A method for controlling a micro-inverter based on active disturbance rejection control

Abstract

The present application relates to the technical field of smart grid, disclose a kind of micro inverter control method based on self-disturbance rejection control, comprising: obtaining inverter operating state and smart grid environmental parameter;Parallel internal slow change and external fast change dual channel are constructed to carry out frequency domain isolation observation, fast channel bandwidth is anchored power grid fundamental angular frequency;Through dual channel, internal low frequency and external high frequency power grid disturbance estimation are extracted respectively;External high frequency power grid disturbance estimation is synchronously mapped to inverter ac, dc side: virtual admittance compensation current correction instruction is generated in ac side, voltage compensation quantity is generated in dc side and is superimposed to the reference given value of previous stage bus, to adjust input power and generate drive instruction, realize across energy domain collaborative control.The present application effectively peels off multiscale interference coupling, smooths transient power range from physical bottom layer, avoids bus over-limit off-grid risk, significantly improves grid-connected stability under extreme working condition.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, specifically to a micro-inverter control method based on active disturbance rejection control. Background Technology

[0002] Microinverters are widely used in photovoltaic power generation systems, typically employing a two-stage topology of DC boost converter in the front stage and AC inverter in the back stage. In conventional control schemes, the front-end and back-end circuits often use an independent decoupled control strategy, meaning the front-end is mainly responsible for DC voltage regulation or maximum power point tracking on the photovoltaic side, while the back-end is responsible for AC grid-connected current regulation.

[0003] In real-world operating environments, microinverters face two types of disturbances: slow-varying parameter perturbations caused by aging and temperature rise of internal power devices, and fast-varying transient disturbances caused by external grid voltage dips, frequency fluctuations, and harmonics. Existing active disturbance rejection (ADRROC) technologies typically employ a single-bandwidth extended state observer to uniformly extract the total system disturbance. Because slow-varying and fast-varying disturbances span large time and frequency domains, a single observer struggles to maintain extraction accuracy. If the observer bandwidth is set too low, the system's response to high-frequency grid transients will lag; if the bandwidth is set too high, sampling noise can easily be introduced, causing system oscillations. Furthermore, the parameters of conventional observers are usually fixed and cannot adaptively adjust to follow the actual frequency shift of the grid, leading to state extraction errors under fluctuating grid conditions.

[0004] Furthermore, because existing two-stage inverters employ independent control architectures for the front and rear stages, when high-frequency transient distortions occur in the external power grid, the regulation system of the rear inverter circuit prioritizes sensing and adjusting the grid-connected current. However, the front-stage boost circuit, limited by its control architecture and communication delays, often lags behind the rear stage in power regulation response. This time difference in the action of the front and rear control loops leads to a mismatch between instantaneous input and output power. This power difference causes energy to rapidly accumulate or overdraw on the intermediate DC bus capacitor, resulting in significant fluctuations or even exceeding limits in the DC bus voltage, ultimately triggering inverter protection shutdown and reducing the grid-connected stability of the system. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a micro-inverter control method based on active disturbance rejection control. This method solves the technical problem that existing micro-inverters, when faced with multi-scale disturbances, suffer from insufficient extraction accuracy of a single observer and asynchronous transient power response due to independent control of upstream and downstream stages, which in turn leads to DC bus energy imbalance and grid instability.

[0006] To achieve the above objectives, the present invention provides a micro-inverter control method based on active disturbance rejection control, comprising the following steps: Obtain the operating status parameters of the microinverter and the environmental parameters of the smart grid; Parallel internal slow-varying disturbance observation channels and external fast-varying adaptive observation channels are constructed to perform frequency-domain isolated observation of the operating state parameters and the smart grid environmental parameters; wherein, the observation bandwidth of the external fast-varying adaptive observation channel is dynamically anchored to the fundamental angular frequency of the power grid in the smart grid environmental parameters. The internal low-frequency disturbance estimate is extracted through the internal slow-varying disturbance observation channel, and the external high-frequency power grid disturbance estimate is extracted through the external fast-varying adaptive observation channel. The external high-frequency grid disturbance estimate is simultaneously mapped to both the AC and DC sides of the micro-inverter: On the AC side, a virtual admittance compensation current is generated using the external high-frequency grid disturbance estimate, and the AC grid connection command is corrected to generate the subsequent inverter drive command; simultaneously on the DC side, a voltage compensation amount is generated using the external high-frequency grid disturbance estimate and directly superimposed on the preceding DC bus voltage reference setpoint to adjust the input power and generate the preceding converter drive command, thereby achieving cross-energy domain coordinated control. Specifically, the acquired operating status parameters include: photovoltaic port DC input voltage, photovoltaic port DC input current, intermediate DC bus actual voltage, and AC grid-connected current; the acquired smart grid environmental parameters include: smart grid actual AC voltage, extracted grid voltage instantaneous phase, and grid fundamental angular frequency.

[0007] Furthermore, in the process of extracting the internal low-frequency disturbance estimate, the AC grid-connected current is combined with the actual inverter control input output within the current control cycle of the system. A low-frequency extended state-space equation is constructed based on the nominal model of the controlled object. State convergence is achieved in the low-frequency bandwidth domain through iterative calculation of the difference between the system control gain and the nonlinear error correction gain, filtering out crosstalk from external high-frequency distortions. This decouples and solves for the internal low-frequency disturbance estimate, denoted as... It is used to characterize parameter perturbations inside a micro-inverter caused by temperature rise or aging.

[0008] Furthermore, in the process of extracting the external high-frequency power grid disturbance estimate, a resonant internal mode is explicitly introduced into the state transition matrix of the external fast-changing adaptive observation channel. The system state error including the resonant internal mode is defined as... ,in For the actual AC grid-connected current, the following high-order adaptive differential equation system is established: ; ; ; In the formula, The high-frequency state estimate of the obtained AC grid-connected current is denoted as . ; The obtained estimate of the external high-frequency power grid disturbance is denoted as... ; To assist in expanding state variables; This is the control gain constant; This is the input for inverter control; , , For error feedback gain scalar; The core operators constituting the resonant internal mode, among which This is the real-time acquisition of the fundamental angular frequency of the power grid. This mechanism enables the observer's central pole to be precisely anchored to the actual operating frequency of the power grid, achieving zero-hysteresis tracking of external high-frequency disturbances.

[0009] In the AC-side computational branch of cross-energy-domain collaborative control, a preset virtual admittance function exhibiting first-order low-pass filtering characteristics is used to estimate the external high-frequency grid disturbance. A mapping process is performed to generate a virtual admittance compensation current; this virtual admittance compensation current is then used to subtract and correct the system's preset AC grid-connected current initial reference command, resulting in a state-corrected reference current. Subsequently, the state-corrected reference current and the high-frequency state estimate are calculated. The difference and using the proportional gain coefficient Magnify; subtract the internal low-frequency perturbation estimate from the result in turn. and compliance adjustment coefficient Attenuated external high-frequency power grid disturbance estimator Generate the nonlinear disturbance rejection control law of the system in the current period. : ; The control law is converted by the modulation generation module into a subsequent inverter drive command used to drive the physical switches of the inverter bridge.

[0010] In the DC-side computational branch of cross-energy domain collaborative control, the control logic synchronously extracts the estimate of the external high-frequency grid disturbance. The absolute value of the value is multiplied by the preset DC-side active power disturbance mapping ratio. The voltage compensation is obtained and then superimposed on the original steady-state front-end DC bus voltage reference value of the system. Above, generate dynamic bus voltage correction commands. : ; Subsequently, the dynamic bus voltage correction command is subtracted from the actual voltage of the intermediate DC bus to form a voltage deviation, which is then input to the regulator for discretized proportional-integral iteration. The duty cycle signal used to control the power switching transistor of the front-stage boost circuit is recalculated, and finally modulated to generate the front-stage conversion drive command.

[0011] Furthermore, the steps of generating the subsequent inverter drive command and generating the preceding converter drive command are executed in parallel within the same control cycle of the control system. After generating and issuing the two sets of drive commands to the corresponding physical switching devices, the system rolls back to execute the parameter acquisition step at the current discrete time node to maintain the continuity of state observation.

[0012] This invention provides a micro-inverter control method based on active disturbance rejection control. It has the following beneficial effects: 1. To address the technical bottleneck of multi-scale interference coupling easily caused by the aging of internal inverter components and external grid transients, this invention constructs a frequency-domain isolated dual-channel observation mechanism. By dynamically anchoring the observation bandwidth of the fast-changing channel to the real-time grid fundamental frequency, the system accurately isolates low-frequency parameter perturbations while achieving zero-hysteresis, high-fidelity decoupling extraction of external high-frequency sudden disturbances. This effectively eliminates crosstalk between different frequency bands and provides an extremely pure feedforward reference for the control loop.

[0013] 2. To address the pain point of traditional independent front-end and rear-end control systems, which are prone to bus energy imbalance due to response delays during transient disturbances, this invention implements a cross-domain collaborative mapping mechanism. The extracted single high-frequency disturbance is synchronously distributed: on the AC side, it generates virtual admittance to flexibly smooth resonance; on the DC side, it is transformed into a dynamic bias to directly intervene in the input power of the front-end. This joint mechanism forces the photovoltaic input and grid-connected output to achieve a highly synchronized response, substantially smoothing out transient power differences at the physical level and effectively avoiding the risk of DC bus disconnection due to exceeding limits. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is the physical topology diagram of the main circuit of the micro inverter of the present invention; Figure 3 This is a block diagram of the overall control system based on the active disturbance rejection architecture of the present invention; Figure 4 This is a block diagram of the internal structure of the dual-channel frequency domain isolation observer of the present invention; Figure 5 This is a block diagram of the cross-domain collaborative mapping control branch logic of the present invention; Figure 6 This is a block diagram of the micro inverter control device of the present invention; Figure 7This is a schematic diagram of the electronic device structure of the present invention. Detailed Implementation

[0015] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Please see the appendix Figure 1 -Appendix Figure 7 This invention provides a micro-inverter control method based on active disturbance rejection control, comprising: Obtain the operating status parameters of the microinverter and the environmental parameters of the smart grid; Parallel internal slow-varying disturbance observation channels and external fast-varying adaptive observation channels are constructed to perform frequency-domain isolated observation of the operating state parameters and the smart grid environmental parameters; wherein, the observation bandwidth of the external fast-varying adaptive observation channel is dynamically anchored to the fundamental angular frequency of the power grid in the smart grid environmental parameters. The internal low-frequency disturbance estimate is extracted through the internal slow-varying disturbance observation channel, and the external high-frequency power grid disturbance estimate is extracted through the external fast-varying adaptive observation channel. The external high-frequency grid disturbance estimate is simultaneously mapped to both the AC and DC sides of the micro-inverter: on the AC side, a virtual admittance compensation current is generated using the external high-frequency grid disturbance estimate, and the AC grid connection command is corrected to generate the subsequent inverter drive command; simultaneously on the DC side, a voltage compensation amount is generated using the external high-frequency grid disturbance estimate, and directly superimposed on the reference setpoint of the preceding DC bus voltage to adjust the input power and generate the preceding conversion drive command, thereby achieving cross-energy domain coordinated control.

[0017] During grid-connected operation, the internal component parameters of a microinverter drift slowly with increasing temperature and usage time. Simultaneously, the smart grid injects rapid disturbances into the microinverter, such as grid voltage dips and harmonic surges. These internal and external disturbances exhibit significant differences in the frequency domain. Conventional control methods treat these disturbances as a whole and observe them centrally, which makes it difficult for the observer bandwidth to simultaneously achieve both low-frequency accuracy and high-frequency response speed.

[0018] To address this technical challenge, this method introduces a frequency-domain isolated observation mechanism. At the start of each control cycle, the controller synchronously acquires the internal electrical parameters of the inverter system and the environmental state parameters of the grid side through a sampling circuit. The internal operating state parameters reflect the real-time steady-state operating point of the power conversion link, while the grid environmental parameters characterize the dynamic boundary conditions of the external grid-connected nodes.

[0019] Based on the acquired parameters, the controller internally operates two state observation channels with independent dynamic characteristics in parallel. The internal slowly varying disturbance observation channel is configured with a low-frequency bandwidth attribute, dedicated to tracking and extracting the system's own parameter perturbations. This channel outputs an estimate of the internal low-frequency disturbance, denoted as... This estimator characterizes the lumped disturbances caused by slow-varying nonlinear factors within the system, such as inductance drift and changes in the on-state voltage drop of the switching transistor.

[0020] The external fast-changing adaptive observation channel is configured with high-frequency bandwidth properties to capture transient high-frequency interference injected from the grid side. To ensure accurate tracking of dynamic frequency fluctuations in the grid, the observation poles of this channel are not fixed as static constants, but are anchored in real time to the extracted grid fundamental angular frequency. The channel outputs an estimate of the external high-frequency power grid disturbance, denoted as... Through a dual-channel architecture, the system achieves decoupling and precise separation of internal and external disturbances.

[0021] After separating and extracting internal and external disturbances, the control system executes cross-energy domain collaborative control logic. The front-end boost circuit and the rear-end inverter bridge of the microinverter achieve physical energy coupling through the intermediate DC bus capacitor. When the external power grid experiences severe fluctuations, the transient power change on the AC side will immediately cause an imbalance in the charging and discharging of the DC bus, thereby triggering the bus voltage to exceed the limit.

[0022] This method breaks away from the conventional decoupled architecture where the front-end independently tracks the maximum power point and the rear-end independently controls the grid-connected current. The controller uses the same external high-frequency grid disturbance estimate. As a collaborative feedforward signal, it is synchronously sent to the AC control branch and the DC control branch.

[0023] In the AC-side control branch, the controller utilizes the established virtual admittance model to... Impedance mapping is performed to generate a virtual admittance compensation current. The compensation current is based on the original AC grid-connected current initial reference command of the system. A differential correction is performed. The corrected state-emission correction reference current is used to generate the final subsequent inverter drive command, enabling the inverter bridge to exhibit damping characteristics that suppress high-frequency resonance of the grid-connected current during grid disturbances, thereby suppressing the high-frequency oscillation of the grid-connected current.

[0024] In the DC-side control branch, the controller synchronously extracts estimates of external high-frequency grid disturbances. The amplitude characteristics are converted into corresponding voltage compensation values. The controller does not wait for an actual physical drop or overshoot in the DC bus voltage, but actively... Superimposed on the system's original front-end DC bus voltage reference setpoint superior.

[0025] The corrected dynamic bus voltage correction command is denoted as Its calculation logic satisfies the formula: ; The closed-loop regulator of the front-end boost circuit responds to the modified dynamic command and recalculates the duty cycle, thereby dynamically adjusting the input power extraction at the photovoltaic port. This feedforward cutoff mechanism forces the micro-inverter to actively deviate from its theoretical maximum power point during grid disturbances, using the reduction or increase of active power on the DC input side to completely offset sudden transient power differences on the AC side, maintaining the energy throughput balance of the intermediate DC bus capacitor, and ultimately generating the front-end conversion drive command.

[0026] The aforementioned dual-sided mapping logic is executed in parallel within the same clock cycle of the control system, realizing synchronous anti-disturbance action of the AC and DC ends of the micro-inverter against a single external disturbance source. This eliminates the risk of grid disconnection and instability caused by complex grid distortion from the underlying logic of system energy conservation.

[0027] In this embodiment, the acquisition of the operating status parameters of the microinverter and the smart grid environment parameters specifically includes synchronously reading real-time electrical data inside and outside the system through the data acquisition bus at the beginning of the system control initialization and each subsequent rolling clock cycle.

[0028] The acquired operating status parameters include: the photovoltaic port DC input voltage, which characterizes the DC energy input capability of the front-end stage. With the DC input current of the photovoltaic port The actual voltage of the intermediate DC bus, representing the energy buffer state. And the AC grid-connected current characterizing the final energy output. .

[0029] The smart grid environmental parameters acquired synchronously include: the actual AC voltage of the smart grid. Instantaneous phase of grid voltage extracted by phase-locked loop algorithm and the fundamental angular frequency of the power grid The two sets of continuous-time parameters mentioned above, after being discretized and sampled, together constitute the state feedback information source and frequency domain anchoring reference of the subsequent nonlinear active disturbance rejection observer.

[0030] In this invention, for physical phenomena such as temperature rise and aging of internal components caused by long-term full-load operation of micro-inverters, the internal low-frequency disturbance estimate is extracted through the internal slow-varying disturbance observation channel.

[0031] Specifically, the AC grid-connected current The inverter control input quantity that is actually output during the current control cycle of the system Combined, these serve as the control and state input matrix for the internal slowly varying disturbance observation channel. Based on the nominal model of the controlled object, the following extended state equation based on low-frequency state observation difference iteration is constructed: ; ; In the formula, These are low-frequency state observations of AC grid-connected current from an internal slow-varying channel. This is the extracted estimate of the internal low-frequency disturbance, denoted as... ; The system control gain constant, whose numerical physical meaning is determined by the hardware parameters of the AC side of the micro-inverter, is specifically configured as follows: ,in This is the rated operating voltage of the intermediate DC bus. This is the nominal inductance value of the AC-side filter inductor; and This is the nonlinear error correction gain for the internal slowly varying observation channel.

[0032] The extracted internal low-frequency disturbance estimate This is used to characterize parameter perturbations in internal components of a microinverter caused by temperature rise or aging, such as temperature drift of the filter inductor and changes in the on-resistance of the switching transistors. The calculation process performs state convergence in the low-frequency bandwidth domain, effectively filtering out crosstalk from high-frequency waveform distortions of the external power grid, and quantifying the nonlinear, slowly varying characteristics of the system itself independently and accurately.

[0033] Furthermore, to achieve zero-hysteresis tracking of high-frequency transient disturbances in the external smart grid (such as voltage sags and harmonic abrupt changes), this invention extracts estimates of external high-frequency grid disturbances through the aforementioned external fast-change adaptive observation channel. Conventional state observers, when processing periodic AC disturbance signals such as sinusoidal signals, exhibit a fundamental steady-state phase lag in their estimation results.

[0034] To eliminate this estimation hysteresis, this invention explicitly introduces a resonant internal mode into the state transition matrix of the external fast-changing adaptive observation channel. First, the system state error including the resonant internal mode is defined as... .

[0035] Based on the system state error including the resonant internal mode Observations and calculations were performed to establish the following system of high-order adaptive differential equations: ; ; ; In the formula, The high-frequency state estimate of the obtained AC grid-connected current is denoted as... ; The obtained estimate of the external high-frequency power grid disturbance is denoted as... ; An auxiliary extended state quantity is introduced to match the resonant internal mode; , as well as This is the error feedback gain scalar for the external rapidly changing adaptive observation channel.

[0036] In the matrix operation logic described above, the state variable differential equation terms This constitutes the core operator of the resonant internal mode. To prevent the observer from diverging due to high-frequency jitter in the phase-locked loop output during severe transient distortions in the power grid, the system introduces the fundamental angular frequency of the power grid. A transient freeze and low-pass filtering mechanism is configured: when the grid voltage change rate is detected to exceed a set threshold, the current is briefly maintained. This represents the steady-state value at the moment before the disturbance. The fundamental angular frequency of the power grid is extracted and updated in real-time by the phase-locked loop. The frequency of the central pole of this external observation channel drifts in real time on the complex plane and is always precisely anchored to the actual operating frequency of the power grid.

[0037] Through the aforementioned difference iteration and internal model mapping, the algorithm completely isolates external disturbances caused by non-internal factors from the overall state error of the system, and finally converges and outputs a high-frequency state estimate of the AC grid-connected current with zero phase error. and the external high-frequency power grid disturbance estimate Thus, the microinverter has completed the precise orthogonal decoupling and extraction of two types of multi-timescale disturbance sources, both internal and external.

[0038] Furthermore, it should be noted that, since the control algorithm of the micro-inverter is implemented in the digital signal processor with a fixed control cycle... Discrete execution: The continuous-domain state-space differential equations constructed above for the internal slowly varying perturbation observation channel and the external rapidly varying adaptive observation channel need to be discretized in actual physical implementation. In this embodiment, the forward Euler method or bilinear transformation method is used to transform the continuous calculus operations into difference equations. State variables... For example, its micro-business items In microcontrollers, it is approximately replaced by The remaining state variables are deduced similarly, thus ensuring that the theoretically designed continuous active disturbance rejection observer can be seamlessly converted into C language digital instruction code that can be iterated over in the microcontroller. At this point, the microinverter has completed the accurate orthogonal decoupling extraction of two types of multi-timescale disturbance sources, both internal and external.

[0039] It should be noted that, to ensure stable convergence of the aforementioned parallel dual-channel observer, this embodiment configures the observation gain based on the bandwidth parameterization tuning method. Specifically, the poles of the internal slowly varying perturbation observation channels are uniformly configured within the slowly varying observation bandwidth. From this, the gain can be derived. , Similarly, the poles of the external fast-changing adaptive observation channel are configured within the fast-changing observation bandwidth. Location, thereby determining the aforementioned , , The numerical relationship greatly reduces the difficulty of engineering debugging and ensures the closed-loop stability of the system.

[0040] In this embodiment, after acquiring and decoupling the estimates of two types of multi-timescale disturbances (internal and external), the micro-inverter control system first performs an active impedance reshaping stage on the AC control branch. To improve the grid-connected compliance of the system when facing sudden changes in grid impedance or high-frequency voltage distortion, this invention uses a preset virtual admittance function to map the external high-frequency grid disturbance estimates, thereby generating a virtual admittance compensation current for smoothing transient power surges on the AC side.

[0041] Specifically, the preset virtual admittance function in the complex frequency domain is defined as follows: To enable the microinverter to exhibit resistive absorption characteristics in the high-frequency distortion band of the power grid, in this embodiment, the preset virtual admittance function is specifically configured as a transfer function that includes virtual conductance parameters and first-order low-pass filter characteristics, and its expression is: ; In the formula, The set steady-state virtual conductance value is used to determine the absorption intensity of high-frequency disturbances; The time constant of the virtual admittance is used to filter out switching noise interference above the switching frequency. The control algorithm uses the external high-frequency power grid disturbance estimate extracted in the preceding steps. The virtual admittance compensation current in the time domain is obtained by discretization techniques such as bilinear transformation or backward difference method, which is considered as an electrical excitation source input to the admittance model, and then obtained through inverse transformation mapping. The compensation current is logically equivalent to connecting a dynamic absorption network in parallel at the AC output port of the micro-inverter, giving the system physical and electrical damping characteristics that it did not originally possess.

[0042] Subsequently, the control system reads the initial reference command for AC grid-connected current issued by the system power outer loop. The initial reference command is corrected by subtraction using the virtual admittance compensation current through the feedforward channel. The mathematical equation for this correction is as follows: ; In the formula, This is the reconstructed state-corrected reference current. This differential correction operation enables the inverter's grid-connected current reference to adaptively bias when encountering transient disturbances in the grid, avoiding control saturation and current runaway oscillations caused by the traditional hard tracking of the original command.

[0043] In this invention, to achieve high disturbance rejection capability with nonlinear current closed-loop control, the comprehensive generation calculation of the subsequent inverter drive command is further performed. The controller calculates the state-corrected reference current. The high-frequency state estimate of the AC grid-connected current obtained by the aforementioned steps. The state tracking error between them is calculated, and a preset proportional gain coefficient is used. The error value is algebraically amplified. The value is calibrated based on the maximum allowable steady-state tracking error of the system and the closed-loop response bandwidth of the AC current loop.

[0044] Based on this, the control logic performs hierarchical compensation and redirection of multi-band disturbances. Using the amplified error result as the base adjustment point, the previously extracted internal low-frequency disturbance estimate is first subtracted without loss. This is to completely eliminate steady-state static errors caused by device aging and temperature drift.

[0045] Next, subtract the preset smoothness adjustment factor again. The external high-frequency power grid disturbance estimate after attenuation processing The compliance adjustment coefficient The value range is constrained within the interval (0,1). By introducing this attenuation operator, the system can effectively avoid over-excitation of the control signal caused by the full-amplitude high-frequency disturbance feedforward. The above-mentioned difference calculation results are divided by the high-frequency control gain constant of the system controlled by the micro-inverter. That is, the nonlinear disturbance rejection control law of the generation system in the current period. .

[0046] The complete mathematical analytical expression of the above nonlinear control law is as follows: ; In the formula, the obtained nonlinear disturbance rejection control law The inverter control input is assigned a value on a rolling basis and used as the system operation in the next control cycle.

[0047] At the end of the AC-side computation branch, the controller will calculate the nonlinear disturbance rejection control law. The signal is transmitted to the underlying modulation generation module. In this embodiment, the nonlinear disturbance rejection control law... In a physical sense, it serves as an equivalent continuous modulated wave signal input; to eliminate the coupling effect of transient fluctuations in the intermediate DC bus voltage on the AC transmission gain, the modulation generation module first inputs this modulated wave signal. Divide by the real-time sampled actual voltage of the intermediate DC bus Normalization is performed to obtain the transient modulation ratio signal. Subsequently, based on the principles of sinusoidal pulse width modulation or space vector pulse width modulation, the modulation generation module compares the normalized transient modulation ratio signal with the internal high-frequency triangular carrier wave in real time, thereby converting the digital control law into a pulse train with corresponding duty cycle width and dead time delay. This generates the subsequent inverter drive command used to directly drive the physical switching action of the inverter bridge of the micro-inverter, and accurately executes the compliant control command for AC grid-connected current.

[0048] In this embodiment, within the same control cycle during the active impedance reshaping of the AC control branch, the micro-inverter control system simultaneously activates the cross-energy domain cutoff and feedforward compensation mechanism of the DC control branch. To effectively suppress the drastic energy throughput and voltage overshooting of the intermediate DC bus caused by high-frequency transient disturbances from the external power grid, this invention forcibly maps the same external high-frequency power grid disturbance estimate extracted from the AC side directly to the photovoltaic front-end power extraction control loop.

[0049] Specifically, the control logic first calculates and obtains the estimate of the external high-frequency power grid disturbance in real time. The absolute value of the disturbance polarity eliminates interference with the unidirectional AC / DC energy transfer logic, accurately characterizing the severity of energy imbalance at both ends of the system caused by transient active power fluctuations on the grid side. To prevent direct injection of high-frequency pulsating signals that could cause oscillations in the front-end closed-loop tracking, the system further performs moving average filtering or low-pass envelope extraction on the aforementioned absolute value signal to obtain a smooth envelope amplitude characterizing the disturbance intensity. Subsequently, the extracted smooth envelope amplitude is multiplied by a preset DC-side active power disturbance mapping coefficient. After linear scaling transformation, the voltage compensation amount specific to the DC side is obtained. Wherein, the mapping scaling factor The upper limit of its value is limited by the steady-state withstand voltage threshold of the physical capacitor of the intermediate DC bus in the system. Its setting principle is: to ensure that, under the maximum allowable amplitude of grid transient disturbances in the system design, the voltage is controlled by... The absolute value of the voltage overshoot generated by the mapping does not trigger the hardware overvoltage protection action.

[0050] In this invention, the system does not passively wait for the actual voltage of the intermediate DC bus to experience a substantial physical drop or surge due to a sudden change in the transient load on the AC side. Instead, it executes proactive defense feedforward intervention logic. The controller calculates the voltage compensation amount. Directly superimposed on the existing front-end DC bus voltage reference setpoint of the system. superior.

[0051] The above-described superposition process forcibly redirects the voltage reference point in the control architecture, generating the refreshed dynamic bus voltage correction command. The mathematical operation relationship is expressed as follows: ; In the formula, This is the reference value of the rated DC voltage issued by the system under steady-state, disturbance-free operating conditions through the maximum power point tracking algorithm or the outer loop of the upper-level steady-state control.

[0052] After completing the dynamic correction of the voltage reference, the control system performs differential calculations for the front-stage DC bus voltage regulation closed loop. The controller then uses the actual voltage of the intermediate DC bus acquired by the hardware sensors in the front-stage circuit. With the refreshed dynamic bus voltage correction command Perform a difference operation. This difference operation generates a voltage deviation signal characterizing the current transient mismatch between DC energy supply and demand. Its equation is: ; Subsequently, the controller will calculate the calculated voltage deviation signal. The input is fed into the closed-loop controller of the preceding system for calculation. In this embodiment, the closed-loop controller is a discretized proportional-integral controller. This controller performs a discretized control law solution on the aforementioned voltage deviation signal, and the solution equation is as follows: ; In the formula, and The duty cycle signals are calculated and output for the current control cycle and the previous control cycle, respectively. and These are the voltage deviation signals for the current cycle and the previous cycle, respectively. and These are the proportional gain and integral gain of the proportional-integral regulator, respectively. Based on the output of the regulator, the system accurately calculates the duty cycle signal used to control the on / off ratio of the power switch transistors in the preceding boost circuit. To ensure the safe operation of the micro-inverter's physical components and resist integral saturation, the controller outputs this duty cycle signal. Previously, physical boundary limiting logic was enforced to strictly clamp its value within the effective duty cycle range allowed by the system hardware. In a typical hardware implementation scenario of this invention, if the front-end boost circuit adopts a standard Boost topology, then the duty cycle signal... This directly determines the boost ratio relationship between the DC bus voltage and the photovoltaic input voltage. Through this compensation mechanism that forces a bias setpoint and directly affects the duty cycle, the micro-inverter actively and briefly deviates from the original optimal photovoltaic power extraction point when facing grid disturbances. By changing the amount of active power input on the DC side, it directly offsets the energy fluctuations on the AC output side in terms of source throughput.

[0053] After completing the above-mentioned underlying logic calculations, based on the determined duty cycle signal... The control system, through a modulation module configured with a carrier frequency matching the system's preset switching frequency, converts the carrier frequency into a high-low level pulse sequence with a precise duty cycle width, thereby generating the pre-stage conversion drive command used to directly drive the physical switching devices of the boost circuit inside the micro-inverter. Thus, this method completely eliminates the physical barrier between the AC disturbance source and the DC input side, achieving synchronous response and coordinated cutoff of a single disturbance source parameter in both energy domains.

[0054] In this embodiment, to ensure absolute synchronization between the physical energy flow at both AC and DC ends and the digital variable flow within the control system, the steps of generating the subsequent inverter drive command and generating the preceding converter drive command are executed in parallel within the same control cycle of the micro-inverter control system. The digital signal processor, based on a unified internal hardware timer triggering interrupt events, reads the operating status parameters and smart grid environment parameters, and then uses the chip's parallel bus or instruction pipeline mechanism to simultaneously calculate the virtual admittance compensation logic on the AC side and the voltage regulation feedforward cutoff logic on the DC side.

[0055] This parallel execution mechanism completely eliminates the inherent phase delay in state calculation in conventional serial computing architectures. By anchoring the AC impedance reshaping action and the DC energy cutoff action to the same time reference plane, the system can ensure that the power throughput gap caused by transient disturbances on the AC side is filled by the active power regulation on the DC input side within the same microsecond time scale, thus guaranteeing zero-time-difference execution of cross-energy domain collaborative control from the underlying timing triggering mechanism.

[0056] In this invention, upon completion of all nonlinear control laws and instruction correction calculations within the current control cycle, the control system generates the subsequent inverter drive instruction and the preceding converter drive instruction. Subsequently, the hardware pulse-width modulation peripheral within the digital controller sends these two sets of drive instructions to the physical switching devices of the AC-side inverter bridge and the DC-side boost circuit of the micro-inverter, respectively. The drive instructions directly change the on / off state of the high-frequency switching transistors, completing the hardware-level action refresh for a single grid disturbance response cycle.

[0057] When a control cycle ends and the next sampling interrupt is triggered immediately, the addressing pointer of the control system scrolls back to the initial operating logic. The time step of the system's discretized control is defined as... The discrete time node of the current control cycle is Then, as the hardware timer is reloaded, the system enters the next discrete time node. The time recursion relationship satisfies: ; In the formula, The system at the physical sampling time corresponding to the currently described operating state parameters is... A new round of timer interrupts is constantly initiated to re-execute the steps of acquiring the operating state parameters of the micro-inverter and the smart grid environment parameters. This rolling iteration mechanism maintains the numerical integration continuity of all state-space differential equations within the active disturbance rejection observation channel, ensuring uninterrupted and high-fidelity capture of the dynamic evolution of the smart grid environment.

[0058] Based on the aforementioned complete closed-loop control method, this invention also provides a micro-inverter control device based on active disturbance rejection control. This virtual functional entity device mainly consists of a parameter acquisition module, a frequency domain isolation observation module, a disturbance decoupling extraction module, and a cross-domain collaborative mapping module. The parameter acquisition module is used to synchronously sense real-time electrical state signals from multiple dimensions inside and outside the inverter. The frequency domain isolation observation module embeds the aforementioned internal slow-varying disturbance observation channel and external fast-varying adaptive observation channel. The disturbance decoupling extraction module is responsible for performing the difference iteration of internal model errors and the internal mode resonance extraction operation. Specifically, this disturbance decoupling extraction module contains a low-frequency feature reconstruction subunit and a high-frequency adaptive decoupling subunit, which are used to output the estimated internal low-frequency disturbance, respectively. With the external high-frequency power grid disturbance estimate The cross-domain collaborative mapping module distributes parallel state-driven instructions to the physical execution layer based on the extracted multi-scale signals. Specifically, this module is configured with parallel processing AC mapping subunits and DC mapping subunits. The former uses a virtual admittance function to smooth grid-connected current instructions, while the latter intervenes in the preceding duty cycle generation stage through dynamic voltage bias. Each functional module interacts with variables and maps addresses via a data bus within the memory.

[0059] Furthermore, this embodiment of the invention also provides an electronic device including core control hardware, which serves as the brain of a microinverter, including a processor and a memory connected via an on-chip system bus. The memory is used to non-volatilely store the system's low-level driver programs and the computer instruction code set that implements the aforementioned complete microinverter control method. The processor is configured as the execution center, and when loading and reading the instruction code from the memory, it sequentially schedules peripheral units such as analog-to-digital converters and hardware multipliers to execute the aforementioned digital control logic from parameter acquisition to frequency domain isolation and AC / DC co-mapping.

[0060] Furthermore, this embodiment of the invention provides a non-transitory computer-readable storage medium, which internally stores a computer program for loading, compiling, and running by a digital control chip. When the computer program is read and executed by a processor core with data processing capabilities, it can completely reproduce all the processing steps of the micro-inverter control method based on active disturbance rejection control disclosed in detail in the foregoing embodiments. This physical medium form ensures that the nonlinear disturbance rejection control algorithm with an adaptive internal model mechanism proposed in this invention can be reliably and standardizedly deployed in batches in various distributed photovoltaic inverter devices in the industry.

[0061] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended technical solutions and their equivalents.

Claims

1. A micro-inverter control method based on active disturbance rejection control, characterized in that, include: Obtain the operating status parameters of the microinverter and the environmental parameters of the smart grid; Parallel internal slow-varying disturbance observation channels and external fast-varying adaptive observation channels are constructed to perform frequency-domain isolated observation of the operating state parameters and the smart grid environmental parameters; wherein, the observation bandwidth of the external fast-varying adaptive observation channel is dynamically anchored to the fundamental angular frequency of the power grid in the smart grid environmental parameters. The internal low-frequency disturbance estimate is extracted through the internal slow-varying disturbance observation channel, and the external high-frequency power grid disturbance estimate is extracted through the external fast-varying adaptive observation channel. The external high-frequency grid disturbance estimate is simultaneously mapped to both the AC and DC sides of the micro-inverter: on the AC side, a virtual admittance compensation current is generated using the external high-frequency grid disturbance estimate, and the AC grid connection command is corrected to generate the subsequent inverter drive command; simultaneously on the DC side, a voltage compensation amount is generated using the external high-frequency grid disturbance estimate, and directly superimposed on the reference setpoint of the preceding DC bus voltage to adjust the input power and generate the preceding conversion drive command, thereby achieving cross-energy domain coordinated control.

2. The micro-inverter control method based on active disturbance rejection control according to claim 1, characterized in that, The acquisition of the operating status parameters of the microinverter and the smart grid environmental parameters includes: The acquired operating status parameters include: photovoltaic port DC input voltage, photovoltaic port DC input current, intermediate DC bus actual voltage, and AC grid-connected current; The acquired smart grid environmental parameters include: the actual AC voltage of the smart grid, the extracted instantaneous phase of the grid voltage, and the fundamental angular frequency of the grid.

3. The micro-inverter control method based on active disturbance rejection control according to claim 2, characterized in that, The extraction of the internal low-frequency disturbance estimate through the internal slowly varying disturbance observation channel includes: The AC grid-connected current is combined with the actual inverter control input output during the current control cycle of the system, and used as the input of the internal slow-varying disturbance observation channel; The internal low-frequency disturbance estimate is extracted by iterating the low-frequency state observation difference. The internal low-frequency disturbance estimate is used to characterize the parameter perturbation of the internal components of the micro-inverter caused by temperature rise or aging.

4. The micro-inverter control method based on active disturbance rejection control according to claim 3, characterized in that, The extraction of external high-frequency power grid disturbance estimates through the external fast-changing adaptive observation channel includes: A resonant internal mode is introduced into the external fast-changing adaptive observation channel; Based on the system state error including the resonant internal mode, observation calculations are performed to remove external disturbances from the system state error, thereby obtaining the high-frequency state estimate of the AC grid-connected current and the estimate of the external high-frequency grid disturbance.

5. The micro-inverter control method based on active disturbance rejection control according to claim 4, characterized in that, The step of generating a virtual admittance compensation current using the external high-frequency grid disturbance estimate and correcting the AC grid connection command includes: The estimated external high-frequency power grid disturbance is mapped using a preset virtual admittance function to generate the virtual admittance compensation current. The virtual admittance compensation current is used to perform differential correction on the system's preset AC grid-connected current initial reference command to obtain the state-corrected reference current.

6. The micro-inverter control method based on active disturbance rejection control according to claim 5, characterized in that, The generation of subsequent inverter drive instructions includes: The difference between the state-corrected reference current and the high-frequency state estimate of the AC grid-connected current is calculated and amplified using a preset proportional gain coefficient. Subtract the internal low-frequency disturbance estimate and the external high-frequency power grid disturbance estimate after attenuation by a preset compliance adjustment coefficient from the amplified result in turn. Divide the above difference calculation result by the high-frequency control gain constant of the system to obtain the nonlinear disturbance rejection control law, and use the nonlinear disturbance rejection control law as the inverter control input for the next control cycle. According to the nonlinear disturbance rejection control law, the modulation generation module generates the subsequent inverter drive command for driving the inverter bridge of the micro-inverter.

7. The micro-inverter control method based on active disturbance rejection control according to claim 2, characterized in that, The step of generating a voltage compensation amount using the external high-frequency power grid disturbance estimate and directly superimposing it onto the upstream DC bus voltage reference setpoint includes: Obtain the absolute value of the external high-frequency power grid disturbance estimate and multiply it by a preset DC-side active power disturbance mapping ratio coefficient to obtain the voltage compensation amount; The voltage compensation amount is superimposed on the original reference value of the front-end DC bus voltage of the system to generate a dynamic bus voltage correction command.

8. The micro-inverter control method based on active disturbance rejection control according to claim 7, characterized in that, The method of generating the pre-conversion drive command by adjusting the input power includes: The voltage deviation is generated by subtracting the dynamic bus voltage correction command from the actual voltage of the intermediate DC bus, and the voltage deviation is input to the regulator. Calculate the duty cycle signal based on the output of the regulator; Based on the duty cycle signal, the modulation module generates the pre-stage conversion drive command for driving the boost circuit.

9. The micro-inverter control method based on active disturbance rejection control according to claim 1, characterized in that, The steps of generating the subsequent inverter drive command and generating the preceding converter drive command are executed in parallel within the same control cycle of the control system.

10. The micro-inverter control method based on active disturbance rejection control according to claim 1, characterized in that, After generating the subsequent inverter drive command and the preceding conversion drive command, the drive commands are sent to the physical switching devices on the AC and DC sides of the microinverter, respectively. After one control cycle, the process rolls back to execute the steps of obtaining the operating status parameters of the microinverter and the smart grid environment parameters.

Images