Abstract: In the Inductive Power Transfer (IPT) common energy launch platform for kitchen appliances, a low efficiency will appear under some light loads. In order to make the designed IPT system achieve the high efficiency under all different power capacity levels, the main losses of selected IPT system have been analyzed and then the efficiency calculation model is presented. Then, the control strategy of efficiency optimization as well as the segmented optimal frequency dynamic tracking method has been proposed. Finally, experimental results show the overall efficiency of IPT system for kitchen appliances using the proposed control strategy is increased significantly under all light loads. Meanwhile, the reliability and stability of system under heavy loads aren’t declined.
INTRODUCTION
The convenience and safety of power supply for kitchen appliances has been widely concerned. The traditional power supply way of kitchen appliances may easily result in electrical safety issues such as loose-contact, electric-spark and short-circuit. Therefore, a more safe, convenient and reliable electricity access solution is needed urgently in the field of kitchen appliances. Fortunately, a wireless power transmission solution based on the inductive power transfer technology has emerged as the times require.
Inductive power transfer technology is a practical and flexible technology for delivering power efficiently from a stationary power supply to one or more movable loads. Such safe and reliable technology overcomes disadvantages of the traditional power transfer methods and is developed rapidly nowadays (Graham et al., 2011; Sun et al., 2012; Zhong et al., 2011), moreover, it has been widely used where the electrical isolation is essential for power supplies (Zhang et al., 2013; Tian et al., 2012; Li et al., 2012; Kissin et al., 2011). But in the field of kitchen appliances, there have been only few application products so far.
Kitchen appliances are known for their variety of types and differences in the power capacity and character of loads. Therefore, the stability of the energy transfer magnetic field and power transfer capability, the constant current in primary winding and the constant output voltage across loads generally must be required in the common energy launch platform of IPT system (Sun et al., 2013). As a result, a very low efficiency will appear under some light loads. Therefore, it is very important to analyze and optimize the efficiency under all different power capacity kitchen appliances. In this way, the efficiency can reach to the target value according to the requirements of consumers.
A Transcutaneous Energy Transfer (TET) system has been researched by Chen et al. (2007). The quantitative analysis and comparison of parameter affecting efficiency have been discussed from the aspects of conduction losses, switching losses and core losses. But there isn’t a detailed analysis about how to optimize the parameters of IPT system so as to make the efficiency reach to the target value.
There is a comprehensive efficiency analysis for a meaningful comparison between main competing solutions by Pinuela et al. (2013) and key differences between link and dc-to-load efficiencies have been highlighted, most of all, the efficiency of the driver has been taken into account. Finally, to improve the efficiency, the coil system design and the suitable driver design have also been implemented.
In this paper, focusing on the high-efficiency and high-stability common energy launch platform of IPT system for different power capacity kitchen appliances, the circuit structure characteristic and main losses of IPT system have been analyzed, with the help of efficiency calculation model and designed capacitance array, the control strategy of efficiency optimization (the IPT system is allowed to operate at different segmented optimal frequency modes according to the power level of loads) has been proposed. Finally, a rated power of 1000 W experimental IPT system similar to the IPT system for kitchen appliances has been set up in order to verify the designed efficiency optimization control strategy.
FUNDAMENTAL ANALYSIS
Fundamental structure: Figure 1 illustrates the fundamental schematic structure of an IPT system that is designed for kitchen appliances with different power capacity. This system essentially is comprised of one stationary primary common energy launch platform and several movable secondary pickups with different power capacity. At any given time, only one pickup is allowed to operate on this common energy launch platform. Kitchen appliances can work properly as long as their pickup winding is within the effective area of the energy launch platform’s energy transmitting winding.
Specifically, in the primary common energy launch platform for kitchen appliances, there is an EMC protector, a power-frequency rectifier and filter, a high-frequency inverter and a resonant compensator. In the secondary side, a receiving winding, a resonant compensator, a high-frequency rectifier and filter and a switched-mode controller are necessary to drive the power consumer. It must be noted that the energy pickup winding, secondary capacitor compensation and power conversion circuits are installed inside the kitchen appliance. Most of all, in this paper, appropriate secondary compensating capacitors have been fore-designed permanently according to the power capacity of load.
Basic circuit topology: At any time, as only one pickup (kitchen appliance) is allowed to operate in the IPT system shown in Fig. 1, such system can be simplified for analysis. Therefore, only a secondary pickup side with a kitchen appliance will be discussed temporarily. According to the power supply requirement described before, the full-bridge inverter circuit of PS-type (parallel primary/series secondary) IPT system is chosen for the main circuit, its detailed circuit schematic can be shown in Fig. 2.
In practical, for a full bridge rectifier connected with a voltage load (a large capacitive filter in parallel with a load resistor Ri) shown in Fig. 2, the AC equivalent resistance R can be modeled as (Li et al., 2012; Sun et al., 2013; Wang et al., 2004; Hu, 2001):
| (1) |
| Fig. 1: | Schematic diagram of IPT system for kitchen appliances |
| Fig. 2: | Full-bridge inverter circuit of PS-type IPT system for kitchen appliances |
In the following discussion in this paper, the circuit in the dashed line frame in Fig. 2 is equivalent to a pure resistance R.
Mutual inductance: For the proposed IPT system for different power capacity kitchen appliances described in Fig. 2, However, once system parameters and circuit structure design are complete and if the system is operated at the rated condition, the distance and relative position between the primary and secondary windings are fixed, that is to say, the mutual inductance M is a constant value. Therefore, in the following discussion of IPT system for different power capacity kitchen appliances, the influence of mutual inductance M is not taken into consideration, the load impedance R is the unique disturbance.
EFFICIENCY OPTIMIZATION
Losses analysis: The loss components of selected IPT system for kitchen appliances shown in Fig. 2 can be classified into the following main categories:
| • | Conduction losses: Copper losses of primary and secondary windings and losses of reverse parallel diodes during conduction |
| • | Switching losses: Turn-on and turn-off losses of Insulated Gate Bipolar Transistors (IGBTs) and losses of reverse parallel diodes in the primary inverter network |
| • | Radiation losses |
The switching losses of the primary inverter network are very small under the Zero Phase Angle (ZPA) condition and the radiation loss is usually negligible because of the low operating frequency. It is obvious that the conduction losses in the primary and secondary windings contribute to the largest proportion of the total losses of whole IPT system shown in Fig. 2. With the help of conclusions by (Chen et al., 2007; Pinuela et al., 2013), the copper losses in the primary and secondary windings are only discussed. Hence, in the process of efficiency analysis, the following assumptions should be made (Hsu et al., 2009; Huang et al., 2009):
| • | The devices in the primary inverter network are ideal devices |
| • | The selected IPT system is designed to be operated at the ZPA condition |
Efficiency calculation model: According to the characteristics of selected IPT system and losses analysis shown above, the energy consumption is mainly divided into three parts: Losses of primary and secondary windings and the output load power. The efficiency of the designed IPT system can be expressed as:
| (2) |
where η is the efficiency of system, P0, P1 and P2 are the load power, the copper losses of primary and secondary windings, respectively and they are given by:
| (3) |
| (4) |
| (5) |
Substituting Eq. 3, 4 and 5 into Eq. 2, the efficiency of the designed IPT system can be rewritten as:
| (6) |
According to design guidelines of IPT system proposed in (Sun et al., 2013), generally, in the PS-type IPT system, we have R>>Rs, therefore, the Eq. 6 can be further simplified as:
| (7) |
As shown in Eq. 7, it is effective to increase the efficiency of IPT system shown in Fig. 2 by improving the mutual inductance M and reducing the inherent resistance of primary energy transmitting winding. It is necessary to improve the mutual inductance and reduce the inherent resistance of primary winding appropriately when the parameters of IPT system are being designed. However, once the process of system parameter design is complete, these two parameters are fixed. Therefore, the efficiency is only related to the load and operating frequency.
According to Eq. 7, the three dimensional map between the operating frequency, load and efficiency can be shown in Fig. 3, it can also be seen from the figure that when the efficiency value is not in the saturation state, it can be increased by improving the operating frequency under the same load R.
| Fig. 3: | Three-dimensional map between the operating frequency, load and efficiency |
While the lighter load, the lower efficiency under the same operating frequency.
Efficiency optimization strategy: It is indicated that the “efficiency optimization” presented in this study is not to make the efficiency reach a maximum value, but achieve a target value. Once the efficiency of selected IPT system has achieved the target value, the efficiency optimization design process ends correspondingly.
If the required efficiency must be higher than the target value η0, so that:
| (8) |
Substituting (8-7), the operating frequency must be satisfied as:
| (9) |
As shown in Eq. 9, the IPT system with a larger load impedance R (a lighter load) is required to operate at the higher frequency to achieve the target efficiency. The efficiency of IPT system with a smaller load impedance R (a heavier load) can also be increased by improving the operating frequency when the efficiency value is not in the saturation state. However, the narrower constant operating frequency, primary resonant current and output voltage area of IPT system will appear with the higher operating frequency, as shown in Fig. 4.
In Fig. 4, fi (i = 1-5) is the inherent frequency. Under different inherent frequency, all resonant parameters of IPT system are consistent apart from primary and secondary compensating capacitor Cp, Cs and a necessary criteria LpCp = LsCs must be established.
| Fig. 4(a-c): | Curves of normalized operating frequency, primary resonant current, output voltage and power transfer capability varying with load under different inherent frequency, (a) Operating frequency (f1>f2>f3>f4>f5, To facilitate mapping and analyzing, the practical operating frequencies f in are normalized using the inherent resonant frequency fi as u = f/fi (i = 1-5), (b) Output voltage (f1>f2>f3>f4>f5) and (c) Power transfer capability (f1>f2>f3>f4>f5) |
Meanwhile, the operating frequency of IPT system can be changed with the change of the inherent frequency.
Taking load R = 50Ω for example, when the inherent resonant frequency of IPT system is allowed to increase to f1, the practical operating frequency becomes greatly unstable. On the contrary, the practical operating frequency remains approximately constant if the inherent resonant frequency is f5. Meanwhile, the primary resonant current and output voltage at the inherent frequency f1 are lower and more unstable than f5. Therefore, the power transfer capability of IPT system is minimized at the inherent frequency f1.
Hence, if the inherent frequency of IPT system with a smaller load is increased, the unstable operating frequency, primary resonant current, output voltage area appears with the disturbance of load, moreover, the stability and power transfer capability of IPT system would be decreased.
Though analyzing about characteristic, losses and efficiency in the IPT system for kitchen appliances, a segmented optimal frequency dynamic tracking method can be proposed for optimizing target efficiency using changing the inherent resonant frequency, accordingly, the operating frequency is also changed. Consequently, the power transfer system is allowed to operate at different frequency modes dynamically according to the power level of loads and the lighter load requires a higher frequency than the heavier load. Moreover, the designed IPT system must be within the constant frequency and voltage area shown in Fig. 4 under different frequency modes.
The phased-control inductor (Hu and Hussmann, 2003) and the switching-capacitor (Si et al., 2008) have been used generally to change the inherent resonant frequency, but it is difficult to wind and control the phased-control inductor, so an online capacitor array different from switching capacitor shown by Si et al. (2008) is constructed to change the inherent resonant frequency in section IV. In this way, the IPT system with a lighter load is allowed to operate at a higher frequency mode than a heavier load, the target efficiency of IPT system for kitchen appliances shown in Fig. 2 can be achieved effectively.
DESIGN IMPLEMENT
Capacitor array in the primary side: The online capacitor array can be constructed to change the inherent resonant frequency, it is shown in Fig. 5 and paralleled with the primary resonant inductor of the common energy launch platform, the capacitance in the primary resonance tank can be changed real-timely by controlling switches Sk (k = 1~m) to be on and off.
| Fig. 5: | Working diagram of capacitor array in the primary side |
And these switches are composed of two reverse series semiconductor devices such as IGBTs.
According to different combinations of the additional capacitor array shown in Fig. 5, the optional primary compensating capacitance fuzzy set can be given by:
| (10) |
If the required target efficiency of designed IPT system for kitchen appliances must be higher than the given value η0, making use of Eq. 9, the optimizationprimary compensating capacitance must be satisfied as:
| (11) |
The efficiency optimization control subsystem real-timely selects capacitance in the capacitance fuzzy set described in Eq. 10 according to inequality (11).
Secondary compensating capacitor: In practice, before kitchen appliances receive from the power transmitting part at the rated condition, different from the real-time and online choice of primary compensating capacitor, appropriate secondary compensating capacitors have been fore-designed and installed inside the kitchen appliances permanently according to the power capacity of load and the capacitance must be satisfied as:
| (12) |
Identification: As shown above, the power capacity of load should be identified to select the primary capacitance. According to the energy relation of selected IPT system, the following equation exists in the form of:
| (13) |
As shown in Eq. 13, the injection current idc includes information of the load condition R. With the identification result of injection current idc, we can identify the load parameter. The control subsystem of IPT system needs the real time sampled data about injection current idc. Similar to the primary compensating capacitance fuzzy set, the injection current identification fuzzy set can be written as:
| (14) |
Start-up and working frequency: As shown in Fig. 4, the constant operating frequency and output voltage area becomes narrower at the higher inherent frequency, meanwhile, the power transfer capability is larger at the lower inherent frequency and it is easy to control at the lower operating frequency. Therefore, a low frequency is selected as the start-up frequency of designed IPT system for kitchen appliances. Meanwhile, the variable frequency control is more feasible and favorable than the fixed frequency control in the proposed control strategy. However, in the starting phase of the variable frequency control, the fixed frequency is essential.
Control: The control design of IPT systems relies strongly on experience and experimental verification because of the complexity regarding the interactions of the primary and secondary resonant circuits. Combing (10) with (14), the control strategy of efficiency optimization can be obtained as:
| (15) |
Based on the control strategy shown in Eq. 15, the control block diagram of variable frequency control and efficiency optimization control subsystem can be designed in Fig. 6.
Normally, the variable frequency control subsystem needs the real time sampled data about the resonant voltage across the primary compensating capacitor, such voltage can be divided by a voltage dividing resistor network. Then, the divided voltage signal should be zero crossing sampled by the LM311 chip. The field programmable gate array (FPGA) chip (EP2C5T144C8N) sends switching gate signal out according to the zero crossing sampled signal. The inverter network composed of four IGBTs (FGA25N120ANTD) is essential to drive the resonant tank and generate the required high-frequency primary current.
In principle, the efficiency optimization control subsystem acquires the information of injection current (idc) pouring into the inverter network with the help of sampling resistor and AD574 chip, then the sampled digital signal is sent into the FPGA chip, in this chip, the current identification, control strategy implement of efficiency optimization and capacitance choosing are executed.
EXPERIMENTAL STUDY
Experimental circuit and parameters: In order to verify the control strategy of efficiency optimization, a wireless power transfer platform for kitchen appliances capable of providing power of 1000 W has been prepared.
| Fig. 6: | Control block diagram of variable frequency control and efficiency optimization control subsystem |
| Fig. 7: | Circuit model of fore-designed secondary pickup side in the experimental system (Csi = {Cs1, Cs2}, Ri = {R1, R2}) |
| Table 1: | Main parameters of experimental system |
The main parameters of such system are shown in Table 1 (The capacitor array shown in Fig. 5 is composed of C1 and C2. With the aid of two loads R1 (the power is P1) and R2 (the power is P2), two kitchen appliances of different power capacity can be imitated. Accordingly, different compensating capacitors Cs1 and Cs2 are fore-selected so that the IPT system can be operated at the secondary inherent resonant frequency as soon as possible. The physical meanings of the other parameters are shown in Fig. 2 and the primary side as well as the common energy launch platform is same with the IPT system shown in Fig. 2, but two circuit models prepared for light load R1 and heavy load R2, respectively are applied at the secondary side, as shown in Fig. 7.
Without efficiency optimization control: The first experiment has been allowed to operate at the lower inherent frequency mode (about 20 kHz) without the optimization control strategy. Therefore, compensating capacitance of two secondary pickup circuits can be fore-designed as Cs1 = Cs2 = 0.11 μF and the compensating capacitance in the primary resonant tank can be selected as Cp = 0.52 μF. Consequently, steady-state waveforms without the control strategy under both light load R1 and heavy load R2 are shown in Fig. 8.
The efficiency of designed IPT system for kitchen appliances is 70.7% at the operating frequency of 20.6 kHz (lower frequency) with an injection current idc of 0.47A and output voltage Uo of 227V under light load R1. Meanwhile, the efficiency of system is 90.0% with an injection current idc of 1.75A and output voltage Uo of 221V under heavy load R2, while the operating frequency under heavy load R2 is same with the operating frequency under light load R1.
| Fig. 8(a-b): | Steady-state waveforms without control strategy under both light load and heavy load (Ch1-injection current idc, Ch2-primary resonant current Ip, Ch3-output voltage Uo) (a) Light load R1 and (b) Heavy load R2 |
It can be clearly seen that the efficiency of IPT system for kitchen appliances under light load R1 is lower relatively. Therefore, it is necessary to optimize the efficiency of IPT system for kitchen appliances under light load R1.
With efficiency optimization control: Anyway, it can be assumed that the target efficiency of IPT system for kitchen appliances is η0 = 80%. Combining (10) with (11), the capacitance array fuzzy set in efficiency optimization control subsystem can be given by:
Cp = {Cp1, Cp2} =
{0.3 μF, 0.52 μF} |
At the same time, according to Eq. 12, compensating capacitances of two secondary pickup circuits can be fore-designed as Cs1 = 0.063 μF for the light load R1 and Cs2 = 0.11 μF for the heavy load R2, respectively.
| Fig. 9(a-b): | Waveforms with efficiency optimization control strategy under the light load (Ch1-injection current idc, Ch2-primary resonant current Ip, Ch3-output voltage Uo) (a) Operating procedure waveforms and (b) Steady-state waveforms |
Meanwhile, relying on the experience and experimental verification, the injection current identification fuzzy set can be selected as:
idc = {idc1} = {1A} |
Hence, the control strategy of efficiency optimization can also be designed as:
Operating procedure waveforms of IPT system with efficiency optimization control algorithm are illustrated in Fig. 9a under light load R1. First of all, the IPT system starts working at the lower inherent frequency mode f1 = 20.6 kHz at time t = to, because the practical identification current idc is less than 1A under light load R1, then, at time t = t1, the primary compensating capacitance is reduced from Cp = Cp2 = 0.52 μF to Cp = Cp1 = 0.3 μF through closing the switch S2, the IPT system under light load R1 is allowed to operate at the higher inherent frequency mode f2 = 26.7 kHz. In this case, the efficiency of system is 82.6% at the operating resonant frequency of 26.7 kHz with an injection current idc of 0.406 A and output voltage Uo of 228 V, steady-state waveforms are shown in Fig. 9b. It is obvious that the efficiency of IPT system has been increased by 12% and achieved the target efficiency under light load R1. The results shown above have verified the effectiveness of proposed control method.
However, according to the control strategy, steady-state waveforms with efficiency optimization are same with steady-state waveforms without efficiency optimization control strategy under heavy load R2 shown in Fig. 8b and they are omitted.
Further discussions: Furthermore, a new experiment used efficiency optimization control strategy has been done. This time, the assumed target efficiency is η0 = 85%. Figure 10a shows the curves between experimental measured efficiency and loads under different inherent frequency modes (f1<f2<f3), it is greatly significant that increasing efficiency through improving inherent frequency mode under light load R = 500 Ω, but this isn’t suitable for heavy load R = 50 Ω, what’s worse, the practical operating frequency becomes unstable and power transfer capability drops dramatically under heavy load. Compared to 4% at the higher frequency f3, the drift value of frequency is only 0.54% at the lower frequency f1 when the load jumps from 100-50 Ω (Fig. 10b). Moreover, Compared to 942 W at f1, the power transfer capability of system is only 648 W at f3 when the load is R = 50 Ω (Fig. 10c). Hence, the IPT system with the heavier load is allowed to operate at the lower inherent frequency mode as soon as possible. On the contrary, the efficiency of IPT system with the lighter load can be boosted efficiently through improving inherent frequency mode.
Therefore, as shown in Fig. 10a, If the required target efficiency must be higher than the given value η0 = 85%, according to the control method proposed in this paper, when the equivalent resistance of load R>R2, the system is operated under the frequency f3. And when R1<R<R2, the operating frequency can be f2. Finally, when R<R1, the operating frequency is f1.
Apparently, with the aid of the capacitor array and the proposed efficiency optimization strategy, the efficiency of the IPT system for kitchen appliances remains a high level under all light and heavy loads rather than a fixed load.
| Fig. 10(a-c): | Curves of efficiency, practical operating frequency and output voltage varying with load at different experimental frequency modes, (a) Efficiency (f1<f2<f3), (b) Practical operating frequency (f1<f2<f3) and (c) Power transfer capability (f1<f2<f3) |
Such result is superior to the ones by (Chen et al., 2007; Pinuela et al., 2013; Sun et al., 2010).
In summary, all the experimental results shown above have verified the rationality of the segmented optimal frequency dynamic tracking method and realized the high-efficiency and high-quality wireless power supply for all different power capacity kitchen appliances.
CONCLUSION
To realize the high-efficiency and high-quality wireless power supply for all different capacity kitchen appliances, in this paper, the structure characteristic, main losses and efficiency calculation model of designed IPT system have been analyzed. With the aid of such efficiency model and designed online capacitance array, the control strategy of efficiency optimization has been proposed. Also, in order to verify the designed efficiency optimization control strategy, a rated power of 1000W experimental system similar to the contactless power transfer system for kitchen appliances has been set up. Experimental results indicate that the total efficiency is increased obviously when the system under a light load is allowed to operate at a higher inherent frequency mode, but this rule is not suitable for the heavy load.
ACKNOWLEDGMENT
This research work is financially supported by The Research Fund for the Doctoral Program of Higher Education (No.20100191120024) and China Postdoctoral Science Foundation (No. 20110490799). Fundamental Research Funds for Central Universities (CDJZR10170004). And the authors also would like to give their special thanks to the reviewers of this paper for their contributions to this work.