hydrogen Energy, vol. 38, no. 2, pp. 709–722, Jan. 2013. [2] K. Bennaceur, B. Clark, F. M. Orr, T. S. Ramakrishnan, C. Roulet, and E. Stout, “Hydrogen: A future energy carrier?,” Gas, vol. 25, no. 20, pp. 15–3, 2005. [3] S. De-León Almaraz, C. Azzaro-Pantel, L. Montastruc, and S. Domenech, “Hydrogen supply chain optimization for deployment scenarios in the Midi-Pyrénées region, France,” Int. J. Hydrogen Energy, vol. 39, no. 23, pp. 11831–11845, 2014. [4] A. Ajanovic, “On the economics of hydrogen from renewable energy sources as an alternative fuel in transport sector in Austria,” Int. J. Hydrogen Energy, vol. 33, no. 16, pp. 4223–4234, Aug. 2008. [5] A. Hugo, P. Rutter, S. Pistikopoulos, A. Amorelli, and G. Zoia, “Hydrogen infrastructure strategic planning using multi-objective optimization,” Int. J. Hydrogen Energy, vol. 30, no. 15, pp. 1523–1534, 2005. [6] C. E. Thomas, I. F. Kuhn, B. D. James, F. D. Lomax, and G. N. Baum, “Affordable hydrogen supply pathways for fuel cell vehicles,” Int. J. Hydrogen Energy, vol. 23, no. 6, pp. 507–516, Jun. 1998. [7] D. J. Thomas and P. M. Griffin, “Coordinated supply chain management,” Eur. J. 62  Oper. Res., vol. 94, no. 1, pp. 1–15, Oct. 1996. [8] D. Joffe, D. Hart, and A. Bauen, “Modelling of hydrogen infrastructure for vehicle refuelling in London,” J. Power Sources, vol. 131, no. 1–2, pp. 13–22, May 2004. [9] O. Herrera, M. Taiebat, F. Sassani, and W. Merida, “Implications of transportation electrification in metro Vancouver,” in 2016 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE), 2016, pp. 1–4. [10] S. Dunn, “Hydrogen futures: toward a sustainable energy system,” Int. J. Hydrogen Energy, vol. 27, no. 3, pp. 235–264, Mar. 2002. [11] J. H. Han, J. H. Ryu, and I. B. Lee, “Modeling the operation of hydrogen supply networks considering facility location,” Int. J. Hydrogen Energy, vol. 37, no. 6, pp. 5328–5346, 2012. [12] A. Almansoori and N. Shah, “Design and operation of a future hydrogen supply chain: Multi-period model,” Int. J. Hydrogen Energy, vol. 34, no. 19, pp. 7883–7897, 2009. [13] J. Kim and I. Moon, “Strategic design of hydrogen infrastructure considering cost and safety using multiobjective optimization,” Int. J. Hydrogen Energy, vol. 33, no. 21, pp. 5887–5896, 2008. [14] J. Kim, Y. Lee, and I. Moon, “Optimization of a hydrogen supply chain under demand uncertainty,” Int. J. Hydrogen Energy, vol. 33, no. 18, pp. 4715–4729, 2008. [15] M. Dayhim, M. A. Jafari, and M. Mazurek, “Planning sustainable hydrogen supply 63  chain infrastructure with uncertain demand,” Int. J. Hydrogen Energy, vol. 39, no. 13, pp. 6789–6801, 2014. [16] G. Romero and L. Martinez, “Modelling, simulation and optimization,” Int. J. Hydrog. EnergynTech, 2010. [17] A. Moazzami, M. Rahiminezhad Galankashi, and A. Khademi, “Simulation, modeling and analysis of a petrol station,” Int. Rev. Model. Simulations, vol. 6, no. 1, pp. 246–253, 2013. [18] M. Pidd, “Computer simulation in Management Science,” 4th Ed. , John Wiley Sons Ltd, Chichester, 1998. [19] C. Cassandras and S. Lafortune, “Introduction to discrete event systems,” Springer, 2008. [20] A. Law, Simulation modeling and analysis. 2014. [21] J. Banks, Handbook of simulation: Principles, methodology, advances, applications, and practice. 1998. [22] P. White and R. G. Ingalls, “Introduction to simulation,” in Winter Simulation Conference., 2009, pp. 12–23. [23] A. M. Law and M. G. McComas, “Simulation optimization: simulation-based optimization,” Proc. Proc. 34th Conf. Winter Simul. Explor. new Front. 2002 Winter Simul. Conf., pp. 41–44, 2002. [24] A. Eriksson, Improving the efficiency of forest fuel supply chains. 2016. [25] A. Thesen and L. E. Travis, Simulation for decision making. 1992. 64  [26] M. Mobini, “Forest biomass supply logistics for a power plant using the discrete-event simulation approach,” Appl. Energy, vol. 88, no. 4, pp. 1241–1250, 2011. [27] M. Mobini, T. Sowlati, and S. Sokhansanj, “A simulation model for the design and analysis of wood pellet supply chains,” Appl. Energy, vol. 111, pp. 1239–1249, 2013. [28] K. Karttunen, K. Väätäinen, A. Asikainen, and T. Ranta, “The operational efficiency of waterway transport of forest chips on Finland’s Lake Saimaa.,” Silva Fenn., vol. 46, no. 3, pp. 395–413, 2012. [29] S. Sokhansanj, S. Mani, S. Tagore, and A. Turhollow, “Techno-economic analysis of using corn stover to supply heat and power to a corn ethanol plant–Part 1: Cost of feedstock supply logistics,” Biomass and Bioenergy, vol. 34, no. 1, pp. 75–81, 2010. [30] T. Dyckman, “An aggregate demand model for automobiles,” J. Bus., vol. 38, no. 3, 1965. [31] X. Cao and P. Mokhtarian, “Future demand for alternative fuel passenger vehicles: A diffusion of innovation approach: UC Davis-Caltrans air quality project,” Res. report, UC Davis-Caltrans Air Qual. Proj. Davis Univ. Calif., 2004. [32] J. Chung and D. Lee, “Structural model of automobile demand in Korea,” Transp. Res. Rec. J. Transp. Res. Board, vol. 1807, pp. 87–91, Jan. 2002. [33] J. C. Tanner, “Long-term forecasting of vehicle ownership and road traffic,” J. R. Stat. Soc. Ser. A, vol. 141, no. 1, p. 14, 1978. 65  [34] K. Train, Qualitative choice analysis: Theory, econometrics, and an application to automobile demand. The MIT Press, 1993. [35] C. F. Manski, “Aggregate demand and discrete choice models of the U.S.A. automobile market: A critical comparison,” Transp. Policy Decis. Mak., vol. 1, pp. 313–326, 1980. [36] G. Dargay, “Income’s effect on car and vehicle ownership, worldwide 1960-2015,” Transp Res, vol. 33, no. 2, 1990. [37] Government of Canada, “Statistics Canada: Canada’s national statistical agency / Statistique Canada : Organisme statistique national du Canada.” [Online]. Available: https://www.statcan.gc.ca/. [Accessed: 13-Jul-2018]. [38] N. Johnson and J. Ogden, “Conceptual design of optimized fossil energy systems with capture and sequestration of carbon dioxide,” Pittsburgh, PA, and Morgantown, WV (United States), 2010. [39] M. Melendez and A. Milbrandt, “Regional consumer hydrogen demand and optimal hydrogen refueling station siting,” 2008. [40] National Research Council (NRC), “The hydrogen economy opportunities, costs, barriers, and R&D needs,” Natl. Acad. Press, p. 240, 2004. [41] D. Hart, A. Bauen, R. Fouquet, M. Leach, P. Pearson, and D. Anderson, “Hydrogen supply for SPFC vehicles,” Imp. Coll. London, pp. 1–69, 2000. [42] D. Hart, M. Leach, R. Fouquet, P. Pearson, and A. Bauen, “Methanol infrastructure - will it affect the introduction of SPFC vehicles?,” J. Power 66  Sources, vol. 86, pp. 542–547, 2000. [43] J. C. Mendez, “Modeling market adoption in Excel with a simplified s-curve,” https://archive.is/H2Zg3, 2007. . [44] A. Almansoori and N. Shah, “Design and operation of a stochastic hydrogen supply chain network under demand uncertainty,” Int. J. Hydrogen Energy, vol. 37, no. 5, pp. 3965–3977, 2012. [45] “Hydrogen | H2energy Now.” [Online]. Available: https://h2energynow.com/wordpress/news2/. [46] Government of Canada, “British Columbia low carbon fuels compliance pathway assessment,” Pathw. Assess. 2107 Rep., 2017. [47] “British Columbia Map.” [Online]. Available: https://yourfreetemplates.com/british-columbia-map-templates. [48] “Government of Canada.” [Online]. Available: https://www2.gov.bc.ca/gov/content/home. [49] “Greenhouse Gas Industrial Reporting and Control Act.” [Online]. Available: http://www.bclaws.ca/civix/document/id/lc/statreg/250_2015. [50] “Government of Canada.” [Online]. Available: https://www2.gov.bc.ca/gov/content/environment/climate-change/planning-and-action/legislation. [51] “Canadian hydrogen and fuel cell association.” [Online]. Available: http://www.chfca.ca/. 67  [52] “Vancouver municipialities.” [Online]. Available: http://www.writeopinions.com/metro-vancouver. [53] D. Sperling and S. Yeh, “An assessment of the near-term costs of hydrogen refueling stations and station components,” Inst. Transp. Stud., vol. Issues in, no. 530, pp. 57–66, 2009.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0378353/manifest

Comment

Related Items