A method for short-term load forecasting of main power grid using hybrid active filter

Abstract

The present invention discloses a method for short-term load forecasting of a main power grid using a hybrid active filter, specifically relating to the technical field of load forecasting. By obtaining the active control parameters and passive branch parameters of the hybrid active filter, combined with the real-time load fluctuation signal of the main power grid, the harmonic suppression characteristics and phase distortion characteristics are extracted, and a time-scale compensation factor is generated through dynamic coupling effect analysis. A phase correction model is constructed based on the transient harmonic residual component and the time-scale compensation factor, and the real-time load fluctuation signal is jointly reconstructed in the frequency domain and time domain to obtain a corrected load fluctuation feature sequence. The short-term load forecasting model is calibrated through a collaborative mapping relationship, and the load power forecast value for the future preset time window of the main power grid is output. This effectively distinguishes between real load fluctuations and sub-high-frequency interference signals generated by the filter's self-resonance, solving the problem of coupling interference of the harmonic residual component on the load forecasting model in scenarios where the grid impedance suddenly changes.

Description

Technical Field

[0001] The present invention relates to the technical field of load forecasting, and more particularly to a method for forecasting short-term load of a main power grid using a hybrid active filter. Background Art

[0002] Hybrid active filters are widely used in power grid environments with nonlinear loads due to their harmonic suppression capabilities. Existing technologies face significant challenges in scenarios where a high proportion of new energy access and large-scale power electronic equipment are connected to the grid: when the grid impedance changes rapidly due to sudden load changes or topology reconstruction, dynamic coupling interference is easily generated between the passive branch resonant characteristics of the hybrid active filter and the active control parameters. This interference not only causes sub-high frequency oscillations in the filter itself, but also superimposes pseudo-fluctuation components on the grid-side voltage and current signals that overlap with the actual load fluctuation frequency band. Traditional load forecasting methods lack a collaborative analysis mechanism for the dynamic parameters of the filter and the grid impedance matching characteristics, making it difficult to effectively separate such pseudo-fluctuation signals. This results in harmonic residual component coupling errors in short-term load forecasting models in impedance mutation scenarios, seriously affecting the prediction accuracy and the reliability of grid dispatching decisions.

[0003] In order to solve the above problems, a technical solution is now provided. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a main power grid short-term load forecasting method using a hybrid active filter to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The main power grid short-term load forecasting method using a hybrid active filter includes the following steps:

[0007] Obtain active control parameters and passive branch parameters of the hybrid active filter, and synchronously obtain real-time load fluctuation signals of the main power grid;

[0008] Extract the harmonic suppression characteristics corresponding to the dynamic compensation bandwidth in the active control parameters and the phase distortion characteristics corresponding to the resonance point offset in the passive branch parameters;

[0009] Perform dynamic coupling effect analysis on harmonic suppression characteristics and phase distortion characteristics to generate time-scale compensation factors;

[0010] Based on the real-time load fluctuation signal, the transient harmonic residual component generated by the hybrid active filter when the grid impedance changes is extracted. Combined with the time-scale compensation factor, a phase correction model for the load forecast window is constructed.

[0011] According to the phase correction model, the real-time load fluctuation signal is reconstructed in both frequency and time domains to generate a corrected load fluctuation feature sequence.

[0012] A collaborative mapping relationship between the hybrid active filter parameters and the load forecast time series is established. The short-term load forecast model is calibrated based on the corrected load fluctuation characteristic sequence, and the short-term load forecast results of the main power grid are output.

[0013] In a preferred embodiment, the active control parameters and passive branch parameters of the hybrid active filter are obtained, and the real-time load fluctuation signal of the main power grid is obtained synchronously, specifically:

[0014] The dynamic compensation bandwidth, harmonic detection gain and current tracking response time among the active control parameters are collected through the control unit of the hybrid active filter;

[0015] The resonant frequency, inductance and capacitance values ​​of the passive branch parameters are collected by an impedance measuring device in the passive branch;

[0016] The real-time load fluctuation signal of the main power grid is collected by the power grid monitoring device; the real-time load fluctuation signal includes a voltage amplitude fluctuation sequence, a current phase jump feature and a power grid impedance change.

[0017] In a preferred embodiment, the harmonic suppression characteristics corresponding to the dynamic compensation bandwidth in the active control parameters and the phase distortion characteristics corresponding to the resonance point offset in the passive branch parameters are extracted, specifically:

[0018] Based on the dynamic compensation bandwidth and harmonic detection gain, the harmonic suppression characteristics of the hybrid active filter within a preset frequency band are extracted through time-frequency analysis. The harmonic suppression characteristics include the attenuation rate of each harmonic and the compensation current tracking error.

[0019] Based on the resonant frequency in the passive branch and the grid impedance change in the real-time load fluctuation signal, the phase distortion characteristics caused by the resonance point offset are extracted through the impedance scanning method; the phase distortion characteristics include the resonant frequency offset and the phase difference change gradient of the passive branch voltage and current.

[0020] In a preferred embodiment, a dynamic coupling effect analysis is performed on the harmonic suppression characteristics and the phase distortion characteristics to generate a time scale compensation factor, specifically:

[0021] According to the harmonic attenuation rate in the harmonic suppression feature and the phase difference change gradient in the phase distortion feature, a dynamic coupling effect matrix is ​​constructed;

[0022] Based on the dynamic coupling effect matrix, the coupling weight coefficients of harmonic suppression and phase distortion are calculated through the transfer function correction model;

[0023] The coupling weight coefficient and the current tracking response time are combined to generate a time scale compensation factor; the time scale compensation factor includes a frequency domain compensation gain and a timing compensation delay amount.

[0024] In a preferred embodiment, based on the real-time load fluctuation signal, the transient harmonic residual component generated by the hybrid active filter when the grid impedance changes is extracted, and combined with the time-scale compensation factor, a phase correction model for the load forecast window is constructed, specifically:

[0025] Based on the voltage amplitude fluctuation sequence and grid impedance variation in the real-time load fluctuation signal, the transient harmonic residual component of the hybrid active filter when the impedance changes is calculated using the state space equation;

[0026] Dynamically couple the transient harmonic residual component with the frequency domain compensation gain in the time-scale compensation factor, and generate the initial parameters of the phase correction model based on a preset transfer function;

[0027] The initial parameters are adaptively modified through a time-varying transfer function to construct a phase correction model for each time node in the load forecast window; the phase correction model includes a frequency domain compensation weight and a time delay compensation coefficient.

[0028] In a preferred embodiment, the real-time load fluctuation signal is reconstructed in both frequency and time domains according to the phase correction model to generate a corrected load fluctuation feature sequence, specifically:

[0029] The frequency domain compensation weight is used to compensate the current phase jump characteristics of the real-time load fluctuation signal in the frequency domain, and the frequency domain reconstructed signal is generated through the inverse Fourier transform;

[0030] The frequency domain reconstructed signal is superimposed on the timing compensation delay in the time-scale compensation factor according to the time delay compensation coefficient, and time domain smoothing is performed using the wavelet transform method to generate a corrected load fluctuation characteristic sequence; the load fluctuation characteristic sequence includes the compensated voltage fluctuation trajectory, the corrected current phase, and the load time series data with suppressed harmonic content.

[0031] In a preferred embodiment, a collaborative mapping relationship between the hybrid active filter parameters and the load forecast time series is established, the short-term load forecast model is calibrated based on the corrected load fluctuation characteristic sequence, and the main grid short-term load forecast result is output, specifically:

[0032] The current tracking response time in the active control parameters and the inductance value in the passive branch parameters are used as input variables, and a collaborative mapping relationship with the load forecasting time series is established through a dynamic weight allocation algorithm.

[0033] The corrected load fluctuation characteristic sequence is input into the short-term load forecasting model, and the short-term load forecasting model is calibrated according to the collaborative mapping relationship;

[0034] The main grid short-term load forecast result is output based on the calibrated short-term load forecast model; the main grid short-term load forecast result is a load power forecast value within a future preset time window.

[0035] In another aspect, the present invention provides a main power grid short-term load forecasting system using a hybrid active filter, comprising:

[0036] Parameter synchronization acquisition module: obtains the active control parameters and passive branch parameters of the hybrid active filter, and synchronously obtains the real-time load fluctuation signal of the main power grid;

[0037] Feature extraction module: extracts the harmonic suppression characteristics corresponding to the dynamic compensation bandwidth in the active control parameters and the phase distortion characteristics corresponding to the resonance point offset in the passive branch parameters;

[0038] Dynamic coupling analysis module: performs dynamic coupling effect analysis on harmonic suppression characteristics and phase distortion characteristics, and generates time-scale compensation factors;

[0039] Phase correction modeling module: Based on real-time load fluctuation signals, it extracts the transient harmonic residual components generated by the hybrid active filter when the grid impedance changes, and combines them with the time-scale compensation factor to build a phase correction model for the load forecast window;

[0040] Frequency-time domain joint reconstruction module: Based on the phase correction model, the real-time load fluctuation signal is jointly reconstructed in the frequency and time domains to generate a corrected load fluctuation feature sequence;

[0041] Load forecast calibration module: establishes a coordinated mapping relationship between the hybrid active filter parameters and the load forecast time series, calibrates the short-term load forecast model based on the corrected load fluctuation characteristic sequence, and outputs the short-term load forecast results of the main power grid.

[0042] On the other hand, the present invention provides a computer-readable storage medium storing a program or instruction. When the program or instruction is executed by a processor, a method for short-term load forecasting of a main power grid using a hybrid active filter is implemented.

[0043] The technical effects and advantages of the main power grid short-term load forecasting method using a hybrid active filter of the present invention are as follows:

[0044] 1. By establishing a deep synergy between the dynamic parameters of the hybrid active filter and the load forecasting model, the problem of short-term power fluctuation perception distortion caused by the self-resonance of the hybrid active filter in large-scale nonlinear load scenarios is effectively resolved. By extracting harmonic suppression and phase distortion characteristics and combining them with the time-scale compensation factor generated by dynamic coupling effect analysis, the self-resonance frequency component of the hybrid active filter and the real-time load fluctuation signal are accurately separated. Based on frequency domain compensation and timing delay correction of transient harmonic residual components, the superimposed interference of sub-high frequency pseudo-fluctuation components generated during the harmonic suppression process when the grid impedance changes suddenly on the voltage and current measurement signals is eliminated, thereby improving the time-frequency domain fidelity of the load fluctuation feature sequence.

[0045] 2. By establishing a coordinated mapping relationship between hybrid active filter parameters and load forecasting time series, the short-term load forecasting model's adaptability to dynamic changes in grid impedance is enhanced, avoiding residual harmonic coupling errors caused by ignoring the dynamic characteristics of the hybrid active filter. This not only suppresses contamination of the measurement signal by the hybrid active filter's self-resonance, but also reconstructs the true load fluctuation trajectory through a phase correction model, providing highly reliable short-term load forecasting results for grid dispatch. This has outstanding engineering application value in complex grid scenarios with high penetration of new energy sources and dense integration of power electronic equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a schematic diagram of a main power grid short-term load forecasting method using a hybrid active filter according to the present invention. DETAILED DESCRIPTION

[0047] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0048] Example 1, Figure 1 The present invention provides a main power grid short-term load forecasting method using a hybrid active filter, which includes the following steps:

[0049] Obtain active control parameters and passive branch parameters of the hybrid active filter, and synchronously obtain real-time load fluctuation signals of the main power grid;

[0050] Extract the harmonic suppression characteristics corresponding to the dynamic compensation bandwidth in the active control parameters and the phase distortion characteristics corresponding to the resonance point offset in the passive branch parameters;

[0051] Perform dynamic coupling effect analysis on harmonic suppression characteristics and phase distortion characteristics to generate time-scale compensation factors;

[0052] Based on the real-time load fluctuation signal, the transient harmonic residual component generated by the hybrid active filter when the grid impedance changes is extracted. Combined with the time-scale compensation factor, a phase correction model for the load forecast window is constructed.

[0053] According to the phase correction model, the real-time load fluctuation signal is reconstructed in both frequency and time domains to generate a corrected load fluctuation feature sequence.

[0054] A collaborative mapping relationship between the hybrid active filter parameters and the load forecast time series is established. The short-term load forecast model is calibrated based on the corrected load fluctuation characteristic sequence, and the short-term load forecast results of the main power grid are output.

[0055] Specifically, the active control parameters and passive branch parameters of the hybrid active filter are obtained, and the real-time load fluctuation signal of the main power grid is obtained synchronously, including:

[0056] The dynamic compensation bandwidth, harmonic detection gain and current tracking response time among the active control parameters are collected through the control unit of the hybrid active filter;

[0057] Specifically, the dynamic compensation bandwidth refers to the frequency range within which the hybrid active filter can effectively suppress harmonics. For example, a dynamic compensation bandwidth of 0-2kHz indicates that the hybrid active filter can suppress harmonics within this frequency band. The control unit measures this bandwidth in real time using a built-in frequency response analysis module.

[0058] The harmonic detection gain indicates the active circuit's sensitivity to harmonic currents. For example, a gain of 1.2 means the detection stage amplifies the actual harmonic current amplitude by 1.2 before feeding it into the compensation control algorithm. This value can be read through the controller's configuration registers.

[0059] Current tracking response time refers to the time delay between the active circuit detecting harmonics and outputting the compensation current. For example, a response time of 50 microseconds means that harmonic compensation is completed within 50 microseconds. This can be measured using the control unit's timestamp recording function.

[0060] The resonant frequency, inductance and capacitance values ​​of the passive branch parameters are collected by an impedance measuring device in the passive branch;

[0061] Specifically, the resonant frequency refers to the frequency of the LC series resonance point of the passive branch. The impedance measurement device determines the resonance point by using a frequency sweep method (i.e., applying test signals of different frequencies in the range of 0-2kHz and measuring the impedance amplitude).

[0062] Inductance and capacitance: Directly measure the inductor and capacitor parameters of the passive branch using an LCR meter.

[0063] The real-time load fluctuation signal of the main power grid is collected by the power grid monitoring device; the real-time load fluctuation signal includes a voltage amplitude fluctuation sequence, a current phase jump characteristic and a power grid impedance change;

[0064] Specifically, the voltage amplitude fluctuation sequence refers to recording the instantaneous changes in the effective value of the bus voltage. For example, the voltage fluctuates from 220V to 215V and then returns to 218V within 1 second.

[0065] The current phase jump characteristic refers to a sudden change in the current phase angle caused by a sudden load change. For example, when a motor starts, the current phase lag angle suddenly changes from 30° to 45°.

[0066] Grid impedance change refers to the change in grid equivalent impedance caused by load switching. For example, disconnecting a branch causes the grid impedance to increase from 0.1 ohms to 0.15 ohms.

[0067] All parameters and signals are aligned with the same timestamp. For example, active control parameters, passive branch parameters, and load fluctuation signals are recorded simultaneously at the same time.

[0068] The load fluctuation signal is decomposed into fundamental current amplitude (such as 50Hz component amplitude), frequency offset (such as fundamental frequency shifted from 50Hz to 49.8Hz) and harmonic distortion rate (such as total harmonic distortion rate THD = 5%) through fast Fourier transform.

[0069] Specifically, extracting the harmonic suppression characteristics corresponding to the dynamic compensation bandwidth in the active control parameters and the phase distortion characteristics corresponding to the resonance point offset in the passive branch parameters includes:

[0070] Based on the dynamic compensation bandwidth and harmonic detection gain, the harmonic suppression characteristics of the hybrid active filter within a preset frequency band are extracted through time-frequency analysis. The harmonic suppression characteristics include the attenuation rate of each harmonic and the compensation current tracking error.

[0071] Specifically, the time-frequency analysis method uses short-time Fourier transform to analyze the harmonic suppression effect of the active filter within a preset frequency band. For example, within the 0-2kHz range, the spectrum amplitude of each harmonic (such as the 5th, 7th, and 11th) is calculated every 10ms.

[0072] The harmonic attenuation rate is defined as the ratio of the harmonic current amplitudes before and after compensation.

[0073] The compensation current tracking error refers to the root mean square error between the active part output current and the target compensation current.

[0074] Based on the resonant frequency in the passive branch and the grid impedance change in the real-time load fluctuation signal, the phase distortion characteristics caused by the resonance point offset are extracted through the impedance scanning method; the phase distortion characteristics include the resonant frequency offset and the phase difference change gradient of the passive branch voltage and current;

[0075] Specifically, the impedance sweep method injects a 0-2kHz sweep signal into the passive branch and measures how the phase difference between voltage and current changes with frequency. For example, near the resonant frequency of 712Hz, the phase difference jumps from -90° to +90°.

[0076] Resonant frequency offset refers to the deviation of the actual resonant frequency from the nominal value. For example, if the nominal resonant frequency is 700Hz and the measured value is 712Hz, the offset is +12Hz.

[0077] The phase difference gradient is defined as the slope of the phase difference changing with frequency.

[0078] For example, if the phase difference changes linearly from -80° to +80° within the range of 700-720 Hz, the gradient is (80-(-80)) / (720-700)=8° / Hz.

[0079] Specifically, the dynamic coupling effect analysis of the harmonic suppression characteristics and phase distortion characteristics is performed to generate a time-scale compensation factor, including:

[0080] According to the harmonic attenuation rate in the harmonic suppression feature and the phase difference change gradient in the phase distortion feature, a dynamic coupling effect matrix is ​​constructed;

[0081] Specifically, the rows of the dynamic coupling effect matrix represent harmonic suppression characteristics (such as the attenuation rate of each harmonic), and the columns of the dynamic coupling effect matrix represent phase distortion characteristics (such as the phase difference change gradient in each frequency band). The matrix elements of the dynamic coupling effect matrix represent the coupling strength of the harmonic attenuation rate in the harmonic suppression characteristics and the phase difference change gradient in the phase distortion characteristics in a specific frequency band. For example, the matrix element Indicates the The subharmonic attenuation rate and Correlation of phase difference gradients across frequency bands.

[0082] Based on the dynamic coupling effect matrix, the coupling weight coefficients of harmonic suppression and phase distortion are calculated through the transfer function correction model;

[0083] Specifically, the transfer function correction model inputs the harmonic suppression and phase distortion characteristics into the state-space equations to solve for the dynamic interaction between these two characteristics. The transfer function and the state-space equations are essentially different mathematical descriptions of the same dynamic system. The transfer function describes the input-output relationship of the system in the frequency domain (Laplace domain) and is suitable for analyzing steady-state harmonic suppression characteristics. The state-space equations describe the dynamic evolution of the system's internal state variables in the time domain and are suitable for analyzing the interaction process of transient phase distortion.

[0084] Transfer functions and state-space equations can be converted to and from each other. For example, a state-space equation can be converted to a transfer function using the Laplace transform, or a transfer function can be expanded to a state-space model by defining state variables. A modified transfer function model is a frequency-domain representation of the state-space equation.

[0085] The weight coefficients are calculated by solving the coefficient matrix in the equation using the least squares method to determine the primary and secondary effects of harmonic suppression and phase distortion. For example, if the weight coefficient for harmonic attenuation is 0.8 and the weight coefficient for phase distortion is 0.2, it indicates that harmonic suppression plays a dominant role.

[0086] Combining the coupling weight coefficient with the current tracking response time to generate a time scale compensation factor; the time scale compensation factor includes a frequency domain compensation gain and a timing compensation delay amount;

[0087] Specifically, the frequency domain compensation gain is to adjust the compensation strength of each frequency band according to the weight coefficient. For example, a higher gain (such as 1.5 times) is set for a frequency band with a high weight coefficient (such as the 5th harmonic).

[0088] The timing compensation delay is calculated based on the current tracking response time. For example, if the response time is 50 microseconds, the compensation delay is 50 microseconds.

[0089] Specifically, based on the real-time load fluctuation signal, the transient harmonic residual component generated by the hybrid active filter when the grid impedance changes is extracted. Combined with the time-scale compensation factor, a phase correction model for the load forecast window is constructed, including:

[0090] Based on the voltage amplitude fluctuation sequence and grid impedance variation in the real-time load fluctuation signal, the transient harmonic residual component of the hybrid active filter when the impedance changes is calculated using the state space equation;

[0091] Specifically, the state-space equations are a set of differential equations describing the dynamics of the hybrid active filter, constructed with grid voltage, current, and impedance as state variables. Calculation of transient harmonic residual components: Solving the differential equations yields the transient harmonic amplitudes.

[0092] Build with grid voltage , grid current and grid impedance is the system of differential equations for the state variables:

[0093] ;

[0094] ;

[0095] .

[0096] in, Indicates the instantaneous value or voltage amplitude of the voltage at the connection point between the main grid and the hybrid active filter; It represents the instantaneous value or current amplitude of the current flowing through the node connecting the hybrid active filter and the main grid, that is, at a certain moment The current state at Indicates the instantaneous value of the equivalent impedance of the main grid at the interface of the hybrid active filter, that is, at a certain moment Impedance of power grid lines and equipment; Indicates the current specific moment.

[0097] By adopting numerical integration methods such as the fourth-order Runge-Kutta method and solving the differential equations through software tools (such as MATLAB), the transient harmonic residual components of the hybrid active filter when the impedance changes are obtained.

[0098] Dynamically couple the transient harmonic residual component with the frequency domain compensation gain in the time-scale compensation factor, and generate the initial parameters of the phase correction model based on a preset transfer function;

[0099] Specifically, dynamic coupling multiplies the residual component by the compensation gain by frequency band. For example, if the 5th harmonic residual is 2A and the compensation gain is 1.5, the compensation target is 3A.

[0100] Transfer function generation: Design a filter transfer function with the compensation target as the coefficient. For example, the transfer function ;in, ; represents the transfer function; ; represents the amplification factor of the transfer function in the low frequency band, ; is the cutoff frequency; ; represents the frequency domain characteristics and is a complex variable in the Laplace transform.

[0101] Adaptively modify the initial parameters through a time-varying transfer function to construct a phase correction model for each time node within the load forecast window; the phase correction model includes a frequency domain compensation weight and a delay compensation coefficient;

[0102] Specifically, the initial parameters are adjusted online through the time-varying transfer function:

[0103] The time-varying transfer function updates the transfer function parameters according to the real-time load fluctuation signal. For example, the cutoff frequency ω is recalculated every 100ms.

[0104] Phase correction model output: The phase correction model includes frequency domain compensation weights (such as the gain of each harmonic) and delay compensation coefficients (such as the delay time of each frequency band).

[0105] Specifically, based on the phase correction model, the real-time load fluctuation signal is jointly reconstructed in the frequency domain and time domain to generate a corrected load fluctuation feature sequence, including:

[0106] The frequency domain compensation weight is used to compensate the current phase jump characteristics of the real-time load fluctuation signal in the frequency domain, and the frequency domain reconstructed signal is generated through the inverse Fourier transform;

[0107] Specifically, the frequency domain compensation method involves performing a Fourier transform on the load fluctuation signal, multiplying it by the frequency domain compensation weight matrix, and then performing an inverse Fourier transform to restore the time domain signal. For example, the 5th harmonic amplitude is multiplied by a gain of 1.5.

[0108] The frequency domain reconstructed signal is superimposed with the time series compensation delay in the time scale compensation factor according to the time delay compensation coefficient, and time domain smoothing is performed using a wavelet transform method to generate a corrected load fluctuation characteristic sequence; the load fluctuation characteristic sequence includes the compensated voltage fluctuation trajectory, the corrected current phase, and the load time series data with suppressed harmonic content;

[0109] Specifically, the wavelet transform method selects an appropriate wavelet basis function (such as the Dobesi wavelet) to perform multi-scale decomposition on the frequency domain reconstructed signal, and superimposes delay compensation at specific scales. For example, a 50μs time shift is added to the detail coefficients.

[0110] The frequency domain and time domain processing results are superimposed to generate the final feature sequence:

[0111] Weighted superposition: The frequency domain reconstructed signal accounts for 70% of the weight, and the time domain compensation signal accounts for 30% of the weight.

[0112] Characteristic sequence content: including the corrected voltage fluctuation trajectory (such as 220V±2V), current phase (such as lag angle 45°±1°), and harmonic content (such as THD=3%).

[0113] Specifically, a coordinated mapping relationship between the hybrid active filter parameters and the load forecast time series is established. The short-term load forecast model is calibrated based on the corrected load fluctuation characteristic sequence, and the main grid short-term load forecast results are output, including:

[0114] The current tracking response time in the active control parameters and the inductance value in the passive branch parameters are used as input variables, and a collaborative mapping relationship with the load forecasting time series is established through a dynamic weight allocation algorithm.

[0115] Specifically, the dynamic weight allocation algorithm uses a linear regression model to determine the weight associated with the parameter and the load. For example, the weight of the current tracking response time is 0.6, and the weight of the inductance value is 0.4.

[0116] Parameter-load association mapping table: This table stores the relationship between parameters and load values. For example, a response time of 50 microseconds corresponds to a load value of 100kW, and an inductance of 5mH corresponds to a load value of 105kW.

[0117] The corrected load fluctuation characteristic sequence is input into the short-term load forecasting model, and the short-term load forecasting model is calibrated according to the collaborative mapping relationship;

[0118] Specifically, the short-term load forecasting model adopts a long short-term memory network model, the input layer receives feature sequences (such as voltage, current, and harmonic content), and the hidden layer learns temporal dependencies.

[0119] Calibration method: Adjust the model input weights based on the collaborative mapping table. For example, if the inductance value weight is increased, the short-term load forecasting model becomes more sensitive to inductance-related characteristics.

[0120] Outputting a short-term load forecast result of the main grid based on the calibrated short-term load forecast model; the short-term load forecast result of the main grid is a load power forecast value within a future preset time window;

[0121] Specifically, the load power forecast value within the future preset time window is output:

[0122] Forecast time window: For example, a forecast point every 1 minute in the next 15 minutes.

[0123] Result form: time series data, including predicted time points and load power forecast values.

[0124] Embodiment 2: The difference between Embodiment 2 of the present invention and Embodiment 1 is that this embodiment introduces a main power grid short-term load forecasting system using a hybrid active filter.

[0125] The main power grid short-term load forecasting system using a hybrid active filter includes:

[0126] Parameter synchronization acquisition module: obtains the active control parameters and passive branch parameters of the hybrid active filter, and synchronously obtains the real-time load fluctuation signal of the main power grid;

[0127] Feature extraction module: extracts the harmonic suppression characteristics corresponding to the dynamic compensation bandwidth in the active control parameters and the phase distortion characteristics corresponding to the resonance point offset in the passive branch parameters;

[0128] Dynamic coupling analysis module: performs dynamic coupling effect analysis on harmonic suppression characteristics and phase distortion characteristics, and generates time-scale compensation factors;

[0129] Phase correction modeling module: Based on real-time load fluctuation signals, it extracts the transient harmonic residual components generated by the hybrid active filter when the grid impedance changes, and combines them with the time-scale compensation factor to build a phase correction model for the load forecast window;

[0130] Frequency-time domain joint reconstruction module: Based on the phase correction model, the real-time load fluctuation signal is jointly reconstructed in the frequency and time domains to generate a corrected load fluctuation feature sequence;

[0131] Load forecast calibration module: establishes a coordinated mapping relationship between the hybrid active filter parameters and the load forecast time series, calibrates the short-term load forecast model based on the corrected load fluctuation characteristic sequence, and outputs the short-term load forecast results of the main power grid.

[0132] Embodiment 3, a computer-readable storage medium stores a program or instruction. When the program or instruction is executed by a processor, a method for short-term load forecasting of a main power grid using a hybrid active filter is implemented.

[0133] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0134] The above embodiments can be implemented in whole or in part via software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in the embodiments of this application are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired means (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0135] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0136] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and modules described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0137] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0138] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, and may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0139] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0140] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0141] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0142] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for short-term load forecasting of a main power grid using a hybrid active filter, characterized in that: The steps include: Obtain active control parameters and passive branch parameters of the hybrid active filter, and synchronously obtain real-time load fluctuation signals of the main power grid; Extract the harmonic suppression characteristics corresponding to the dynamic compensation bandwidth in the active control parameters and the phase distortion characteristics corresponding to the resonance point offset in the passive branch parameters; Based on the dynamic compensation bandwidth and harmonic detection gain, the harmonic suppression characteristics of the hybrid active filter within the preset frequency band are extracted through time-frequency analysis method. The harmonic suppression characteristics include the attenuation rate of each harmonic and the compensation current tracking error; Based on the resonant frequency in the passive branch and the grid impedance change in the real-time load fluctuation signal, the phase distortion characteristics caused by the resonance point offset are extracted through the impedance scanning method; the phase distortion characteristics include the resonant frequency offset and the phase difference change gradient of the passive branch voltage and current; Perform dynamic coupling effect analysis on harmonic suppression characteristics and phase distortion characteristics to generate time-scale compensation factors; According to the harmonic attenuation rate in the harmonic suppression feature and the phase difference change gradient in the phase distortion feature, a dynamic coupling effect matrix is ​​constructed; Based on the dynamic coupling effect matrix, the coupling weight coefficients of harmonic suppression and phase distortion are calculated through the transfer function correction model; Combining the coupling weight coefficient with the current tracking response time to generate a time scale compensation factor; the time scale compensation factor includes a frequency domain compensation gain and a timing compensation delay amount; Based on the real-time load fluctuation signal, the transient harmonic residual component generated by the hybrid active filter when the grid impedance changes is extracted. Combined with the time-scale compensation factor, a phase correction model for the load forecast window is constructed. Based on the voltage amplitude fluctuation sequence and grid impedance variation in the real-time load fluctuation signal, the transient harmonic residual component of the hybrid active filter when the impedance changes is calculated using the state space equation; Dynamically couple the transient harmonic residual component with the frequency domain compensation gain in the time-scale compensation factor, and generate the initial parameters of the phase correction model based on a preset transfer function; Adaptively modify the initial parameters through a time-varying transfer function to construct a phase correction model for each time node within the load forecast window; the phase correction model includes a frequency domain compensation weight and a delay compensation coefficient; According to the phase correction model, the real-time load fluctuation signal is reconstructed in both frequency and time domains to generate a corrected load fluctuation feature sequence. A collaborative mapping relationship between the hybrid active filter parameters and the load forecast time series is established. The short-term load forecast model is calibrated based on the corrected load fluctuation characteristic sequence, and the short-term load forecast results of the main power grid are output.

2. The main power grid short-term load forecasting method using a hybrid active filter according to claim 1, characterized in that: Obtain the active control parameters and passive branch parameters of the hybrid active filter, and synchronously obtain the real-time load fluctuation signal of the main power grid, specifically: The dynamic compensation bandwidth, harmonic detection gain and current tracking response time among the active control parameters are collected through the control unit of the hybrid active filter; The resonant frequency, inductance and capacitance values ​​of the passive branch parameters are collected by an impedance measuring device in the passive branch; The real-time load fluctuation signal of the main power grid is collected by the power grid monitoring device; the real-time load fluctuation signal includes a voltage amplitude fluctuation sequence, a current phase jump feature and a power grid impedance change.

3. The main power grid short-term load forecasting method using a hybrid active filter according to claim 2, characterized in that: According to the phase correction model, the real-time load fluctuation signal is reconstructed in both frequency and time domains to generate a corrected load fluctuation feature sequence, specifically: The frequency domain compensation weight is used to compensate the current phase jump characteristics of the real-time load fluctuation signal in the frequency domain, and the frequency domain reconstructed signal is generated through the inverse Fourier transform; The frequency domain reconstructed signal is superimposed on the timing compensation delay in the time-scale compensation factor according to the time delay compensation coefficient, and time domain smoothing is performed using the wavelet transform method to generate a corrected load fluctuation characteristic sequence; the load fluctuation characteristic sequence includes the compensated voltage fluctuation trajectory, the corrected current phase, and the load time series data with suppressed harmonic content.

4. The main power grid short-term load forecasting method using a hybrid active filter according to claim 3, characterized in that: A collaborative mapping relationship between the hybrid active filter parameters and the load forecast time series is established. The short-term load forecast model is calibrated based on the corrected load fluctuation characteristic sequence, and the short-term load forecast results of the main power grid are output. Specifically: The current tracking response time in the active control parameters and the inductance value in the passive branch parameters are used as input variables, and a collaborative mapping relationship with the load forecasting time series is established through a dynamic weight allocation algorithm. The corrected load fluctuation characteristic sequence is input into the short-term load forecasting model, and the short-term load forecasting model is calibrated according to the collaborative mapping relationship; Output the main power grid short-term load forecast result based on the calibrated short-term load forecast model; The main power grid short-term load forecast result is the load power forecast value within a future preset time window.

5. A mains power grid short-term load forecasting system using a hybrid active filter, used to implement the mains power grid short-term load forecasting method using a hybrid active filter according to any one of claims 1 to 4, characterized in that: include: Parameter synchronization acquisition module: obtains the active control parameters and passive branch parameters of the hybrid active filter, and synchronously obtains the real-time load fluctuation signal of the main power grid; Feature extraction module: extracts the harmonic suppression characteristics corresponding to the dynamic compensation bandwidth in the active control parameters and the phase distortion characteristics corresponding to the resonance point offset in the passive branch parameters; Dynamic coupling analysis module: performs dynamic coupling effect analysis on harmonic suppression characteristics and phase distortion characteristics, and generates time-scale compensation factors; Phase correction modeling module: Based on real-time load fluctuation signals, it extracts the transient harmonic residual components generated by the hybrid active filter when the grid impedance changes, and combines them with the time-scale compensation factor to build a phase correction model for the load forecast window; Frequency-time domain joint reconstruction module: Based on the phase correction model, the real-time load fluctuation signal is jointly reconstructed in the frequency and time domains to generate a corrected load fluctuation feature sequence; Load forecast calibration module: establishes a coordinated mapping relationship between the hybrid active filter parameters and the load forecast time series, calibrates the short-term load forecast model based on the corrected load fluctuation characteristic sequence, and outputs the short-term load forecast results of the main power grid.

6. A computer-readable storage medium, characterized in that A program or instruction is stored on a computer-readable storage medium, and when the program or instruction is executed by a processor, the method for short-term load forecasting of a main power grid using a hybrid active filter according to any one of claims 1 to 4 is implemented.

Images