Flexible voltage regulation control method and system based on power electronic conversion

Abstract

The application discloses a flexible voltage regulation control method and system based on power electronic conversion, and relates to the technical field of power electronics. The method comprises the following steps: triggering a sensing unit to perform high-speed sampling on N three-phase alternating current power supplies, and obtaining real-time signals; performing dq coordinate system PLL phase decomposition on the real-time signals, and obtaining dynamic power quality components; outputting scores according to multi-dimensional evaluation of the components; extracting target power supplies in descending order; outputting compensation instructions according to flexible voltage regulation of target components; and driving multi-level PWM modulation according to the instructions, and then delivering energy to target electric appliances through an H-bridge converter topology after optimization. The technical problems of slow response, limited regulation range and inability to intelligently select the optimal power supply for collaborative compensation in a multi-path power supply scene in the traditional voltage regulation system, which leads to insufficient voltage control precision and reliability, are solved, and the technical effects of achieving millisecond-level dynamic response through high-speed sampling and PLL phase locking, effectively suppressing harmonics, improving power quality and power supply stability are achieved.

Description

Technical Field

[0001] This invention relates to the field of power electronics technology, and more specifically to a flexible voltage regulation control method and system based on power electronic conversion. Background Technology

[0002] With the continuous expansion of power system scale and the rapid increase in the proportion of renewable energy integration, the power grid operating environment is becoming increasingly complex, and power quality problems such as voltage fluctuations, harmonic distortion, and inrush currents are becoming increasingly prominent. In scenarios sensitive to power quality, such as industrial production, precision manufacturing, data centers, and rail transportation, voltage deviations, frequent fluctuations, or unstable power quality will directly affect the normal operation of electrical equipment, and may even lead to shutdowns, damage, or safety accidents. Therefore, the stability and flexible control of power quality has become one of the key technical issues in power system operation.

[0003] Existing voltage regulation control methods mainly include mechanical voltage regulation, tap changer voltage regulation, and static reactive power compensation and active filtering based on power electronic devices. Mechanical and tap changer voltage regulation have slow response speeds and are difficult to cope with rapid voltage disturbances; traditional power electronic compensation devices are mostly single-objective oriented, such as voltage stabilization or filtering, and lack the ability to coordinate and adjust multi-dimensional power quality parameters under complex fluctuating environments, making it difficult to achieve millisecond-level flexible dynamic voltage regulation control. Summary of the Invention

[0004] This application provides a flexible voltage regulation control method and system based on power electronic conversion, which solves the technical problems of slow response, limited adjustment range, and inability to intelligently select the optimal power supply for collaborative compensation in multi-power supply scenarios, resulting in insufficient voltage control accuracy and reliability of traditional voltage regulation systems.

[0005] The first aspect of this application provides a flexible voltage regulation control method based on power electronic conversion, the method comprising:

[0006] N current and voltage sensing units are triggered to perform high-speed sampling of N three-phase AC power supplies, obtaining N real-time current and voltage signals. The N current and voltage sensing units are directly coupled to each phase of the N three-phase AC power supplies. PLL phase decomposition based on a dq rotating coordinate system is performed on the N real-time current and voltage signals to obtain N dynamic power quality components. Multi-dimensional power supply stability is quantitatively evaluated based on the N dynamic power quality components, outputting N comprehensive stability scores. The target AC power supply corresponding to the highest score is extracted based on the descending order of the N comprehensive stability scores. Flexible stepless voltage regulation analysis is performed on the target AC power supply's target dynamic power quality components, outputting dynamic compensation commands. After multi-level PWM modulation driven by the dynamic compensation commands to optimize the target AC power supply's dynamic voltage at multiple levels, the power from the target AC power supply is delivered to the target appliance via an H-bridge converter topology.

[0007] A second aspect of this application provides a flexible voltage regulation control system based on power electronic conversion, the system comprising:

[0008] Signal Sampling Module: Triggers N current and voltage sensing units to perform high-speed sampling of N three-phase AC power supplies, obtaining N real-time current and voltage signals, wherein the N current and voltage sensing units are directly coupled to each phase of the N three-phase AC power supplies; Phase Decomposition Module: Performs PLL phase decomposition based on the dq rotating coordinate system on the N real-time current and voltage signals, obtaining N dynamic power quality components; Stability Evaluation Module: Performs multi-dimensional quantitative evaluation of power supply stability based on the N dynamic power quality components, outputting N comprehensive stability scores; Descending Order Module: Extracts the target AC power supply corresponding to the maximum score based on the descending order of the N comprehensive stability scores; Voltage Regulation Analysis Module: Performs flexible stepless voltage regulation analysis based on the target dynamic power quality components of the target AC power supply, outputting dynamic compensation commands; Voltage Optimization Module: Drives multi-level PWM modulation to perform multi-level dynamic voltage optimization of the target AC power supply according to the dynamic compensation commands, and then delivers the power of the target AC power supply to the target appliance through an H-bridge converter topology.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] First, multiple current and voltage sensing units are used to sample each phase of the multi-phase three-phase AC power supply at high speed, acquiring voltage and current signals in real time. Then, using PLL phase-locked loop technology based on a dq rotating coordinate system, the acquired signals are decomposed into dynamic components reflecting power quality. Next, these power quality components undergo multi-dimensional quantitative analysis and comprehensive scoring, selecting the most stable power supply from multiple sources as the target AC power supply. Then, flexible stepless voltage regulation analysis is performed on the power quality characteristics of the target power supply, generating corresponding dynamic compensation control commands. Finally, multi-level PWM modulation technology is used to dynamically optimize the output voltage of the target power supply at multiple levels. Finally, an H-bridge converter topology is used to efficiently and stably deliver power to the load-side equipment, achieving fast and precise flexible voltage regulation control. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A schematic flowchart of a flexible voltage regulation control method based on power electronic conversion provided in an embodiment of this application.

[0013] Figure 2 This is a schematic diagram of a flexible voltage regulation control system based on power electronic conversion, provided in an embodiment of this application.

[0014] Explanation of reference numerals in the attached diagram: Signal sampling module 11, Phase decomposition module 12, Stability evaluation module 13, Descending order sorting module 14, Voltage regulation analysis module 15, Voltage optimization module 16. Detailed Implementation

[0015] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0016] Example 1, as Figure 1 As shown, this application provides a flexible voltage regulation control method based on power electronic conversion, the method including:

[0017] N current and voltage sensing units are triggered to perform high-speed sampling of N three-phase AC power supplies to obtain N real-time current and voltage signals. The N current and voltage sensing units are directly coupled to each phase of the N three-phase AC power supply.

[0018] In this embodiment, N pre-set current and voltage sensing units are first triggered. These units correspond to N independent three-phase AC power supplies, and each unit is directly coupled to the A, B, and C phases of the power supply via voltage and current transformers to achieve real-time sensing and sampling of voltage and current in each phase. Each unit employs a high-speed analog-to-digital converter with a sampling frequency no less than 256 times the fundamental AC frequency to ensure high-precision capture of transient voltage and current changes. Subsequently, the system issues a unified sampling command to the N current and voltage sensing units through a synchronous triggering mechanism to ensure the time consistency of the sampled data. The analog voltage and current signals collected by each sensing unit are filtered, isolated, and then input to the high-speed analog-to-digital converter for digital processing to obtain N sets of real-time three-phase voltage and current signals. After acquisition, the signals are transmitted via a bus to form N complete real-time current and voltage signal data streams, providing high-precision, synchronized basic data input for subsequent phase-locked analysis, dynamic power quality decomposition, and flexible voltage regulation control.

[0019] Perform PLL phase decomposition based on the dq rotating coordinate system on the N real-time current and voltage signals to obtain N dynamic power quality components.

[0020] In one embodiment, the acquired three-phase voltage signal is first transformed into a coordinate system using a Clarke transform, converting it from a static abc three-phase coordinate system to an αβ two-phase orthogonal coordinate system for subsequent phase locking. Then, based on the voltage vector in the αβ coordinate system, a synchronous coordinate rotating PLL control loop is constructed. A phase detector calculates the phase error between the voltage vector and the reference vector in real time. A loop filter performs PI control filtering on the phase error signal to eliminate high-frequency disturbances. A numerically controlled oscillator adjusts the internal phase angle frequency according to the filtered output to synchronize the rotating coordinate system with the input voltage vector, thus outputting the real-time grid synchronization angle. Subsequently, using the phase-locked real-time grid synchronization angle, a Park transform is performed to further convert the αβ coordinate system signal to a rotating dq coordinate system. Through this transformation, the original three-phase alternating signal is converted into d-axis DC and q-axis DC components. Harmonics and transient components are represented as alternating signals, where the d-axis DC component represents the active power component and the q-axis DC component represents the reactive power component. Subsequently, the converted dq component signals are analyzed and filtered in real time to extract dynamic power quality elements such as the fundamental positive sequence component, active power component, reactive power component, and harmonic spectrum component of each AC power source. By summarizing these components, N dynamic power quality components are formed, which serve as input parameters for subsequent power supply stability evaluation and flexible voltage regulation analysis, helping to achieve quantitative and controllable power quality processing.

[0021] Furthermore, each dynamic power quality component includes a fundamental positive sequence component, an active power component, a reactive power component, and a harmonic spectrum component.

[0022] Optionally, each dynamic power quality component consists of a fundamental positive-sequence component, an active power component, a reactive power component, and a harmonic spectrum component. This comprehensively characterizes the power quality status of each AC power source at a specific moment. The fundamental positive-sequence component is extracted by performing Fourier decomposition on the three-phase voltage and current signals. It reflects the stable output characteristics of the power source at normal power frequency and is the primary basis for judging voltage amplitude, phase angle, and frequency deviation. The active power component is obtained by multiplying the d-axis DC quantity in the dq coordinate system with the current signal, representing the effective power actually transmitted to the load and used to measure power conversion efficiency. The reactive power component is calculated from the q-axis DC quantity in the dq coordinate system, reflecting the periodic interaction characteristics of electromagnetic energy and is of great significance for voltage support and power factor regulation. The harmonic spectrum component is extracted in the frequency domain using fast Fourier transform or weighted spectrum analysis methods, covering the amplitude and phase information of the 3rd, 5th, 7th, and higher harmonics, used to reflect waveform distortion caused by nonlinear loads, power electronic devices, etc. By jointly calculating the above four types of components, the system can obtain a complete power quality feature set for each power source, providing accurate basic parameter inputs for subsequent stability quantification evaluation and flexible voltage regulation control, ensuring high sensitivity and high responsiveness under multi-disturbance environments.

[0023] Based on the N dynamic power quality components, a multi-dimensional quantitative evaluation of power supply stability is performed, and N comprehensive stability scores are output.

[0024] In one embodiment, after obtaining N dynamic power quality components, the fundamental positive-sequence component, active power component, reactive power component, and harmonic spectrum component of each dynamic power quality component are normalized to their maximum and minimum values ​​to eliminate differences in voltage level, load scale, and measurement sensitivity among different power supplies. Subsequently, based on the corresponding N scoring regression functions, multi-dimensional power supply stability quantification calculations are performed on each normalized dynamic power quality component, mapping each dynamic power quality component to a single comprehensive stability score. These comprehensive stability scores reflect the power supply stability of the corresponding power supply under real-time operating conditions. Higher scores indicate more stable voltage output, smaller power fluctuations, and better power quality, providing a basis for subsequent selection of target power supplies.

[0025] Furthermore, based on the aforementioned N dynamic power quality components, a multi-dimensional quantitative evaluation of power supply stability is performed, outputting N comprehensive stability scores. The method includes:

[0026] Based on the load type codes of the N three-phase AC power supplies, N sample power quality component sets and N sample stability scores are retrieved; multiple regression analysis is performed on the N sample power quality component sets and N sample stability scores to obtain N score regression functions; by loading the N dynamic power quality components into the N score regression functions, multi-dimensional power supply stability quantification calculation is performed, and the N comprehensive stability scores are output.

[0027] Preferably, for N three-phase AC power sources, based on the load type code corresponding to each power source, N sample power quality component sets and N sample stability scores are retrieved from a preset historical operation database. The sample power quality component sets include time-series data such as the sample fundamental positive-sequence component, sample active power component, sample reactive power component, and sample harmonic spectrum component. The sample stability scores are standard scores obtained by evaluating power supply stability under historical operating conditions. The load type code is used to distinguish different load characteristics such as resistive, inductive, capacitive, or mixed types to ensure the model's accuracy and adaptability for different power supply application scenarios. Subsequently, multiple regression analysis is performed on each sample data source. In this process, the power quality component sets are used as independent variables, and the stability scores as dependent variables. Mathematical methods such as least squares, multiple linear regression, or partial least squares regression are used to establish the functional relationship between power quality parameters and stability scores, resulting in N score regression functions. Each score regression function corresponds to one AC power source and can describe the influence of voltage and current characteristic parameters on the stability score under specific load and grid fluctuation conditions. After training N scoring regression functions, the N currently collected dynamic power quality components are loaded into their respective scoring regression functions to perform multi-dimensional power supply stability calculations. By substituting parameters such as the current fundamental positive sequence component, active power component, reactive power component, and harmonic spectrum component, the predicted score values ​​of each scoring regression function are calculated as N comprehensive stability scores. These comprehensive stability scores accurately reflect the power supply stability of each AC power source under actual operating conditions at the current moment, providing a reliable quantitative basis for subsequent flexible voltage regulation to select the optimal power source.

[0028] Furthermore, the method also includes:

[0029] Based on the fluctuation attributes of the N real-time current and voltage signals, state transition prediction is performed to obtain N updated power quality components; the N updated power quality components are loaded into the N scoring regression functions to perform multi-dimensional power supply stability quantification calculation, and N updated stability scores are output; the N updated stability scores are superimposed on the state transition probabilities of the N updated power quality components to obtain the N comprehensive stability scores, resulting in N comprehensive prediction scores; the N comprehensive prediction scores are used to screen the target AC power source.

[0030] Optionally, firstly, based on the fluctuation attributes of the N real-time current and voltage signals collected, corresponding fluctuation feature vectors are constructed. These fluctuation feature vectors typically include features such as voltage fluctuation intensity and harmonic mutation energy. Then, a preset state transition probability matrix is ​​invoked, and state prediction calculations are performed based on these fluctuation feature vectors to obtain the real-time state probability vector for the next moment. On this basis, through multi-step iteration, state evolution prediction is performed within a short timescale to obtain N updated power quality components. These updated power quality components reflect the possible changing trends of voltage, active power, reactive power, and harmonic distribution under the future predicted state. Next, the N updated power quality components are loaded into N scoring regression functions, and multi-dimensional power supply stability quantification calculations are performed again to obtain N updated stability scores. These updated stability scores are used to reflect the potential stability changes of each power source under the predicted state. Finally, based on the state transition probability, the updated stability scores and the current comprehensive stability score are superimposed and weighted to obtain N comprehensive prediction scores. This process comprehensively considers the current power quality level and future state evolution trends, achieving dynamic prediction of power supply stability. Finally, based on the descending order of these comprehensive prediction scores, the system selects the AC power supply with the highest score as the target AC power supply for subsequent flexible voltage regulation control and power optimization output.

[0031] Furthermore, based on the fluctuation attributes of the N real-time current and voltage signals, state transition prediction is performed to obtain N updated power quality components. The method includes:

[0032] A fluctuation feature vector is constructed based on the first real-time current and voltage signal, wherein the fluctuation feature vector includes voltage fluctuation intensity, harmonic mutation energy, voltage sag label, and current surge label; the fluctuation feature vector is input into the state space model for real-time discrete state definition; state prediction is performed on the state transition probability matrix according to the real-time discrete state, and a real-time state probability vector is output; starting from the fluctuation feature vector and using the real-time state probability vector as the initial state vector, multi-step state prediction is performed, and the first updated power quality component is output.

[0033] Optionally, firstly, a real-time current and voltage signal from one AC power source is randomly selected from N AC power sources as the first real-time current and voltage signal. Dynamic features are extracted from this signal within a time window. Specifically, by performing statistical and spectral analysis on rapidly sampled voltage and current data, a fluctuation feature vector reflecting instantaneous disturbance characteristics is constructed. This fluctuation feature vector includes voltage fluctuation intensity, harmonic mutation energy, voltage sag label, and current surge label. Voltage fluctuation intensity is obtained by calculating the deviation between the fluctuation amplitude of the effective voltage value or peak value within the time window and the reference value. Harmonic mutation energy is calculated by integrating the rate of change of spectral energy, reflecting the suddenness of nonlinear disturbances. The voltage sag label is obtained by setting a voltage threshold to monitor sudden drop events and recording them with binary flags. The current surge label is obtained by detecting the current rise slope or surge peak exceeding limits and recording it with binary flags. Subsequently, this fluctuation feature vector is input into a preset state-space model. Based on its feature value range and combination relationship, the current power grid operating state is mapped to a discrete state category, such as stable, mild disturbance, moderate disturbance, or severe disturbance, thus completing the real-time discrete state definition. Subsequently, based on the defined discrete states, state prediction operations are performed within a pre-trained or pre-defined state transition probability matrix. This matrix describes the evolution of different disturbance states over a short timescale. Through matrix multiplication, the current discrete state is mapped to a set of real-time state probability vectors, representing the probability distribution of transitions from the current state to various possible disturbance states. Finally, starting with the fluctuation feature vector, the real-time state probability vector is used as the initial state vector, and multi-step iterative predictions are performed on the state transition probability matrix. By progressively expanding the prediction step size, the dynamic evolution trends of voltage, active power, reactive power, and harmonic components are obtained. After each prediction step, feature aggregation and filtering are performed, ultimately outputting the first updated power quality component. In summary, this process realizes the predictive transformation from instantaneous fluctuation information to future power quality change trends, enabling the system to complete response preparation before disturbances occur, providing reliable predictive input and decision-making basis for flexible voltage regulation control.

[0034] Furthermore, the method involves performing state prediction based on the real-time discrete state in the state transition probability matrix and outputting a real-time state probability vector.

[0035] The real-time discrete state is binary encoded to output a real-time state vector; the state transition probability matrix is ​​retrieved according to the industrial application scenario of the target AC power supply; the state transition product of the real-time state vector is calculated from the state transition probability matrix to output an initial probability vector; the initial probability vector is weighted and corrected according to the load type encoding of the target AC power supply to output the real-time state probability vector.

[0036] Optionally, after discretizing and classifying the current AC power supply fluctuation characteristics, each real-time discrete state is represented by a unique binary code. For example, a stable state is encoded as 00, a minor disturbance as 01, a moderate disturbance as 10, and a severe disturbance as 11. This allows for unified modeling and rapid computation of complex disturbance states at the digital level. The encoded state data is encapsulated into a real-time state vector, representing the current discrete operating state of the power supply. Subsequently, based on the industrial application scenario of the target AC power supply, such as metallurgical processing, motor drive, data center power supply, or rail transit traction, the corresponding state transition probability matrix is ​​retrieved from the database. These state transition probability matrices are obtained by training a Markov chain probability model with long-term operating data and are used to describe the transition patterns between different disturbance states under specific industrial conditions. For example, in a motor drive scenario, the probability of a minor disturbance transitioning to a moderate disturbance is relatively high, while in a data center scenario, this probability is relatively low. Through scenario-based matrix retrieval, the system can improve the targeting and accuracy of state prediction. Next, a state transition product operation is performed on the real-time state vector within the state transition probability matrix. Specifically, the real-time state vector is left-multiplied by the state transition probability matrix, and a set of initial probability vectors is obtained through matrix multiplication. This initial probability vector reflects the probability distribution of transitions from the current state to each possible state in the next time step. Finally, based on the load type encoding of the target AC power supply, a pre-set correction coefficient matrix is ​​retrieved, and this correction coefficient matrix is ​​multiplied element-wise with the initial probability vector to output the final real-time state probability vector. This real-time state probability vector comprehensively reflects the current power supply state, scenario characteristics, and load response features, providing a probabilistic basis input for subsequent multi-step state prediction and power quality component updates, thereby improving the prediction accuracy and stability of the system under complex operating conditions.

[0037] Furthermore, starting with the fluctuation feature vector and using the real-time state probability vector as the initial state vector, multi-step state prediction is performed to output the first updated power quality component. The method includes:

[0038] Using the real-time state probability vector as the initial state vector, multi-step state prediction is iteratively performed on the state transition probability matrix to output a multi-step predicted state probability vector; the multi-step predicted state probability vector is aggregated to extract steady-state probability components, transient probability components, fluctuation probability components, and harmonic probability components; based on the steady-state probability components, transient probability components, fluctuation probability components, and harmonic probability components, the fundamental voltage prediction component, active power prediction component, reactive power prediction component, and harmonic spectrum prediction component are mapped and calculated to form the first updated power quality component.

[0039] Optionally, the real-time state probability vector calculated in the previous stage is first used as the initial state vector and input into the state transition probability matrix corresponding to the target industrial scenario. By performing multiple iterative multiplication operations on the matrix, for example, 3 to 10 time steps, the state probability distribution at multiple time steps is obtained, forming a multi-step predicted state probability vector sequence. This multi-step predicted state probability vector sequence describes the possibility of transitioning from the current state to different disturbance states in a short period of time in the future, and is the core foundation of power quality prediction. Subsequently, the steady-state probability component, transient probability component, fluctuation probability component, and harmonic probability component obtained from each prediction are extracted from the multi-step prediction state probability vector. These components are then aggregated using a maximum value filtering method to obtain the required steady-state probability component, transient probability component, fluctuation probability component, and harmonic probability component. Among them, the steady-state probability component reflects the highest probability of maintaining a stable operating state in future multi-step predictions; the transient probability component reflects the maximum probability of occurrence under short-term disturbances or transitional states; the fluctuation probability component reflects the strongest probability trend of voltage periodic fluctuations or random fluctuations within the prediction period; and the harmonic probability component reflects the maximum risk of harmonic energy enhancement or spectral line diffusion during future disturbance evolution. Subsequently, based on the obtained steady-state probability components, transient probability components, fluctuation probability components, and harmonic probability components, the current fundamental voltage prediction component is calculated using the formula: Fundamental voltage prediction component = Current measured fundamental voltage value × (1 + Voltage fluctuation intensity) × (1 − 0.4 × Fluctuation probability component); the current harmonic spectrum prediction component is calculated using the formula: Harmonic spectrum prediction component = Harmonic mutation energy × Harmonic probability component; the current harmonic spectrum prediction component is calculated using the formula: Active power prediction component = Fundamental voltage prediction component × Fundamental current d-axis component × Phase angle offset cosine value; the current harmonic spectrum prediction component is calculated using the formula: Reactive power prediction component = Fundamental voltage prediction component × Fundamental current q-axis component × Phase angle offset sine value. Finally, the obtained fundamental voltage prediction component, active power prediction component, reactive power prediction component, and harmonic spectrum prediction component are vector-concatenated in a set order to generate a structured first updated power quality component. This first updated power quality component fully depicts the evolution trend of voltage, power, harmonic disturbances, etc. in the short term, providing an accurate predictive input basis for flexible stepless voltage regulation, thereby improving the ability to perceive complex disturbances in advance and the speed of dynamic control response.

[0040] Based on the descending order of the N comprehensive stability scores, the target AC power source corresponding to the maximum score is extracted.

[0041] In one embodiment, after obtaining N comprehensive stability scores, these scores are sorted in descending order. Based on the sorting result, the highest comprehensive stability score is extracted, and the three-phase AC power supply corresponding to this comprehensive stability score is marked as the target AC power supply, providing a reliable basis for subsequent voltage dynamic optimization and energy allocation.

[0042] Based on the target dynamic power quality components of the target AC power supply, a flexible stepless voltage regulation analysis is performed, and a dynamic compensation command is output.

[0043] In one embodiment, after identifying the target AC power source, the target dynamic power quality components of the target AC power source are quantized to extract the deviation information required for adjustment. Corresponding PWM commands are generated for different types of power quality deviations. Then, a three-dimensional spatial vector modulation method is used to fuse the generated PWM commands in both the time domain and amplitude domain, thereby integrating the various commands into a unified dynamic control signal, forming the final dynamic compensation command. This dynamic compensation command can be directly issued to achieve flexible, stepless, and rapid adjustment of the target AC power source's output voltage, thereby ensuring voltage stability and improving power quality.

[0044] Furthermore, based on the target dynamic power quality components of the target AC power supply, a flexible stepless voltage regulation analysis is performed, and a dynamic compensation command is output. The method includes:

[0045] The target dynamic power quality components are matched with differentiated PWM commands on a component-by-component basis, and multi-component PWM commands are output. The multi-component PWM commands are fused by three-dimensional space vector modulation, and the dynamic compensation command is output.

[0046] Optionally, the fundamental voltage component, active power component, reactive power component, and harmonic spectrum component are first extracted from the dynamic power quality components of the target AC power source. These components correspond to different control objectives such as voltage amplitude regulation, power balance regulation, and harmonic suppression. To achieve refined control, independent PWM command generation channels are configured for different power quality components in the controller, and corresponding control strategy mapping tables are established. For the fundamental voltage component, a main voltage regulation PWM command is generated based on the relationship between the voltage deviation and the reference value to control the rise and fall and steady-state maintenance of the output voltage. For the active power component, a power compensation type PWM command is matched according to the power demand change to improve the output power response capability and maintain the voltage stability at the load end. For the reactive power component, a corresponding phase correction PWM command is calculated based on the current phase offset angle to improve the power factor and reduce reactive power fluctuations. For the harmonic spectrum component, a frequency domain suppression type PWM command is matched through harmonic frequency identification and amplitude decomposition to reduce high-order harmonics and waveform distortion. By summarizing these commands, a set of multi-component PWM commands classified according to power quality components is obtained. Subsequently, the multi-component PWM commands are input into the three-dimensional space vector modulation unit. This three-dimensional space vector modulation module, based on the three-phase voltage space vector, maps each component command to a spatial coordinate system, achieving coordinated fusion of different control objectives through vector decomposition and weighted superposition. Specifically, the space voltage vector corresponding to each PWM command is first calculated, and then vector synthesis is performed based on the amplitude and phase relationship to generate a single composite PWM signal. This composite PWM signal is then used as a dynamic compensation command to dynamically adjust and optimize the output voltage of the target AC power supply at the millisecond level, providing technical support for high-quality power supply.

[0047] After the target AC power supply is dynamically and multi-level optimized by driving multi-level PWM modulation according to the dynamic compensation command, the power of the target AC power supply is delivered to the target electrical appliance through H-bridge converter topology.

[0048] In one embodiment, the obtained dynamic compensation command is first input to a multi-level PWM control unit. This multi-level PWM control unit adopts a hierarchical modulation architecture, including a main controller, a driver stage, and a power conversion stage. The main controller adjusts the carrier frequency, modulation ratio, and phase difference in real time according to the dynamic compensation command, thereby controlling the turn-on and turn-off timing of power switching devices at each stage, such as IGBTs or MOSFETs, enabling flexible adjustment of the output voltage within a continuous range. Under multi-level PWM control, the fundamental component, harmonic components, and transient fluctuations of the voltage can be optimized hierarchically. In the first level of optimization, the fundamental output is controlled to achieve steady-state voltage amplitude maintenance and active power regulation. In the second level of optimization, higher harmonics are weakened by changing the superimposed waveform of the high-frequency carrier. In the third level of optimization, the output voltage is quickly adjusted during grid disturbances or load changes to ensure power supply continuity and stability. After completing the multi-level PWM optimization, the modulated voltage signal is sent to an H-bridge converter topology. The H-bridge consists of four power switching arms and uses fully controllable devices, such as IGBTs, to achieve AC-DC-AC energy conversion. Specifically, the input AC power is rectified to form a DC bus voltage, which is then synthesized using an H-bridge converter topology according to PWM commands, resulting in an output AC voltage signal with controllable amplitude and phase. Finally, the optimized AC power is delivered to the target appliance through filtering and isolation stages. The output voltage waveform approximates an ideal sine wave, with significantly reduced harmonic content. The voltage amplitude and phase are automatically adjusted according to real-time load demands, ensuring stable power supply, rapid response, and excellent power quality, thereby improving the stability of power transmission and the robustness of system operation.

[0049] Furthermore, the H-bridge converter topology is connected in series at the AC output terminal of the target AC power supply.

[0050] Preferably, the H-bridge converter topology is directly connected in series at the AC output terminal of the target AC power supply. With this arrangement, the H-bridge can superimpose or reduce the output voltage according to dynamic compensation commands without altering the main power supply line structure, achieving flexible voltage boosting, bucking, or harmonic suppression. When the output voltage fluctuates, the H-bridge can quickly respond and inject compensation voltage to keep the load voltage stable.

[0051] In summary, the embodiments of this application have at least the following technical effects:

[0052] First, N current and voltage sensing units are triggered to perform high-speed sampling of N three-phase AC power supplies, obtaining N real-time current and voltage signals. These N current and voltage sensing units are directly coupled to each phase of the N three-phase AC power supplies. Next, PLL phase decomposition based on a dq rotating coordinate system is performed on the N real-time current and voltage signals to obtain N dynamic power quality components. Then, a multi-dimensional power supply stability quantification evaluation is performed based on these N dynamic power quality components, outputting N comprehensive stability scores. Afterward, the target AC power supply corresponding to the highest score is extracted based on the descending order of the N comprehensive stability scores. Then, flexible stepless voltage regulation analysis is performed based on the target dynamic power quality components of the target AC power supply, outputting a dynamic compensation command. Finally, after multi-level PWM modulation driven by the dynamic compensation command to perform multi-level optimization of the dynamic voltage of the target AC power supply, the power from the target AC power supply is delivered to the target appliance through an H-bridge converter topology. It solves the technical problems of slow response, limited adjustment range and inability to intelligently select the optimal power supply for collaborative compensation in multi-power supply scenarios of traditional voltage regulation systems, resulting in insufficient voltage control accuracy and reliability. It achieves the technical effect of achieving millisecond-level dynamic response through high-speed sampling and PLL phase locking, effectively suppressing harmonics and improving power quality and power supply stability.

[0053] Example 2, based on the same inventive concept as the flexible voltage regulation control method based on power electronic conversion in the previous examples, such as... Figure 2 As shown, this application provides a flexible voltage regulation control system based on power electronic conversion, the system comprising:

[0054] Signal sampling module 11: Triggers N current and voltage sensing units to perform high-speed sampling of N three-phase AC power supplies, obtaining N real-time current and voltage signals, wherein the N current and voltage sensing units are directly coupled to each phase line of the N three-phase AC power supplies; Phase decomposition module 12: Performs PLL phase decomposition based on the dq rotating coordinate system on the N real-time current and voltage signals, obtaining N dynamic power quality components; Stability evaluation module 13: Performs multi-dimensional quantitative evaluation of power supply stability based on the N dynamic power quality components, outputting N comprehensive stability scores; Descending order sorting module 14: Extracts the target AC power supply corresponding to the maximum score based on the descending order of the N comprehensive stability scores; Voltage regulation analysis module 15: Performs flexible stepless voltage regulation analysis based on the target dynamic power quality components of the target AC power supply, outputting dynamic compensation commands; Voltage optimization module 16: Drives multi-level PWM modulation to perform multi-level dynamic voltage optimization of the target AC power supply according to the dynamic compensation commands, and then delivers the power of the target AC power supply to the target appliance through an H-bridge converter topology.

[0055] Furthermore, the phase decomposition module 12 is used to perform the following method:

[0056] Each dynamic power quality component includes the fundamental positive sequence component, active power component, reactive power component, and harmonic spectrum component.

[0057] Furthermore, the stability evaluation module 13 is used to perform the following method:

[0058] Based on the load type codes of the N three-phase AC power supplies, N sample power quality component sets and N sample stability scores are retrieved; multiple regression analysis is performed on the N sample power quality component sets and N sample stability scores to obtain N score regression functions; by loading the N dynamic power quality components into the N score regression functions, multi-dimensional power supply stability quantification calculation is performed, and the N comprehensive stability scores are output.

[0059] Furthermore, the stability evaluation module 13 is used to perform the following method:

[0060] Based on the fluctuation attributes of the N real-time current and voltage signals, state transition prediction is performed to obtain N updated power quality components; the N updated power quality components are loaded into the N scoring regression functions to perform multi-dimensional power supply stability quantification calculation, and N updated stability scores are output; the N updated stability scores are superimposed on the state transition probabilities of the N updated power quality components to obtain the N comprehensive stability scores, resulting in N comprehensive prediction scores; the N comprehensive prediction scores are used to screen the target AC power source.

[0061] Furthermore, the stability evaluation module 13 is used to perform the following method:

[0062] A fluctuation feature vector is constructed based on the first real-time current and voltage signal, wherein the fluctuation feature vector includes voltage fluctuation intensity, harmonic mutation energy, voltage sag label, and current surge label; the fluctuation feature vector is input into the state space model for real-time discrete state definition; state prediction is performed on the state transition probability matrix according to the real-time discrete state, and a real-time state probability vector is output; starting from the fluctuation feature vector and using the real-time state probability vector as the initial state vector, multi-step state prediction is performed, and the first updated power quality component is output.

[0063] Furthermore, the stability evaluation module 13 is used to perform the following method:

[0064] The real-time discrete state is binary encoded to output a real-time state vector; the state transition probability matrix is ​​retrieved according to the industrial application scenario of the target AC power supply; the state transition product of the real-time state vector is calculated from the state transition probability matrix to output an initial probability vector; the initial probability vector is weighted and corrected according to the load type encoding of the target AC power supply to output the real-time state probability vector.

[0065] Furthermore, the stability evaluation module 13 is used to perform the following method:

[0066] Using the real-time state probability vector as the initial state vector, multi-step state prediction is iteratively performed on the state transition probability matrix to output a multi-step predicted state probability vector; the multi-step predicted state probability vector is aggregated to extract steady-state probability components, transient probability components, fluctuation probability components, and harmonic probability components; based on the steady-state probability components, transient probability components, fluctuation probability components, and harmonic probability components, the fundamental voltage prediction component, active power prediction component, reactive power prediction component, and harmonic spectrum prediction component are mapped and calculated to form the first updated power quality component.

[0067] Furthermore, the voltage regulation analysis module 15 is used to perform the following method:

[0068] The target dynamic power quality components are matched with differentiated PWM commands on a component-by-component basis, and multi-component PWM commands are output. The multi-component PWM commands are fused by three-dimensional space vector modulation, and the dynamic compensation command is output.

[0069] Furthermore, the voltage optimization module 16 is used to perform the following method:

[0070] The H-bridge converter topology is connected in series at the AC output terminal of the target AC power supply.

[0071] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A flexible voltage regulation control method based on power electronic conversion, characterized in that, The method includes: N current and voltage sensing units are triggered to perform high-speed sampling of N three-phase AC power supplies to obtain N real-time current and voltage signals, wherein the N current and voltage sensing units are directly coupled to each phase line of the N three-phase AC power supplies. Perform PLL phase decomposition based on the dq rotating coordinate system on the N real-time current and voltage signals to obtain N dynamic power quality components, wherein each dynamic power quality component includes a fundamental positive sequence component, an active power component, a reactive power component, and a harmonic spectrum component. Based on the N dynamic power quality components, a multi-dimensional quantitative evaluation of power supply stability is performed, and N comprehensive stability scores are output. Based on the descending order of the N comprehensive stability scores, extract the target AC power source corresponding to the maximum score. Based on the target dynamic power quality components of the target AC power supply, a flexible stepless voltage regulation analysis is performed, and a dynamic compensation command is output. After the target AC power supply is dynamically and multi-level optimized by driving multi-level PWM modulation according to the dynamic compensation command, the power of the target AC power supply is delivered to the target electrical appliance through the H-bridge converter topology. Among them, a multi-dimensional quantitative evaluation of power supply stability is performed based on the N dynamic power quality components, and N comprehensive stability scores are output, including: Based on the load type codes of the N three-phase AC power supplies, retrieve N sample power quality component sets and N sample stability scores; Multiple regression analysis was performed on the N sample power quality component sets and the N sample stability scores to obtain N score regression functions; By loading the N dynamic power quality components onto the N scoring regression functions, multi-dimensional power supply stability quantification calculation is performed, and the N comprehensive stability scores are output. This also includes: Based on the fluctuation attributes of the N real-time current and voltage signals, state transition prediction is performed to obtain N updated power quality components. The N updated power quality components are loaded into the N scoring regression functions to perform multi-dimensional power supply stability quantification calculations and output N updated stability scores. The N comprehensive prediction scores are obtained by superimposing the N updated stability scores on the state transition probabilities of the N updated power quality components to the N comprehensive stability scores; The target AC power source is selected using the N comprehensive prediction scores.

2. The flexible voltage regulation control method based on power electronic conversion as described in claim 1, characterized in that, Based on the fluctuation attributes of the N real-time current and voltage signals, state transition prediction is performed to obtain N updated power quality components. The method includes: A fluctuation feature vector is constructed based on the first real-time current and voltage signal, wherein the fluctuation feature vector includes voltage fluctuation intensity, harmonic abrupt energy, voltage sag label and current surge label; The fluctuation feature vector is input into the state space model for real-time discrete state definition; Based on the real-time discrete state, state prediction is performed on the state transition probability matrix, and a real-time state probability vector is output. Starting with the fluctuation feature vector and the real-time state probability vector as the initial state vector, multi-step state prediction is performed, and the first updated power quality component is output.

3. The flexible voltage regulation control method based on power electronic conversion as described in claim 2, characterized in that, Based on the real-time discrete state, state prediction is performed on the state transition probability matrix, and a real-time state probability vector is output. The method includes: Perform binary encoding on the real-time discrete state and output a real-time state vector; The state transition probability matrix is ​​retrieved based on the industrial application scenario of the target AC power supply; The state transition product of the real-time state vector is calculated using the state transition probability matrix, and the initial probability vector is output. The initial probability vector is weighted and corrected according to the load type encoding of the target AC power supply, and the real-time state probability vector is output.

4. The flexible voltage regulation control method based on power electronic conversion as described in claim 2, characterized in that, Starting with the fluctuation feature vector and using the real-time state probability vector as the initial state vector, a multi-step state prediction is performed to output the first updated power quality component. The method includes: Using the real-time state probability vector as the initial state vector, multi-step state prediction is iteratively performed on the state transition probability matrix to output a multi-step predicted state probability vector. By aggregating the multi-step predicted state probability vector, steady-state probability components, transient probability components, fluctuation probability components, and harmonic probability components are extracted. Based on the steady-state probability component, transient probability component, fluctuation probability component, and harmonic probability component, the fundamental voltage prediction component, active power prediction component, reactive power prediction component, and harmonic spectrum prediction component are mapped and calculated to form the first updated power quality component.

5. The flexible voltage regulation control method based on power electronic conversion as described in claim 1, characterized in that, Based on the target dynamic power quality components of the target AC power supply, a flexible stepless voltage regulation analysis is performed, and a dynamic compensation command is output. The method includes: Perform component-by-component differentiated PWM instruction matching on the target dynamic power quality components and output multi-component PWM instructions; The dynamic compensation command is output by fusing the multi-component PWM command with three-dimensional space vector modulation.

6. The flexible voltage regulation control method based on power electronic conversion as described in claim 1, characterized in that, The H-bridge converter topology is connected in series at the AC output terminal of the target AC power supply.

7. A flexible voltage regulation control system based on power electronic conversion, characterized in that, For implementing the flexible voltage regulation control method based on power electronic conversion as described in any one of claims 1-6, the system comprises: Signal sampling module: triggers N current and voltage sensing units to perform high-speed sampling of N three-phase AC power supplies to obtain N real-time current and voltage signals, wherein the N current and voltage sensing units are directly coupled to each phase line of the N three-phase AC power supplies; Phase decomposition module: Performs PLL phase decomposition based on the dq rotating coordinate system on the N real-time current and voltage signals to obtain N dynamic power quality components; Stability evaluation module: Based on the N dynamic power quality components, perform multi-dimensional quantitative evaluation of power supply stability and output N comprehensive stability scores; Descending order sorting module: Based on the descending order sorting results of the N comprehensive stability scores, extract the target AC power source corresponding to the maximum score; Voltage regulation analysis module: performs flexible stepless voltage regulation analysis based on the target dynamic power quality components of the target AC power supply, and outputs dynamic compensation commands; Voltage optimization module: After performing dynamic voltage multi-level optimization of the target AC power supply according to the dynamic compensation command, the power of the target AC power supply is delivered to the target electrical appliance through the H-bridge converter topology.

Images