Inductive Power Transfer Subsystem for Integrated Motor Drive

: An inductive power transfer subsystem for an integrated motor drive is presented in this paper. First, the concept of an integrated motor drive system is overviewed, and its main components are described. Next, the paper is focused on its inductive power transfer subsystem, which includes a magnetically coupled resonant circuit and two-stage energy conversion with an appropriate control method. Simpliﬁed complex domain analysis of the magnetically coupled resonant circuit is provided and the applied procedure for its component selection is explained. Furthermore, the prototype of the integrated motor drive system with its control is described. Finally, the prototype based on the gallium nitride ﬁeld effect transistors (GaN FET) inductive power transfer subsystem is experimentally tested, conﬁrming the feasibility of the concept.


Introduction
Wireless power transfer (WPT) has been an object of increased interest of researchers in recent years. Contactless energy transmission eliminates cables and connectors, thus increasing reliability. The technology is also naturally safe by providing electrical insulation. Inductive power transfer (IPT), which usually employs two magnetically coupled coils (forming a loosely coupled transformer) and compensation capacitors, is one of multiple technologies used for WPT [1][2][3]. IPT systems allow to achieve relatively long transmission distances with relatively high efficiencies. Their efficiencies depend heavily on the coupling coefficient between primary and secondary coils and their quality factors [4,5]. They can also supply multiple receivers [6]. IPT technology is a very popular solution for a wide range of applications [7][8][9], e.g., contactless charging of electric vehicles [10,11] and mobile devices [5], powering biomedical implants [4], flexible induction heating [12], or fluorescent lighting [13]. Recent advances in this regard relate to the structures of IPT systems [14][15][16], compensation circuits [15,17] and coil designs [5,18,19].
So far, little attention has been paid to the possibility of using IPT technology for motor drive systems. For example, [20] proposes a wireless in-wheel motor drive for electric vehicles to solve the problems with power cables and signal wires exposed to harsh environments. This system applies an additional controller and power converter at the motor side. In turn, in [21] a wireless bidirectional servo motor drive is described and [22] presents a wireless switched reluctance motor drive. In both cases the principles of the so-called "energy encryption" [23] and selective wireless power transfer [24] are applied. The reluctance motor drive in [22] is switched by using receiver coils of different resonant frequencies, while the servo motor drive in [21] is based on the same method for bidirectional motion. In [21] and [22], neither a controller nor a power converter at the secondary-side of the IPT system is used. In contrast to these studies, the solution presented in this paper provides a fully functional electric motor drive system, which is integrated with motor housing. Importantly, the system is powered via a contactless plug which makes it suitable for specific applications. It can easily be adapted to all types of electric motors [25]. The developed system is composed of an electric motor connected to a dedicated drive, powered wirelessly by transmitter and receiver modules-a substitute for a regular plug. Therefore, the system is fully hermetic and dedicated to harsh environments, e.g., the food and beverage (F&B) industry. The average lifetime of motors in the F&B industry is about one month [26]. The main failure reasons are bearings or windings of the motors. Motors that have been in operation for more than a month are often replaced by new ones as part of preventive maintenance [27,28]. The most popular motors are rated in the range of 1 to 2 kW and are used in pumps, conveyors, compressors, and others. The F&B industry (similarly to the chemical industry) has very high requirements for clean and reliable operation due to the continuous production process. Any unpredicted stops and delays may damage semi-products used in production and increase operating costs. The system modularity is a valuable advantage, allowing for simple installation and quick replacement of a damaged module (drive with motor). Due to the complexity of the whole system, this paper is mainly focused on presenting the general concept of the developed system and in more detail on its IPT subsystem. This IPT subsystem has a fully load-independent voltage gain. Similar solutions are analyzed in [29][30][31], but the design method of the IPT subsystem given in this paper is simplified and appropriately adapted. At the same time, additional feedback is applied to ensure the required fully load-independent voltage gain.
Section 2 provides an overview of the integrated motor drive system and its structure. The selection of magnetically coupled resonant circuit components is given in Section 3. Section 4 describes the prototype of the integrated motor drive system, presenting the applied IPT subsystem and its control in detail. The results of experimental tests of the prototype are shown in Section 5. Finally, the conclusions are drawn in Section 6.

Concept of IPT-IMD System
The proposed IPT-IMD system includes the IPT subsystem powering the integrated motor drive (IMD), which consists of a variable-frequency drive (VFD) with an electric motor (M). First, it should be clarified that the IPT and IMD subsystems can be considered individually. The IPT subsystem provides the required power to the IMD subsystem under a variable load, and its DC output voltage aligned with to the DC input voltage contains the information about the set angular velocity of the motor. On the other hand, the IMD subsystem based on this voltage is responsible for obtaining the desired angular velocity. A 3D visualization of the system is presented in Figure 1 and its schematic is depicted in Figure 2.
The applied magnetically coupled coils are designed to be identical and to fit within the plastic casings. They are single-layer and spiral, wound with litz-wire. Additionally, the coils are shielded with 3F3 ferrite and aluminum plates. The casings are built to have the same dimensions as the motor housing. The IPT-IMD system can be divided into the following sections: the input (IN) consisting of the DC supply U I with capacitor bank C I , the primary-side GaN FET inverter (INV), the magnetically coupled resonant circuit (MCR circuit), the secondary-side bridge rectifier (REC) and the output load (OUT/IMD) consisting of a DC link with capacitor bank C O and the VFD with the electric motor (M). The equivalent components for the MCR circuit are also shown on the schematic. Namely, the AC sinusoidal source u 1 on the primary-side, and the AC load resistance R OAC equivalent to the DC load resistance R ODC on the secondary-side.   The applied magnetically coupled coils are designed to be identical and to fit within the plastic casings. They are single-layer and spiral, wound with litz-wire. Additionally, the coils are shielded with 3F3 ferrite and aluminum plates. The casings are built to have the same dimensions as the motor housing. The IPT-IMD system can be divided into the following sections: the input (IN) consisting of the DC supply UI with capacitor bank CI, the primary-side GaN FET inverter (INV), the magnetically coupled resonant circuit (MCR circuit), the secondary-side bridge rectifier (REC) and the output load (OUT/IMD) consisting of a DC link with capacitor bank CO and the VFD with the electric motor (M). The equivalent components for the MCR circuit are also shown on the schematic. Namely, the AC sinusoidal source u1 on the primary-side, and the AC load resistance ROAC equivalent to the DC load resistance RODC on the secondary-side.
The angular velocity of the motor is controlled by the VFD, proportionally to the output voltage UO, which varies with the input voltage UI. In turn, the IPT subsystem is controlled to ensure the equality of UI and UO despite load variation. Therefore, the MCR    The applied magnetically coupled coils are designed to be identical and to fit within the plastic casings. They are single-layer and spiral, wound with litz-wire. Additionally, the coils are shielded with 3F3 ferrite and aluminum plates. The casings are built to have the same dimensions as the motor housing. The IPT-IMD system can be divided into the following sections: the input (IN) consisting of the DC supply UI with capacitor bank CI, the primary-side GaN FET inverter (INV), the magnetically coupled resonant circuit (MCR circuit), the secondary-side bridge rectifier (REC) and the output load (OUT/IMD) consisting of a DC link with capacitor bank CO and the VFD with the electric motor (M). The equivalent components for the MCR circuit are also shown on the schematic. Namely, the AC sinusoidal source u1 on the primary-side, and the AC load resistance ROAC equivalent to the DC load resistance RODC on the secondary-side.
The angular velocity of the motor is controlled by the VFD, proportionally to the output voltage UO, which varies with the input voltage UI. In turn, the IPT subsystem is controlled to ensure the equality of UI and UO despite load variation. Therefore, the MCR The angular velocity of the motor is controlled by the VFD, proportionally to the output voltage U O , which varies with the input voltage U I . In turn, the IPT subsystem is controlled to ensure the equality of U I and U O despite load variation. Therefore, the MCR circuit is designed accordingly, to make the output voltage stable in a wide range of load conditions and to be approximately 10% higher than the input voltage in case of openloop operation. Additionally, an optical link is used as feedback information about the output voltage U O for the IPT subsystem control in case of closed-loop operation. Further considerations are made under the assumption that the primary-side inverter switching frequency is 150 kHz and the input and output voltages should change in the range of 100 to 300 V, which relates to the full range of angular velocity of the considered motor (motor parameters are given in Section 4.3). The selected switching frequency ensures feasibility and acceptable power losses of the IPT-IMD system. It has a direct impact on the parameters of the magnetically coupled coils, which should be flat and properly fill the An additional component of the system is a set of magnets placed in the corners of both casings. In the event of a failure on either side of the system, it can be quickly plugged or unplugged without having to use any wires.

MCR Circuit Component Selection
A schematic of the MCR circuit is shown in Figure 3. It is assumed that passive components of the circuit are linear, invariant in time and ideal, except for the parasitic resistances of the coils. Overall, two series-series compensation capacitors are included in the circuit. The analysis is carried out using fundamental harmonic approximation in the complex domain-sinusoidal voltages and currents. The primary-side inverter is replaced by a voltage source U 1 (1), while the rectifier is substituted by an equivalent resistance R OAC (2)- Figure 2.
circuit is designed accordingly, to make the output voltage stable in a wide range of load conditions and to be approximately 10% higher than the input voltage in case of openloop operation. Additionally, an optical link is used as feedback information about the output voltage UO for the IPT subsystem control in case of closed-loop operation. Further considerations are made under the assumption that the primary-side inverter switching frequency is 150 kHz and the input and output voltages should change in the range of 100 to 300 V, which relates to the full range of angular velocity of the considered motor (motor parameters are given in Section 4.3). The selected switching frequency ensures feasibility and acceptable power losses of the IPT-IMD system. It has a direct impact on the parameters of the magnetically coupled coils, which should be flat and properly fill the plastic casings. Hence, their relatively high quality factors and strong magnetic coupling are obtained. An additional component of the system is a set of magnets placed in the corners of both casings. In the event of a failure on either side of the system, it can be quickly plugged or unplugged without having to use any wires.

MCR Circuit Component Selection
A schematic of the MCR circuit is shown in Figure 3. It is assumed that passive components of the circuit are linear, invariant in time and ideal, except for the parasitic resistances of the coils. Overall, two series-series compensation capacitors are included in the circuit. The analysis is carried out using fundamental harmonic approximation in the complex domain-sinusoidal voltages and currents. The primary-side inverter is replaced by a voltage source U1 (1), while the rectifier is substituted by an equivalent resistance ROAC (2)- Figure 2. (1) U1 and U2 stand for root mean square (RMS) values of the input and output voltages.
The MCR circuit is described by the following equations: where Rp1 and Rp2 are the parasitic resistances of the coils. The mutual reactance XM is expressed by the coils' self-reactance values XL1, XL2 and the coupling coefficient k. Previous works [29][30][31] show that in the analyzed series-series compensated circuit it is possible to obtain the output voltage U2, which is nearly invariant with changes in the load resistance ROAC and proportional to the input voltage U1. Complete invariance can The MCR circuit is described by the following equations: where R p1 and R p2 are the parasitic resistances of the coils. The mutual reactance X M is expressed by the coils' self-reactance values X L1 , X L2 and the coupling coefficient k.
Previous works [29][30][31] show that in the analyzed series-series compensated circuit it is possible to obtain the output voltage U 2 , which is nearly invariant with changes in the load resistance R OAC and proportional to the input voltage U 1 . Complete invariance can only be obtained if the circuit is lossless (R p1 = R p2 = 0). If the coils' quality factors Q 1 and Q 2 (4) are high, the output voltage does not change much. Additionally, if the coils are strongly coupled (k > 0.8- [32]) the efficiency is also high.
Further analysis is done for the special case of the IPT subsystem-the output voltage U O should be approximately 10% higher than the input voltage U I in open-loop operation and the coils should be roughly similar. Initially, the analysis is simplified by assuming the circuit to be lossless (R p1 = R p2 = 0). By applying Thevenin's theorem to the MCR circuit, the equivalent schematic shown in Figure 4 is obtained.
Further analysis is done for the special case of the IPT subsystem-the output voltage UO should be approximately 10% higher than the input voltage UI in open-loop operation and the coils should be roughly similar. Initially, the analysis is simplified by assuming the circuit to be lossless (Rp1 = Rp2 = 0). By applying Thevenin's theorem to the MCR circuit, the equivalent schematic shown in Figure 4 is obtained.
The equivalent circuit is described by the following equations: It is obvious that the output voltage U2 is independent of the load resistance ROAC (U2 = E) when: Condition (7) is met, among others, for: The output to input voltage ratio UO/UI is equal to the ratio of U2 to U1 (1). By using Equations (5), (7), and (8), it is obtained: Therefore, if the coupling coefficient k is in the order of 0.9 and the coils' self-reactance XL1 and XL2 are similar, the output voltage will be approximately 10% higher than the input voltage.
Equation (8) will be used for the design process of the MCR circuit. In this case the input impedance of the circuit is: and remains inductive for any resistance value. This is beneficial due to the use of a resonant voltage source inverter ( Figure 2). Considering the non-zero parasitic resistances of the coils, the efficiency of the MCR circuit can be estimated using the following equation: The equivalent circuit is described by the following equations: It is obvious that the output voltage U 2 is independent of the load resistance R OAC (U 2 = E) when: Condition (7) is met, among others, for: The output to input voltage ratio U O /U I is equal to the ratio of U 2 to U 1 (1). By using Equations (5), (7), and (8), it is obtained: Therefore, if the coupling coefficient k is in the order of 0.9 and the coils' self-reactance X L1 and X L2 are similar, the output voltage will be approximately 10% higher than the input voltage.
Equation (8) will be used for the design process of the MCR circuit. In this case the input impedance of the circuit is: and remains inductive for any resistance value. This is beneficial due to the use of a resonant voltage source inverter ( Figure 2). Considering the non-zero parasitic resistances of the coils, the efficiency of the MCR circuit can be estimated using the following equation: where ∆P is equal to power losses in the coils, P 1AC is equal to the input power of the circuit, I 1 and I 2 are RMS values of input and output currents. The output to input voltage ratio is given as:

MCR Circuit
The MCR circuit has been built based on the analysis results presented in Section 3. The single-layer and spiral coils (outside diameter: 83 mm, inside diameter: 35 mm) were wound with 24 turns of litz wire (Figure 5a). They are well fitted to the plastic casings. 3F3 ferrite plates (PLT38, E series-25 × 38 × 3.8 mm 3 ) were used to shield and increase both the self-and mutual inductances of the coils. The magnetic shields were placed outside of the coils (Figure 5b). Additionally, aluminum plates were used on the outside to improve the shielding effect and at the same time provide more effective cooling (Figure 5c). The aluminum plates were cut accordingly to avoid additional power losses due to eddy currents.
where ΔP is equal to power losses in the coils, P1AC is equal to the input power of the circuit, I1 and I2 are RMS values of input and output currents. The output to input voltage ratio is given as:

MCR Circuit
The MCR circuit has been built based on the analysis results presented in Section 3. The single-layer and spiral coils (outside diameter: 83 mm, inside diameter: 35 mm) were wound with 24 turns of litz wire (Figure 5a). They are well fitted to the plastic casings. 3F3 ferrite plates (PLT38, E series-25 × 38 × 3.8 mm 3 ) were used to shield and increase both the self-and mutual inductances of the coils. The magnetic shields were placed outside of the coils (Figure 5b). Additionally, aluminum plates were used on the outside to improve the shielding effect and at the same time provide more effective cooling (Figure 5c). The aluminum plates were cut accordingly to avoid additional power losses due to eddy currents.  The coils, magnetic shields and aluminum plates were placed in the dedicated plastic casings ( Figure 5). The casings ensure a stable position of the coils and their coaxial arrangement. The distance of 12 mm between the coils results directly from the applied casings and their mounting position. The parameters of the coils were measured with an Agilent 4294A impedance analyzer at the frequency of 150 kHz. The measurement results are summarized in Table 1. Using Equation (8) and assuming the frequency (f = 150 kHz), the coupling coefficient (k = 0.87) and the primary coil inductance (L1 = 229 µH), the required primary-side compensation capacitance was calculated (C1 = 20.2 nF). Based on Equation (12) and Table 1, assuming the input voltage (e.g., UI = 300 V) and the output power (e.g., PO = 1200 W), the output voltage (UO = 330 V, UO/UI = 1.1) was iteratively calculated. This operation point corresponds to RODC = 90.8 Ω and ROAC = 73.6 Ω (2), respectively. Finally, according to Equation (11), the efficiency (η = 97%) was found.

Control Method
The control method of the IPT subsystem is explained in Figure 6. Figure 6a shows a simplified diagram of the IPT subsystem control. The control includes a proportional-integral (PI) controller that aligns the DC output voltage UO with the DC supply voltage UI. The voltage UI is measured and, using an additional rate limiter, is entered into the PI controller as a reference signal UOREF. In turn, the feedback signal (UO) is provided via an optical link. The output of the PI controller is the phase shift angle Φ between control signals of both the inverter half-bridges (pulse width modulation (PWM) control method) as explained in more detail in Figure 6b. The primary-side inverter (T1, T2, T3, T4) is controlled by a field-programmable gate array (FPGA)-based system. The switching fre-  Table 1. Using Equation (8) and assuming the frequency (f = 150 kHz), the coupling coefficient (k = 0.87) and the primary coil inductance (L 1 = 229 µH), the required primary-side compensation capacitance was calculated (C 1 = 20.2 nF). Based on Equation (12) and Table 1, assuming the input voltage (e.g., U I = 300 V) and the output power (e.g., P O = 1200 W), the output voltage (U O = 330 V, U O /U I = 1.1) was iteratively calculated. This operation point corresponds to R ODC = 90.8 Ω and R OAC = 73.6 Ω (2), respectively. Finally, according to Equation (11), the efficiency (η = 97%) was found.

Control Method
The control method of the IPT subsystem is explained in Figure 6. Figure 6a shows a simplified diagram of the IPT subsystem control. The control includes a proportionalintegral (PI) controller that aligns the DC output voltage U O with the DC supply voltage U I . The voltage U I is measured and, using an additional rate limiter, is entered into the PI controller as a reference signal U OREF . In turn, the feedback signal (U O ) is provided via an optical link. The output of the PI controller is the phase shift angle Φ between control signals of both the inverter half-bridges (pulse width modulation (PWM) control method) as explained in more detail in Figure 6b. The primary-side inverter (T 1 , T 2 , T 3 , T 4 ) is controlled by a field-programmable gate array (FPGA)-based system. The switching frequency f is set to 150 kHz. The inverter is inductively loaded, thus hard commutations are avoided. The VFD microprocessor system controls the optical link transmitter according to the PWM method. The signal has a fixed frequency of 20 kHz and the duty cycle carries the information about the output voltage U O . The optical link is placed in the center of the coils, within the primary and secondary casings. The VFD integrated on the secondary-side A soft-start of the system is required to safely pre-charge the capacitors CO on the VFD. This is done when the connected IPT-IMD system is started for the first time, or after the primary coil is put in its place (as for a regular plug). The pre-charge process is realized by limiting the rate of the reference signal UOREF at the input of the PI controller. An overcurrent protection, which turns off the inverter when its load current threshold is exceeded, is also implemented. This can occur, for example, when the motor is overloaded or when the coils are accidentally decoupled. The settings of the PI controller are not critical as the response time is also deliberately limited by the rate limiter during normal operation. It is an order of magnitude faster than for the system soft-start. This solution has been applied due to the relatively large capacitance of the rectifier filter to avoid unnecessary overloading of the IPT system. A soft-start of the system is required to safely pre-charge the capacitors C O on the VFD. This is done when the connected IPT-IMD system is started for the first time, or after the primary coil is put in its place (as for a regular plug). The pre-charge process is realized by limiting the rate of the reference signal U OREF at the input of the PI controller. An over-current protection, which turns off the inverter when its load current threshold is exceeded, is also implemented. This can occur, for example, when the motor is overloaded or when the coils are accidentally decoupled. The settings of the PI controller are not critical as the response time is also deliberately limited by the rate limiter during normal operation. It is an order of magnitude faster than for the system soft-start. This solution has been applied due to the relatively large capacitance of the rectifier filter to avoid unnecessary overloading of the IPT system.

IPT-IMD System
The IPT-IMD system prototype is depicted in photographs in Figure 7. Figure 7a presents its primary-side: the inverter (INV), the resonant capacitor (C 1 ), the primary casing (PC) with the shielded coil and the receiver-side of the optical link (OL). The inverter uses TP65H035WS cascode GaN FETs as power semiconductors. The inverter control system is based on an Artix-7 FPGA (100 MHz, 32 MB). The optical link is placed outside the casings for measurement purposes. Figure 7b presents the secondary-side of the prototype: the electric motor (M) with the VFD and the capacitors C O , the secondary casing (SC) with the shielded coil, the rectifier (REC) and the transceiver-side of the optical link (OL). The rectifier printed circuit board (PCB) is placed outside the casing for measurement purposes only. It is actually mounted on the aluminum plate inside the casing. The main system parameters are summarized in Table 2.
casings for measurement purposes. Figure 7b presents the secondary-side of the prototype: the electric motor (M) with the VFD and the capacitors CO, the secondary casing (SC) with the shielded coil, the rectifier (REC) and the transceiver-side of the optical link (OL). The rectifier printed circuit board (PCB) is placed outside the casing for measurement purposes only. It is actually mounted on the aluminum plate inside the casing. The main system parameters are summarized in Table 2.

Experimental Results
The measurements were carried out in a laboratory setup corresponding to the schematic shown in Figure 2. The main aim of the experimental tests focused on the IPT subsystem was to determine its AC/AC and DC/DC efficiencies and the dependence of its output voltage UO on the output power PO. As this subsystem can be considered individually for much easier and more comprehensive testing, its DC output was loaded with a laboratory resistor RODC. This equivalent resistor replaces the IMD subsystem. Additionally, the VFD was also connected to the DC output, but only its input capacitor bank was used. All measurements were conducted using a Yokogawa WT5000 power analyzer. The following quantities were recorded to illustrate the system performance ( Figure 2):

Experimental Results
The measurements were carried out in a laboratory setup corresponding to the schematic shown in Figure 2. The main aim of the experimental tests focused on the IPT subsystem was to determine its AC/AC and DC/DC efficiencies and the dependence of its output voltage U O on the output power P O . As this subsystem can be considered individually for much easier and more comprehensive testing, its DC output was loaded with a laboratory resistor R ODC . This equivalent resistor replaces the IMD subsystem.
Additionally, the VFD was also connected to the DC output, but only its input capacitor bank was used. All measurements were conducted using a Yokogawa WT5000 power analyzer. The following quantities were recorded to illustrate the system performance ( Figure 2 The PWM control of the primary-side inverter is responsible for this efficiency reduction. In simple terms, the conduction time of the transistors is shorter and the inverter zero output voltage interval, where energy is not converted but power losses are generated, is longer. The depicted efficiency curves have a typical shape for IPT systems [28]. The output power changes are caused by load changes, resulting in energy conversion for a different voltage-current relationship. For the input and output voltages of 300 V and the DC output power of 1200 W in closed-loop operation, the distribution of power losses is as follows: 12 W in the inverter (including 4 W for its control), 24 W in the MCR circuit, and 12 W in the rectifier. The AC/AC efficiency was also calculated analytically using Equations (11) and (12) for the output power P O in the range of 200 to 1200 W at input voltage U I equal to 300 V. These theoretical results were in the range of 97-98%, corresponding to the measurements.  The PWM control of the primary-side inverter is responsible for this efficiency reduction. In simple terms, the conduction time of the transistors is shorter and the inverter zero output voltage interval, where energy is not converted but power losses are generated, is longer. The depicted efficiency curves have a typical shape for IPT systems [28]. The output power changes are caused by load changes, resulting in energy conversion for a different voltage-current relationship. For the input and output voltages of 300 V and the DC output power of 1200 W in closed-loop operation, the distribution of power losses is as follows: 12 W in the inverter (including 4 W for its control), 24 W in the MCR circuit, and 12 W in the rectifier. The AC/AC efficiency was also calculated analytically using Equations (11) and (12) for the output power PO in the range of 200 to 1200 W at input voltage UI equal to 300 V. These theoretical results were in the range of 97-98%, corresponding to the measurements. The output voltage measurements under analogous experimental conditions are shown in Figure 9. Both open-loop (UO_O) and closed-loop (UO_C) operations are included. As predicted by the MCR circuit analysis, UO_O is approximately 10% higher than UO_C. The output voltage UO was also calculated using Equation (12). These theoretical characteristics are within 1% of the measurements (Figure 9). The output voltage measurements under analogous experimental conditions are shown in Figure 9. Both open-loop (U O_O ) and closed-loop (U O_C ) operations are included. As predicted by the MCR circuit analysis, U O_O is approximately 10% higher than U O_C . The output voltage U O was also calculated using Equation (12). These theoretical characteristics are within 1% of the measurements (Figure 9).  Figure 9. Output voltage as a function of output power (UI = 300 and 150 V). Figure 10 presents the efficiency measurements as a function of the input voltage, at a constant load resistance (RODC = 72 Ω) in closed-loop operation. The measurements were recorded for the input voltage UI in the range of 100-300 V. In this case, the AC/AC efficiency is independent of the input voltage. In turn, the DC/DC efficiency varies due to the use of PWM control.  Figure 11 shows the preliminary experimental results for the whole IPT-IMD system prototype-a graph of the motor angular velocity (no load condition) as a function of the input voltage in closed-loop operation. According to the adopted assumptions, this relationship is linear. Detailed experimental studies of the IMD subsystem and the IPT-IMD system prototype will be presented in a separately prepared paper.      Figure 11 shows the preliminary experimental results for the whole IPT-IMD system prototype-a graph of the motor angular velocity (no load condition) as a function of the input voltage in closed-loop operation. According to the adopted assumptions, this relationship is linear. Detailed experimental studies of the IMD subsystem and the IPT-IMD system prototype will be presented in a separately prepared paper.   Figure 11 shows the preliminary experimental results for the whole IPT-IMD system prototype-a graph of the motor angular velocity (no load condition) as a function of the input voltage in closed-loop operation. According to the adopted assumptions, this relationship is linear. Detailed experimental studies of the IMD subsystem and the IPT-IMD system prototype will be presented in a separately prepared paper.  Figure 9. Output voltage as a function of output power (UI = 300 and 150 V). Figure 10 presents the efficiency measurements as a function of the input voltage, at a constant load resistance (RODC = 72 Ω) in closed-loop operation. The measurements were recorded for the input voltage UI in the range of 100-300 V. In this case, the AC/AC efficiency is independent of the input voltage. In turn, the DC/DC efficiency varies due to the use of PWM control.  Figure 11 shows the preliminary experimental results for the whole IPT-IMD system prototype-a graph of the motor angular velocity (no load condition) as a function of the input voltage in closed-loop operation. According to the adopted assumptions, this relationship is linear. Detailed experimental studies of the IMD subsystem and the IPT-IMD system prototype will be presented in a separately prepared paper.  Representative voltage and current waveforms of the primary-side inverter and the secondary-side rectifier of the IPT subsystem in the closed loop case are presented in Figure 12. For the tested voltages and powers, the phase shift angle Φ measured with an oscilloscope changed from 37 • to 51 • , which corresponded to the duty factor D (D = 1 − Φ/180 • ) in the range of 0.79 to 0.72. These changes are minor due to the applied design method of the MCR circuit-natural load-independent voltage gain. Representative voltage and current waveforms of the primary-side inverter and the secondary-side rectifier of the IPT subsystem in the closed loop case are presented in Figure 12. For the tested voltages and powers, the phase shift angle Φ measured with an oscilloscope changed from 37° to 51°, which corresponded to the duty factor D (D = 1 − Φ/180°) in the range of 0.79 to 0.72. These changes are minor due to the applied design method of the MCR circuit-natural load-independent voltage gain.  Table 3 summarizes the performance of three IPT systems, two representative ones described in [29,31] and the third one presented in this paper. Their geometric parameters are similar. The measured coupling coefficients, coils' self-inductances, switching frequencies, output voltages, and maximum output powers and maximum efficiencies are tabulated. Unfortunately, the efficiencies in [25] are omitted. The IPT subsystem developed in this paper is compact and integrated with motor housing. Its performance are comparable with others, also in closed-loop operation (* in Table 3).

Conclusions
The concept of an integrated motor drive has been presented and examined. The developed system is safe, wirelessly powered ("wireless plug"), and allows the angular velocity of the motor to be changed by setting the input voltage. The system is more complex  Table 3 summarizes the performance of three IPT systems, two representative ones described in [29,31] and the third one presented in this paper. Their geometric parameters are similar. The measured coupling coefficients, coils' self-inductances, switching frequencies, output voltages, and maximum output powers and maximum efficiencies are tabulated. Unfortunately, the efficiencies in [25] are omitted. The IPT subsystem developed in this paper is compact and integrated with motor housing. Its performance are comparable with others, also in closed-loop operation (* in Table 3).

Conclusions
The concept of an integrated motor drive has been presented and examined. The developed system is safe, wirelessly powered ("wireless plug"), and allows the angular velocity of the motor to be changed by setting the input voltage. The system is more complex than a regular motor drive and this complexity results in lower efficiency and can make the system more prone to failure. However, this can be accepted in an industrial environment with an additional benefit. This benefit is the introduction of galvanic isolation, where the IPT transmitter and receiver modules can be freely connected and disconnected by workers not fully qualified to service electrical equipment.
For the given assumptions, the simplified procedure for selecting MCR circuit parameters is explained. A compact prototype of the IPT-IMD system has been built and the IPT subsystem has been experimentally tested. The results confirm feasibility of the system. The AC/AC and DC/DC efficiencies of 96-98.6% and 93-97.5% were recorded for the input voltage of 300 and 150 V, respectively. These conditions corresponded to the output power in the ranges of 200-1200 W (300 V) and 60-600 W (150 V). The efficiencies are approximately 1 p.p. lower in closed-loop operation than in the open-loop case. This is due to the PWM control of the primary-side inverter.
Continuation of studies on the IPT-IMD system will be presented in a separate publication. At the same time, further research should be focused on the system optimization, in particular on its reliability and control methods.