Harmonic suppression and power quality adaptive control method for energy storage converter

Abstract

The present application relates to energy storage control technology field, and further relates to a harmonic suppression and power quality adaptive control method of energy storage converter, comprising the following steps: step 1, multi-channel signal acquisition and harmonic component separation: the output current signal, the output voltage signal and the DC bus capacitor voltage signal of the active neutral point clamped three-level inverter are collected; step 2, port-controlled Hamilton energy shaping and super-spiral sliding mode fusion control: the port-controlled Hamilton system framework is constructed with the bridge arm inductance current and the capacitor voltage as the state variable; step 3, inverter modulation and active neutral point potential balance: the basic voltage vector is determined by the space vector modulation of the composite modulation voltage, and the gate drive pulse is generated to drive the active neutral point clamped three-level inverter. The present application realizes the accurate tracking of fundamental current, effective suppression of harmonic current and active balance of neutral point potential, and improves the power quality and operation reliability of the energy storage converter.

Description

Technical Field

[0001] This invention belongs to the field of energy storage control technology, specifically relating to a harmonic suppression and adaptive power quality control method for energy storage converters. Background Technology

[0002] As a core device connecting energy storage devices to the power grid, the energy storage converter plays a crucial role in bidirectional energy flow and power quality regulation. The performance of the energy storage converter directly affects the operating efficiency of the energy storage system, the power quality connected to the grid, and the safety and stability of the power grid. However, energy storage converters inevitably generate harmonic currents during operation. These harmonic currents, when injected into the grid, can lead to a series of power quality problems, such as voltage waveform distortion, equipment overheating, and malfunctions of protection devices. Therefore, effectively suppressing the output harmonics of energy storage converters and improving power quality has become a key research topic in this field.

[0003] Regarding the topology of energy storage converters, traditional two-level voltage source inverters are simple in structure and mature in control, but their output voltage has only two levels, resulting in high harmonic content and requiring bulky filters to meet grid connection standards. Three-level inverters, by increasing the number of output voltage levels, can effectively reduce harmonic content and voltage stress on switching devices, and are widely used in medium- and high-voltage, large-capacity energy storage systems. The midpoint-clamped three-level inverter is the most common three-level topology, but it suffers from midpoint potential fluctuations and uneven distribution of switching losses. Active midpoint-clamped three-level inverters, by replacing the clamping diodes with controllable switching devices, can achieve flexible selection of the commutation path and balanced distribution of switching losses, but their control complexity also increases accordingly. In terms of harmonic suppression control strategies, proportional-integral (PI) control is the most basic method, capable of zero steady-state error tracking of the fundamental current, but it is difficult to achieve precise control of AC harmonic currents. Proportional resonant control, by introducing a resonant element at a specific harmonic frequency, can achieve zero steady-state error regulation of the corresponding harmonic current. However, it requires designing a separate resonant controller for each harmonic, and when there are many harmonics to be suppressed, the order and computational complexity of the control system increase significantly. Repetitive control utilizes the repetitive characteristics of periodic signals to suppress all integer harmonics through the internal mode principle, but its dynamic response is slow, making it difficult to quickly track changes in the reference value during load abrupt changes or grid fluctuations. Sliding mode control has strong robustness to parameter perturbations and external disturbances, but traditional first-order sliding mode control suffers from chattering. High-frequency chattering not only affects control accuracy but also exacerbates losses in switching devices and electromagnetic interference. Summary of the Invention

[0004] The main objective of this invention is to provide a harmonic suppression and power quality adaptive control method for energy storage converters. It combines port-controlled Hamiltonian energy shaping control with super-spiral sliding mode control. The energy shaping method ensures the stability and robustness of the fundamental current control, while the super-spiral algorithm eliminates the chattering problem of traditional sliding mode control while maintaining strong robustness. Combined with the redundant vector selection strategy of an active midpoint clamped three-level inverter, it achieves coordinated control of accurate fundamental current tracking, effective harmonic current suppression, and active midpoint potential balancing, thereby improving the power quality and operational reliability of the energy storage converter.

[0005] To solve the above problems, the technical solution of the present invention is implemented as follows:

[0006] The harmonic suppression and adaptive power quality control method for energy storage converters includes the following steps: Step 1: Multi-channel signal acquisition and harmonic component separation: Acquire the output current signal, output voltage signal, and DC bus capacitor voltage signal of the active neutral clamp three-level inverter. Perform coordinate transformation on the output current signal to obtain the current component in the two-phase stationary coordinate system. Obtain the fundamental current component and each harmonic current component through fundamental wave separation and harmonic extraction. Obtain the grid phase angle through the phase-locked loop module. Step 2: Port-controlled Hamiltonian energy shaping and super-helical sliding mode fusion control: Construct a port-controlled Hamiltonian system framework with bridge arm inductor current and capacitor voltage as state variables. The fundamental current error is controlled by energy shaping and damping injection to obtain the fundamental control voltage. The harmonic compensation voltage is obtained by super-helical sliding mode control of each harmonic current component. The fundamental control voltage and the harmonic compensation voltage are superimposed to obtain the composite modulation voltage. Step 3, Inverter Modulation and Active Neutral Point Potential Balancing: The composite modulation voltage is spatially vector modulated to determine the basic voltage vector. The redundant switch states are selected based on the midpoint potential deviation signal and phase current polarity to achieve midpoint potential balancing. The gate drive pulse is then generated to drive the active midpoint clamped three-level inverter.

[0007] Furthermore, in step 1, the three-phase output current signal of the active neutral clamp three-level inverter is acquired by a current sensor, the three-phase output voltage signal is acquired by a voltage sensor, and the upper half bus capacitor voltage and the lower half bus capacitor voltage are acquired by a DC side voltage sensor respectively; the upper half bus capacitor voltage and the lower half bus capacitor voltage are subtracted from each other to obtain the neutral point potential deviation signal.

[0008] Further, in step 1, the three-phase output current signal is input to the Clarke transform module to obtain the alpha-axis current component and the beta-axis current component in the two-phase stationary coordinate system; the alpha-axis current component is input to the first second-order generalized integrator, and the beta-axis current component is input to the second second-order generalized integrator; the first second-order generalized integrator outputs the first in-phase component that is in phase with the alpha-axis current component and the first quadrature component that lags by 90 degrees; the second second-order generalized integrator outputs the second in-phase component that is in phase with the beta-axis current component and the second quadrature component that lags by 90 degrees.

[0009] Further, in step 1, the first in-phase component is added to the second quadrature component to obtain the positive-sequence alpha-axis current component, and the second in-phase component is subtracted from the first quadrature component to obtain the positive-sequence beta-axis current component; the alpha-axis current component is subtracted from the positive-sequence alpha-axis current component to obtain the harmonic alpha-axis current component, and the beta-axis current component is subtracted from the positive-sequence beta-axis current component to obtain the harmonic beta-axis current component; the positive-sequence alpha-axis current component and the positive-sequence beta-axis current component are input to the phase-locked loop module, and the phase-locked loop module outputs the grid phase angle; based on the grid phase angle, the positive-sequence alpha-axis current component and the positive-sequence beta-axis current component are transformed to a synchronous rotating coordinate system to obtain the d-axis fundamental current component and the q-axis fundamental current component.

[0010] Furthermore, in step 1, the harmonic alpha-axis current component and the harmonic beta-axis current component are input to the 5th harmonic rotation transformation module, the 7th harmonic rotation transformation module, the 11th harmonic rotation transformation module, and the 13th harmonic rotation transformation module, respectively. The 5th harmonic rotation transformation module performs transformation with a transformation angle of -5 times the grid phase angle and extracts the 5th harmonic d-axis current component and the 5th harmonic q-axis current component through a low-pass filter. The 7th harmonic rotation transformation module performs transformation with a transformation angle of +7 times the grid phase angle and extracts the 7th harmonic d-axis current component and the 7th harmonic q-axis current component through a low-pass filter. The 11th harmonic rotation transformation module and the 13th harmonic rotation transformation module extract the 11th harmonic current component and the 13th harmonic current component, respectively, in the same manner.

[0011] Furthermore, in step 2, the bridge arm inductor current, filter capacitor voltage, and DC bus capacitor voltage of the active midpoint clamped three-level inverter are used as system state variables to construct a port-controlled Hamiltonian system framework. In the port-controlled Hamiltonian system framework, the sum of the magnetic field energy stored in the bridge arm inductor, the electric field energy stored in the filter capacitor, and the electric field energy stored in the DC bus capacitor is used as the total system energy function.

[0012] Further, in step 2, the d-axis fundamental current command and the q-axis fundamental current command are set. The d-axis fundamental current error is obtained by subtracting the d-axis fundamental current component from the d-axis fundamental current command, and the q-axis fundamental current error is obtained by subtracting the q-axis fundamental current component from the q-axis fundamental current command. The d-axis fundamental current error is multiplied by the first energy shaping gain to obtain the d-axis energy shaping output, and the q-axis fundamental current error is multiplied by the second energy shaping gain to obtain the q-axis energy shaping output. The d-axis fundamental current component is multiplied by the first damping injection gain to obtain the d-axis damping injection output, and the q-axis fundamental current component is multiplied by the second damping injection gain to obtain the q-axis damping injection output. The d-axis energy shaping output and the d-axis damping injection output are added to obtain the d-axis Hamiltonian control output, and the q-axis energy shaping output and the q-axis damping injection output are added to obtain the q-axis Hamiltonian control output.

[0013] Furthermore, in step 2, the 5th harmonic d-axis current component is used as the 5th d-axis sliding surface variable. The sign of the 5th d-axis sliding surface variable is determined: when the 5th d-axis sliding surface variable is greater than zero, the sign output is positive 1; when the 5th d-axis sliding surface variable is less than zero, the sign output is negative 1; and when the 5th d-axis sliding surface variable is equal to zero, the sign output is zero. The sign output is multiplied by the first superspiral gain and then integrated over time to obtain the 5th d-axis superspiral integral term. The 5th d-axis sliding surface... The square root of the absolute value of the variable is then multiplied by the sign output of the 5th d-axis sliding surface variable and the second superspiral gain to obtain the 5th d-axis superspiral proportional term. The 5th d-axis superspiral integral term is added to the 5th d-axis superspiral proportional term to obtain the 5th d-axis superspiral sliding output. The 5th q-axis superspiral sliding output, 7th d-axis superspiral sliding output, 7th q-axis superspiral sliding output, 11th superspiral sliding output, and 13th superspiral sliding output are obtained in the same way.

[0014] Further, in step 2, the 5th d-axis superspiral sliding mode output and the 5th q-axis superspiral sliding mode output are inversely rotated using a negative 5 times the grid phase angle as the inverse transformation angle to obtain the alpha-axis and beta-axis components of the 5th harmonic compensation voltage; the 7th d-axis superspiral sliding mode output and the 7th q-axis superspiral sliding mode output are inversely rotated using a positive 7 times the grid phase angle as the inverse transformation angle to obtain the alpha-axis and beta-axis components of the 7th harmonic compensation voltage; the alpha-axis and beta-axis components of the 11th and 13th harmonic compensation voltages are obtained in the same way; the alpha-axis and beta-axis components of the 5th harmonic compensation voltage and the alpha-axis components of the 7th harmonic compensation voltage are then combined. The alpha and beta components of the harmonic compensation voltage are superimposed in a two-phase stationary coordinate system to obtain the total alpha and beta components of the harmonic compensation voltage. The d-axis Hamiltonian control output and the q-axis Hamiltonian control output are inversely transformed using the grid phase angle as the inverse transformation angle to obtain the alpha and beta components of the fundamental control voltage. The alpha component of the fundamental control voltage is added to the total alpha component of the harmonic compensation voltage to obtain the alpha component of the composite modulation voltage. The beta component of the fundamental control voltage is added to the total alpha component of the harmonic compensation voltage to obtain the beta component of the composite modulation voltage.

[0015] Further, in step 3, the alpha and beta components of the composite modulation voltage are input to the space vector modulation module to determine the sector number and triangle region number of the voltage vector, and three adjacent basic voltage vectors are selected. For basic voltage vectors with redundant switching states, the polarity of the midpoint potential deviation signal and the current phase current is read. When the midpoint potential deviation signal is positive and the current phase current is positive, a redundant switching state that causes current to flow out of the midpoint is selected; when the midpoint potential deviation signal is positive and the current phase current is negative, a redundant switching state that causes current to flow into the midpoint is selected. When the signal is negative and the current phase current is positive, a redundant switching state that allows current to flow into the midpoint is selected. When the midpoint potential deviation signal is negative and the current phase current is negative, a redundant switching state that allows current to flow out of the midpoint is selected. The action time of each basic voltage vector is calculated according to the volt-second balance principle, and gate drive pulses for four switching devices are generated. The two outer switching devices adopt a low-frequency switching mode, and the two inner switching devices adopt a high-frequency switching mode. Dead time is inserted between the conduction pulses of adjacent switching devices. After dead time compensation according to the phase current polarity, the output is sent to each switching device of the active midpoint clamping three-level inverter.

[0016] The harmonic suppression and adaptive power quality control method for energy storage converters described in this invention has the following advantages: This invention employs a second-order generalized integrator to extract the fundamental component. Compared to traditional low-pass filtering methods, the second-order generalized integrator can achieve rapid separation of the fundamental and harmonic components without introducing phase delay, significantly improving the dynamic response speed of harmonic detection. Simultaneously, by using multiple synchronous rotating coordinate transformations to extract each harmonic current component, each harmonic is represented as a DC quantity in its corresponding synchronous coordinate system, facilitating subsequent independent control and avoiding mutual interference between different harmonic orders.

[0017] This invention employs a port-controlled Hamiltonian energy shaping method, designing the control law from the perspective of system energy flow. The energy shaping stage alters the shape of the system's energy function, making the desired operating point the minimum point of the energy function, thus ensuring system convergence towards the desired operating point. The damping injection stage injects virtual damping into the system, accelerating energy dissipation and improving the convergence speed. This energy-based control method has clear physical meaning and inherent Lyapunov stability guarantees, exhibiting stronger robustness compared to traditional proportional-integral control, and can adapt to changes in system parameters and a wide range of operating point adjustments.

[0018] This invention employs a superspiral sliding mode control method. As a second-order sliding mode control method, the superspiral algorithm introduces an integral element to generate continuous control inputs, effectively eliminating the chattering problem inherent in traditional first-order sliding mode control and avoiding the additional stress caused by high-frequency chattering on switching devices and filters. Simultaneously, the superspiral sliding mode control maintains the strong robustness of sliding mode control to parameter perturbations and external disturbances, and can still maintain good harmonic suppression under conditions such as changes in grid impedance or sudden load changes. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the single-phase bridge arm topology of an active neutral-point clamped three-level inverter provided in an embodiment of the present invention; Figure 2 A comparison diagram of the output current waveform of the energy storage converter provided in the embodiments of the present invention; Figure 3 A comparison diagram of the output current spectrum of the energy storage converter provided in an embodiment of the present invention; Figure 4 This is a diagram illustrating the effect of midpoint potential balance control provided in an embodiment of the present invention. Detailed Implementation

[0020] The harmonic suppression and adaptive power quality control method for energy storage converters includes the following steps: Step 1: Multi-channel signal acquisition and harmonic component separation: Acquire the output current signal, output voltage signal, and DC bus capacitor voltage signal of the active neutral clamp three-level inverter. Perform coordinate transformation on the output current signal to obtain the current component in the two-phase stationary coordinate system. Obtain the fundamental current component and each harmonic current component through fundamental wave separation and harmonic extraction. Obtain the grid phase angle through the phase-locked loop module. Step 2: Port-controlled Hamiltonian energy shaping and super-helical sliding mode fusion control: Construct a port-controlled Hamiltonian system framework with bridge arm inductor current and capacitor voltage as state variables. The fundamental current error is controlled by energy shaping and damping injection to obtain the fundamental control voltage. The harmonic compensation voltage is obtained by super-helical sliding mode control of each harmonic current component. The fundamental control voltage and the harmonic compensation voltage are superimposed to obtain the composite modulation voltage. Step 3, Inverter Modulation and Active Neutral Point Potential Balancing: The composite modulation voltage is spatially vector modulated to determine the basic voltage vector. The redundant switch states are selected based on the midpoint potential deviation signal and phase current polarity to achieve midpoint potential balancing. The gate drive pulse is then generated to drive the active midpoint clamped three-level inverter.

[0021] During grid-connected operation, energy storage converters inevitably contain harmonic components in their output current due to the nonlinear switching characteristics of power devices and the nonlinear electrical equipment on the load side. If these harmonic components are not suppressed, they will lead to problems such as grid voltage distortion, equipment overheating, and malfunctioning protection devices. To achieve precise harmonic suppression, it is first necessary to acquire the electrical signals of the energy storage converter with high precision and accurately separate the fundamental component from each harmonic component.

[0022] refer to Figure 1 The DC side of the active neutral-clamped three-level inverter consists of two electrolytic capacitors connected in series, and the upper bus capacitor... Connected between the positive terminal of the DC bus and the midpoint O, the lower half bus capacitor It is connected between the midpoint O and the negative terminal of the DC bus. Under normal operating conditions, the voltage of the upper half bus capacitor is... With the voltage of the lower half bus capacitor They should be kept equal, each being half of the total DC bus voltage. Each phase arm contains four power switching devices, sequentially labeled as the first switching device from the positive to the negative DC bus terminal. , second switching device , third switching device and the 4th switching device First switching device and the 4th switching device Located on the outer side of the bridge arm, it withstands high voltage stress during switching; the second switching device. and the third switching device Located on the inner side of the bridge arm, it is connected to the midpoint clamping branch. The key difference between the active midpoint clamping topology and the traditional midpoint clamping topology lies in the use of a controllable switching device instead of a clamping diode, i.e., the fifth switching device. and the 6th switching device Fifth switching device One end is connected to the midpoint O of the DC bus, and the other end is connected to the first switching device. With the second switching device The nodes between; the 6th switching device One end is connected to the midpoint O of the DC bus, and the other end is connected to the third switching device. With the 4th switching device The nodes between them. Each power switching device has a freewheeling diode connected in anti-parallel to provide a freewheeling path for the inductor current when the switching device is turned off. The output of the bridge arm is located at the second switching device. With the third switching device Between, via bridge arm inductance Connected to the filter capacitor Filter capacitor With grid-side inductor Together, they form an LCL-type filter, used to filter out high-frequency switching harmonics in the inverter output voltage, making the output current waveform closer to a sine wave. (Grid-side inductor) The other end is connected to the power grid. When the first switching device... and the second switching device When both are simultaneously on, the bridge arm outputs a positive level; when the second switching device... and the third switching device When both are simultaneously on, the bridge arm outputs a zero level; when the third switching device... and the 4th switching device When both are simultaneously on, the bridge arm outputs a negative level. The active midpoint clamping topology allows for flexible control of the fifth switching device. and the 6th switching device The turn-on timing can enable active selection of the commutation path, thereby evenly distributing switching losses between the outer and inner switching devices and extending the service life of power devices.

[0023] In the signal acquisition stage, a closed-loop Hall current sensor is used to measure the three-phase output current of the active neutral-point clamped three-level inverter. The closed-loop Hall current sensor features fast response and high linearity, with a bandwidth exceeding 100 kHz, enabling accurate capture of high-frequency harmonic components in the current. After the three-phase output current signal is converted into a voltage signal by the sensor, it undergoes digital processing via an analog-to-digital converter. The sampling frequency is set to 20 kHz, which is 400 times the power frequency of 50 Hz. According to the Nyquist sampling theorem, this frequency can accurately acquire harmonic components up to the 40th order without spectral aliasing. In another implementation, the sampling frequency can be increased to 40 kHz to achieve even higher harmonic detection accuracy.

[0024] For DC-side voltage acquisition, voltage sensors are installed on both the upper and lower half of the bus capacitor. The voltage of the upper half bus capacitor is denoted as... The voltage of the lower half bus capacitor is denoted as ,in This represents the voltage value between the positive terminal and the midpoint of the DC bus. This represents the voltage value between the midpoint and the negative terminal of the DC bus. The voltage of the upper half of the bus capacitor... With the voltage of the lower half bus capacitor Subtracting the values ​​yields the midpoint potential deviation signal. A positive midpoint potential deviation signal indicates that the voltage of the upper bus capacitor is higher than that of the lower bus capacitor, and the midpoint potential is lower than the ideal value; conversely, a negative midpoint potential deviation signal indicates the opposite. This midpoint potential deviation signal will be used in subsequent modulation stages to guide the selection of the redundancy vector, ensuring DC bus capacitor voltage balance. In typical applications, the total DC bus voltage is 750 volts, and the nominal values ​​of both the upper and lower bus capacitor voltages are 375 volts. The absolute value of the midpoint potential deviation signal should be controlled within 10 volts to ensure the symmetry of the inverter output voltage waveform.

[0025] After signal acquisition, the three-phase output current signals need to be transformed from a three-phase stationary coordinate system to a two-phase stationary coordinate system. This transformation process is called the Clarke transform. In a three-phase current system, there is a constraint relationship between the three-phase currents, meaning the sum of their instantaneous values ​​is always equal to zero. Therefore, only two of the three current variables are independent. The Clarke transform converts the three-phase currents into two orthogonal current components through linear combination, thus simplifying subsequent analysis and control. Let the three-phase output current signals be... , , ,in This represents the instantaneous value of the output current of phase a. This represents the instantaneous value of the output current of phase b. This represents the instantaneous value of the c-phase output current. The alpha-axis current component is obtained after the Clarke transform. and beta-axis current components ,in Coaxial with phase a current, Advanced The phase is 90 degrees. The transformation relationship is as follows: ; In the above transformation formula, It is directly equal to the phase a current, and It consists of a linear combination of phase a current and phase b current, with coefficients... Ensure that the power of the transformed two-phase system is equal to that of the three-phase system. In another equivalent implementation, an equal-amplitude transformation can be used, in which case the coefficient is adjusted to... The transformed current amplitude is equal to the phase current amplitude, which facilitates an intuitive understanding of the physical quantity.

[0026] After obtaining the current components in the two-phase stationary coordinate system, it is necessary to separate the fundamental positive-sequence component and harmonic components. While traditional low-pass filtering methods are simple, they suffer from phase delay and slow dynamic response, making it difficult to meet the rapid power regulation requirements of energy storage converters. Second-order generalized integrators can extract specific frequency components without introducing phase delay, and are therefore adopted as the core method for fundamental separation.

[0027] The working mechanism of the second-order generalized integrator is based on the idea of ​​adaptive notch filtering. It incorporates the alpha axis current component. The input signal is fed into a first-order second-order generalized integrator, which internally contains a bandpass filter with its center frequency set to the fundamental frequency of the power grid, i.e., 50 Hz. After processing the input signal, the first-order second-order generalized integrator outputs two signals: a first in-phase component and a first quadrature component. The first in-phase component is in phase with the fundamental component of the input signal and has the same amplitude, while the first quadrature component lags the first in-phase component by 90 degrees. This characteristic of simultaneously outputting in-phase and quadrature components is a key advantage of the second-order generalized integrator compared to a conventional bandpass filter, providing the necessary phase information for subsequent positive and negative sequence separation.

[0028] Similarly, the beta-axis current component The input is fed into the second-order generalized integrator to obtain the second in-phase component and the second quadrature component. The second in-phase component is in phase with the fundamental component of the beta-axis current component, and the second quadrature component lags the second in-phase component by 90 degrees.

[0029] The resonant characteristics of a second-order generalized integrator are determined by its internal parameters. Let the gain coefficient of the second-order generalized integrator be... This coefficient determines the filter's bandwidth and dynamic response speed. The larger the value, the wider the bandwidth and the faster the dynamic response, but the ability to suppress harmonics decreases. A smaller value results in a narrower bandwidth and enhanced harmonic suppression, but also a slower dynamic response. In engineering practice, The value is typically set to 1.414, a value that achieves a good balance between dynamic response and harmonic suppression. In applications with significant grid frequency fluctuations, The value can be appropriately increased to 2.0 to improve the ability to track frequency changes.

[0030] After extracting the fundamental component, the positive and negative sequence components are separated using their phase relationship. In a three-phase balanced system, the positive sequence component corresponds to the normal phase rotation direction, while the negative sequence component corresponds to the opposite rotation direction. Negative sequence components are generated when the grid voltage is unbalanced; if not separated, this will affect the accuracy of the control system. The positive sequence alpha axis current component is obtained by adding the first in-phase component to the second quadrature component, and the positive sequence beta axis current component is obtained by subtracting the second in-phase component from the first quadrature component. This cross-operation is based on the fact that the beta axis signal of the positive sequence component lags the alpha axis signal by 90 degrees, while the beta axis signal of the negative sequence component leads the alpha axis signal by 90 degrees. Through these addition and subtraction operations, the positive sequence component is enhanced, and the negative sequence component is canceled out.

[0031] The harmonic alpha-axis current components and harmonic beta-axis current components are obtained by subtracting the positive-sequence fundamental component from the original current components. Specifically, the alpha-axis current components are... Subtracting the positive-sequence alpha-axis current component from the positive-sequence alpha-axis current component yields the harmonic alpha-axis current component; the beta-axis current component... Subtracting the positive-sequence beta-axis current component from the positive-sequence beta-axis current component yields the harmonic beta-axis current component. The physical meaning of this subtraction operation is: the original current contains the fundamental frequency and harmonics, and the remaining component after subtracting the fundamental frequency is the harmonic component.

[0032] Accurate acquisition of the grid phase angle is crucial for subsequent coordinate transformation. The positive-sequence alpha-axis current component and the positive-sequence beta-axis current component are input into a phase-locked loop (PLL), which outputs the grid phase angle. ,in This represents the angle between the voltage vector of phase a of the power grid and the alpha axis, with a value ranging from 0 to... The phase angle varies periodically between radians. The working principle of a phase-locked loop (PLL) is based on closed-loop feedback regulation: when there is a deviation between the estimated phase angle inside the PLL and the actual phase angle, a phase error signal is generated. This error signal is processed by a proportional-integral (PI) regulator to adjust the frequency estimation of the PLL, ultimately causing the estimated phase to converge to the actual phase. At a grid frequency of 50 Hz, the grid phase angle increases at a rate of 314.159 radians per second. The locking time of the PLL is typically within 20 milliseconds, enabling rapid tracking of changes in the grid phase.

[0033] Based on the grid phase angle, the positive-sequence alpha-axis current component and the positive-sequence beta-axis current component are transformed to a synchronous rotating coordinate system using the Park transformation to obtain the d-axis fundamental current component. and q-axis fundamental current component The core idea of ​​the Parker transform is to establish a coordinate system that rotates synchronously with the fundamental voltage vector of the power grid. In this coordinate system, the fundamental component is represented as a DC component, which facilitates zero steady-state error control using a simple proportional-integral controller. The transformation relationship is as follows: ; In the above formula, This represents the positive-sequence alpha-axis current component. Represents the positive-sequence beta-axis current component. Indicates the phase angle of the power grid. and These are the cosine and sine values ​​of the grid phase angle, respectively. The d-axis fundamental current component. Corresponding to the active current, its value is proportional to the active power output of the energy storage converter; q-axis fundamental current component Corresponding to reactive current, its value is proportional to the reactive power output of the energy storage converter. This can be achieved by adjusting... and With a given value, independent control of the active and reactive power of the energy storage converter can be achieved.

[0034] For harmonic component extraction, a multiple synchronous rotating coordinate transformation method is employed. The harmonics in the output current of the energy storage converter are predominantly odd harmonics, with the 5th, 7th, 11th, and 13th harmonics being the most significant. Different harmonic orders have different rotation directions and velocities: the 5th harmonic is a negative-sequence component, rotating in the opposite direction to the fundamental frequency, with a rotational angular velocity five times that of the fundamental frequency; the 7th harmonic is a positive-sequence component, rotating in the same direction as the fundamental frequency, with a rotational angular velocity seven times that of the fundamental frequency; the 11th harmonic is a negative-sequence component; and the 13th harmonic is a positive-sequence component.

[0035] The harmonic alpha-axis current component and harmonic beta-axis current component are input into the 5th harmonic rotation transformer, and the transformation angle is taken as the grid phase angle. Five times negative, that is The reason for using a negative transformation angle is that the 5th harmonic is a negative sequence component, and its rotation direction is opposite to that of the fundamental wave. In a coordinate system with a rotation angle, the 5th harmonic component appears as a DC quantity, while other harmonic components still appear as AC quantities. By using a low-pass filter with a cutoff frequency of 10 Hz, the 5th harmonic component can be extracted from the mixed signal, yielding the 5th harmonic d-axis current component and the 5th harmonic q-axis current component. Setting the cutoff frequency of the low-pass filter to 10 Hz effectively filters out interference from other frequency components while ensuring sufficient dynamic response speed. In applications requiring higher dynamic performance, the cutoff frequency can be appropriately increased to 20 Hz.

[0036] The harmonic alpha-axis current component and harmonic beta-axis current component are input into the 7th harmonic rotation transformer, and the transformation angle is taken as the grid phase angle. 7 times that, The 7th harmonic is a positive-sequence component, and its rotation direction is the same as the fundamental frequency; therefore, a positive transformation angle is used. After coordinate transformation and low-pass filtering, the d-axis current component and q-axis current component of the 7th harmonic are obtained.

[0037] The transformation angle for the 11th harmonic rotation transformation is taken as -11 times the grid phase angle, and the transformation angle for the 13th harmonic rotation transformation is taken as +13 times the grid phase angle. Following the same processing procedure, the 11th harmonic current component and the 13th harmonic current component are extracted respectively.

[0038] In certain application scenarios, the number of harmonics to be extracted can be increased or decreased depending on the actual harmonic distribution. For example, when the load harmonic characteristics are mainly 6-pulse rectification, the main harmonics are the 5th and 7th harmonics, and only these two harmonics can be extracted to simplify the control algorithm; when the load harmonic characteristics are complex, the extraction of higher harmonics such as the 17th and 19th can be increased.

[0039] The control objectives of energy storage converters encompass two levels: first, to achieve rapid and accurate tracking of the fundamental current, ensuring that active and reactive power are output according to commands; and second, to suppress harmonic components in the output current, improving power quality. Traditional proportional-integral (PI) control methods can achieve zero steady-state error tracking when handling the fundamental current, but their effectiveness in suppressing harmonic currents is limited, and they lack robustness when system parameters change. Port-controlled Hamiltonian control methods design control laws from an energy perspective, possessing clear physical meaning and good stability guarantees; while super-spiral sliding mode control methods exhibit strong robustness against parameter perturbations and external disturbances, and can generate continuous control quantities, avoiding the chattering problem of traditional sliding mode control. Combining these two control methods can balance the stability of fundamental current control and the robustness of harmonic current suppression.

[0040] When constructing a port-controlled Hamiltonian system framework, it is necessary to select appropriate state variables to describe the system's energy storage state. The energy storage elements of an active neutral-point clamped three-level inverter include the bridge arm inductor, filter capacitor, and DC bus capacitor. The bridge arm inductor current, filter capacitor voltage, and DC bus capacitor voltage are used as system state variables. Let the bridge arm inductance value be... ,in This represents the inductance of each phase arm, typically 3 millihenries; assuming the filter capacitor value is... ,in This represents the capacitance of the filter capacitor, typically 10 microfarads; let the DC bus capacitance be... ,in This indicates the capacitance of a single DC bus capacitor, typically 4700 microfarads.

[0041] The total system energy function is the sum of the energy stored in each energy storage element. The magnetic field energy stored in the bridge arm inductor is proportional to the square of the inductor current; the electric field energy stored in the filter capacitor is proportional to the square of the capacitor voltage; and the electric field energy stored in the DC bus capacitor is proportional to the square of the capacitor voltage. In the port-controlled Hamiltonian system framework, the energy function, as a candidate Lyapunov function, determines the system's stability through its rate of change with time. When the time derivative of the energy function is negative, the system energy continuously decreases, and the state variables will converge to the equilibrium point.

[0042] The core idea of ​​energy shaping control is to change the shape of the system's energy function through control actions, so that the desired operating point becomes the minimum point of the energy function. This involves setting the d-axis fundamental current command. and q-axis fundamental current command ,in It is calculated by the upper-level power control loop based on the active power command. It is calculated by the upper-level power control loop based on the reactive power command. When the energy storage converter operates as a power factor correction device, It is usually set to zero to achieve unity power factor operation.

[0043] The d-axis fundamental wave current component With d-axis fundamental current command Subtraction yields the d-axis fundamental current error. ,Right now The q-axis fundamental wave current component With q-axis fundamental current command Subtraction yields the q-axis fundamental current error. ,Right now Current error reflects the deviation between the actual current and the desired current, and the design goal of the control law is to make the current error approach zero.

[0044] For the control of the fundamental current, two control loops are designed: energy shaping and damping injection. The function of the energy shaping loop is to shape the system's energy function to a form that minimizes the desired operating point. This reduces the d-axis fundamental current error. Multiplied by the first energy shaping gain The d-axis energy shaping output is obtained, where This represents the gain coefficient for d-axis energy shaping control, and its physical meaning is similar to the proportional gain of a proportional controller, typically ranging from 50 to 200. The q-axis fundamental current error... Multiplied by the second energy shaping gain The q-axis energy shaping output is obtained, where This represents the gain coefficient for q-axis energy shaping control, and its value is related to... The same or similar. Energy shaping output provides the driving force to move the system state toward the desired operating point.

[0045] The role of damping injection is to inject virtual damping into the system, accelerating energy dissipation and thus speeding up the convergence of the state variables to the desired operating point. Without damping injection, although the system will eventually converge to the desired operating point under energy shaping, oscillations may occur. The d-axis fundamental current component... Multiply by the first damped injection gain The d-axis damped injection output is obtained, where This represents the d-axis damping injection coefficient, which is physically equivalent to a virtual resistor connected in series in the circuit, typically ranging from 0.5 to 5 ohms. The q-axis fundamental current component... Multiply by the second damped injection gain The q-axis damped injection output is obtained, where This represents the q-axis damping injection coefficient. The selection of the damping injection coefficient requires a trade-off between convergence speed and anti-interference capability: an excessively large damping coefficient will lead to a sluggish response to changes in commands, while an excessively small damping coefficient may cause oscillations.

[0046] refer to Figure 2The figure includes the current waveform before and after control. The current waveform before control shows the output current of the energy storage converter without the harmonic suppression method of this invention. The horizontal axis represents time in milliseconds, ranging from 0 to 60 milliseconds, covering three complete power frequency cycles. The vertical axis represents the current amplitude, expressed in per-unit values, ranging from -1.5 to +1.5. The solid line in the figure represents the actual output current waveform, and the dashed line represents the ideal fundamental current waveform. It can be observed from the current waveform before control that there is a significant deviation between the actual output current and the ideal fundamental current, and the current waveform exhibits obvious distortion characteristics. The waveform distortion is particularly significant near the current zero-crossing point and near the current peak, which is the result of the superposition of low-order harmonics such as the 5th, 7th, 11th, and 13th harmonics on the fundamental wave. The presence of these harmonic currents will have an adverse effect on the grid voltage quality and may cause additional losses and overheating in electrical equipment such as transformers and motors. The current waveform after control demonstrates the output current status of the energy storage converter after adopting the port-controlled Hamiltonian energy shaping and super-spiral sliding mode fusion control method of this invention. It can be observed from the current waveform after control that the actual output current closely matches the ideal fundamental wave, and the waveform distortion is significantly improved. Near the current zero-crossing point and near the current peak, the current waveform after control smoothly tracks the ideal fundamental wave trajectory, without the obvious step and glitches seen before control. This indicates that the super-spiral sliding mode control method used in this invention can effectively suppress harmonic current components, making the output current approach a pure sine wave. The super-spiral sliding mode control, by introducing an integral element to generate continuous control quantity, avoids the high-frequency chattering problem of traditional sliding mode control; therefore, the current waveform after control is smooth and free of high-frequency oscillations. The port-controlled Hamiltonian energy shaping control ensures rapid and accurate tracking of the fundamental current, ensuring that both the current amplitude and phase meet the commanded values.

[0047] The d-axis energy shaping output is added to the d-axis damped injection output to obtain the d-axis Hamiltonian control output; the q-axis energy shaping output is added to the q-axis damped injection output to obtain the q-axis Hamiltonian control output. The Hamiltonian control output represents the fundamental control voltage component in the synchronous rotating coordinate system.

[0048] To suppress harmonic currents, a superspiral sliding mode control method is employed. The basic idea of ​​sliding mode control is to design a sliding surface, and through control actions, drive the system state to reach and maintain it on the sliding surface. Traditional first-order sliding mode control suffers from chattering, where the control quantity switches at high frequencies near the sliding surface, generating undesirable high-frequency oscillations. Superspiral sliding mode control, a second-order sliding mode control, introduces an integral term to generate a continuous control quantity, effectively eliminating chattering while maintaining the strong robustness of sliding mode control to uncertainties.

[0049] The design process of superspiral sliding mode control is illustrated using the suppression of the 5th harmonic d-axis current component as an example. Since the expected value of the harmonic current is zero, the 5th harmonic d-axis current component itself serves as the 5th d-axis sliding surface variable, denoted as... .when When the current is equal to zero, the 5th harmonic d-axis current is completely suppressed, and the system is on the sliding surface.

[0050] The superspiral sliding mode control law consists of two parts: an integral term and a proportional term. The integral term provides continuous control to overcome unknown disturbances and parameter uncertainties in the system. For the 5th order d-axis sliding surface variable... Perform sign determination: when When the value is greater than zero, the sign output is positive 1; when... When less than zero, the sign output is negative 1; when When the sum is zero, the sign output is zero. Multiply the sign output by the first superspiral gain. Time integration was then performed, yielding a 5th d-axis superspiral integral term. The first superspiral gain... The rate of change of the integral term is determined, and its value needs to satisfy certain conditions to ensure the finite-time convergence of the system. According to the stability conditions of the superspiral algorithm, It should be greater than the upper bound of the system uncertainty boundary. In energy storage converter applications, The typical value ranges from 1000 to 5000.

[0051] The proportional term provides immediate corrective control based on the degree of deviation of the sliding surface variable. For the 5th order d-axis sliding surface variable... Taking the square root of the absolute value, we get... The square root operation automatically reduces the control gain near the sliding surface, which helps suppress chattering. Multiplied by 5 times the d-axis sliding surface variable The symbolic output and the second superspiral gain The fifth d-axis superhelical scaling term was obtained. The second superhelical gain... This determines the strength of the proportional term, and its value must also satisfy stability conditions; typical values ​​range from 50 to 200. Gain and The selection of [the elements] involves a coupling relationship, and is usually based on [the principle that]... The initial design is based on empirical formulas, followed by fine-tuning according to the response characteristics of the actual system. The 5th d-axis superhelical integral term is added to the 5th d-axis superhelical proportional term to obtain the 5th d-axis superhelical sliding mode output. This output represents the control voltage required to suppress the 5th harmonic d-axis current component in a 5th harmonic synchronous rotating coordinate system.

[0052] refer to Figure 3The current spectrum includes the current spectrum before and after control. The horizontal axis represents the harmonic order, from the fundamental to the 19th harmonic; the vertical axis represents the percentage of each harmonic current amplitude relative to the fundamental current amplitude, in percentage form. The current spectrum before control shows the distribution of each harmonic current without the harmonic suppression method of this invention. The fundamental current amplitude is set to 100% as a reference. The 5th harmonic current amplitude is 15% of the fundamental current, which is the most abundant harmonic component in the output current of the energy storage converter. The 7th harmonic current amplitude is 10% of the fundamental current, ranking second. The 11th harmonic current amplitude is 6% of the fundamental current, and the 13th harmonic current amplitude is 4% of the fundamental current. The 17th and 19th harmonic current amplitudes are 2% and 1.5% of the fundamental current, respectively. The total harmonic distortion rate of the current calculated based on the harmonic content is 19.2%, far exceeding the 5% limit requirement stipulated by the national standard. The current spectrum after control shows the distribution of each harmonic current after using the method of this invention. The fundamental current amplitude remains at 100%. The amplitude of the 5th harmonic current is significantly reduced to 2% of the fundamental current, a decrease of 86.7% compared to before control. The amplitude of the 7th harmonic current is reduced to 1.5% of the fundamental current, a decrease of 85% compared to before control. The amplitude of the 11th harmonic current is reduced to 1% of the fundamental current, and the amplitude of the 13th harmonic current is reduced to 0.8% of the fundamental current. The amplitudes of the 17th and 19th harmonic currents are reduced to 0.5% and 0.3% of the fundamental current, respectively. The total harmonic distortion rate of the current calculated based on the harmonic content after control is 2.8%, meeting the 5% limit requirement stipulated by the national standard, and the power quality is significantly improved. This invention achieves independent and precise suppression of the main harmonic components such as the 5th, 7th, 11th, and 13th harmonics by designing a super-helical sliding mode controller in a synchronous rotating coordinate system for each harmonic. Each harmonic is represented as a DC quantity in its corresponding synchronous coordinate system, which facilitates zero steady-state error adjustment by the controller. This is the key reason for the significant harmonic suppression effect.

[0053] Following the same design methodology, superspiral sliding mode controllers were designed for the 5th harmonic q-axis current component, the 7th harmonic d-axis current component, the 7th harmonic q-axis current component, the 11th harmonic current component, and the 13th harmonic current component, respectively, resulting in 5th q-axis superspiral sliding mode outputs, 7th d-axis superspiral sliding mode outputs, 7th q-axis superspiral sliding mode outputs, 11th superspiral sliding mode outputs, and 13th superspiral sliding mode outputs. The superspiral gain for different harmonic orders can be set according to the suppression priority and control difficulty of each harmonic. Generally, lower-order harmonics have a larger content and are easier to suppress, so a smaller gain can be used; higher-order harmonics have a smaller content but change more rapidly, so a larger gain can be used to improve the response speed.

[0054] The superhelical sliding mode outputs of each harmonic lie in their respective harmonic synchronous rotating coordinate systems, requiring inverse rotation transformation back to the two-phase stationary coordinate system. The 5th d-axis superhelical sliding mode outputs and the 5th q-axis superhelical sliding mode outputs are then compared using the grid phase angle. Using a negative 5 times the inverse transformation angle, an inverse rotation transformation is performed to obtain the alpha-axis and beta-axis components of the 5th harmonic compensation voltage. The inverse rotation transformation is the inverse operation of the forward rotation transformation, and its transformation angle is the opposite of the forward transformation angle. Since the forward transformation angle of the 5th harmonic is... Therefore, the inverse transformation angle is also... (Rotation in the opposite direction is called forward rotation) (The inverse of). The transformation relationship is as follows: ; In the above formula, This indicates the output of 5 d-axis super spiral sliding mode cycles. This indicates 5 cycles of q-axis super-spiral sliding mode output. The alpha axis component represents the 5th harmonic compensation voltage. This represents the beta-axis component of the 5th harmonic compensation voltage.

[0055] The 7th d-axis superspiral sliding mode output and the 7th q-axis superspiral sliding mode output are used with the grid phase angle. Using a positive 7 times the inverse transformation angle, an inverse rotation transformation is performed to obtain the alpha-axis and beta-axis components of the 7th harmonic compensation voltage. The alpha-axis and beta-axis components of the 11th and 13th harmonic compensation voltages are obtained in the same manner.

[0056] The alpha-axis components of the 5th, 7th, 11th, and 13th harmonic compensation voltages are superimposed in a two-phase stationary coordinate system to obtain the total alpha-axis component of the harmonic compensation voltage. Similarly, the beta-axis components of the 5th, 7th, 11th, and 13th harmonic compensation voltages are superimposed in a two-phase stationary coordinate system to obtain the total beta-axis component of the harmonic compensation voltage.

[0057] The d-axis Hamiltonian control output and the q-axis Hamiltonian control output are set at the grid phase angle. The inverse Parker transformation is performed using the inverse transformation angle to obtain the alpha and beta components of the fundamental control voltage. The inverse Parker transformation is the inverse operation of the Parker transformation, converting the voltage components in the synchronous rotating coordinate system back to the two-phase stationary coordinate system.

[0058] The alpha-axis component of the fundamental control voltage is added to the total alpha-axis component of the harmonic compensation voltage to obtain the alpha-axis component of the composite modulation voltage; the beta-axis component of the fundamental control voltage is added to the total beta-axis component of the harmonic compensation voltage to obtain the beta-axis component of the composite modulation voltage. The composite modulation voltage simultaneously includes the fundamental control component and the harmonic compensation component; the former ensures accurate tracking of the fundamental current, while the latter achieves active suppression of harmonic currents.

[0059] In another implementation, a feedforward decoupling stage can be added to the energy shaping control to compensate for the coupling between the d-axis and q-axis. In a synchronous rotating coordinate system, there is a cross-coupling term between the d-axis and q-axis currents, which is proportional to the product of the grid angular frequency and the inductance value. By subtracting the component of the q-axis current multiplied by the grid angular frequency and inductance value from the d-axis control output, and adding the component of the d-axis current multiplied by the grid angular frequency and inductance value to the q-axis control output, decoupling control of the d-axis and q-axis can be achieved, improving the bandwidth and dynamic response speed of the control system.

[0060] In another implementation, the gain parameter of the superspiral sliding mode controller can be adaptively adjusted based on the estimated grid impedance. A higher grid impedance results in a slower system response to control actions, requiring an increase in control gain to maintain the same response speed; conversely, a lower grid impedance leads to a faster system response, allowing for a reduction in control gain to avoid overreaction. The grid impedance can be estimated by injecting a small test signal and analyzing the response; the estimation period can be set to 1 second.

[0061] Compared to traditional two-level inverters, active midpoint clamped three-level inverters offer more voltage levels in their output voltage, better approximating a sinusoidal waveform and thus reducing harmonic content and voltage stress on switching devices. However, the three-level topology introduces a midpoint potential balance problem: the DC bus is divided by two series-connected capacitors. Uneven voltage distribution between these capacitors can distort the output voltage waveform and, in severe cases, damage power devices. Space vector modulation (SVM) technology, through careful selection of the voltage vector and its duration, can maintain midpoint potential balance while achieving high-quality voltage output.

[0062] The alpha and beta components of the composite modulated voltage together constitute the voltage vector in a two-phase stationary coordinate system. The magnitude and phase of this voltage vector determine the output voltage of the energy storage converter. Let the alpha component of the composite modulated voltage be... The beta-axis component is ,in This represents the projection of the voltage vector onto the alpha axis. This represents the projection of the voltage vector onto the beta axis. The magnitude of the voltage vector. and phase angle This can be determined through the following relationship: ; In the above formula, This represents the amplitude of the composite modulated voltage vector. This represents the angle between the composite modulated voltage vector and the alpha axis. During steady-state operation of the energy storage converter, the amplitude of the voltage vector is basically constant, the phase angle increases uniformly at the grid angular frequency, and the trajectory of the endpoints of the voltage vector forms a circle.

[0063] Each phase arm of a three-level inverter can output three voltage levels: positive, zero, and negative. A positive level indicates that the phase output is connected to the positive terminal of the DC bus, and the output voltage is half the DC bus voltage. A zero level indicates that the phase output is connected to the midpoint of the DC bus, and the output voltage is zero. A negative level indicates that the phase output is connected to the negative terminal of the DC bus, and the output voltage is half the negative DC bus voltage. There are 27 possible voltage level combinations for the three phase arms, corresponding to 27 space voltage vectors. These 27 voltage vectors are categorized into four types based on their amplitude: three zero vectors, six small vectors, six medium vectors, and six large vectors. The zero vector has zero amplitude and is located at the origin of the complex plane; the large vector has the largest amplitude and is located at the vertex of the regular hexagon; the medium vector has a moderate amplitude and is located at the midpoint of the side of the regular hexagon; and the small vector has the smallest amplitude and is located at the vertex of the inner regular hexagon.

[0064] The first step in space vector modulation is to determine the sector containing the voltage vector. Using the alpha axis as a reference, the complex plane is divided into six sectors, each occupying a 60-degree angular range. The sectors are numbered from 1 to 6, with sector 1 ranging from 0 to 60 degrees, sector 2 from 60 to 120 degrees, and so on. This is based on the phase angle of the composite modulated voltage vector. This allows us to determine the sector number it belongs to. : when hour, ;when hour, ;when hour, ;when hour, ;when hour, ;when hour, .

[0065] After determining the sectors, it is necessary to further determine the triangular regions within the sectors where the voltage vectors reside. The spatial vector diagram of a three-level inverter differs from that of a two-level inverter; each sector is divided into four triangular regions. Taking sector 1 as an example, the four triangular regions are located as follows: Region 1 is close to the origin, enclosed by the zero vector and two smaller vectors; Region 2 is located outside Region 1, enclosed by one smaller vector, one medium vector, and one large vector; Regions 3 and 4 are located on either side of the sector, each enclosed by adjacent vectors. The division of the triangular regions is based on the volt-second balance principle, ensuring that the weighted average of the selected vectors equals the reference voltage vector in each switching cycle.

[0066] The triangular region is determined using the following method. First, the composite modulation voltage vector is projected onto the local coordinate system of sector 1, and then... and Two intermediate variables: ; In the above formula, This represents the total voltage of the DC bus. It represents half of the DC bus voltage, that is, the nominal value of the upper half bus capacitor voltage or the lower half bus capacitor voltage. and These are normalized coordinate components, with values ​​ranging from 0 to 1. According to... and Determine the numerical relationship of the triangle region numbering: when and When, the area number is 1; when and and When, the area number is 2; when and and When, the area number is 3; when At that time, the area code was 4.

[0067] After determining the sector number and the triangular region number, three adjacent basic voltage vectors can be selected to synthesize the reference voltage vector. Each vertex of the triangular region corresponds to three basic voltage vectors. By adjusting the duration of these three vectors within one switching cycle, a reference voltage vector located at any position inside the triangle can be equivalently synthesized. This synthesis method is based on the volt-second balance principle: within one switching cycle... Within this range, the sum of the products of the duration of action of each basic vector and its voltage amplitude is equal to the product of the reference voltage vector and the switching period.

[0068] Let the three selected basic voltage vectors be respectively , , Their action times are respectively , , The volt-second equilibrium condition is: ; In the above formula, This represents the reference voltage vector, i.e., the composite modulated voltage vector. This represents the switching period, typically 50 microseconds, corresponding to a switching frequency of 20 kHz. By solving the above system of equations, the duration of each fundamental vector can be obtained. In practical calculations, the reference voltage vector is usually decomposed onto the sides of a selected triangle, and the duty cycle is directly calculated using geometric relationships.

[0069] The key feature of space vector modulation in a three-level inverter lies in the redundant switching states of the small vectors. Each small vector can be implemented by two different switching states, which produce the same output voltage vector but have opposite effects on the midpoint potential. Taking the small vector corresponding to phase a positive level, phase b zero level, and phase c zero level as an example, this vector can be implemented by the switching states POO or ONN, where P represents positive level, O represents zero level, and N represents negative level. In the POO state, when the phase a current is positive, the current flows out from the midpoint, causing the midpoint potential to drop; in the ONN state, when the phase a current is positive, the current flows into the midpoint, causing the midpoint potential to rise. It is this redundancy characteristic that makes active balancing of the midpoint potential possible.

[0070] The polarity of the midpoint potential deviation signal indicates the direction of the current midpoint potential offset from the ideal value. When the midpoint potential deviation signal is positive, it indicates that the voltage of the upper half bus capacitor is higher than that of the lower half bus capacitor, requiring the selection of a redundant switching state that can reduce the voltage of the upper half bus capacitor (or increase the voltage of the lower half bus capacitor); the opposite is true when the midpoint potential deviation signal is negative. The polarity of the phase current determines the direction of current flow through the midpoint under a specific switching state. By comprehensively considering the polarity of the midpoint potential deviation signal and the polarity of the phase current, a selection rule for redundant switching states can be established.

[0071] When the midpoint potential deviation signal is positive and the current phase current is positive, the redundant switching state that allows current to flow out of the midpoint is selected. Current flowing out of the midpoint means the upper half of the bus capacitor is discharging or the lower half of the bus capacitor is charging, thereby reducing the voltage of the upper half of the bus capacitor and increasing the voltage of the lower half of the bus capacitor, thus reducing the midpoint potential deviation. When the midpoint potential deviation signal is positive and the current phase current is negative, the redundant switching state that allows current to flow into the midpoint is selected. When the phase current is negative, the actual direction of the current is opposite to the positive direction; in this case, selecting the state that allows current to flow into the midpoint also has the actual effect of reducing the midpoint potential deviation. When the midpoint potential deviation signal is negative and the current phase current is positive, the redundant switching state that allows current to flow into the midpoint is selected; when the midpoint potential deviation signal is negative and the current phase current is negative, the redundant switching state that allows current to flow out of the midpoint is selected. The selection logic for these four cases can be summarized as follows: when the product of the midpoint potential deviation signal and the phase current is positive, the redundant state that allows current to flow out of the midpoint is selected; when the product is negative, the redundant state that allows current to flow into the midpoint is selected.

[0072] Besides small vectors, zero vectors also have redundant switching states. Zero vectors can be implemented using three switching states: PPP, OOO, and NNN. In the PPP state, all three output terminals are connected to the positive terminal of the DC bus. In the NNN state, all three output terminals are connected to the negative terminal of the DC bus. These two states do not involve the neutral point current and have no effect on the neutral point potential. In the OOO state, all three output terminals are connected to the neutral point, and the neutral point current is the sum of the three phase currents. In a three-phase balanced system, this current is zero, and similarly, it has no effect on the neutral point potential. Therefore, the selection of the redundant state for a zero vector mainly considers minimizing the number of switching operations and balancing losses.

[0073] Each phase arm of an active neutral-point clamped three-level inverter contains four switching devices. These are labeled as switching device 1, switching device 2, switching device 3, and switching device 4, sequentially from the positive to the negative DC bus. Switching devices 1 and 4 are located on the outer side of the arm and bear the main voltage stress during the switching process; switching devices 2 and 3 are located on the inner side of the arm, connected in parallel with the neutral-point clamping diode, and bear less voltage stress. In the active neutral-point clamped topology, the inner switching devices not only perform clamping but also participate in the active commutation process, actively controlling the current commutation path.

[0074] Different voltage levels correspond to different switching device conduction combinations. When the output is positive, switches 1 and 2 are on, while switches 3 and 4 are off. Current flows from the positive terminal of the DC bus through switches 1 and 2 to the output. When the output is zero, there are two conduction modes depending on the current direction: when the current is positive (flowing out of the inverter), switches 2 and 3 are on, and current flows from the midpoint through either switch 2 or 3 to the output; when the current is negative (flowing into the inverter), switches 2 and 3 are on, and current flows from the output through either switch 2 or 3 to the midpoint. When the output is negative, switches 3 and 4 are on, while switches 1 and 2 are off.

[0075] The advantage of active midpoint clamping topology over traditional midpoint clamping topology lies in its ability to flexibly select commutation paths, achieving a balanced distribution of switching losses. In traditional midpoint clamping topology, the outer switching devices bear all switching losses, while the inner switching devices only bear conduction losses, resulting in higher temperature rise and shorter lifespan for the outer devices. Active midpoint clamping topology, by adjusting the commutation timing of the inner switching devices, can transfer some switching losses to the inner devices, making the losses and temperature rise of the four switching devices more balanced.

[0076] The generation of the gate drive pulse is based on the calculated vector action time and the selected switching state. Within a switching cycle, the action sequence of the vectors is arranged according to a seven-segment or five-segment modulation scheme. Seven-segment modulation symmetrically arranges the voltage vectors within each switching cycle, resulting in more switching operations but lower output harmonic content; five-segment modulation reduces the number of switching operations and is suitable for high switching frequencies. Taking seven-segment modulation as an example, assuming the three vectors selected within a certain switching cycle are as follows... (Zero Vector) (Small vector) (For large vectors), the order of vector action is as follows: This forms a symmetrical structure.

[0077] For each phase arm of an active midpoint clamped three-level inverter, gate drive pulses for four switching devices are generated based on the vector action time. The two outer switching devices, namely the first and fourth switching devices, employ a low-frequency switching mode, with their switching action occurring when the voltage level crosses zero, and the switching frequency approximately twice the fundamental frequency, i.e., around 100 Hz. The two inner switching devices, namely the second and third switching devices, employ a high-frequency switching mode, with their switching action occurring at the level switching moments within each modulation cycle, and the switching frequency being comparable to the carrier frequency, i.e., 20 kHz. This differentiated switching frequency allocation utilizes the characteristics of the switching devices in different positions: the outer devices withstand high voltage and use low-frequency switching to reduce switching losses; the inner devices withstand low voltage and use high-frequency switching to achieve fine voltage regulation.

[0078] Dead time must be inserted between the conduction pulses of adjacent switching devices to prevent bridge arm shoot-through short circuits. Dead time This refers to the protection interval between the drive signals of adjacent switching devices, typically ranging from 2 to 5 microseconds. Let the turn-off time of the first switching device be... Therefore, the turn-on time of the second switching device must not be earlier than The dead time setting needs to take into account the turn-off delay time and current drop time of the switching devices, ensuring that the next device only starts to conduct after the previous device has been completely turned off.

[0079] Dead time introduces output voltage errors. During the dead time, the bridge arm output level is determined by the direction of the load current, not the control signal. When the phase current is positive, the output level is clamped to a lower value during the dead time; when the phase current is negative, the output level is clamped to a higher value. This undesirable level offset causes output voltage waveform distortion and generates low-order harmonics. Dead time compensation is required to eliminate the effects of dead time.

[0080] Dead-time compensation is achieved by adjusting the leading edge position of the turn-on pulse based on the polarity of the phase current. The polarity of the current phase current is detected; when the phase current is positive, the leading edge of the turn-on pulse is advanced by half the dead time. When the phase current is negative, the leading edge of the conduction pulse is delayed by half of the dead time. This symmetrical compensation method is based on the fact that the voltage error introduced by the dead time is essentially a change in the conduction pulse width, which can be offset by adjusting the pulse width in the opposite direction. With a dead time of 4 microseconds, the compensation is to advance or delay the leading edge by 2 microseconds.

[0081] Phase current polarity can be detected using either a hardware comparator or software. The hardware method uses a high-speed comparator to compare the current sampling signal with a zero-level signal, outputting a digital signal indicating the current polarity with a response time within 100 nanoseconds. The software method directly uses the current sampling value from the digital control system for sign determination; the response time depends on the sampling period, typically 50 microseconds. Near the current zero-crossing point, polarity determination may fluctuate due to sampling noise. In this case, a dead-time window can be set, maintaining the previous polarity determination result when the absolute current value is less than a threshold. This threshold is typically set to 5% of the rated current.

[0082] After dead-time compensation, the gate drive pulses are transmitted to the switching devices of the active neutral-clamped three-level inverter via optocouplers or magnetic isolators. The isolators provide electrical isolation between the control circuit and the power circuit, preventing damage to the control circuit from high voltage and strong interference in the power circuit. The propagation delay of an optocoupler is approximately 500 nanoseconds, and that of a magnetic isolator is approximately 100 nanoseconds. The delay difference between different isolation channels should be controlled within 50 nanoseconds to ensure precise timing coordination of the switching devices in the same phase arm.

[0083] After the drive pulse reaches the gate of the switching device, the device turns on or off according to the drive signal. During turn-on, the gate voltage rises above the threshold, the device transitions from the blocking state to the conducting state, and current begins to flow through the device. During turn-off, the gate voltage drops below the threshold, the device transitions from the conducting state to the blocking state, and current is cut off. The turn-on and turn-off actions of the switching device, in conjunction with the filtering inductor and capacitor, ultimately generate a three-phase AC power with low harmonic content and optimized power quality at the output of the energy storage converter.

[0084] In another implementation, carrier stacking modulation can be used instead of space vector modulation. Carrier stacking modulation uses two triangular carriers with the same amplitude but different DC biases, which are compared with the modulating wave to generate switching signals. The upper carrier has a positive bias, corresponding to the switching between positive and zero levels; the lower carrier has a negative bias, corresponding to the switching between zero and negative levels. Carrier stacking modulation is relatively simple to implement and suitable for situations with limited hardware resources, but it is less flexible than space vector modulation in terms of midpoint potential balance control.

[0085] refer to Figure 4 This includes the DC bus capacitor voltage waveform and the midpoint potential deviation signal waveform. In the DC bus capacitor voltage waveform diagram, the horizontal axis represents time in milliseconds, ranging from 0 to 200 milliseconds; the vertical axis represents capacitor voltage in volts, ranging from 350 to 400 volts. The blue curve in the diagram represents the upper half of the bus capacitor voltage. The red curve represents the voltage of the lower half bus capacitor. The green dashed line represents the nominal capacitor voltage of 375 volts. During the period from 0 milliseconds to 50 milliseconds, the midpoint potential balance control has not yet been activated, and the upper bus capacitor voltage... The voltage of the lower half bus capacitor is maintained at approximately 390 volts. The voltage remains at approximately 360 volts, with a deviation of about 30 volts between the two values. This imbalance will cause an asymmetrical inverter output voltage waveform. At 50 milliseconds, the midpoint potential balancing control is activated. Subsequently, the upper bus capacitor voltage can be observed... The voltage of the lower half bus capacitor gradually decreases. Gradually rising, the two curves approach the nominal value of 375 volts. At approximately 100 milliseconds, the two capacitor voltages converge to near their nominal values, effectively correcting the midpoint potential deviation. Between 100 and 200 milliseconds, the two capacitor voltages operate stably around 375 volts with only minor fluctuations. In the midpoint potential deviation signal waveform graph, the horizontal axis represents time, and the vertical axis represents the midpoint potential deviation in volts, ranging from -20 volts to +40 volts. The green shaded area represents the allowable deviation range, within ±10 volts. The orange dashed line indicates the deviation limit. Between 0 and 50 milliseconds, the midpoint potential deviation signal remains at approximately 30 volts, exceeding the allowable range. After the midpoint potential balance control is activated at 50 milliseconds, the midpoint potential deviation signal rapidly decreases, entering the allowable range at approximately 80 milliseconds, and then stably remains within ±5 volts for the subsequent time. This invention achieves rapid and active balancing of the midpoint potential by selecting redundant switch states based on the polarity of the midpoint potential deviation signal and the phase current polarity, thus ensuring the symmetry of the output voltage waveform of the three-level inverter.

[0086] In another implementation, the allocation ratio of the redundancy vector can be adaptively adjusted according to the magnitude of the midpoint potential deviation. When the midpoint potential deviation is small, the duration of the two redundancy states is evenly distributed to reduce low-frequency ripple in the output voltage; when the midpoint potential deviation is large, the proportion of the duration of the redundancy state that is conducive to balance is increased to accelerate the recovery of the midpoint potential. Let the midpoint potential deviation signal be... The deviation threshold is ,when At that time, the duration of each of the two redundant states is 50%; when At this time, the proportion of time spent in redundant states that are conducive to balance increases to 80%. Deviation threshold. The typical value is 2% of the DC bus voltage, which is 15 volts at a DC bus voltage of 750 volts.

[0087] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for harmonic suppression and adaptive power quality control of an energy storage converter, characterized in that, Includes the following steps: Step 1: Multi-channel signal acquisition and harmonic component separation: Acquire the output current signal, output voltage signal, and DC bus capacitor voltage signal of the active neutral clamp three-level inverter. Perform coordinate transformation on the output current signal to obtain the current component in the two-phase stationary coordinate system. Obtain the fundamental current component and each harmonic current component through fundamental wave separation and harmonic extraction. Obtain the grid phase angle through the phase-locked loop module. Step 2: Port-controlled Hamiltonian energy shaping and super-helical sliding mode fusion control: Construct a port-controlled Hamiltonian system framework with bridge arm inductor current and capacitor voltage as state variables. The fundamental current error is controlled by energy shaping and damping injection to obtain the fundamental control voltage. The harmonic compensation voltage is obtained by super-helical sliding mode control of each harmonic current component. The fundamental control voltage and the harmonic compensation voltage are superimposed to obtain the composite modulation voltage. Step 3: Inverter Modulation and Active Neutral Point Potential Balancing: Space vector modulation is applied to the composite modulation voltage. The sector number and triangle region number of the voltage vector are determined. Three adjacent basic voltage vectors are selected, where the smaller vectors have redundant switching states. Each smaller vector is implemented by two different switching states, producing the same output voltage vector, but with opposite effects on the neutral point potential. For basic voltage vectors with redundant switching states, the polarity of the neutral point potential deviation signal and the current phase current is read. When the neutral point potential deviation signal is positive and the current phase current is positive, the redundant switching state that causes current to flow out of the neutral point is selected. When the neutral point potential deviation signal is positive and the current phase current is negative, the state that causes current to flow into the neutral point is selected. The system employs redundant switching states for the current. When the midpoint potential deviation signal is negative and the current phase current is positive, a redundant switching state that allows current to flow into the midpoint is selected. When the midpoint potential deviation signal is negative and the current phase current is negative, a redundant switching state that allows current to flow out of the midpoint is selected, thus achieving midpoint potential balance. Based on the volt-second balance principle, the action time of each basic voltage vector is calculated, generating gate drive pulses for four switching devices. The two outer switching devices adopt a low-frequency switching mode, and the two inner switching devices adopt a high-frequency switching mode. Dead time is inserted between the conduction pulses of adjacent switching devices. After dead time compensation based on the phase current polarity, the pulses are output to each switching device of the active midpoint clamping three-level inverter to drive the active midpoint clamping three-level inverter.

2. The method according to claim 1, characterized in that, In step 1, the three-phase output current signal of the active neutral clamp three-level inverter is acquired by the current sensor, the three-phase output voltage signal is acquired by the voltage sensor, and the upper half bus capacitor voltage and the lower half bus capacitor voltage are acquired by the DC side voltage sensor respectively. The upper half bus capacitor voltage and the lower half bus capacitor voltage are subtracted to obtain the neutral point potential deviation signal.

3. The method according to claim 2, characterized in that, In step 1, the three-phase output current signal is input to the Clarke transform module to obtain the alpha-axis current component and the beta-axis current component in the two-phase stationary coordinate system; the alpha-axis current component is input to the first second-order generalized integrator, and the beta-axis current component is input to the second second-order generalized integrator; the first second-order generalized integrator outputs the first in-phase component that is in phase with the alpha-axis current component and the first quadrature component that lags by 90 degrees; the second second-order generalized integrator outputs the second in-phase component that is in phase with the beta-axis current component and the second quadrature component that lags by 90 degrees.

4. The method according to claim 3, characterized in that, In step 1, the first in-phase component is added to the second quadrature component to obtain the positive sequence alpha axis current component, and the second in-phase component is subtracted from the first quadrature component to obtain the positive sequence beta axis current component; the alpha axis current component is subtracted from the positive sequence alpha axis current component to obtain the harmonic alpha axis current component, and the beta axis current component is subtracted from the positive sequence beta axis current component to obtain the harmonic beta axis current component. The positive sequence alpha axis current component and the positive sequence beta axis current component are input to the phase-locked loop module, and the phase-locked loop module outputs the grid phase angle. Based on the grid phase angle, the positive sequence alpha axis current component and the positive sequence beta axis current component are transformed to the synchronous rotating coordinate system to obtain the d-axis fundamental current component and the q-axis fundamental current component.

5. The method according to claim 4, characterized in that, In step 1, the harmonic alpha-axis current component and the harmonic beta-axis current component are input to the 5th harmonic rotation transformation module, the 7th harmonic rotation transformation module, the 11th harmonic rotation transformation module, and the 13th harmonic rotation transformation module, respectively. The 5th harmonic rotation transformation module performs transformation with a transformation angle of -5 times the grid phase angle and extracts the 5th harmonic d-axis current component and the 5th harmonic q-axis current component through a low-pass filter. The 7th harmonic rotation transformation module performs transformation with a transformation angle of +7 times the grid phase angle and extracts the 7th harmonic d-axis current component and the 7th harmonic q-axis current component through a low-pass filter. The 11th harmonic rotation transformation module and the 13th harmonic rotation transformation module extract the 11th harmonic current component and the 13th harmonic current component, respectively, in the same manner.

6. The method according to claim 5, characterized in that, In step 2, the bridge arm inductor current, filter capacitor voltage, and DC bus capacitor voltage of the active midpoint clamped three-level inverter are used as system state variables to construct a port-controlled Hamiltonian system framework. In the port-controlled Hamiltonian system framework, the sum of the magnetic field energy stored in the bridge arm inductor, the electric field energy stored in the filter capacitor, and the electric field energy stored in the DC bus capacitor is used as the total system energy function.

7. The method according to claim 6, characterized in that, In step 2, the d-axis fundamental current command and the q-axis fundamental current command are set. The d-axis fundamental current error is obtained by subtracting the d-axis fundamental current component from the d-axis fundamental current command. The q-axis fundamental current error is obtained by subtracting the q-axis fundamental current component from the q-axis fundamental current command. Multiply the d-axis fundamental current error by the first energy shaping gain to obtain the d-axis energy-shaped output, and multiply the q-axis fundamental current error by the second energy shaping gain to obtain the q-axis energy-shaped output; Multiplying the d-axis fundamental current component by the first damping injection gain yields the d-axis damped injection output, and multiplying the q-axis fundamental current component by the second damping injection gain yields the q-axis damped injection output. The d-axis energy shaping output is added to the d-axis damping injection output to obtain the d-axis Hamiltonian control output, and the q-axis energy shaping output is added to the q-axis damping injection output to obtain the q-axis Hamiltonian control output.

8. The method according to claim 7, characterized in that, In step 2, the 5th harmonic d-axis current component is used as the 5th d-axis sliding surface variable. The sign of the 5th d-axis sliding surface variable is determined: when the 5th d-axis sliding surface variable is greater than zero, the sign output is positive 1; when the 5th d-axis sliding surface variable is less than zero, the sign output is negative 1; and when the 5th d-axis sliding surface variable is equal to zero, the sign output is zero. The sign output is multiplied by the first superspiral gain and then integrated over time to obtain the 5th d-axis superspiral integral term. The absolute value of the 5th d-axis sliding surface variable is taken, the square root is taken, and then multiplied by the sign output of the 5th d-axis sliding surface variable and the second superspiral gain to obtain the 5th d-axis superspiral proportional term. The 5th d-axis superspiral integral term is added to the 5th d-axis superspiral proportional term to obtain the 5th d-axis superspiral sliding output. The 5th q-axis superspiral sliding output, 7th d-axis superspiral sliding output, 7th q-axis superspiral sliding output, 11th superspiral sliding output, and 13th superspiral sliding output are obtained in the same way.

9. The method according to claim 8, characterized in that, In step 2, the 5th d-axis superspiral sliding mode output and the 5th q-axis superspiral sliding mode output are inversely rotated using a transformation angle of -5 times the grid phase angle to obtain the alpha-axis and beta-axis components of the 5th harmonic compensation voltage. Similarly, the 7th d-axis superspiral sliding mode output and the 7th q-axis superspiral sliding mode output are inversely rotated using a transformation angle of +7 times the grid phase angle to obtain the alpha-axis and beta-axis components of the 7th harmonic compensation voltage. The 11th harmonic compensation voltage is obtained in the same manner. The alpha and beta components of the voltage, as well as the alpha and beta components of the 13th harmonic compensation voltage; the alpha and beta components of the 5th, 7th, 11th, and 13th harmonic compensation voltages are superimposed in a two-phase stationary coordinate system to obtain the total alpha and beta components of the harmonic compensation voltage. The d-axis Hamiltonian control output and the q-axis Hamiltonian control output are subjected to Park inverse transformation with the grid phase angle as the inverse transformation angle to obtain the alpha-axis component and beta-axis component of the fundamental control voltage. The alpha-axis component of the fundamental control voltage is added to the total alpha-axis component of the harmonic compensation voltage to obtain the alpha-axis component of the composite modulation voltage. The beta-axis component of the fundamental control voltage is added to the total beta-axis component of the harmonic compensation voltage to obtain the beta-axis component of the composite modulation voltage.

Images