The sensor coil acts as a sensor with a small loop. As shown in Fig. 2, sensor coils are also affected by Faraday law, like a load coil, because they form a loop. The sensor coils for WPT systems have previously been proposed in other papers [10–15]. In one paper [10], the author proposed a method to detect lateral misalignment of the load coil using two sensor coils. The direction and magnitude of the misalignment were detected through the voltage difference induced on two sensor coils wound ahead of the load coil. In a paper published a few years later, a single sensor coil was used to detect lateral misalignment of the load coil [11]. In addition, unlike the previous studies, the sensor coils were also wound around the source coil rather than the load coil to confirm the location of a train [12]. The location of the train was determined by the voltage change that occurred in the sensor coil when the load coil attached to the train passed the source coil.

The main difference from the use of sensor coils in previous studies is the shape of the sensor coil and its role. In previous studies, the sensor coils were wound in the load coil or source coil, meaning that only a small amount of magnetic flux passes through the sensor coils. Therefore, in another paper [13], the author used an amplifier to check the induced voltage on the sensor coil. However, as shown in Fig. 3(a), the proposed system uses attached sensor coils with planes perpendicular to the magnetic flux to increase the induced voltage on the sensor coil. When comparing the previous sensor coil and the proposed sensor coil in a situation with an identical amount of inductance and same number of turns in a simulation, it was confirmed that the mutual inductance with the source coil was 60 times larger compared with when no misalignment occurred and at least 40 times larger when a misalignment occurred. In addition, the proposed sensor coil has the advantage of simply being attached to the existing load coil. In the proposed system, two sensor coils are attached under the load coil to detect misalignment and to receive information.

As mentioned in the introduction, a misalignment between the source coil and the load coil reduces the power transfer efficiency and causes EMF problems. Misalignment does not occur when a train travels along the rail, such as in a railway system, but misalignment must be considered when a vehicle travels along a road. Therefore, the proposed system can be applied to the dynamic WPT system of an EV. When no misalignment occurs, the sensor coil is symmetrically attached under the load coil, resulting in induced voltage at an identical magnitude. However, when a misalignment occurs to the left or right, the source coil and sensor coil are not symmetrical, resulting in induced voltage at different magnitudes on the two sensor coils. Therefore, the induced voltage on each sensor coil can be used to detect a misalignment of the vehicle in motion.

In this paper, sensor coils are used for information reception as well as for misalignment detection. The voltage induced in the sensor coils can be used to receive information from the source coil part to the load coil part. The amount of magnetic flux reaching the sensor coils can be adjusted using ferrite bars, which will be described in detail later in the paper. A small ferrite bar exists between the source coil and sensor coil and it reduces the magnetic flux transferred to the sensor coil, thereby reducing the induced voltage on the sensor coil. The number of ferrite bars determines the number of induced voltage reductions on the sensor coil, which can be used to transfer information. When applied to EVs, information about the lane in which the vehicle is currently being driven and about the driving environment can be transferred.

Voltage of the load coil and sensor coil is induced when the magnetic flux Φ from the source coil passes through the area of the load coil and sensor coil. The magnetic flux generated from the source coil can be determined as shown in Fig. 3, as follows:

Here, ℱ is the magnetomotive force, an ℛ_air and ℛ_ferrite are the reluctance values of air and ferrite, respectively. In addition, the reluctance values of air and ferrite can be obtained as follows:

In these equations, l_air is the air gap, l_ferrite is the height of the ferrite, μ₀ and μ₀μ_r are correspondingly the permeability of air and ferrite, and S_air and S_ferrite are the cross-section area perpendicular to the flux of the air and the ferrite, respectively. According to (1)–(3), the two magnetic fluxes Φ₁ and Φ₂ are equal in the absence of a ferrite bar, resulting in the same magnitude of the voltage on the two sensor coils. However, if a ferrite bar exists between the source coil and sensor coil, the magnetic flux moves towards the ferrite bar with a low level of reluctance, as shown in Fig. 3(b) and Fig. 4(b). A ferrite bar can be used simply by positioning it in a system with a W-shaped source coil ferrite component [16]. Owing to the ferrite bar, the amount of magnetic flux passing through the sensor coil located above the ferrite bar decreases, which also reduces the induced voltage on the sensor coil. Because the width of the ferrite bar is as small as the size of the sensor coil, the voltage induced in the load coil decreases very little compared with that in the sensor coil. The equation below confirms that a small decrease in the induced voltage of the load coil does not cause a significant change in the mutual inductance between the source coil and load coil:

Here, ω is the operating frequency, M_s–l is the mutual inductance between the source coil and load coil, and I_source is the current of the source coil. Also, from the following WPT system efficiency equation, it can be seen that small changes in the mutual inductance do not significantly change power transfer efficiency:

In this equation, R_source and R_load are the resistance values of the source coil and load coil, respectively. The efficiency does not change significantly because the decrease in the mutual inductance M_s–l is small. In other words, even if a small ferrite bar exists between the source coil and sensor coil for information transfer purpose, there is no significant change in the mutual inductance between the source coil and load coil, meaning that there is no significant decrease in power transfer efficiency. However, if it takes a long time to pass through the ferrite bar due to the slow speed of the moving load coil, a reduction in the efficiency can be further exasperated compared with when it quickly passes through.

A ferrite position identification (FPID) system, which uses ferrite to identify locations, was proposed in a previous study [13]. The system used FPID blocks to determine the location of a running train. Using the FPID block, this system changes the reluctance ℛ of (1)–(3) to generate higher magnetic flux Φ and higher induced voltage on the sensor coil. However, the system can generate higher magnetic flux Φ precisely to the target location (sensor coil) only in systems where no misalignment occurs, such as a train. In systems where misalignment occurs, such as a vehicle, the increased magnetic flux Φ does not reach the sensor coil, making the system inoperable. Furthermore, according to the EV wireless charging regulation SAE-J2954, there is no space to position FPID blocks due to large number of turns per layer of the source coil, and even if located, there is a limit to reducing the reluctance ℛ. However, the ferrite bar system proposed in this paper can play not only a role in information transfer but also a role in detecting misalignment even in systems where misalignment occurs, such as in vehicles. Furthermore, the proposed system simply positions a ferrite bar on top, meaning that even if the source coil for dynamic charging in the future has a large number of turns, as in the current static charging model of SAE-J2954, it can still be used.

The proposed system can play two roles. In sections where no ferrite bars exist, as shown in “section 1” in Fig. 5, sensor coils detect misalignment of a moving vehicle. When no misalignment occurs, voltage of the same magnitude is induced in the two sensor coils such that there is no significant difference. However, the magnitude of the voltage induced in the two sensor coils differs when misalignment occurs. The difference in the induced voltage in the two sensor coils can be used to determine not only whether a misalignment has occurred but also the direction of the misalignment. That is, as shown in Fig. 6, when the difference between the two induced voltages is zero, there is no misalignment, and the sign (+ or −) of the difference between the two induced voltages indicates the direction of the misalignment. Furthermore, the higher the voltage difference, the greater the misalignment from the center.

In sections where ferrite bars exist, indicated as “section 2” in Fig. 5, sensor coils receive information. As shown in Fig. 3(b) and Fig. 4(b), the magnetic flux moves to a ferrite bar with a low level of reluctance, hence reducing the magnetic flux reaching the sensor coil. As a result, the voltage induced in the sensor coil decreases when the sensor coil passes over the ferrite bar, as shown in Fig. 7(a). As shown in Fig. 5, when the load coil and sensor coils pass section 1 and enter section 2, and sensor coil No. 1 passes over the ferrite bar. Subsequently, sensor coil No. 2 passes over two ferrite bars, and section 2 ends as sensor coil No. 1 passes over the ferrite bar. As a result, as shown in Fig. 7(b), the magnitude of induced voltage on sensor coil No. 1 decreases first, after which the magnitude of induced voltage on sensor coil No. 2 decreases two times, with the magnitude of induced voltage on sensor coil No. 1 then decreasing again. The number of induced voltage reductions serves as important information. For example, if information is given to the vehicle about the lane in which the vehicle is being driven, the result in Fig. 7 can be interpreted as information indicating that the vehicle is currently driving in the second lane, because there are two induced voltage reductions in sensor coil No. 2 between the induced voltage reductions in sensor coil No. 1. If there is one ferrite bar passing through sensor coil No. 2, there will be one induced voltage reduction, which can be interpreted as the vehicle being in the first lane, whereas it is interpreted as being in the third lane if there are three ferrite bars. In addition to vehicle lane information, the induced voltage reduction can be recognized as a 0 (zero), allowing various forms of information, such as the current location, to be delivered in binary form.

In addition, even if the vehicles enter section 2 when misalignment is occurring, information is received. Due to the occurrence of the misalignment, only the reference voltage level that checks for a decrease in the voltage changes, with the reference voltage level determined based on the induced voltage in the sensor coil at the entry point.

Simulations were conducted to identify the two roles (misalignment detection and information transfer) of the proposed sensor coil system. The simulations were conducted using ANSYS Maxwell 3D, and the size of the load coil was designed to match the WPT3/Z2 dimension in accordance with the EV wireless charging regulation SAE-J2954. The dimensions of the air gap and operating frequency were also determined in accordance with SAE-J2954, and the dimensions of the source coil was designed only according to the width and thickness of the static charging source coil model in the SAE-J2954 regulation, because there is no model of dynamic charging. The dimensions and electrical parameters of the source coil, load coil, sensor coil, and ferrite bar are shown in Fig. 8 and Table 1.

The simulation is shown in Fig. 8(b), where the load coil and sensor coils pass through section 1, section 2, and section 1 again, in that order. Because the ferrite bar does not exist in section 1, the sensor coils detect a misalignment between the load coil and source coil, and in section 2, the ferrite bars are positioned to transfer information to the sensor coil. In section 2, the ferrite bars are located at 1,640 mm and 2,840 mm on the right side of the source coil and at 2,440 mm on the left side of the source coil.

During section 1 (0–1,640 mm and 2,840–5,800 mm), simulations were conducted for no misalignment and for 20 mm and 40 mm misalignment to the left and right, respectively. As mentioned earlier, if a misalignment occurs, it can be seen that the EMF problem is exacerbated in the direction of the occurrence of the misalignment, as shown in Fig. 9. Furthermore, as shown in Table 2, the mutual inductance between the source coil and load coil is reduced, leading to a reduction in the power transfer efficiency.

Sensor coils were used to solve those problems caused by a misalignment, and when no misalignment occurred, the sensor coils attached to both sides of the load coil were induced at the same voltage magnitude, indicating that the difference between the two induced voltages was zero, as shown in Fig. 10(a). However, if a misalignment occurs, the induced voltage difference between the two sensor coils is not zero, as shown in Fig. 10(b)–10(d). A comparison between Fig. 10(b) and 10(c) confirms that the phase of the induced voltage difference is opposite depending on the direction of the misalignment. In addition, a comparison between Fig. 10(c) and 10(d) shows that the larger the misalignment distance generated from the center is, the higher the induced voltage difference becomes. In other words, the voltage induced in the two sensor coils attached under the load coil can be used to determine the occurrence of a misalignment and the direction and magnitude of the misalignment.

In section 2 (1,640–2,840 mm), simulations were conducted to position the ferrite bars at a specific location and assess the possibility of information transfer. When the sensor coil passes over the ferrite bar, the path of the magnetic flux produced from the source coil is determined, as shown in Fig. 11. Unlike the right side of Fig. 11, a ferrite bar with low reluctance is located on the left side such that the magnetic flux moves through the ferrite bar. This reduces the magnetic flux passing through the sensor coil and the induced voltage every time it passes over the ferrite bar.

In the simulation, a ferrite bar is located 1,640 mm, 2,440 mm and 2,840 mm, as shown in Fig. 8(b). Sensor coil No. 2 attached to the right bottom of the load coil will pass through the ferrite bar at 1,640 mm and 2,840 mm points, while sensor coil No. 1 attached to the left bottom of the load coil will pass through the ferrite bar at 2,440 mm. As shown in Fig. 12, when the sensor coil passes over the ferrite bar, the magnitude of the induced voltage on the sensor coil decreases. In the case of sensor coil No. 1, one ferrite bar is located at 2,440 mm, so only the induced voltage on sensor coil No. 1 is reduced at 2,440 mm. Also, with regard to sensor coil No. 2, two ferrite bars are located at 1,640 mm and 2,840 mm such that the induced voltage in sensor coil No. 2 is reduced at 1,640 mm and 2,840 mm. Even when misalignment occurs, the reference voltage level changes, as shown in Fig. 12(c), but it can be seen that the induced voltage reduction for an information transfer still occurs. If the proposed system is used for lane recognition, this system will transfer information indicating that the vehicle’s driving lane is the first lane because the induced voltage reduction in sensor coil No. 1 occurred once between the induced voltage reductions in sensor coil No. 2. In addition, as the load coil passes over the ferrite bar, the mutual inductance between the source coil and load coil decrease by 3%, resulting in a 0.1% reduction in efficiency. Given these findings, it can be confirmed that information can be delivered without significant reduction in the efficiency.

Experiments were conducted to verify the operation of the proposed sensor coil system. The dimensions and electrical parameters of the source coil, load coil, sensor coil, air gap, and operating frequency used in the experiment are shown in Fig. 13 and Table 3. The experiments were conducted by downscaling to about a 1:2 ratio compared with the simulation model in Fig. 8. Also, simulations for the model that were downscaled to a 1:2 ratio were carried out, and these results are presented in Fig. 14.

As in the simulations, the load coil and sensor coils pass through section 1, where the ferrite bar does not exist, and then section 2, where the ferrite bars exist. In section 2 in the experiments, as shown in Fig. 13(d), the ferrite bars are located at 70 mm and 110 mm on the right side of the source coil and at 90 mm on the left side of the source coil.

The induced voltage difference between the two sensor coils when left and right misalignment occurs in section 1 is shown in Fig. 14(a) and 14(b), respectively. As in the simulation, the phase of the induced voltage difference between the two sensor coils is the opposite depending on the direction in which the misalignment occurs. Furthermore, the magnitude of the induced voltage difference is similar in the event of a misalignment at the same distance from the center. That is, the experiments show that two small sensor coils attached under both sides of the load coil detect the direction and magnitude of the misalignment.

The magnitudes of the induced voltage in sensor coil No. 1 and sensor coil No. 2 are shown in Fig. 14(c) and 14(d), respectively. Sensor coil No. 1 attached to the bottom left of the load coil passes the ferrite bar at 90 mm, thus reducing the induced voltage at 90 mm. Also, sensor coil No. 2 attached to the bottom right of the load coil passes the ferrite bar at 70 mm and 110 mm, thus reducing the induced voltage twice at 70 mm and 110 mm. In other words, as in the simulation, the experiments showed that when the sensor coil passes over the ferrite bar, the magnetic flux passing through the sensor coil decreases, hence reducing the induced voltage.