**5. Discussion**

The study reconfirmed several factors, including a number of challenges that support the success and potential benefits of the DWPT system. The deployment of appropriate charging infrastructure is deemed a prerequisite for the wider uptake of EVs. Several charging infrastructure solutions are being installed across the country, such as rapid chargers, fast chargers and slow chargers [36]. They form a charging ecosystem that responds to the requirements for destination and opportunity charging. The success of the DWPT charging infrastructure will ultimately be based on the number and type of users if the initial costs are to be sufficiently amortised. The participants mentioned that the initial deployment scenarios need to maximise the usage of the technology. The commercial users who drive long distances on fixed repeatable routes are suggested as initial target users. This is supported by the literature, with Meijer [37] suggesting the deployment of DWPT systems along urban bus routes as well as short and long haul national and international freight corridors.

In order to create stakeholder acceptance of DWPT charging infrastructure, it is necessary to predict the energy demand from the EVs that potentially use the facility. From the focus group transcripts, and supported by the literature, key externalities were identified, and these externalities were classified into a taxonomy consisting of six categories:


The externalities noted above are factors that will affect the uptake of DWPT either positively or negatively, and it is the combination of these factors that is important in defining demand. Figure 2 shows the taxonomy generated from the analysis of the focusgroup transcripts and how the factors combine to determine demand. The structure of the taxonomy is derived from the analysis of the focus group discussions. Its relation to the published literature is discussed in the following section.

available on the route options for the user.

published literature is discussed in the following section.

system.

**Figure 2.** Taxonomy—linking Factors identified from focus groups to demand. **Figure 2.** Taxonomy—linking Factors identified from focus groups to demand.

The focus group participants mentioned that the EV drivers are usually sensitive to costs associated with charging, concurring with previous studies [38–42]. Two concerns arose regarding cost. One was the investment cost and the recognition that, in order to mitigate the high investment cost of DWPT being passed on to the users, supportive government policies and financial subsidies, and incentives would be essential for wider uptake of the charging infrastructure, again concurring with existing studies [41]. Second was that the focus group suggested that EV users' charging behaviour is influenced by cost-driven factors such as (a) the actual cost to charge and (b) the impact of the charging method on operational efficiency. It is recognised in the literature that the actual cost to charge includes one or more components such as access fee, a kWh-based charge, usually varying with the time of use (ToU) and payment processing fees [42]. Studies have shown that smart pricing strategies based on ToU can shift charging to the off-peak period that is beneficial to the grid and, most consumers have been willing to accept this costing method. However, there is a lack of uncertainty in the willingness to use smart charging schemes between private and commercial users. Unlike static charging solutions, DWPT allows a vehicle to charge while in motion, avoiding financial losses that may be incurred due to vehicle downtime associated with stationary charging solutions. Indeed, Oliveira et al. (2020) [24] concluded that taxi drivers are more likely to lose earnings due to charging time associated with wired solutions. Furthermore, time-sensitive services like freight and public transport systems may not have enough time to get the required energy with stationary solutions [43,44]. By reducing the vehicle downtime owing to charging stops, The focus group participants mentioned that the EV drivers are usually sensitive to costs associated with charging, concurring with previous studies [38–42]. Two concerns arose regarding cost. One was the investment cost and the recognition that, in order to mitigate the high investment cost of DWPT being passed on to the users, supportive government policies and financial subsidies, and incentives would be essential for wider uptake of the charging infrastructure, again concurring with existing studies [41]. Second was that the focus group suggested that EV users' charging behaviour is influenced by cost-driven factors such as (a) the actual cost to charge and (b) the impact of the charging method on operational efficiency. It is recognised in the literature that the actual cost to charge includes one or more components such as access fee, a kWh-based charge, usually varying with the time of use (ToU) and payment processing fees [42]. Studies have shown that smart pricing strategies based on ToU can shift charging to the off-peak period that is beneficial to the grid and, most consumers have been willing to accept this costing method. However, there is a lack of uncertainty in the willingness to use smart charging schemes between private and commercial users. Unlike static charging solutions, DWPT allows a vehicle to charge while in motion, avoiding financial losses that may be incurred due to vehicle downtime associated with stationary charging solutions. Indeed, Oliveira et al. (2020) [24] concluded that taxi drivers are more likely to lose earnings due to charging time associated with wired solutions. Furthermore, time-sensitive services like freight and public transport systems may not have enough time to get the required energy with stationary solutions [43,44]. By reducing the vehicle downtime owing to charging stops, dynamic wireless charging can be an effective solution in such scenarios [45].

V. Traffic: these externalities determine the demand upon the DWPT system that will

VI. Infrastructure: these externalities determine the broader user choice to adopt DWPT

result from external factors, e.g., the density and mix of traffic that flows across the

based on the availability of the system both geographically and capability, e.g., is it

The externalities noted above are factors that will affect the uptake of DWPT either positively or negatively, and it is the combination of these factors that is important in defining demand. Figure 2 shows the taxonomy generated from the analysis of the focusgroup transcripts and how the factors combine to determine demand. The structure of the taxonomy is derived from the analysis of the focus group discussions. Its relation to the

dynamic wireless charging can be an effective solution in such scenarios [45]. Batteries constitute a significant proportion of the EV cost [46]. DWPT offers the opportunity to reduce the battery size whilst reducing vehicle cost, increasing range, and vehicle's utility (for example, capacity). Participants were also concerned about the costs associated with installing the hardware and software that are required for DWPT charging. The participants mentioned that customers expect these costs to be recouped either with a low charging price or improved operational efficiency. A similar opinion was reported in Oliveira et al. (2020) [24].

The focus group discussions suggested that the use case for the DWPT infrastructure has interdependencies with factors relating to journeys undertaken by the users, such as origin and destination of the trip, purpose of the trip and routing preference. The EV users who generally drive fewer miles than the battery range prefer destination charging at home or work [38]. Private EV users generally fall in this category. On the other hand, long-distance and energy-intensive operations such as freight and public transport services such as buses, coaches and taxis require public charging solutions at different locations. Moreover, fixed-route services such as buses and coaches cannot detour; therefore, they rely on charging facilities along the route in addition to the infrastructure within the depots [45,47]. In general, these journey-related factors influence the willingness to alter the route for accessing a charging solution. Philipsen et al. (2015) [48] reported that users are more willing to accept a detour of 5 km or 10 min to a fast-charging station. Furthermore, Philipsen et al. (2016) [49] indicate that participants prefer to make a detour rather than accept waiting times for charging.

The participants mentioned that the design of the system needs to be interoperable, catering for a wider type of vehicle and operating conditions. However, a challenge is that there are several factors relating to vehicle characteristics as well as the technology and type of the charging system that affect the energy demand of an EV. Battery capacity is a primary factor that determines the range of an EV. A larger battery capacity enables the vehicle to travel longer distances. Therefore, when faced with a specific journey requirement, they may seek fewer charging opportunities than vehicles with smaller battery capacities. Previous research found SoC to be a key factor that determines the energy demand of an EV during a charging event. It is well recognised that specific real energy consumption that determines the SoC depends on several parameters, including battery temperature, utilisation of in-vehicle systems such as heating and ventilation, vehicle load, vehicle acceleration/deceleration, rolling friction, aerodynamic drag, and road gradient. Mishra (2018) [50] recommends operating EV batteries within a threshold range rather than taking advantage of the full range between 0% and 100%. For Li-ion batteries, the optimum SoC area is determined to be between 20% and 80%. Maintaining SoC at these levels reduces the rate of battery degradation and expands operation lifetime. Respecting the SoC range is essential; however, a driver or fleet owner may choose to operate outside this range.

A particular challenge is to develop a DWPT system capable of higher power transfer efficiencies with a wider range of lateral and vertical misalignments (air gap) between the primary source coils embedded within the road and secondary on-board coils (receiver) [51]. Naberezhnykh [52] recommends the system design to consider driver lateral misalignment of up to 15 cm for optimum usage of the charging system. The power transfer efficiency is shown to drop gradually as vertical misalignment grows [28]. Further, the air gap differs with the vehicle type, and it can vary with the loading conditions. The speed of the vehicle is expected to affect the power transfer efficiency. With an increase in the vehicle speed, the interaction time between the primary and secondary coils reduces, resulting in lower power transfer efficiency [52]. Other vehicle characteristics such as mass and vehicle length affect the energy demand of EVs. Further simulation and/or demonstrator work is required to demonstrate the relationship between the vehicle speeds and power transfer efficiencies. Heavier vehicles require more mechanical power, consuming higher energy. Vehicle type (length) determines the maximum number of vehicles in a section. It also influences the number of secondary coils (receivers) that can be fitted to the vehicle.

These findings reiterate the need for standardisation of DWPT systems and associated technologies to ensure that the deployed systems are safe, efficient and interoperable. Furthermore, standards allow manufacturers to develop and optimise their systems to the infrastructure [52]. Several organisations such as the International Organisation for Standardisation (ISO), International Electrotechnical Commission (IEC), and Society of Automotive Engineers (SAE) are currently developing standards related to DWPT, irrespective of the barriers faced due to the technical complexity and current level of maturity of the technology [53].

Studies have shown that charging behaviour is heterogeneous [54,55] among drivers. In general, factors including range anxiety, user comfort and other individual preferences contribute to charging decision making [39]. Range anxiety among EV users is a psychological barrier that can be induced by insufficient range to reach a destination or complete daily trips within a stipulated time, associated with the time required for charging and charger location. An EV driver with a higher level of range anxiety may access a charging facility when the SoC is higher than the recommended threshold. Furthermore, a higher

level of range anxiety can potentially lead to dangerous driving behaviour and negatively affect drivers' emotions. As the driver's experience with the EV increases, the driver may correctly predict the EV's range concerning their range requirements, thus reducing the range anxiety [56]. The focus group participants mentioned that an in-vehicle system capable of providing a real-time charging recommendation for the driver could be helpful to overcome this challenge [57].

The participants raised issues concerning willingness to adapt driving style to suit the traffic and improve charging efficiency. The alterations to normal driving behaviour may cause inconvenience to drivers and affect their charging decision making. For example, whilst accessing the charging lane, the driver may not be able to overtake other vehicles and need to drive within a specific speed range. In addition, the driver may need to maintain distance with the leading vehicle (headway) and vehicle alignment, which can be challenging even for an experienced driver [37]. Yang and Lu (2018) [56] acknowledged that it is necessary to account for these behavioural and psychological factors for a successful mass adaptation of EV infrastructure. In those vehicles with advanced driver-assist or automated driving systems, the vehicle's alignment can be controlled by electronic systems, thus improving driver comfort. Further measures such as driver training, road markings, and lane guides can improve charging efficiency and user comfort [42]. However, further user-based research is expected to yield satisfactory measures for users to choose DWPT charging mode.

These categories and factors within the categories can be used to support the development of DWPT. Based on the above discussion, it is recognised that there will be a logical flow that will determine if a charge event takes place, the power demand and the energy that is transferred. The gateway will be the vehicle condition; the capability of the vehicle—if it is equipped with a receiver and the ability to store energy or consume energy—will determine a request to charge and the power transferred. Following this, the journey or mission of the vehicle will determine the probability of a charge event—the longer the journey, the higher energy consumption, etc., the greater the probability of a charge request. This probability will be moderated to an extent by the immediate cost– benefit of charging dynamically or statically. As described in the preceding discussion, if a charge event is required, then the choice of how to meet that charge event—static or dynamic—is a complex interplay between access costs, energy cost, monetarisation of time, etc. The behaviour of the user and the traffic environment then determines the transit time—a quicker transit time means less energy transferred, but further, a transit event that is permeated by stop-start traffic will also impact upon energy transfer. Finally, the availability of infrastructure, either geographically—leading to different vehicle routing—or having a cap based on power supply—leading to a limit on vehicles that can be serviced—will impact power levels and energy transfer.

The identified parameters, those that determine the behaviours outlined above, can be further used to estimate demand on the network using simulation modelling techniques, such as the Agent-based model (ABM) or Bayesian network model. An Agent-based model (ABM) is a class of simulation in which a system is modelled as a collection of autonomous decision-making entities called agents that interact with each other, allowing exploration of emerging behaviour of the system, usually difficult to predict in the real world [58]. Bayesian network models are structured based on Bayes' theorem, capable of updating the prior probability of some unknown variable when some evidence describing that variable exists [59].
