The AC charging pile is the main energy supply facility for household electric vehicles, which uses a vehicle mounted charger to charge the power battery. The current standard of the State Grid Corporation of China clearly stipulates the function of the AC charging pile and does not take into account the impact of the harmonics of the vehicle mounted charger on the power grid. Therefore, in view of the deficiency that AC charging piles cannot suppress the current harmonics of the vehicle mounted charger, application of the active power filtering technology to the design of AC charging piles is proposed to form a new type of AC charging pile with better functions. In the experimental prototype that was built, for vehicle mounted chargers with two load characteristics, the composite control method of traditional PI control and repetitive control is adopted, where the new AC charging pile effectively suppresses the harmonics of the vehicle-mounted charger. Experiments show that the AC charging pile using active power filtering technology cannot only improve the power quality of the grid side but also reduce the impact of harmonics on the power metering and billing system, ensuring the stability of the charging communication system.

## I. INTRODUCTION

At present, the problems of environmental pollution and energy shortage are becoming more serious. Due to the high energy consumption and serious environmental pollution of traditional fuel vehicles, more people are concerned and worried.^{1,2} Therefore, countries around the world are actively implementing energy conservation and emission reduction strategies.^{3,4} As an important part of it, electric vehicles have been developed rapidly in recent years. The power source of the drive system in electric vehicles is all- or partly electric. Compared with traditional fuel vehicles, electric vehicles have obvious advantages such as environmental protection, cleanliness, energy saving, and so on. Electric vehicles will become the main direction of contemporary automobile development and the most potential means of transportation in the 21st century.^{5,6} The energy supply device of electric vehicles is an indispensable and important piece of equipment for the electric vehicle industry, mainly including DC chargers and AC charging piles. The DC charger has a large power (about 100 kW) and a short charging time. Due to its large size, it is generally installed in a special electric vehicle charging station. The AC charging pile directly provides AC mains power and uses a vehicle mounted charger to charge the power battery.^{7,8} Generally, the AC charging pile has a small power (about 10 kW) and a long charging time. Due to its small size and small carbon footprint, it can be installed in every corner of the city.

Electric vehicle DC chargers (including vehicle mounted chargers) are actually switching power supplies using power electronics technology, which inevitably generate harmonics and reactive currents during the charging process, affecting the power quality of the power grid. From the current construction of the charging station to the harmonic and reactive power interference generated by the high-power DC charger, the charging station is equipped with a special harmonic control and reactive power compensation device.^{9,10} Ordinary low power vehicle mounted chargers, in consideration of quality, volume, and cost, generally do not deal with their own harmonic problems. As for the power grid, with the widespread use of AC charging piles, the vehicle mounted charger will be a source of harmonics, which will affect the power quality and stability of the power grid. Moreover, the existence of harmonics will inevitably affect the accuracy of the charging pile metering and billing system and the stability of the communication system.

Judging from the current standards and actual technical level of AC charging piles, the existing charging piles have perfect metering and billing measures, perfect communication capabilities, strong monitoring capabilities, and simple power supply functions, and they have not considered the effective harmonic control of the vehicle mounted charger. Therefore, this paper proposes to apply active power filter (APF) technology to AC charging piles and develops a new single-phase AC charging pile based on single-phase parallel active filter technology.

## II. APF FUNCTION SINGLE-PHASE AC CHARGING PILE CIRCUIT TOPOLOGY AND SINGLE-PHASE APF MODEL CONSTRUCTION

### A. Topological structure analysis of a single-phase AC charging pile with an APF function

The circuit structure of a single-phase AC charging pile with an APF function is similar to the single-phase APF circuit of the internal output power line of the existing AC charging pile. The power supply of the metering system, the communication system, and the control system are taken from the front stage of the APF circuit branch, which can eliminate harmonic interference of the vehicle mounted charger. The fact that whether the single-phase APF circuit works normally or not does not affect the normal power supply function of the charging pile. Its primary circuit topology and secondary control function are shown in Fig. 1. In Fig. 1, SD is the signal detection, MBCC is the metering and billing and communication circuit, CP is the control power, DSP CC is the DSP control circuit, IC PC is the isolation circuit and protection circuit, and APF D is the APF drive. The main circuit of the APF is composed of a full-bridge circuit composed of switch tubes PT_{1} to PT_{4}, an AC-side inductor *L*, and a DC-side capacitor *C*, which is consistent with a high-frequency pulse width modulation (PWM) converter. Among them, *R*_{s}, *D*_{s}, and *C*_{s} constitute the DC bus buffer circuit, which is used to suppress the overvoltage of the switching tube. *R*_{s} is the snubber resistance; *D*_{s} is the snubber diode; *C*_{s} is the snubber capacitor; K_{1} is the power grid input control switch; K_{2} is the charging pile output control switch; K_{3} is the APF switching control switch.

In Fig. 1, *u*_{s} represents the grid voltage; *i*_{s} is the grid current; *i*_{L} is the output current of the charging pile, that is, the input current of the vehicle mounted charger; *i*_{sh} is the output current of the APF used to compensate the harmonic and reactive current generated in the charging process of the vehicle mounted charger; and *u*_{L} is the output voltage of the charging pile. The voltage Hall element VHL and the current Hall element CHL are used to detect the voltage and current of the system, respectively, in real time. The DSP control circuit collects the voltage and current signals in real time for the metering and billing and communication systems and outputs PWM driving pulses according to the control algorithm. The protection and drive isolation circuit can photoelectrically isolate the primary circuit and the secondary circuit.

### B. Mathematical model of the APF

*u*

_{AB}is the AC side voltage of the APF,

*r*represents the equivalent impedance of the inductor internal resistance and the bridge arm dead zone voltage drop, and

*u*

_{AB}=

*nU*

_{dc}is the controlled AC side voltage, where

*U*

_{dc}is the APF DC side voltage and

*n*is the switching function.

*u*

_{s}can be replaced by its instantaneous value, the period average value of

*u*

_{AB}is consistent with its instantaneous value, and the influence of

*r*is ignored; the period average model of the APF is obtained as

The APF output current $ishtk+1$ in Eq. (2) is replaced by the harmonic current command, and the output voltage command value $uABk$ of the converter at time *k* can be calculated to realize the digital control of the APF.

*u*

_{c}is the triangular carrier wave signal and

*u*

_{r}is the standard modulated sinusoidal signal. The expression of the switch function

*n*is

## III. APF CONTROL SYSTEM

*z*-transform of the continuous system s function, and s is the complex frequency in the system frequency domain analysis. Due to the factors of sample holding and lagging one beat, the use of traditional PI control algorithms will affect the stability of the system. Figure 3 shows the Bode diagram analysis of the closed-loop transfer function of the converter controlled object after passing through the PI controller with or without lagging one beat. In Fig. 3, $G1z$ is the amplitude phase frequency characteristic of the lagging one beat system, $G2z$ is the amplitude phase frequency characteristic of the system without lagging one beat, the ordinate A represents the amplitude frequency characteristic,

*φ*represents the phase frequency characteristic, and the abscissa

*f*represents the current inner loop error signal frequency. If the proportional parameter of the PI controller is increased to widen the bandwidth of the system, the $G1z$ system will tend to be critically stable or even destroy the stability of the system. From the analysis of phase frequency characteristics, it can be seen that in the medium frequency band, the phase shift of the system is serious, which will affect the performance of the system in tracking high-order harmonics.

The load characteristic of the AC charging pile is mainly the uncontrolled rectifier circuit inside the vehicle mounted charger, which is a nonlinear load, including a small amount of reactive current and a large amount of harmonic current. Therefore, the current command of the APF is often a signal superimposed by the power frequency and its multiple frequency signals. As can be seen in Fig. 3, under the premise that the system is stable, as the frequency increases, the amplitude–frequency characteristics of the system are gradually attenuated, which will affect the magnitude of harmonic compensation. However, the phase–frequency characteristic of the system has a large attenuation, which will affect the phase accuracy of the harmonic compensation. Therefore, the traditional PI controller is difficult to meet the application requirements of the APF, which must be compensated by more effective methods.

## IV. DESIGN OF A COMPOSITE CONTROL SYSTEM BASED ON REPETITIVE CONTROL THEORY

### A. The principle of repetitive control

The basic idea of repetitive control is that under the premise that the system is stable, in order to achieve good tracking ability (that is, the steady-state error tends to zero), it must satisfy the mathematical model containing external signals in its open-loop transfer function and form a closed-loop control system.

*N*is the number of samples per fundamental frequency period, which is equal to the modulation wave period divided by the carrier wave period.

Due to the influence of lagging one beat and the accuracy of the APF compensation harmonic current, the traditional PI controller will generate a certain harmonic compensation tracking error, the integral I control coefficient is not very effective in reducing the tracking error, and it has an adverse impact on the stability of the system. Therefore, on the basis of the original PI controller, only the proportional P controller is used, and the compound controller is composed of parallel repetitive controllers to improve the ability of harmonic current tracking.

*k*

_{p}is the expression of the proportional controller. In the dashed box is the repetitive controller, which includes the repetitive signal generator, the repetitive control gain link

*k*

_{r}, and the compensation link $Sz$. $Qz$ is used to weaken the effect of the integral and increase the stability margin of the system, and it generally takes a constant close to 1. $ish*$ represents the harmonic current command,

*i*

_{sh1}is the harmonic current output by the proportional control, and

*i*

_{sh2}is the harmonic current output by the repetitive control. Simultaneously, the transfer function of the current closed loop is obtained as

_{2}is in the unit circle or not; the criterion expression is

As shown in Fig. 5, the stability of the system requires the magnitude of the vector *a* to be less than 1. When $QejwT=1$, the center of the unit circle is fixed at point (1,0), the left arc of the unit circle will be tangent to the imaginary axis at the origin, and the area contained in the unit circle is limited to part of the first and fourth quadrants. In the middle and low frequency bands, due to the effect of the compensator, it can be considered that the vector *b* has approximately zero gain and zero phase shift. However, in the high frequency bands, due to the modeling errors, the compensation effect cannot be guaranteed. If at a certain frequency the phase angle of vector *b* is close to ±90°, even if its amplitude is small, the end of the vector *a* will stay on the unit circle, and the system becomes critically stable. Therefore, if $QejwT$ is set to a constant less than 1, the whole unit circle is shifted to the left so that even if the phase angle of vector *b* exceeds ±90° in the mid-to-high frequency band, the end of vector *a* can be adjusted in the unit circle to ensure the stability of the system. Therefore, the addition of $Qz$ makes the stability region of the system cover the four quadrants of the complex plane. Even if a serious phase error is caused by modeling errors in the middle and high frequency bands, the stability of the system can be ensured, and the robustness of the system is enhanced.

### B. Design of a compound controller

*i*

_{sh2}without error, and superimpose it on the original output

*i*

_{sh1}so that the system can finally realize the tracking without static error. Therefore, it can be seen from Eq. (11) that the expression of the equivalent control object of the repetitive controller is

*a*

_{1}is designed as the pole of the equivalent controlled object to reduce the order of the system. The compensated system should meet the characteristics of zero gain in the low and middle frequency bands, so when

*ω*→ 0, $z=ej\omega T\u21921$, the compensated system satisfies the following relation:

The value of *c*_{1} affects the bandwidth of the corrected system, and the value of *b*_{1} is used to correct the gain characteristics of the system. The system designed in this paper can have zero gain characteristics between 50 and 1200 Hz.

*z*

^{2}. Therefore, according to Eq. (8), it can be deduced that the value range of the repetitive controller gain

*k*

_{r}is

Figure 6 shows the Bode plot of the transfer function of the current inner loop with the constructed composite control. Compared with the simple PI controller, the phase shift and gain attenuation of the system at the fundamental frequency and its multiplier within 1200 Hz are smaller, which can meet the application requirements of APF tracking the harmonics of vehicle mounted chargers.

## V. EXPERIMENTAL VERIFICATION

According to Fig. 1, the AC charging pile circuit with a single phase APF function is built for experimental verification. Table I shows the circuit structure parameters of the single phase parallel APF. The uncontrolled rectifier is connected to RC and the uncontrolled rectifier is connected to RL to simulate the front stage rectification link of the vehicle mounted charger, as shown in Fig. 7, and the harmonic compensation test is carried out.

Category . | Parameter . |
---|---|

Grid voltage rms u_{s} (V) | 220 |

Grid frequency f_{s} (Hz) | 50 |

APF output inductor L (mH) | 3 |

DC bus capacitance C (μF) | 4700 |

DC bus voltage setting U_{dc} (V) | 400 |

Switching frequency f_{c} (kHz) | 10 |

Category . | Parameter . |
---|---|

Grid voltage rms u_{s} (V) | 220 |

Grid frequency f_{s} (Hz) | 50 |

APF output inductor L (mH) | 3 |

DC bus capacitance C (μF) | 4700 |

DC bus voltage setting U_{dc} (V) | 400 |

Switching frequency f_{c} (kHz) | 10 |

As shown in Fig. 7, the DC side inductance *L* of the harmonic source is 100 mH, the resistance *R* is 20 Ω, and the DC side capacitor *C* is 4700 *µ*F.

Figure 8 shows the compensation current waveform of the uncontrolled rectifier connected to RL load with the APF. Figure 9 shows the Fast Fourier Transform (FFT) frequency spectrum of the load current, and Fig. 10 shows the power grid current frequency spectrum. It can be seen that the Total Harmonic Distortion (THD) of the grid current decreases from 41.7% to 4.6%. The abscissa m in Figs. 9 and 10 represents the harmonic order, and the ordinate represents the percentage of the harmonic current amplitude *i*_{m} in the fundamental current *i*_{1}.

## VI. DISCUSSION

In this paper, the construction and stability criteria of the single phase active filter compound controller are theoretically derived, and the feasibility of harmonic suppression is verified from the prototype experiment, but the effect of harmonic suppression depends on the accuracy and rapidity of the harmonic current command. The harmonic current command comes from the extraction of the harmonic components of the input current of the vehicle mounted charger. The main methods applied to the single phase system are the fundamental wave component method, the FFT analysis method, the adaptive detection method, and the wavelet theory detection method. Due to the small amount of calculation of the adaptive detection method, it is easy to realize the digital system, and its dynamic detection performance can meet the actual requirements of charging, so this paper adopts the adaptive harmonic detection method to uniformly compensate all harmonic components of the charger except the fundamental wave. If the FFT analysis method can accurately extract each harmonic component, it can accurately compensate the specified harmonic.

In the experiment carried out in this paper, the harmonic current component and the output of the voltage outer loop controller are given as the current inner loop. The phenomenon in the experiment is that the DC voltage on the DC bus side of the APF main circuit is stable, and the AC side output compensates the harmonic current of the vehicle mounted charger. If the reactive component of the vehicle mounted charger is included in the given current inner loop, the APF can also compensate the reactive current of the charger. For the power grid, the whole charging device is a resistive device.

## VII. CONCLUSION

In this paper, the topological structure of the new AC charging pile with an APF function is analyzed, and the state period average model of a single phase APF is established to realize the digital control of the APF.

The control performance of the current inner loop in the control system is analyzed, a compound control system combining proportional control and repetitive control is designed, and the expression of the system stability criterion and the method of geometric vector judgment are derived.

The prototype experiment shows that the method of using the AC charging pile with an APF function to compensate the harmonics of the vehicle charger is feasible. Through the comparative analysis of system Bode plots in the frequency domain, it can be seen that the compound control system increases the system response bandwidth and improves the robustness of the system. Therefore, the compound control method can compensate the harmonics of the vehicle mounted charger better so that the harmonic content of the current on the power grid side can be reduced to less than 5%, which meets the requirements of national standards.

The harmonics of the vehicle mounted charger can be effectively suppressed, which will help improve the efficiency of the whole charging system, reduce system losses, ensure the accuracy of electric energy metering and billing, and ensure the stability of the charging pile communication system. The charging device meets the requirements of a strong smart grid and promotes the development of the electric vehicle industry better.

## ACKNOWLEDGMENTS

This work was supported by the National Key R&D Program of China under Grant (Grant No. 2016YFF0201201).

## AUTHOR DECLARATIONS

### Conflict of Interest

The authors have no conflicts to disclose.

### Author Contributions

**Jin Bao**: Data curation (equal). **Meimei Duan**: Project administration (equal). **Jun Li**: Formal analysis (equal).

## DATA AVAILABILITY

The data that support the findings of this study are available from the corresponding author upon reasonable request.