Dynamic power factor correction method and system for subway DC power supply system

Abstract

The present invention relates to the field of data processing technology, and discloses a method and system for dynamic power factor correction of a subway DC power supply system, which is used to improve the accuracy and efficiency of dynamic power factor correction. The method includes: performing fast Fourier transform processing on the original data set to obtain voltage spectrum data and current spectrum data and performing dynamic power factor calculation and harmonic distortion analysis to obtain a real-time power factor value sequence and a harmonic distortion degree sequence; performing multi-parameter correlation analysis and trend prediction to obtain influencing factor data and expected change trends for power factor optimization; performing parameter sensitivity calculation and priority sorting according to the influencing factor data and expected change trends, and generating a regulation strategy for the substation rectifier trigger angle and the on-board inverter modulation index; based on the regulation strategy, adjusting the substation rectifier trigger angle and the on-board inverter modulation index in real time, dynamically correcting the real-time power factor value sequence, and obtaining a corrected power factor value sequence.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a dynamic power factor correction method and system for a subway direct current power supply system. Background Art

[0002] The DC power supply system of the subway is a key infrastructure of urban rail transit, and its performance directly affects the safety, reliability and energy efficiency of train operation. Traditional power factor correction methods mainly rely on static compensation devices such as reactive power compensators and active filters. These devices improve the power factor by injecting reactive power or filtering harmonics at fixed locations. At the same time, some advanced control strategies, such as fuzzy control and neural network control, have also been applied to power factor optimization. These methods have improved the power quality of the system to a certain extent, reduced power loss, and provided support for the efficient operation of the subway system.

[0003] However, the existing technologies have some significant shortcomings. First, static compensation devices are difficult to adapt to the rapid changes and complexity of subway system loads, resulting in unstable compensation effects. Second, traditional control strategies often only focus on the optimization of a single device or a local system, lacking a global consideration of the entire power supply network. Furthermore, existing methods usually use fixed parameters or preset models, which cannot effectively cope with the dynamic changes and uncertainties of system parameters. In addition, most power factor correction methods fail to make full use of real-time data and advanced data analysis techniques, resulting in insufficient timeliness and accuracy of optimization decisions. Finally, existing technologies often find it difficult to achieve an ideal balance when dealing with the trade-off between harmonic pollution and power factor optimization. Summary of the invention

[0004] In view of this, an embodiment of the present invention provides a method and system for dynamic power factor correction of a subway DC power supply system, which are used to improve the accuracy and efficiency of dynamic power factor correction.

[0005] The present invention provides a method for dynamic power factor correction of a subway DC power supply system, comprising: real-time acquisition of key parameters of the subway DC power supply system to obtain an original data set; fast Fourier transform processing of the original data set to obtain spectrum data containing voltage and current harmonic components; dynamic power factor calculation and harmonic distortion analysis of the spectrum data to obtain a real-time power factor value sequence and a harmonic distortion degree sequence; based on the real-time power factor value sequence and the harmonic distortion degree sequence, multi-parameter correlation analysis and trend prediction are performed to obtain influencing factor data and expected change trends of power factor optimization; according to the influencing factor data and the expected change trends, parameter sensitivity calculation and priority sorting are performed to generate a regulation strategy for a substation rectifier trigger angle and a vehicle-mounted inverter modulation index; based on the regulation strategy, the substation rectifier trigger angle and the vehicle-mounted inverter modulation index are adjusted in real time, the real-time power factor value sequence is dynamically corrected, and a corrected power factor value sequence is obtained.

[0006] In the present invention, the step of collecting the key parameters of the subway DC power supply system in real time to obtain the original data set includes: installing a voltage sensor and a current sensor at the output end of the traction substation of the subway DC power supply system, collecting initial DC voltage and initial current data, and converting the initial DC voltage and the initial current data into digital signals to obtain digital voltage data and digital current data; collecting the real-time position coordinates and running status information of the train, and encoding the real-time position coordinates and the running status information into a preset standard format to obtain coded position information and coded status information; performing noise filtering and outlier detection on the digital voltage data and the digital current data, eliminating interference signals and outliers, and obtaining cleaned voltage data and cleaned current data; performing timestamp alignment and data completion on the cleaned voltage data and the cleaned current data to obtain synchronous voltage data and synchronous current data; performing coordinate conversion on the coded position information, mapping the physical coordinates to the power supply network topology of the subway DC power supply system, obtaining train position information, and performing data decoding on the coded status information to obtain train status data; merging the synchronous voltage data, the synchronous current data, the train position information and the train status data in time series to obtain the original data set.

[0007] In the present invention, the step of performing fast Fourier transform processing on the original data set to obtain voltage spectrum data and current spectrum data includes: performing segmentation processing on the synchronous voltage data in the original data set to obtain segmented voltage data, and performing segmentation processing on the synchronous current data in the original data set to obtain segmented current data; performing windowing processing on the segmented voltage data and the segmented current data respectively to obtain windowed voltage data and windowed current data; performing zero-filling on the windowed voltage data and the windowed current data to obtain filled voltage data and filled current data; performing fast Fourier transform on the filled voltage data to obtain first spectrum data, and performing fast Fourier transform on the filled current data to obtain second spectrum data; performing amplitude normalization and phase correction on the first spectrum data and the second spectrum data to obtain first corrected spectrum data and second corrected spectrum data; performing frequency resolution correction on the first corrected spectrum data and the second corrected spectrum data respectively to obtain voltage spectrum data containing voltage harmonic components and current spectrum data containing current harmonic components.

[0008] In the present invention, the steps of performing dynamic power factor calculation and harmonic distortion analysis on the voltage spectrum data and the current spectrum data to obtain a real-time power factor value sequence and a harmonic distortion degree sequence include: extracting the fundamental frequency component of the voltage spectrum data to obtain voltage fundamental frequency amplitude data; extracting the fundamental frequency component of the current spectrum data to obtain current fundamental frequency amplitude data; calculating the ratio of the voltage fundamental frequency amplitude data to the current fundamental frequency amplitude data to obtain system impedance characteristic data; extracting the amplitude of each harmonic component in the voltage spectrum data to obtain voltage harmonic data; and extracting the fundamental frequency component of the current spectrum data to obtain the current fundamental frequency amplitude data. The amplitude of each harmonic component in the current spectrum data is extracted to obtain current harmonic data; the power contribution of each harmonic is calculated according to the voltage harmonic data and the current harmonic data to obtain harmonic power data; based on the system impedance characteristic data and the harmonic power data, the active power and reactive power of the subway DC power supply system are calculated to obtain a dynamic power factor value; the voltage harmonic data and the current harmonic data are frequency-weighted, the total harmonic distortion is calculated, and a harmonic distortion sequence is obtained; the dynamic power factor values ​​are combined in chronological order to obtain the real-time power factor value sequence.

[0009] In the present invention, the step of performing multi-parameter correlation analysis and trend prediction based on the real-time power factor value sequence and the harmonic distortion sequence to obtain the influencing factor data and the expected change trend of power factor optimization includes: performing time series decomposition on the real-time power factor value sequence to obtain power factor decomposition data; performing time series decomposition on the harmonic distortion sequence to obtain harmonic distortion decomposition data; performing cross-correlation analysis on the power factor decomposition data and the harmonic distortion decomposition data to obtain correlation data; performing spectrum analysis on the power factor decomposition data and the harmonic distortion decomposition data to obtain frequency characteristic data; performing combined analysis on the correlation data and the frequency characteristic data to obtain multi-scale correlation indicators; performing principal component analysis on the multi-scale correlation indicators to obtain the influencing factor data of power factor optimization; performing time series prediction on the power factor decomposition data and the harmonic distortion decomposition data to obtain the expected change trend.

[0010] In the present invention, the steps of calculating parameter sensitivity and prioritizing according to the influencing factor data and the expected change trend, and generating an adjustment strategy for the trigger angle of the substation rectifier and the modulation index of the on-board inverter include: normalizing the influencing factor data to obtain standardized influencing factor data; performing principal component analysis on the standardized influencing factor data to obtain reduced dimension feature data; performing correlation analysis on the reduced dimension feature data and the expected change trend to obtain parameter sensitivity indicators; sorting the parameter sensitivity indicators to obtain a priority adjustment parameter set; performing constraint condition analysis on the priority adjustment parameter set to obtain a parameter adjustment range; mapping and converting the parameter adjustment range to obtain a substation rectifier trigger angle adjustment interval and a vehicle inverter modulation index adjustment interval; based on the trigger angle adjustment interval and the modulation index adjustment interval, obtaining an adjustment strategy for the substation rectifier trigger angle and the vehicle inverter modulation index.

[0011] In the present invention, the step of adjusting the trigger angle of the substation rectifier and the modulation index of the on-board inverter in real time based on the regulation strategy, dynamically correcting the real-time power factor value sequence, and obtaining the corrected power factor value sequence includes: performing time-series decomposition on the regulation strategy to obtain a discrete regulation sequence, and interpolating the discrete regulation sequence to obtain a continuous regulation function; calculating the trigger angle adjustment amount of the substation rectifier according to the continuous regulation function to obtain an updated trigger angle value; calculating the modulation index change of the on-board inverter according to the continuous regulation function to obtain an updated modulation index value; performing harmonic analysis on the updated trigger angle value and the updated modulation index value to obtain a target power factor value; comparing the target power factor value with the real-time power factor value sequence to obtain a power factor improvement amount; and updating the real-time power factor value sequence based on the power factor improvement amount to obtain the corrected power factor value sequence.

[0012] The present invention also provides a dynamic power factor correction system for a subway DC power supply system, comprising:

[0013] The acquisition module is used to collect key parameters of the subway DC power supply system in real time to obtain the original data set;

[0014] A processing module, used for performing fast Fourier transform processing on the original data set to obtain voltage spectrum data and current spectrum data;

[0015] An analysis module is used to perform dynamic power factor calculation and harmonic distortion analysis on the voltage spectrum data and the current spectrum data to obtain a real-time power factor value sequence and a harmonic distortion degree sequence;

[0016] A prediction module, used to perform multi-parameter correlation analysis and trend prediction based on the real-time power factor value sequence and harmonic distortion sequence, to obtain the influencing factor data and expected change trend of power factor optimization;

[0017] A sorting module, used to perform parameter sensitivity calculation and priority sorting according to the influencing factor data and expected change trends, and generate a regulation strategy for the substation rectifier trigger angle and the on-board inverter modulation index;

[0018] The correction module is used to adjust the substation rectifier trigger angle and the vehicle inverter modulation index in real time based on the adjustment strategy, dynamically correct the real-time power factor value sequence, and obtain a corrected power factor value sequence.

[0019] In the technical solution provided by the present invention, the key parameters of the system are collected in real time and fast Fourier transform processing is performed, which provides high-precision spectrum data for power factor calculation and greatly improves the accuracy of correction. Dynamic power factor calculation and harmonic distortion analysis can reflect the changes in system status in real time, making the correction process more timely and accurate. Multi-parameter correlation analysis and trend prediction technology deeply explore the intrinsic correlation between data, effectively identify the key factors affecting the power factor, and improve the pertinence and efficiency of correction. Parameter sensitivity calculation and priority sorting ensure that resources are reasonably allocated to the most critical parameters and optimize the resource utilization efficiency of the correction process. The real-time adjustment strategy of the substation rectifier trigger angle and the on-board inverter modulation index takes into account the dynamic characteristics and equipment limitations of the system, and realizes fast and smooth power factor correction. In addition, the adaptability of this method enables it to respond quickly to load changes, continuously optimize the correction strategy, and keep the system running in the optimal power factor state, which not only significantly improves the accuracy of power factor correction, but also greatly improves the correction efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 The present invention is a flowchart of a method for dynamic power factor correction of a subway DC power supply system according to an embodiment of the present invention.

[0022] Figure 2 The figure is a schematic diagram of a dynamic power factor correction system of a subway DC power supply system according to an embodiment of the present invention.

[0023] Reference numerals:

[0024] 201, acquisition module; 202, processing module; 203, analysis module; 204, prediction module; 205, sorting module; 206, correction module. DETAILED DESCRIPTION

[0025] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0027] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0028] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , Figure 1 : is a flow chart of a method for dynamic power factor correction of a subway DC power supply system according to an embodiment of the present invention, such as Figure 1 As shown, the following steps are included:

[0029] S101, collecting key parameters of the subway DC power supply system in real time to obtain an original data set;

[0030] S102, performing fast Fourier transform processing on the original data set to obtain voltage spectrum data and current spectrum data;

[0031] S103, performing dynamic power factor calculation and harmonic distortion analysis on the voltage spectrum data and the current spectrum data to obtain a real-time power factor value sequence and a harmonic distortion degree sequence;

[0032] S104, based on the real-time power factor value sequence and harmonic distortion sequence, multi-parameter correlation analysis and trend prediction are performed to obtain the influencing factor data and expected change trend of power factor optimization;

[0033] S105, performing parameter sensitivity calculation and priority sorting according to the influencing factor data and expected change trends, and generating a regulation strategy for the substation rectifier trigger angle and the on-board inverter modulation index;

[0034] S106. Based on the regulation strategy, the substation rectifier trigger angle and the vehicle inverter modulation index are adjusted in real time, and the real-time power factor value sequence is dynamically corrected to obtain a corrected power factor value sequence.

[0035] It should be noted that the dynamic power factor correction method of the subway DC power supply system first collects the key parameters of the system in real time, obtains DC voltage and current data through the voltage sensor and current sensor installed at the output end of the traction substation, and collects the train position information and running status at the same time. After digital conversion, noise filtering and outlier detection, these data are time-stamp aligned and data completed, and finally form an original data set containing synchronous voltage and current data, train position information and running status. The original data set is processed by fast Fourier transform to convert the time domain signal into frequency domain information. This step first performs segmentation processing on the voltage and current data, then applies the windowing function to reduce spectrum leakage, and then expands the data length by zero filling. The fast Fourier transform algorithm converts these processed data into the initial spectrum, and after amplitude normalization and phase correction, the final voltage spectrum data and current spectrum data are obtained.

[0036] Based on the obtained spectrum data, dynamic power factor calculation and harmonic distortion analysis are performed. This process includes extracting the fundamental frequency components of voltage and current, calculating the system impedance characteristics, and analyzing the power contribution of each harmonic. The dynamic power factor value is obtained by calculating the active power and reactive power. At the same time, the harmonic data is frequency-weighted and the total harmonic distortion is calculated. These results are combined in time sequence to form a real-time power factor value sequence and a harmonic distortion sequence. Subsequently, multi-parameter correlation analysis and trend prediction are performed based on these two sequences. First, the sequence is decomposed into a time series to obtain decomposed data. Then, correlation data and frequency characteristic data are obtained through cross-correlation analysis and spectrum analysis. These data are combined and analyzed to form a multi-scale correlation index, and the influencing factor data of power factor optimization are extracted through principal component analysis. At the same time, the expected change trend is obtained using time series prediction technology.

[0037] According to the influencing factor data and expected change trends, parameter sensitivity calculation and priority sorting are performed. This process first normalizes and reduces the dimension of the influencing factor data, and then performs correlation analysis with the expected change trend to obtain the parameter sensitivity index. These indicators are sorted and constraint conditions are analyzed, and finally the adjustment range of the substation rectifier trigger angle and the on-board inverter modulation index is obtained to form an adjustment strategy. Finally, based on the generated adjustment strategy, the substation rectifier trigger angle and the on-board inverter modulation index are adjusted in real time. This process includes converting the adjustment strategy into a continuous adjustment function, calculating the adjustment amount of the trigger angle and the modulation index, and performing harmonic analysis to obtain a new power factor value. By comparing the new and old power factor values, the improvement amount is calculated, and the real-time power factor value sequence is updated accordingly, and finally the corrected power factor value sequence is obtained.

[0038] For example, in the raw data collected from the traction substation of a subway line, the voltage value is 1500V, the current value is 3000A, and the train is located 5 kilometers away from the substation and is in an accelerating state. After fast Fourier transform processing, it is found that the fundamental frequency component of the voltage is 1450V, the third harmonic component is 50V, the fundamental frequency component of the current is 2900A, and the third harmonic component is 100A. The calculated dynamic power factor is 0.95 and the total harmonic distortion is 3.5%. Through multi-parameter correlation analysis, it is found that there is a strong correlation between the acceleration state of the train and the decrease in the power factor, and it is predicted that the power factor may drop to 0.92 in the next 10 minutes. Based on this prediction, the regulation strategy recommends adjusting the trigger angle of the substation rectifier by 2 degrees and increasing the modulation index of the on-board inverter by 0.05. After implementing this regulation strategy, the power factor is increased to 0.97, which effectively improves the energy efficiency of the system.

[0039] By executing the above steps, the key parameters of the system are collected in real time and processed by fast Fourier transform, which provides high-precision spectrum data for power factor calculation and greatly improves the accuracy of correction. Dynamic power factor calculation and harmonic distortion analysis can reflect the changes in system status in real time, making the correction process more timely and accurate. Multi-parameter correlation analysis and trend prediction technology deeply explore the intrinsic correlation between data, effectively identify the key factors affecting power factor, and improve the pertinence and efficiency of correction. Parameter sensitivity calculation and priority sorting ensure that resources are reasonably allocated to the most critical parameters and optimize the resource utilization efficiency of the correction process. The real-time adjustment strategy of the substation rectifier trigger angle and the on-board inverter modulation index takes into account the dynamic characteristics and equipment limitations of the system, and realizes fast and smooth power factor correction. In addition, the adaptability of this method enables it to respond quickly to load changes, continuously optimize the correction strategy, and keep the system running in the optimal power factor state, which not only significantly improves the accuracy of power factor correction, but also greatly improves the correction efficiency.

[0040] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0041] (1) Install a voltage sensor and a current sensor at the output end of the traction substation of the subway DC power supply system to collect initial DC voltage and initial current data, and convert the initial DC voltage and initial current data into digital signals to obtain digital voltage data and digital current data;

[0042] (2) collecting the real-time position coordinates and running status information of the train, and encoding the real-time position coordinates and running status information into a preset standard format to obtain encoded position information and encoded status information;

[0043] (3) Perform noise filtering and outlier detection on the digital voltage data and digital current data, remove interference signals and outliers, and obtain cleaned voltage data and cleaned current data;

[0044] (4) Performing timestamp alignment and data completion on the cleaning voltage data and the cleaning current data to obtain synchronized voltage data and synchronized current data;

[0045] (5) Perform coordinate conversion on the coded position information, map the physical coordinates to the power supply network topology of the subway DC power supply system, obtain the train position information, and decode the coded status information to obtain the train status data;

[0046] (6) The synchronous voltage data, synchronous current data, train position information, and train status data are merged in time series to obtain the original data set.

[0047] Specifically, the dynamic power factor correction method of the subway DC power supply system first installs high-precision voltage sensors and current sensors at the output end of the traction substation. These sensors use the Hall effect principle and can collect initial DC voltage and initial current data in real time. The collected analog signal is then converted into a digital signal through an analog-to-digital converter to obtain digital voltage data and digital current data. This process is usually carried out at a sampling rate of thousands of times per second to ensure that the rapid changes in voltage and current are captured. At the same time, the real-time position coordinates of the train are obtained through the on-board GPS system, and the operating status information is provided by the train control system. This information is standardized and converted into a preset data format. The position information usually includes latitude and longitude coordinates and altitude, while the operating status information includes parameters such as speed, acceleration, and traction. These data are encoded to form coded position information and coded status information, which is convenient for subsequent processing and transmission.

[0048] Next, the digital voltage data and digital current data are subjected to noise filtering and outlier detection. Noise filtering uses digital filters, such as Butterworth filters or Chebyshev filters, to effectively remove high-frequency noise. Outlier detection uses statistical methods, such as the 3σ principle or box plot method, to identify and remove outliers. This step obtains cleaned voltage data and current data, which greatly improves data quality. The cleaned voltage data and current data also need to be timestamp aligned and data completed. Timestamp alignment ensures that data from different sources accurately match in the time dimension, which is usually achieved using an interpolation algorithm. Data completion fills in possible data gaps, and common methods include linear interpolation or more complex time series interpolation algorithms. This step outputs synchronized voltage data and synchronized current data, laying the foundation for subsequent analysis.

[0049] The encoded position information requires coordinate transformation to map the GPS coordinates to the power supply network topology of the subway DC power supply system. This process involves a complex coordinate transformation algorithm that takes into account factors such as the curvature and slope of the subway line. At the same time, the encoded status information is decoded and converted into directly usable train status data, including current speed, acceleration, traction mode and other information. Finally, the synchronous voltage data, synchronous current data, train position information and train status data are merged in time series. This step uses the timestamp as the key to integrate different types of data into a unified data structure to form the final original data set. This data set is a multidimensional time series that contains complete system status information at each time point.

[0050] For example, the data of a sampling point on a subway line is as follows: the initial DC voltage collected by the traction substation is 1500V and the initial current is 3000A. These analog signals are converted into digital signals by a 16-bit analog-to-digital converter with a resolution of 0.1V and 0.1A. At the same time, the train GPS positioning shows that it is located at 30.5432 degrees north latitude, 114.3456 degrees east longitude, and 50 meters above sea level. These data are encoded as a 64-bit integer. The current speed of the train is 60km / h, it is in an accelerating state, and the traction force is 200kN. This information is encoded as a 32-bit status word. After noise filtering, the voltage data is smoothed to 1498V and the current data is corrected to 2995A. The timestamp alignment process unifies all data to the same time point, such as 10:00:00.000 on August 1, 2023. Coordinate transformation converts the GPS position to a line position 5.2 kilometers away from the traction substation.

[0051] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0052] (1) segmenting the synchronous voltage data in the original data set to obtain segmented voltage data, and segmenting the synchronous current data in the original data set to obtain segmented current data;

[0053] (2) performing windowing processing on the segmented voltage data and the segmented current data respectively to obtain windowed voltage data and windowed current data;

[0054] (3) performing zero-filling on the windowed voltage data and the windowed current data to obtain filled voltage data and filled current data;

[0055] (4) performing a fast Fourier transform on the filling voltage data to obtain first spectrum data, and performing a fast Fourier transform on the filling current data to obtain second spectrum data;

[0056] (5) performing amplitude normalization and phase correction on the first spectrum data and the second spectrum data to obtain first corrected spectrum data and second corrected spectrum data;

[0057] (6) Performing frequency resolution correction on the first corrected spectrum data and the second corrected spectrum data respectively to obtain voltage spectrum data including voltage harmonic components and current spectrum data including current harmonic components.

[0058] Specifically, the synchronous voltage data and synchronous current data in the original data set are segmented. This process divides the long time series data into several shorter data segments, each of which usually contains 1024 or 2048 sampling points. Such a segment length can ensure sufficient frequency resolution and capture the dynamic characteristics of the system. The segmentation process uses a sliding window technology, and there is a 50% overlap between adjacent data segments to ensure that important transient information is not missed. The segmented voltage data and current data are windowed separately. The purpose of windowing is to reduce the spectrum leakage effect and improve the accuracy of spectrum analysis. Commonly used window functions include Hanning window, Hamming window and Blackman window. Taking the Hanning window as an example, it can achieve a good balance between the main lobe width and sidelobe suppression. The windowing process multiplies the original data with the window function to obtain the windowed voltage data and the windowed current data.

[0059] In order to further improve the accuracy of spectrum analysis, the windowed voltage data and current data are zero-filled. Zero-filling is to add a certain number of zeros at the end of the original data sequence, usually extending the data length to 2 or 4 times the original. This operation can increase the number of interpolation points of the spectrum and improve the frequency resolution, but will not introduce additional spectrum information. After zero-filling, fast Fourier transform (FFT) is performed on the filled voltage data and filled current data respectively. FFT is an efficient discrete Fourier transform algorithm that can convert time domain signals to frequency domain. The computational complexity of FFT is O(NlogN), where N is the number of data points, which greatly improves the computational efficiency. The result of FFT is complex spectrum data, which contains amplitude and phase information.

[0060] After obtaining the initial spectrum data, amplitude normalization and phase correction are required. Amplitude normalization divides the spectrum amplitude by the number of data points so that the spectrum amplitude corresponds to the actual signal amplitude. Phase correction takes into account the impact of windowing and zero-filling on the phase to ensure the accuracy of the phase information. This step obtains the first corrected spectrum data and the second corrected spectrum data, which correspond to voltage and current, respectively. Finally, the frequency resolution of the corrected spectrum data is corrected. Since the frequency resolution obtained by FFT is fixed, it may not meet the actual analysis requirements. Frequency resolution correction remaps the spectrum data to the required frequency grid through interpolation algorithms, such as linear interpolation or spline interpolation. This step ultimately obtains voltage spectrum data containing voltage harmonic components and current spectrum data containing current harmonic components.

[0061] For example, the synchronous voltage data and synchronous current data collected by the traction substation of a subway line have a sampling frequency of 10kHz and a data length of 10 seconds. First, the data is divided into 20 data segments with a length of 0.5 seconds (5000 points), and adjacent segments overlap by 2500 points. The Hanning window function is applied to each data segment, and the window function expression is 0.5*(1-cos(2πn / N)), where n is the sample number and N is the window length. After windowing, the data is zero-filled to 8192 points. Then, an FFT operation is performed to obtain a complex spectrum of 4097 frequency points. The spectrum is amplitude normalized and divided by 8192 / 2. Phase correction takes into account the group delay of the Hanning window and compensates the phase of each frequency point. Finally, the spectrum is resampled to 1000 equally spaced frequency points in the range of 0-5kHz through cubic spline interpolation. The processing results show that the fundamental component of voltage is 1500V, the third harmonic is 45V, and the fifth harmonic is 30V; the fundamental component of current is 3000A, the third harmonic is 150A, and the fifth harmonic is 90A.

[0062] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0063] (1) Extract the fundamental frequency component of the voltage spectrum data to obtain the voltage fundamental frequency amplitude data;

[0064] (2) Extract the fundamental frequency component of the current spectrum data to obtain the current fundamental frequency amplitude data;

[0065] Calculate the ratio of voltage fundamental frequency amplitude data to current fundamental frequency amplitude data to obtain system impedance characteristic data;

[0066] (3) Extract the amplitude of each harmonic component in the voltage spectrum data to obtain voltage harmonic data;

[0067] (4) Extract the amplitude of each harmonic component in the current spectrum data to obtain current harmonic data;

[0068] (5) Calculate the power contribution of each harmonic based on the voltage harmonic data and the current harmonic data to obtain the harmonic power data;

[0069] (6) Based on the system impedance characteristic data and harmonic power data, calculate the active power and reactive power of the subway DC power supply system to obtain the dynamic power factor value;

[0070] (7) Perform frequency-weighted processing on the voltage harmonic data and the current harmonic data, calculate the total harmonic distortion, and obtain the harmonic distortion sequence;

[0071] (8) Combine the dynamic power factor values ​​in chronological order to obtain a real-time power factor value sequence.

[0072] Specifically, the fundamental frequency component is extracted from the voltage spectrum data, which is usually the amplitude at 0Hz or very low frequency, to obtain the voltage fundamental frequency amplitude data. Similarly, the fundamental frequency component is extracted from the current spectrum data to obtain the current fundamental frequency amplitude data. This process is achieved by locating the maximum amplitude point in the spectrum data. Considering the characteristics of the DC system, the fundamental frequency is usually located at the lowest frequency. The ratio of the voltage fundamental frequency amplitude data to the current fundamental frequency amplitude data is calculated to obtain the system impedance characteristic data. This step reflects the basic electrical characteristics of the system and is crucial to understanding the operating status of the system. The calculation formula for the system impedance characteristic data is as follows:

[0073] ;

[0074] Where Z is the system impedance, is the voltage fundamental frequency amplitude, is the current fundamental frequency amplitude.

[0075] Subsequently, the amplitude of each harmonic component in the voltage and current spectrum data is extracted to obtain voltage harmonic data and current harmonic data. This process usually takes into account the 25th harmonic, because higher-order harmonics have less impact on the system. Harmonic data extraction is achieved by locating the amplitude of the corresponding frequency point in the spectrum.

[0076] Based on the voltage harmonic data and current harmonic data, the power contribution of each harmonic is calculated to obtain the harmonic power data. The calculation formula for harmonic power is as follows:

[0077] ;

[0078] in, For the h The power of the subharmonics, and Respectively h The voltage and current amplitudes of the subharmonics, is the phase difference of the hth harmonic.

[0079] Using the system impedance characteristic data and harmonic power data, the active power and reactive power of the subway DC power supply system are calculated, and then the dynamic power factor value is obtained. The calculation formula of the dynamic power factor is:

[0080] ;

[0081] in, is the power factor, P is the total active power, and Q is the total reactive power.

[0082] In addition, the voltage harmonic data and current harmonic data are frequency-weighted, and the total harmonic distortion is calculated to obtain a harmonic distortion sequence. This step takes into account the differences in the impact of different frequency harmonics on the system, and usually uses the weighting coefficient recommended by the International Electrotechnical Commission (IEC). Finally, the calculated dynamic power factor values ​​are combined in chronological order to form a real-time power factor value sequence, which provides basic data for subsequent dynamic correction.

[0083] For example, the voltage spectrum data obtained from the traction substation of a subway line in a sampling period shows that the amplitude of the fundamental frequency (0Hz) is 1500V, the amplitude of the third harmonic (150Hz) is 45V, and the amplitude of the fifth harmonic (250Hz) is 30V. The current spectrum data of the same period shows that the amplitude of the fundamental frequency is 3000A, the amplitude of the third harmonic is 150A, and the amplitude of the fifth harmonic is 90A. The calculated system impedance characteristic is 0.5Ω. Assuming that the phase difference of each harmonic is 0°, 30° and 45° respectively, the fundamental power is 4.5MW, the third harmonic power is 5.84kW, and the fifth harmonic power is 1.91kW. The total active power is about 4.5078MW. Assuming that the total reactive power is 0.9016MW, the calculated dynamic power factor value is 0.9804.

[0084] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0085] (1) Perform time series decomposition on the real-time power factor value series to obtain power factor decomposition data;

[0086] (2) Perform time series decomposition on the harmonic distortion series to obtain harmonic distortion decomposition data;

[0087] (3) Perform cross-correlation analysis on the power factor decomposition data and the harmonic distortion decomposition data to obtain correlation data;

[0088] (4) Perform spectrum analysis on the power factor decomposition data and the harmonic distortion decomposition data to obtain frequency characteristic data;

[0089] (5) Combine and analyze the correlation data and frequency characteristic data to obtain multi-scale correlation indicators;

[0090] (6) Perform principal component analysis on multi-scale correlation indicators to obtain data on influencing factors for power factor optimization;

[0091] (7) Perform time series prediction on the power factor decomposition data and harmonic distortion decomposition data to obtain the expected change trend.

[0092] Specifically, the real-time power factor value series is decomposed into a time series. This process uses the seasonal and trend decomposition (STL) method to decompose the original series into trend terms, seasonal terms, and residual terms. The trend term reflects the long-term trend, the seasonal term captures periodic fluctuations, and the residual term contains short-term fluctuations and noise. The same decomposition process is also applied to the harmonic distortion series to obtain the corresponding trend terms, seasonal terms, and residual terms. This decomposition allows the various components of the data to be analyzed separately, thereby gaining a deeper understanding of the dynamic characteristics of the system. Cross-correlation analysis is performed on the power factor decomposition data and the harmonic distortion decomposition data. Cross-correlation analysis calculates the correlation between two time series at different time delays, which can reveal the time lag relationship and mutual influence between the two series. This step uses the sliding window technique to calculate the correlation coefficient at different time scales to obtain a multi-dimensional correlation data matrix. Spectral analysis is performed on the power factor decomposition data and the harmonic distortion decomposition data. This process uses fast Fourier transform (FFT) or wavelet transform to convert time domain signals to frequency domain. Spectral analysis can identify the main periodic components in the data, including intraday fluctuations, weekly fluctuations, etc. By analyzing the peak value and energy distribution of the spectrum, frequency characteristic data is obtained, which reflects the periodic characteristics of the system operation.

[0093] The correlation data and frequency characteristic data are combined and analyzed to obtain multi-scale correlation indicators. This step uses a multi-resolution analysis method to comprehensively consider the relationship between power factor and harmonic distortion at different time and frequency scales. Multi-scale correlation indicators include short-term, medium-term and long-term correlations, as well as correlation strengths in different frequency bands. The multi-scale correlation indicators are subjected to principal component analysis (PCA), which is a dimensionality reduction technique that can extract the most important features from high-dimensional data. PCA finds the main direction of change in the data by calculating the eigenvalues ​​and eigenvectors of the covariance matrix. By selecting the first few principal components, the influencing factor data for power factor optimization can be obtained, which represent the aspects that have the greatest impact on the power factor change.

[0094] Finally, the power factor decomposition data and harmonic distortion decomposition data are predicted in time series. This step uses the autoregressive integrated moving average (ARIMA) model or the long short-term memory (LSTM) neural network. The prediction model predicts the power factor and harmonic distortion in the future based on the pattern of historical data to obtain the expected change trend.

[0095] For example, 10,080 power factor and harmonic distortion data points (one per minute) were collected for a subway line in a week. STL decomposition of the power factor series yielded seasonal terms showing weekday and weekend differences, as well as a slowly increasing trend term. The decomposition results for the harmonic distortion series showed similar seasonality, but the trend term was volatile. Cross-correlation analysis found that power factor changes lagged behind harmonic distortion changes by about 5 minutes, with a correlation coefficient of -0.75. Spectral analysis revealed a strong 24-hour cycle and a weaker 7-day cycle. Multiscale correlation indicators showed that the short-term (within 1 hour) correlation was the strongest, while the long-term (more than 1 day) correlation was weaker. Principal component analysis extracted three main influencing factors that explained 75%, 15%, and 5% of the variance, respectively, corresponding to daily load changes, temperature effects, and equipment aging. The ARIMA model predicted that the average power factor for the next day would increase from 0.95 to 0.96, while the harmonic distortion would decrease from 3.5% to 3.3%.

[0096] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0097] (1) Normalize the influencing factor data to obtain standardized influencing factor data;

[0098] (2) Perform principal component analysis on the standardized influencing factor data to obtain dimension-reduced feature data;

[0099] (3) Perform correlation analysis on the reduced dimension feature data and expected change trends to obtain parameter sensitivity indicators;

[0100] (4) Sort the parameter sensitivity indicators to obtain the priority adjustment parameter set;

[0101] (5) Analyze the constraints of the priority adjustment parameter set to obtain the parameter adjustment range;

[0102] (6) Mapping and converting the parameter adjustment range to obtain the substation rectifier trigger angle adjustment range and the vehicle inverter modulation index adjustment range;

[0103] (7) Based on the trigger angle adjustment range and the modulation index adjustment range, the adjustment strategies for the substation rectifier trigger angle and the on-board inverter modulation index are obtained.

[0104] Specifically, the influencing factor data is normalized. This step converts data of different dimensions to the same scale for subsequent analysis. The normalization process uses the maximum and minimum method to map the value of each feature between 0 and 1, thereby obtaining standardized influencing factor data. Next, principal component analysis (PCA) is performed on the standardized influencing factor data. PCA is a dimensionality reduction technique that projects the original high-dimensional data into a low-dimensional space through linear transformation while retaining the main information of the data. PCA first calculates the covariance matrix of the data, and then solves the eigenvalues ​​and eigenvectors. The eigenvectors corresponding to the first few largest eigenvalues ​​are selected as principal components, and the original data are projected onto these principal components to obtain reduced-dimensional feature data.

[0105] Perform correlation analysis on the reduced dimension feature data and the expected change trend. This step uses the Pearson correlation coefficient calculation method to analyze the linear correlation between each reduced dimension feature and the expected change trend. The larger the absolute value of the correlation coefficient, the more significant the impact of the feature on the expected change trend. Through this analysis, the parameter sensitivity index is obtained, which reflects the sensitivity of each factor to the power factor change. Sort the parameter sensitivity indexes and arrange the sensitivity indexes from large to small according to the absolute value to obtain the priority adjustment parameter set. This step determines the parameters that should be given priority in the power factor correction process, which helps to improve the correction efficiency.

[0106] Constraint analysis is performed on the priority adjustment parameter set. This process takes into account factors such as equipment physical limitations, safety boundaries, and operating specifications to determine the adjustable range of each parameter. Constraint analysis ensures the feasibility and safety of the adjustment strategy and prevents over-adjustment from causing system instability or equipment damage. Then, the parameter adjustment range is mapped and converted to obtain the substation rectifier trigger angle adjustment range and the on-board inverter modulation index adjustment range. This step converts the abstract parameter adjustment range into specific control parameter values, providing direct guidance for actual operations. The mapping conversion takes into account the equipment characteristics and system response curve to ensure the optimization of the adjustment effect.

[0107] Finally, based on the trigger angle adjustment interval and the modulation index adjustment interval, the adjustment strategy of the substation rectifier trigger angle and the on-board inverter modulation index is formulated. The adjustment strategy includes not only the target value of the parameter, but also the timing and step size of the adjustment to achieve smooth and effective power factor correction.

[0108] For example, in the power factor optimization analysis of a subway line, the initial influencing factor data included 10 factors such as load level, ambient temperature, and equipment aging. After normalization, the load level was converted from 70% of the original value to 0.7, and the ambient temperature was converted from 25°C to 0.5. The results of principal component analysis showed that the first three principal components explained 85% of the variance of the data, so these three principal components were selected as dimensionality reduction feature data. Correlation analysis found that the correlation coefficient between the first principal component and the expected power factor change trend was 0.85, which was much higher than other principal components. The parameter sensitivity ranking results showed that the load level, rectifier trigger angle, and inverter modulation index were the three most influential factors. Considering equipment limitations, the adjustment range of the rectifier trigger angle was determined to be 15° to 30°, and the adjustment range of the inverter modulation index was 0.8 to 0.95. The final regulation strategy is: when the load level increases, the rectifier trigger angle is gradually adjusted from the current 20° to 25°, and the inverter modulation index is increased from 0.85 to 0.9, with an adjustment step of 0.5° and 0.01 per minute.

[0109] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0110] (1) Decompose the regulation strategy in time series to obtain a discrete regulation sequence, and interpolate the discrete regulation sequence to obtain a continuous regulation function;

[0111] (2) Calculate the trigger angle adjustment of the substation rectifier according to the continuous adjustment function to obtain the updated trigger angle value;

[0112] (3) Calculate the modulation index change of the vehicle inverter according to the continuous regulation function to obtain an updated modulation index value;

[0113] (4) Perform harmonic analysis on the updated trigger angle value and the updated modulation index value to obtain the target power factor value;

[0114] (5) Compare the target power factor value with the real-time power factor value sequence to obtain the power factor improvement amount;

[0115] (6) Based on the power factor improvement amount, the real-time power factor value sequence is updated to obtain a corrected power factor value sequence.

[0116] It should be noted that the regulation strategy is decomposed in time series, and the continuous regulation strategy is converted into the control quantity at discrete time points to obtain a discrete regulation sequence. This process takes into account the dynamic response characteristics and control cycle of the system. Then, the discrete regulation sequence is interpolated, usually using the cubic spline interpolation method, to obtain a continuous regulation function. The continuous regulation function can be expressed as:

[0117] ;

[0118] in, t For time, , , , is the interpolation coefficient, is a discrete time point.

[0119] The trigger angle adjustment of the substation rectifier is calculated according to the continuous adjustment function to obtain the updated trigger angle value. The calculation formula of the trigger angle adjustment is:

[0120] ;

[0121] in, is the trigger angle adjustment, is the firing angle adjustment coefficient, is the adjustment function value at the current moment, is the regulation function value at the initial moment. The modulation index change of the vehicle inverter is calculated according to the continuous regulation function to obtain the updated modulation index value. The calculation formula of the modulation index change is:

[0122] ;

[0123] in, is the modulation index change, is the modulation index adjustment coefficient.

[0124] Perform harmonic analysis on the updated trigger angle value and the updated modulation index value, and use the fast Fourier transform (FFT) algorithm to obtain the target power factor value. Then, compare the target power factor value with the real-time power factor value sequence to calculate the power factor improvement. Finally, based on the power factor improvement, update the real-time power factor value sequence to obtain the corrected power factor value sequence.

[0125] For example, the regulation strategy of a subway line stipulates that the rectifier trigger angle is adjusted from 20° to 25° and the inverter modulation index is adjusted from 0.85 to 0.9 within 5 minutes. First, divide these 5 minutes into 30 discrete time points, one point every 10 seconds. Perform cubic spline interpolation on these 30 points to obtain a continuous regulation function. Assume that at the second minute, the value of the continuous regulation function is 0.4. Trigger angle adjustment coefficient is 10°, then the trigger angle adjustment is = 10×(0.4 - 0) = 4°, the updated trigger angle value is 24°. Modulation index adjustment coefficient is 0.1, then the modulation index change is = 0.1×(0.4 - 0)= 0.04, and the updated modulation index value is 0.89. Perform harmonic analysis on these updated values ​​and assume that the target power factor value is 0.94. Compare 0.94 with the current real-time power factor value of 0.92 to obtain the power factor improvement of 0.02. Finally, apply this improvement to the real-time power factor value sequence to update the power factor value at the corresponding time point.

[0126] The embodiment of the present invention also provides a dynamic power factor correction system for a subway DC power supply system, such as Figure 2 As shown, the dynamic power factor correction system of the subway DC power supply system specifically includes:

[0127] The acquisition module 201 is used to acquire key parameters of the subway DC power supply system in real time to obtain an original data set;

[0128] The processing module 202 is used to perform fast Fourier transform processing on the original data set to obtain voltage spectrum data and current spectrum data;

[0129] An analysis module 203 is used to perform dynamic power factor calculation and harmonic distortion analysis on the voltage spectrum data and the current spectrum data to obtain a real-time power factor value sequence and a harmonic distortion degree sequence;

[0130] A prediction module 204 is used to perform multi-parameter correlation analysis and trend prediction based on the real-time power factor value sequence and harmonic distortion sequence to obtain influencing factor data and expected change trend of power factor optimization;

[0131] A sorting module 205 is used to perform parameter sensitivity calculation and priority sorting according to the influencing factor data and expected change trend, and generate a regulation strategy for the substation rectifier trigger angle and the vehicle inverter modulation index;

[0132] The correction module 206 is used to adjust the substation rectifier trigger angle and the vehicle inverter modulation index in real time based on the adjustment strategy, dynamically correct the real-time power factor value sequence, and obtain a corrected power factor value sequence.

[0133] Through the collaborative work of the above modules, the key parameters of the system are collected in real time and fast Fourier transform processing is performed, providing high-precision spectrum data for power factor calculation, greatly improving the accuracy of correction. Dynamic power factor calculation and harmonic distortion analysis can reflect the changes in system status in real time, making the correction process more timely and accurate. Multi-parameter correlation analysis and trend prediction technology deeply explore the inherent correlation between data, effectively identify the key factors affecting the power factor, and improve the pertinence and efficiency of correction. Parameter sensitivity calculation and priority sorting ensure that resources are reasonably allocated to the most critical parameters and optimize the resource utilization efficiency of the correction process. The real-time adjustment strategy of the substation rectifier trigger angle and the on-board inverter modulation index takes into account the dynamic characteristics and equipment limitations of the system, and realizes fast and smooth power factor correction. In addition, the adaptability of this method enables it to respond quickly to load changes, continuously optimize the correction strategy, and keep the system running in the optimal power factor state, which not only significantly improves the accuracy of power factor correction, but also greatly improves the correction efficiency.

[0134] The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the same. Although the present invention has been described in detail with reference to the embodiments, a person skilled in the art should understand that the specific implementation modes of the present invention can still be modified or replaced by equivalents, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A dynamic power factor correction method for a subway DC power supply system, characterized in that: include: Collect key parameters of the subway DC power supply system in real time to obtain the original data set; Performing fast Fourier transform processing on the original data set to obtain voltage spectrum data and current spectrum data; Perform dynamic power factor calculation and harmonic distortion analysis on the voltage spectrum data and the current spectrum data to obtain a real-time power factor value sequence and a harmonic distortion degree sequence; Based on the real-time power factor value sequence and the harmonic distortion sequence, multi-parameter correlation analysis and trend prediction are performed to obtain the influencing factor data and expected change trend of power factor optimization. The step of performing multi-parameter correlation analysis and trend prediction based on the real-time power factor value sequence and the harmonic distortion sequence to obtain the influencing factor data and expected change trend of power factor optimization includes: performing time series decomposition on the real-time power factor value sequence to obtain power factor decomposition data; performing time series decomposition on the harmonic distortion sequence to obtain harmonic distortion decomposition data; performing cross-correlation analysis on the power factor decomposition data and the harmonic distortion decomposition data to obtain correlation data; performing spectrum analysis on the power factor decomposition data and the harmonic distortion decomposition data to obtain frequency characteristic data; performing combined analysis on the correlation data and the frequency characteristic data to obtain multi-scale correlation indicators; performing principal component analysis on the multi-scale correlation indicators to obtain the influencing factor data of power factor optimization; performing time series prediction on the power factor decomposition data and the harmonic distortion decomposition data to obtain the expected change trend; According to the influencing factor data and the expected change trend, parameter sensitivity calculation and priority sorting are performed to generate an adjustment strategy for the substation rectifier trigger angle and the on-board inverter modulation index. The steps of performing parameter sensitivity calculation and priority sorting according to the influencing factor data and the expected change trend to generate an adjustment strategy for the substation rectifier trigger angle and the on-board inverter modulation index include: normalizing the influencing factor data to obtain standardized influencing factor data; performing principal component analysis on the standardized influencing factor data to obtain reduced dimension feature data; performing correlation analysis on the reduced dimension feature data and the expected change trend to obtain parameter sensitivity indexes; sorting the parameter sensitivity indexes to obtain a priority adjustment parameter set; performing constraint condition analysis on the priority adjustment parameter set to obtain a parameter adjustment range; mapping and converting the parameter adjustment range to obtain a substation rectifier trigger angle adjustment interval and a on-board inverter modulation index adjustment interval; based on the trigger angle adjustment interval and the modulation index adjustment interval, obtaining an adjustment strategy for the substation rectifier trigger angle and the on-board inverter modulation index; Based on the regulation strategy, the substation rectifier trigger angle and the vehicle inverter modulation index are adjusted in real time, and the real-time power factor value sequence is dynamically corrected to obtain a corrected power factor value sequence.

2. The dynamic power factor correction method for a subway DC power supply system according to claim 1, characterized in that: The step of collecting the key parameters of the subway DC power supply system in real time to obtain the original data set includes: Install a voltage sensor and a current sensor at the output end of the traction substation of the subway DC power supply system to collect initial DC voltage and initial current data, and convert the initial DC voltage and initial current data into digital signals to obtain digital voltage data and digital current data; Collecting the real-time position coordinates and running status information of the train, and encoding the real-time position coordinates and running status information into a preset standard format to obtain encoded position information and encoded status information; Performing noise filtering and outlier detection on the digital voltage data and the digital current data, removing interference signals and outliers, and obtaining cleaning voltage data and cleaning current data; Performing timestamp alignment and data completion on the cleaning voltage data and the cleaning current data to obtain synchronous voltage data and synchronous current data; Performing coordinate conversion on the coded position information, mapping the physical coordinates to the power supply network topology of the subway DC power supply system, obtaining train position information, and performing data decoding on the coded status information to obtain train status data; The synchronous voltage data, the synchronous current data, the train position information and the train status data are combined in time series to obtain the original data set.

3. The dynamic power factor correction method for a subway DC power supply system according to claim 1, characterized in that: The step of performing fast Fourier transform processing on the original data set to obtain voltage spectrum data and current spectrum data includes: Performing segment processing on the synchronous voltage data in the original data set to obtain segment voltage data, and performing segment processing on the synchronous current data in the original data set to obtain segment current data; Performing windowing processing on the segmented voltage data and the segmented current data respectively to obtain windowed voltage data and windowed current data; Performing zero-filling on the windowed voltage data and the windowed current data to obtain filled voltage data and filled current data; Performing a fast Fourier transform on the filling voltage data to obtain first spectrum data, and performing a fast Fourier transform on the filling current data to obtain second spectrum data; Performing amplitude normalization and phase correction on the first spectrum data and the second spectrum data to obtain first corrected spectrum data and second corrected spectrum data; The frequency resolution of the first corrected spectrum data and the second corrected spectrum data are respectively corrected to obtain voltage spectrum data including voltage harmonic components and current spectrum data including current harmonic components.

4. The dynamic power factor correction method for a subway DC power supply system according to claim 1, characterized in that: The step of performing dynamic power factor calculation and harmonic distortion analysis on the voltage spectrum data and the current spectrum data to obtain a real-time power factor value sequence and a harmonic distortion degree sequence includes: Extracting the fundamental frequency component of the voltage spectrum data to obtain voltage fundamental frequency amplitude data; Extracting the fundamental frequency component of the current spectrum data to obtain current fundamental frequency amplitude data; Calculating the ratio of the voltage fundamental frequency amplitude data to the current fundamental frequency amplitude data to obtain system impedance characteristic data; Extracting the amplitude of each harmonic component in the voltage spectrum data to obtain voltage harmonic data; Extracting the amplitude of each harmonic component in the current spectrum data to obtain current harmonic data; Calculating the power contribution of each harmonic according to the voltage harmonic data and the current harmonic data to obtain harmonic power data; Based on the system impedance characteristic data and the harmonic power data, the active power and reactive power of the subway DC power supply system are calculated to obtain a dynamic power factor value; Performing frequency-weighted processing on the voltage harmonic data and the current harmonic data, calculating total harmonic distortion, and obtaining a harmonic distortion sequence; The dynamic power factor values ​​are combined in time sequence to obtain the real-time power factor value sequence.

5. The dynamic power factor correction method for a subway DC power supply system according to claim 1, characterized in that: The step of adjusting the substation rectifier trigger angle and the vehicle inverter modulation index in real time based on the regulation strategy, dynamically correcting the real-time power factor value sequence, and obtaining the corrected power factor value sequence includes: Decomposing the regulation strategy in time series to obtain a discrete regulation sequence, and interpolating the discrete regulation sequence to obtain a continuous regulation function; Calculate the trigger angle adjustment amount of the substation rectifier according to the continuous adjustment function to obtain an updated trigger angle value; Calculating a modulation index change of the vehicle inverter according to the continuous adjustment function to obtain an updated modulation index value; Performing harmonic analysis on the updated trigger angle value and the updated modulation index value to obtain a target power factor value; Comparing the target power factor value with the real-time power factor value sequence to obtain a power factor improvement amount; Based on the power factor improvement amount, the real-time power factor value sequence is updated to obtain the corrected power factor value sequence.

6. A dynamic power factor correction system for a subway DC power supply system, used to execute the dynamic power factor correction method for a subway DC power supply system according to any one of claims 1 to 5, characterized in that: include: The acquisition module is used to collect key parameters of the subway DC power supply system in real time to obtain the original data set; A processing module, used for performing fast Fourier transform processing on the original data set to obtain voltage spectrum data and current spectrum data; An analysis module is used to perform dynamic power factor calculation and harmonic distortion analysis on the voltage spectrum data and the current spectrum data to obtain a real-time power factor value sequence and a harmonic distortion degree sequence; A prediction module is used to perform multi-parameter correlation analysis and trend prediction based on the real-time power factor value sequence and the harmonic distortion sequence to obtain the influencing factor data and the expected change trend of power factor optimization. The step of performing multi-parameter correlation analysis and trend prediction based on the real-time power factor value sequence and the harmonic distortion sequence to obtain the influencing factor data and the expected change trend of power factor optimization includes: performing time series decomposition on the real-time power factor value sequence to obtain power factor decomposition data; performing time series decomposition on the harmonic distortion sequence to obtain harmonic distortion decomposition data; performing cross-correlation analysis on the power factor decomposition data and the harmonic distortion decomposition data to obtain correlation data; performing spectrum analysis on the power factor decomposition data and the harmonic distortion decomposition data to obtain frequency characteristic data; performing combined analysis on the correlation data and the frequency characteristic data to obtain multi-scale correlation indicators; performing principal component analysis on the multi-scale correlation indicators to obtain the influencing factor data of power factor optimization; performing time series prediction on the power factor decomposition data and the harmonic distortion decomposition data to obtain the expected change trend; A sorting module is used to perform parameter sensitivity calculation and priority sorting according to the influencing factor data and the expected change trend, and generate an adjustment strategy for the substation rectifier trigger angle and the vehicle inverter modulation index. The steps of performing parameter sensitivity calculation and priority sorting according to the influencing factor data and the expected change trend, and generating an adjustment strategy for the substation rectifier trigger angle and the vehicle inverter modulation index, include: normalizing the influencing factor data to obtain standardized influencing factor data; performing principal component analysis on the standardized influencing factor data to obtain reduced dimension feature data; performing correlation analysis on the reduced dimension feature data and the expected change trend to obtain parameter sensitivity index; sorting the parameter sensitivity index to obtain a priority adjustment parameter set; performing constraint condition analysis on the priority adjustment parameter set to obtain a parameter adjustment range; mapping and converting the parameter adjustment range to obtain a substation rectifier trigger angle adjustment interval and a vehicle inverter modulation index adjustment interval; based on the trigger angle adjustment interval and the modulation index adjustment interval, obtaining the adjustment strategy for the substation rectifier trigger angle and the vehicle inverter modulation index; The correction module is used to adjust the substation rectifier trigger angle and the vehicle inverter modulation index in real time based on the adjustment strategy, dynamically correct the real-time power factor value sequence, and obtain a corrected power factor value sequence.

Images