Magnetic resistance type dynamic voltage recovery device and method based on particle swarm optimization

Abstract

The application provides a magnetic resistance type dynamic voltage recovery device and method based on particle swarm optimization, relates to the technical field of dynamic voltage recovery devices, and through deep time sequence analysis on historical voltage drop, the amplitude and duration of voltage drop are dynamically predicted, so that the amplitude interval and duration of injected voltage are adaptively determined; a particle swarm algorithm is combined to perform multi-objective optimization on a control scheme, and the performance of an evaluation scheme is fed back, so that precise design and real-time adjustment of the injected voltage waveform are realized; a multi-scheme screening and switching mechanism effectively guarantees stable operation and safety of the system under abnormal working conditions, improves the intelligent level and compensation quality of the dynamic voltage recovery device as a whole, meets the requirements of rapid response and efficient compensation under a complex power grid environment, and enhances the stability and reliability of the system as a whole.

Description

Technical Field

[0001] This invention relates to the field of dynamic voltage recovery device technology, specifically to a magnetic reactive dynamic voltage recovery device and method based on particle swarm optimization. Background Technology

[0002] With the continuous development of power systems, power grids often encounter voltage dips when dealing with load fluctuations, equipment maintenance, and transient disturbances, which seriously affect power supply quality and user safety. Traditional dynamic voltage restoration devices typically employ compensation strategies based on fixed parameters or experience, which are difficult to adapt to the variability and complexity of voltage dips. In particular, they cannot achieve precise compensation control under different operating conditions, leading to unstable compensation effects, increased equipment load, and even safety risks.

[0003] In the prior art, CN120999649A discloses an optimized configuration method and device for a magnetic reactive dynamic voltage restorer, relating to the field of optimized configuration technology. Specifically, it includes: acquiring historical voltage sag event data of the target distribution bus and the electrical parameters of the downstream load motor; screening a set of events to be compensated based on the principle that the sag depth exceeds the load tolerance threshold and the duration is less than the critical stall time, and determining a compensation strategy based on the comparison between the maximum phase jump angle and the threshold; determining the maximum compensation voltage based on the maximum sag depth, determining the rated output current based on the load parameters, calculating the theoretical energy required to compensate the longest event based on the compensation strategy, and determining the actual required energy based on the energy correction coefficient of the device topology; finally, generating an optimized configuration scheme based on the compensation strategy, the maximum compensation voltage, the rated output current, and the actual required energy. While this scheme can achieve accurate mapping from measured grid data to key device parameters, it focuses on static optimization and has a single compensation strategy, failing to achieve adaptive control under different voltage drop events. This results in insufficient compensation accuracy and limited applicability in complex grid environments.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a magnetic reactive dynamic voltage recovery device and method based on particle swarm optimization to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] The particle swarm optimization-based magnetic reactive dynamic voltage recovery device specifically includes a control device and a voltage recovery device body, wherein:

[0008] The control device includes:

[0009] The data acquisition module is used to collect historical voltage drops and transmit them to the time series analysis module.

[0010] The timing analysis module has a built-in timing analysis model for performing timing analysis on historical voltage drops to obtain predicted voltage drops, and generating the amplitude range of the injection voltage based on the predicted voltage drops.

[0011] The algorithm optimization module includes a data detection unit, an algorithm optimization unit, and a scheme selection unit, wherein:

[0012] The data detection unit is used to detect the temperature and electrical parameter changes generated when the voltage recovery device body generates injection voltages of different amplitudes.

[0013] The algorithm optimization unit has a built-in particle swarm optimization algorithm, which is used to search within the amplitude range of the injected voltage and generate several sets of control schemes for the injected voltage.

[0014] The scheme screening unit is used to select the optimal scheme and at least one backup scheme from several control schemes, and transmit the optimal scheme and the backup scheme to the voltage recovery device body.

[0015] The signal feedback module includes a voltage detection unit and a scheme switching unit, wherein:

[0016] The voltage detection unit is used to detect the injection voltage generated by the voltage recovery device body in real time, obtain the measured value of the injection voltage, and transmit it to the scheme switching unit.

[0017] The scheme switching unit is used to generate an abnormal index based on the measured value of the injected voltage, and when the abnormal index exceeds the preset index threshold, control the voltage recovery device to switch the injection voltage scheme.

[0018] The voltage recovery device body is used to generate injection voltages of different amplitudes by adjusting its own magnetic circuit and injection current.

[0019] Preferably, the prediction logic for the amplitude range of the injected voltage is as follows:

[0020] The collected historical voltage drops are preprocessed, including but not limited to noise reduction, outlier removal, and normalization, and a continuous time series dataset is generated.

[0021] A time series analysis is performed on the time series dataset based on historical drop voltages, and the predicted drop voltage and predicted drop duration are obtained based on the time series analysis results. The maximum and minimum values ​​of the predicted drop voltage are used as the upper and lower limits of the amplitude range of the injected voltage, respectively.

[0022] Preferably, the working logic of the algorithm optimization module is as follows:

[0023] The duration of the injected voltage is determined based on the predicted drop duration. Several sets of control points are evenly set within the duration, with each set of control points corresponding to an injected voltage of different amplitudes. All control points together constitute a control scheme for the injected voltage.

[0024] Each time the algorithm optimization unit generates a control scheme using the particle swarm optimization algorithm, it inputs it into the voltage recovery device body and uses the data detection unit to detect the temperature changes and electrical parameter changes generated therefrom. The electrical parameters include, but are not limited to, the injected current and output voltage of the voltage recovery device body.

[0025] The algorithm optimization unit scores each control scheme based on the detection results output by the data detection unit, and transmits the control scheme and its corresponding score value to the scheme selection unit.

[0026] The scheme selection unit ranks the control schemes in descending order of their scores, designating the top-ranked scheme as the optimal scheme and the second-ranked or subsequent schemes as backup schemes.

[0027] Preferably, the algorithm optimization unit scores the control scheme by constructing an objective function;

[0028] The objective function is obtained based on the detection results output by the data detection unit, and the objective function is composed of voltage error term, current impact term, energy loss term, and temperature rise constraint term multiplied by their respective normalized weights.

[0029] Preferably, the voltage error term is the reciprocal of the mean square error between the output voltage and the preset target voltage;

[0030] The current surge term is the reciprocal of the peak value of the injected current;

[0031] The energy loss term is the reciprocal of the total energy over the duration, and the total energy is calculated from the injection voltage, injection current, and time interval between two adjacent control points.

[0032] The temperature rise constraint term is a conditional function. When the temperature change in the detection result does not exceed the preset temperature threshold, the temperature rise constraint term is a fixed constant. When the temperature change in the detection result exceeds the preset temperature threshold, the magnitude of the temperature rise constraint term is negatively correlated with the amount of temperature change exceeding the threshold.

[0033] Preferably, the anomaly indicator is the percentage error of the output voltage compared to the target voltage.

[0034] Preferably, the dynamic voltage recovery device further includes an alarm module. When the abnormal indicator exceeds a preset indicator threshold, the backup scheme is switched sequentially according to the score value from high to low until the abnormal indicator does not exceed the preset indicator threshold. If the abnormal indicator still exceeds the preset indicator threshold after going through all backup schemes, the alarm module issues an alarm.

[0035] A magnetic reactive dynamic voltage recovery method based on particle swarm optimization, wherein the dynamic voltage recovery method is performed by the aforementioned dynamic voltage recovery device, includes the following specific steps:

[0036] S1: Collect historical voltage dips in the power grid, perform time-series analysis to generate predicted voltage dips, and generate the amplitude range of the injected voltage based on the predicted voltage dips.

[0037] S2: Detect the temperature and electrical parameter changes of the voltage recovery device body when generating different injection voltages. Combine the detection results with the particle swarm algorithm, use the amplitude range of the injection voltage as the search range, generate several control schemes for the injection voltage, and score them.

[0038] S3: Sort the control schemes based on their scores and select the optimal scheme and at least one backup scheme from several groups of control schemes;

[0039] S4: Execute the optimal solution to generate the injection voltage, monitor the injection voltage in real time, generate anomaly indicators based on the measured values ​​of the injection voltage, and switch solutions based on the anomaly indicators.

[0040] Compared with the prior art, the beneficial effects of the present invention are:

[0041] This invention performs in-depth time-series analysis of historical voltage dips to dynamically predict the amplitude and duration of voltage dips, thereby adaptively determining the amplitude range and duration of the injected voltage. It combines particle swarm optimization (PSO) with multi-objective optimization of the control scheme and provides feedback to evaluate scheme performance, achieving precise design and real-time adjustment of the injected voltage waveform. The multi-scheme screening and switching mechanism effectively ensures the stable operation and safety of the system under abnormal conditions, comprehensively improving the intelligence level and compensation quality of the dynamic voltage recovery device. This meets the requirements for rapid response and efficient compensation in complex power grid environments, enhancing the overall stability and reliability of the system. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the module structure of the present invention;

[0043] Figure 2 This is a schematic diagram of the overall method flow of the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0045] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0046] Example:

[0047] Please see Figure 1 The present invention provides a technical solution:

[0048] The magnetic reactive dynamic voltage recovery device based on particle swarm optimization specifically includes a control device and a voltage recovery device body. The control device further includes: a data acquisition module, a timing analysis module, an algorithm optimization module, and a signal feedback module.

[0049] The data acquisition module can use high-precision voltage sensors (such as voltage transformers VT or voltage sampling circuits) to collect historical voltage drops in the power grid. After analog-to-digital conversion (ADC), the signal is transmitted in real time to the time series analysis module through an industrial-grade communication interface (such as Ethernet, CAN bus, or fiber optic). The collected data can be stored in an embedded memory or cloud database for historical data accumulation and subsequent processing.

[0050] The timing analysis module incorporates a timing analysis model to perform timing analysis on historical voltage drops to obtain predicted voltage drops, and generates the amplitude range of the injected voltage based on the predicted voltage drops. From a hardware perspective, the timing analysis module can be built using an embedded industrial computer or a high-performance DSP / FPGA, while from a software perspective, it can be implemented using an LSTM neural network or an ARIMA model.

[0051] The algorithm optimization module includes a data detection unit, an algorithm optimization unit, and a scheme selection unit. It can also be built using an embedded industrial computer or a high-performance DSP / FPGA, along with corresponding software. Among these:

[0052] The data detection unit is used to detect the temperature changes and electrical parameter changes generated when the voltage recovery device body generates injection voltages of different amplitudes. The electrical parameters include, but are not limited to, the injection current and output voltage of the voltage recovery device body, and can be composed of temperature sensors, high-precision current sensors, and high-precision voltage sensors.

[0053] The algorithm optimization unit incorporates a particle swarm optimization algorithm to search within the amplitude range of the injected voltage and generate several control schemes for the injected voltage. This allows for the dynamic determination of the upper and lower limits of the injected voltage, ensuring that the amplitude range is neither too wide to avoid invalid searches nor too narrow to guarantee compensation capability.

[0054] The scheme selection unit is used to select the optimal scheme and at least one backup scheme from several control schemes, and transmit the optimal scheme and the backup scheme to the voltage recovery device body.

[0055] The signal feedback module includes a voltage detection unit and a scheme switching unit, wherein:

[0056] The voltage detection unit is used to detect the injected voltage generated by the voltage recovery device in real time, obtain the measured value of the injected voltage, and transmit it to the scheme switching unit. Similar to the data acquisition module, the voltage detection unit can also use a high-precision voltage sensor for real-time detection.

[0057] The scheme switching unit is used to generate an abnormal index based on the measured value of the injected voltage, and when the abnormal index exceeds the preset index threshold, it controls the voltage recovery device to switch the injection voltage scheme.

[0058] The voltage recovery device mainly consists of a magnetic reactance element (controllable saturable magnetic core reactor), an inverter bridge arm (IGBT or SiC MOSFET device), and a DC energy storage unit (large capacity capacitor or supercapacitor). By saturating and regulating the magnetic permeability of the magnetic core and controlling the inverter current command, it adjusts its own magnetic circuit and injection current to generate injection voltages of different amplitudes. Modular devices can be used.

[0059] The prediction logic for the amplitude range of the injected voltage is as follows:

[0060] The collected historical voltage drops are preprocessed, including but not limited to noise reduction, outlier removal, and normalization, and a continuous time series dataset is formed. This ensures the quality of the prediction input data, thereby obtaining more accurate voltage drop and duration predictions and reducing misjudgments or overcompensation.

[0061] A time series analysis is performed on the time series dataset based on historical drop voltages, and the predicted drop voltage and predicted drop duration are obtained based on the time series analysis results. The maximum and minimum values ​​of the predicted drop voltage are used as the upper and lower limits of the amplitude range of the injected voltage, respectively.

[0062] Understandably, voltage dips in the power grid are often triggered by specific events, such as sudden load changes, maintenance, and oscillations. Examples include peak daytime electricity consumption, off-peak nighttime consumption, and differences in electricity consumption between weekdays and holidays, leading to sudden load changes in the grid; routine maintenance activities such as generator start-ups and shutdowns, and line switching; and resonance, oscillations, and transient processes existing in the power grid. These events typically exhibit periodicity; therefore, time-series analysis can be used to capture the periodicity, suddenness, and slow changes in voltage dips, thereby improving the ability to predict injected voltage and enabling adaptive compensation for different voltage dip scenarios.

[0063] The working logic of the algorithm optimization module is as follows:

[0064] The duration of the injected voltage is determined based on the predicted drop duration. Within this duration, several sets of control points are evenly distributed, each corresponding to an injected voltage of different amplitudes. All control points together constitute a control scheme for the injected voltage. Simply put, the waveform of the required injected voltage is discretized into multiple control points, thereby greatly reducing the control complexity in engineering systems.

[0065] Each time the algorithm optimization unit generates a control scheme using the particle swarm optimization algorithm, it inputs the scheme into the voltage recovery device and uses the data detection unit to detect the resulting temperature and electrical parameter changes.

[0066] The algorithm optimization unit scores each control scheme based on the detection results output by the data detection unit, and transmits the control scheme and its corresponding score value to the scheme selection unit.

[0067] The scheme selection unit ranks the control schemes in descending order of their scores, designating the top-ranked scheme as the optimal scheme and the second-ranked or subsequent schemes as backup schemes.

[0068] In other words, the predicted drop duration is the same as the duration of the injected voltage. Assuming the duration is... Evenly divided into Group of control points, the time interval between adjacent control points is ,have: The injection voltage corresponding to each control point is expressed as follows: subscript Let the index of the control point be used. Then a set of control schemes can be represented in vector form as follows:

[0069] ;

[0070] This is equivalent to needing the voltage restoration device body to sequentially... The injected voltage is output, and the duration of each output is... The total output time is .

[0071] For particle swarm optimization, a vector representing a set of control schemes is a particle, with dimension 1. In this embodiment, the number of particles in the particle swarm optimization algorithm is set to 20 to balance search accuracy and computational efficiency, and the maximum number of iterations is set to 50. That is, during random initialization, 20 control schemes are generated, and new control schemes are generated with each iteration. The entire iteration process updates the positions of the 20 particles 50 times, calculating the score values ​​corresponding to these control schemes to obtain the optimal scheme (equivalent to the global optimum) and backup schemes.

[0072] In this step, by discretizing the continuous injection voltage waveform into a finite number of control points, the complex waveform design is transformed into an optimization problem of a finite-dimensional vector. This significantly reduces the continuity and dimensionality of the optimization variables, avoids the computational burden of directly optimizing continuous functions, and improves the convergence speed and stability of the optimization algorithm. Furthermore, the control commands corresponding to the discrete control points at the sampling time points directly match the sampling / execution cycle of the digital controller, facilitating implementation and debugging in the hardware control system. Moreover, each control scheme is verified by a high-precision data detection unit to ensure that the scoring is based on real electrical parameters and temperature characteristics, thereby enhancing the engineering reliability and practical value of the optimization results.

[0073] The algorithm optimization unit scores the control scheme by constructing an objective function;

[0074] The objective function is obtained based on the detection results output by the data detection unit, and the objective function is composed of voltage error term, current impact term, energy loss term, and temperature rise constraint term multiplied by their respective normalized weights.

[0075] The voltage error term is the reciprocal of the mean square error between the output voltage and the preset target voltage, and its expression is:

[0076] ;

[0077] In the formula Indicates the first The voltage error term corresponding to the group control scheme. Indicates the index of the control scheme. Indicates the first The first group control scheme Injection voltage at group control points, This represents the output voltage of the voltage recovery device. It's understandable that the control scheme is used to control the voltage recovery device itself; therefore, the injected voltage at each control point is the voltage the voltage recovery device is intended to achieve, which is the "preset target voltage," and the two are equivalent. However, due to differences in equipment, environmental influences, and other factors, the actual output voltage of the voltage recovery device (i.e., the output terminal voltage) will differ from the target voltage. The smaller the difference, the better the control effect on the voltage recovery device, the smaller the denominator of the voltage error term, and the larger the value of this term.

[0078] The current surge term is the reciprocal of the peak value of the injected current, and its expression is:

[0079] ;

[0080] In the formula Indicates the first Current surge term corresponding to the group control scheme Indicates the first In the group control scheme, if the peak value of the injected current is too large, it is considered that the risk of current surge in the voltage recovery device body is greater, and the value of the current surge term will be smaller.

[0081] The energy loss term is the reciprocal of the total energy over the duration. The total energy is calculated from the injection voltage, injection current, and time interval between two adjacent control points, and its expression is:

[0082] ;

[0083] In the formula Indicates the first Energy loss items corresponding to the group control scheme. Indicates the voltage recovery device body in the first The first group control scheme The injected current at each control point. As can be seen from the denominator of the energy loss term, it represents the sum of the theoretical power consumption of the voltage recovery device at each control point, which is the total energy consumed over the duration.

[0084] The temperature rise constraint term is a conditional function. When the temperature change in the detection result does not exceed the preset temperature threshold, the temperature rise constraint term is a fixed constant. When the temperature change in the detection result exceeds the preset temperature threshold, the magnitude of the temperature rise constraint term is negatively correlated with the amount of temperature change exceeding the threshold. Its expression is as follows:

[0085] ;

[0086] In the formula Indicates the first Temperature rise constraints corresponding to the group control scheme Indicates the amount of temperature change. Indicates the temperature threshold. This represents a fixed constant, the value of which is determined by expert experience. As can be seen from the expression of the temperature rise constraint term, it is actually a penalty term. When the temperature change does not exceed the temperature threshold, the voltage recovery device is considered to be operating normally, the value of the temperature rise constraint term remains unchanged, and it will not affect the control scheme score. However, when the temperature change exceeds the temperature threshold, the greater the exceedance, the smaller the value of the temperature rise constraint term becomes, resulting in a lower score for the control scheme. Furthermore, the use of an exponential function is to improve the sensitivity to temperature and avoid safety risks caused by excessively high operating temperatures of the voltage recovery device.

[0087] The final objective function expression can be represented as:

[0088] ;

[0089] in Indicates the first The score value corresponding to the group control scheme ~ Represents the normalized weights, i.e. The weighting of each component is adjusted according to design requirements. For example, if the design requirement is to prioritize accuracy, followed by risk control, and lastly power consumption, then the weights for each component can be set to 0.5, 0.2, 0.1, and 0.2, respectively.

[0090] It is understandable that the voltage recovery device and the load terminal can be controlled by setting a switch. That is to say, when the optimal or backup plan is not executed, the voltage recovery device is not connected to the load terminal, and the injected voltage is only used for data acquisition to generate a control plan. After the optimal or backup plan is determined, the switch can be opened to connect to the load terminal when a voltage drop occurs in the power grid, thereby realizing dynamic voltage compensation.

[0091] The dynamic voltage recovery device also includes an alarm module. The abnormality indicator is the percentage error between the output voltage and the target voltage, calculated as follows:

[0092] ;

[0093] In the formula Indicates the first Abnormal indicators corresponding to the group control scheme.

[0094] When an abnormal indicator exceeds a preset threshold, the backup plan is switched sequentially in descending order of score until the abnormal indicator does not exceed the preset threshold. If the abnormal indicator still exceeds the preset threshold after going through all backup plans, the alarm module will issue an alarm.

[0095] In this step, the percentage error between the output voltage and the target voltage is used as an anomaly indicator. This can reflect the deviation between the actual compensation effect of the DVR and the expected target in real time, promptly detect control scheme failures or abnormalities, and also allow the system to quickly switch to a better-performing backup scheme when the main scheme is abnormal, ensuring the continuity and stability of voltage recovery, thereby improving the system's anti-interference and anti-fault capabilities.

[0096] Please see Figure 2 This embodiment also provides a magnetic reactive dynamic voltage recovery method based on particle swarm optimization. The dynamic voltage recovery method is obtained by the aforementioned dynamic voltage recovery device, and the specific steps include:

[0097] S1: Collect historical voltage dips in the power grid, perform time-series analysis to generate predicted voltage dips, and generate the amplitude range of the injected voltage based on the predicted voltage dips.

[0098] S2: Detect the temperature and electrical parameter changes of the voltage recovery device body when generating different injection voltages. Combine the detection results with the particle swarm algorithm, use the amplitude range of the injection voltage as the search range, generate several control schemes for the injection voltage, and score them.

[0099] S3: Sort the control schemes based on their scores and select the optimal scheme and at least one backup scheme from several groups of control schemes;

[0100] S4: Execute the optimal solution to generate the injection voltage, monitor the injection voltage in real time, generate anomaly indicators based on the measured values ​​of the injection voltage, and switch solutions based on the anomaly indicators.

[0101] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0102] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0103] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0104] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A magnetic reactive dynamic voltage recovery device based on particle swarm optimization, characterized in that, Specifically, it includes a control device and a voltage recovery device body, wherein: The control device includes: The data acquisition module is used to collect historical voltage drops and transmit them to the time series analysis module. The timing analysis module has a built-in timing analysis model for performing timing analysis on historical voltage drops to obtain predicted voltage drops, and generating the amplitude range of the injection voltage based on the predicted voltage drops. The algorithm optimization module includes a data detection unit, an algorithm optimization unit, and a scheme selection unit, wherein: The data detection unit is used to detect the temperature and electrical parameter changes generated when the voltage recovery device body generates injection voltages of different amplitudes. The algorithm optimization unit has a built-in particle swarm optimization algorithm, which is used to search within the amplitude range of the injected voltage and generate several sets of control schemes for the injected voltage. The scheme screening unit is used to select the optimal scheme and at least one backup scheme from several control schemes, and transmit the optimal scheme and the backup scheme to the voltage recovery device body together; The signal feedback module includes a voltage detection unit and a scheme switching unit, wherein: The voltage detection unit is used to detect the injection voltage generated by the voltage recovery device body in real time, obtain the measured value of the injection voltage, and transmit it to the scheme switching unit. The scheme switching unit is used to generate an abnormal index based on the measured value of the injected voltage, and when the abnormal index exceeds the preset index threshold, control the voltage recovery device to switch the injection voltage scheme. The voltage recovery device body is used to generate injection voltages of different amplitudes by adjusting its own magnetic circuit and injection current.

2. The magnetic reactive dynamic voltage recovery device based on particle swarm optimization according to claim 1, characterized in that: The prediction logic for the amplitude range of the injected voltage is as follows: The collected historical voltage drops are preprocessed, including noise reduction, outlier removal, and normalization, and a continuous time series dataset is generated. A time series analysis is performed on the time series dataset based on historical drop voltages, and the predicted drop voltage and predicted drop duration are obtained based on the time series analysis results. The maximum and minimum values ​​of the predicted drop voltage are used as the upper and lower limits of the amplitude range of the injected voltage, respectively.

3. The magnetic reactive dynamic voltage recovery device based on particle swarm optimization according to claim 2, characterized in that: The working logic of the algorithm optimization module is as follows: The duration of the injected voltage is determined based on the predicted drop duration. Several sets of control points are evenly set within the duration, with each set of control points corresponding to an injected voltage of different amplitudes. All control points together constitute a control scheme for the injected voltage. Each time the algorithm optimization unit generates a control scheme using the particle swarm optimization algorithm, it inputs it into the voltage recovery device body and uses the data detection unit to detect the temperature changes and electrical parameter changes generated therefrom. The electrical parameters include the injected current and output voltage of the voltage recovery device body. The algorithm optimization unit scores each control scheme based on the detection results output by the data detection unit, and transmits the control scheme and its corresponding score value to the scheme selection unit. The scheme selection unit ranks the control schemes in descending order of their scores, designating the top-ranked scheme as the optimal scheme and the second-ranked or subsequent schemes as backup schemes.

4. The magnetic reactive dynamic voltage recovery device based on particle swarm optimization according to claim 3, characterized in that: The algorithm optimization unit scores the control scheme by constructing an objective function; The objective function is obtained based on the detection results output by the data detection unit, and the objective function is composed of voltage error term, current impact term, energy loss term, and temperature rise constraint term multiplied by their respective normalized weights.

5. The magnetic reactive dynamic voltage recovery device based on particle swarm optimization according to claim 4, characterized in that: The voltage error term is the reciprocal of the mean square error between the output voltage and the preset target voltage; The current surge term is the reciprocal of the peak value of the injected current; The energy loss term is the reciprocal of the total energy over the duration, and the total energy is calculated from the injection voltage, injection current, and time interval between two adjacent control points. The temperature rise constraint term is a conditional function. When the temperature change in the detection result does not exceed the preset temperature threshold, the temperature rise constraint term is a fixed constant. When the temperature change in the detection result exceeds the preset temperature threshold, the magnitude of the temperature rise constraint term is negatively correlated with the amount of temperature change exceeding the threshold.

6. The magnetic reactive dynamic voltage recovery device based on particle swarm optimization according to claim 5, characterized in that: The abnormality indicator is the percentage error between the output voltage and the target voltage.

7. The magnetic reactive dynamic voltage recovery device based on particle swarm optimization according to claim 6, characterized in that: The dynamic voltage recovery device also includes an alarm module. When the abnormal indicator exceeds the preset indicator threshold, the backup scheme is switched sequentially in descending order of the score value until the abnormal indicator does not exceed the preset indicator threshold. If the abnormal indicator still exceeds the preset indicator threshold after going through all backup schemes, the alarm module will issue an alarm.

8. A magnetic reactive dynamic voltage recovery method based on particle swarm optimization, characterized in that: The dynamic voltage recovery method is performed by the dynamic voltage recovery device according to any one of claims 1-7, and the specific steps include: S1: Collect historical voltage dips in the power grid, perform time-series analysis to generate predicted voltage dips, and generate the amplitude range of the injected voltage based on the predicted voltage dips. S2: Detect the temperature and electrical parameter changes of the voltage recovery device body when generating different injection voltages. Combine the detection results with the particle swarm algorithm, use the amplitude range of the injection voltage as the search range, generate several control schemes for the injection voltage, and score them. S3: Sort the control schemes based on their scores and select the optimal scheme and at least one backup scheme from several groups of control schemes; S4: Execute the optimal solution to generate the injection voltage, monitor the injection voltage in real time, generate anomaly indicators based on the measured values ​​of the injection voltage, and switch solutions based on the anomaly indicators.

Images