Broadband impedance analysis method based on network structure in DC power distribution network

Abstract

The invention discloses a broadband impedance analysis method based on a network structure in a DC power distribution network, and relates to the technical field of broadband impedance, and the method comprises the steps: based on time-frequency domain impedance features, extracting frequency domain-time domain joint features through a lightweight convolutional neural network, and generating a fault detection result through a multi-layer perceptron classifier; on the basis of fault detection scores, data synchronization is achieved on node detection results through a distributed network protocol, weighted aggregation is conducted on frequency domain-time domain joint feature extraction weight parameters through a federated learning framework, and a power grid health report containing regional impedance deviation indexes is generated; according to the method, node data are synchronized and aggregated through a distributed network protocol and a federated learning framework, the power grid health report containing the regional impedance deviation index is generated, and the cooperative work capability of power grid systems at different geographic positions is greatly enhanced in the process.

Description

Technical Field

[0001] The present invention relates to the technical field of broadband impedance, in particular to a broadband impedance analysis method based on a network structure in a DC power distribution network. Background Art

[0002] With the rapid development of renewable energy generation, DC microgrids, and flexible DC transmission technologies, the complexity and dynamic characteristics of DC distribution networks are becoming increasingly significant. Their broadband impedance characteristics directly impact stability and power quality. Traditional impedance analysis methods, primarily based on frequency domain scanning or small-signal modeling, offer some accuracy in offline analysis but struggle to adapt to the dynamic operating conditions of DC distribution networks.

[0003] In recent years, impedance analysis methods based on online monitoring have become a research hotspot, such as the use of fast Fourier transforms for real-time impedance estimation. However, these methods typically focus only on a single frequency band or steady-state conditions, making it difficult to fully capture impedance variations across a wide frequency band, particularly transient issues such as high-frequency resonance and harmonic coupling. Furthermore, existing technologies often rely on centralized computing architectures, making it difficult to meet the real-time requirements of distributed DC distribution networks and lacking optimized mechanisms for multi-node collaborative analysis. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a wide-band impedance analysis method based on the network structure in the DC distribution network, which solves the problem of insufficient accuracy of distributed collaborative analysis.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a wideband impedance analysis method based on the network structure in a DC distribution network, which includes: real-time collection of voltage and current data and preprocessing, while performing time-frequency analysis through short-time Fourier transform to generate time-frequency domain impedance features; based on the time-frequency domain impedance features, extracting frequency domain-time domain joint features through a lightweight convolutional neural network, and generating fault detection results through a multi-layer perceptron classifier; based on the fault detection score, data synchronization of node detection results is achieved through a distributed network protocol, and weighted aggregation of frequency domain-time domain joint feature extraction weight parameters is performed using a federated learning framework to generate a power grid health report including a regional impedance deviation index; based on the power grid health report, feature space mapping and classification decision-making are performed on the time-frequency domain impedance features and the regional impedance deviation index through a support vector machine algorithm to generate a wideband impedance spectrum and graded warning information; based on the wideband impedance spectrum and graded warning information, the converter PWM control parameters are adjusted in real time through an adaptive control algorithm, and an equipment maintenance strategy is generated in combination with the power grid health report.

[0008] As a preferred solution of the broadband impedance analysis method based on the network structure in the DC distribution network of the present invention, wherein: the preprocessing includes eliminating power frequency harmonics and random noise through a Gaussian filter and performing signal normalization processing.

[0009] As a preferred solution of the broadband impedance analysis method based on the network structure in the DC distribution network of the present invention, wherein: the time-frequency analysis is performed through short-time Fourier transform to generate time-frequency domain impedance characteristics, and the specific steps are as follows.

[0010] Perform time-frequency decomposition on the preprocessed voltage and current data through short-time Fourier transform to generate a time-frequency spectrum matrix.

[0011] Based on the time-frequency spectrum matrix, perform impedance characteristic conversion through the frequency-domain Ohm's law to generate time-frequency domain impedance characteristics.

[0012] As a preferred solution of the broadband impedance analysis method based on the network structure in the DC distribution network of the present invention, wherein: based on the time-frequency domain impedance characteristics, extract frequency-domain and time-domain joint characteristics through a lightweight convolutional neural network, and generate a fault detection result through a multi-layer perceptron classifier. The specific steps are as follows.

[0013] Based on the time-frequency domain impedance characteristic matrix, generate a multi-dimensional feature tensor with spatio-temporal correlation through sliding calculation with a three-dimensional convolution kernel.

[0014] Based on the multi-dimensional feature tensor, generate a frequency-domain and time-domain joint feature vector through a cross-channel attention fusion mechanism.

[0015] Based on the frequency-domain and time-domain joint feature vector, generate a fault detection result y through a multi-layer perceptron classifier.

[0016] Define a fault threshold A based on historical data.

[0017] When y < A, it is considered that the device is operating normally, and the current monitoring mode is maintained.

[0018] When y ≥ A, it is considered that the device is in an abnormal state, and the operation and maintenance personnel need to be notified immediately for repair.

[0019] As a preferred solution of the broadband impedance analysis method based on the network structure in the DC distribution network of the present invention, wherein: based on the fault detection result, synchronize node data through a distributed protocol, and use a federated learning framework to extract model parameters for the frequency-domain and time-domain joint characteristics, and generate a power grid health report including regional impedance deviation indexes. The specific steps are as follows.

[0020] Through a distributed consistency protocol, perform time alignment and quality verification on the node detection data to generate a synchronized data set.

[0021] Based on the synchronized data set, the federated learning framework aggregates the neural network parameter updates of each monitoring terminal and uses a weighted average algorithm to generate global shared parameters.

[0022] Based on global shared parameters, a grid health report is generated through impedance deviation analysis and health score mapping.

[0023] As a preferred solution of the broadband impedance analysis method based on the network structure in the DC distribution network described in the present invention, wherein: based on the grid health report, the time-frequency domain impedance characteristics and regional impedance deviation index are mapped into feature space and classified by the support vector machine algorithm to generate broadband impedance spectrum and graded warning information. The specific steps are as follows:

[0024] Based on the power grid health report, the time-frequency domain impedance characteristics and regional impedance deviation index are mapped to a high-dimensional space through kernel space mapping to generate a high-dimensional feature representation.

[0025] Based on high-dimensional feature representation, the optimal classification decision boundary is generated through the support vector machine optimization algorithm;

[0026] Based on the classification decision boundary, the spectrum is reconstructed by clustering impedance features and locating abnormal areas to generate a wide-band impedance spectrum and graded warning information.

[0027] As a preferred solution of the broadband impedance analysis method based on the network structure in the DC distribution network of the present invention, wherein: based on the broadband impedance spectrum and the graded warning information, the PWM control parameters of the converter are adjusted in real time through the adaptive control algorithm, the specific steps are as follows

[0028] Based on the broadband impedance spectrum and graded warning information, the dominant resonant frequency band characteristics are generated through Hilbert spectrum analysis;

[0029] Based on the dominant resonant frequency band characteristics, the PWM carrier frequency correction is generated through the PID regulator;

[0030] Based on the graded warning information, a safe adjustment range of the PWM modulation ratio is generated through a dynamic limiting strategy.

[0031] As a preferred solution of the broadband impedance analysis method based on the network structure in the DC distribution network of the present invention, wherein: the device maintenance strategy is generated in combination with the power grid health report, the specific steps are as follows:

[0032] Based on the PWM carrier frequency correction value and the safe adjustment range of the PWM modulation ratio, the remaining service life of the DC link capacitor is estimated using a life prediction algorithm.

[0033] Based on the remaining usage time evaluation results, a preventive maintenance strategy with current control parameters is generated through the maintenance decision optimization method.

[0034] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the wide-band impedance analysis method based on the network structure in the DC distribution network as described in the first aspect of the present invention is implemented.

[0035] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the wide-band impedance analysis method based on the network structure in the DC distribution network as described in the first aspect of the present invention is implemented.

[0036] The present invention achieves the following benefits: by synchronizing and aggregating data from each node through a distributed network protocol and a federated learning framework, a grid health report containing a regional impedance deviation index is generated. This process significantly enhances the interoperability of grid systems in different locations. This method not only accurately assesses the overall health of the grid but also dynamically generates a regional impedance deviation index, providing a key basis for subsequent maintenance decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 This is a flow chart of the broadband impedance analysis method based on the network structure in the DC distribution network in Example 1.

[0039] Figure 2 This is a diagram of the time-frequency domain impedance feature generation process in Example 1.

[0040] Figure 3 This is a flowchart of fault detection and decision-making in Example 1.

[0041] Figure 4 Generate a graph for the adaptive control and maintenance strategy in Example 1. DETAILED DESCRIPTION

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0043] Example 1, reference Figures 1 to 4This embodiment provides a broadband impedance analysis method based on the network structure in a DC distribution network, comprising the following steps:

[0044] S1, real-time acquisition of voltage and current data and preprocessing, while performing time-frequency analysis through short-time Fourier transform to generate time-frequency domain impedance characteristics;

[0045] Preprocessing includes eliminating power frequency harmonics and random noise through Gaussian filters and performing signal normalization;

[0046] It should be noted that the bandwidth of the Gaussian filter transfer function is set to a specific proportion of the fundamental frequency according to the power frequency harmonic characteristics. The original voltage signal and current signal are respectively subjected to discrete Fourier transform to obtain frequency domain representation, and frequency domain filtering is realized by performing point-by-point complex multiplication operation with the filter function in the frequency domain. Subsequently, random noise suppression processing is performed, and a fixed-length sliding window is used to calculate the energy ratio of the filtered signal to the original signal. When the ratio is lower than the preset judgment threshold, it is identified as a noise point and corrected using linear interpolation of adjacent sampling points. Finally, the signal normalization processing is completed by calculating the effective values of the voltage and current within the complete power frequency cycle, and dividing the denoised signal by its corresponding effective value to realize amplitude normalization. The entire preprocessing process ensures that the output signal meets the requirements of power frequency harmonic suppression, random noise elimination and amplitude normalization at the same time.

[0047] The pre-processed voltage and current data are decomposed into time and frequency using short-time Fourier transform to generate a time-frequency spectrum matrix;

[0048] It should be noted that the pre-processed voltage and current signals are windowed using the Hanning window function, and the window length is set to cover an integer multiple of the power frequency period. The windowed signal is segmented along the time axis at a fixed overlap rate, and a discrete Fourier transform is performed on the signal segment within each window to obtain the spectrum distribution of each time period. All window spectra of the voltage and current signals are arranged and combined in chronological order to form a voltage time-frequency spectrum matrix and a current time-frequency spectrum matrix, respectively. The row vectors of the time-frequency spectrum matrix represent the frequency component distribution, the column vectors represent the time-varying process, and the matrix elements contain the complex spectrum values at the corresponding time and frequency points. The voltage time-frequency spectrum matrix and the current time-frequency spectrum matrix have the same time resolution and frequency resolution to ensure the accuracy of the subsequent impedance characteristic conversion, and finally output the time-frequency spectrum matrix.

[0049] Based on the time-frequency spectrum matrix, the impedance characteristics are converted using Ohm's law in the frequency domain to generate the time-frequency domain impedance characteristics.

[0050] It should be noted that the complex spectrum values corresponding to the time points and frequency points in the time-spectrum matrix are extracted, and the element-by-element complex division operation is performed according to Ohm's law in the frequency domain to calculate the complex impedance value at each time and frequency point. The real part of the complex impedance value represents the resistance component, and the imaginary part represents the reactance component. The calculation results are rearranged and combined according to the time-frequency structure of the original time-spectrum matrix to form an impedance time-spectrum matrix. The row vectors of the impedance time-spectrum matrix represent the change characteristics of the impedance with frequency, and the column vectors represent the change process of the impedance with time. Each element in the impedance time-spectrum matrix contains the complex impedance value at the corresponding time and frequency point, which fully characterizes the impedance characteristics of the object under test in the time-frequency domain. The impedance time-spectrum matrix finally generated is the impedance characteristic in the time-frequency domain.

[0051] S2. Based on the impedance characteristics in the time-frequency domain, a lightweight convolutional neural network is used to extract the frequency-time domain joint features, and a multi-layer perceptron classifier is used to generate the fault detection results.

[0052] Based on the impedance characteristic matrix in the time-frequency domain, a multidimensional feature tensor with spatiotemporal correlation is generated through three-dimensional convolution kernel sliding calculation;

[0053] It should be noted that the size of the convolution kernel is defined along the three dimensions of the time-frequency domain impedance characteristic matrix, where the first dimension corresponds to the frequency component, the second dimension corresponds to the time series, and the third dimension corresponds to the spatial distribution characteristics. The three-dimensional convolution kernel is slid along the frequency dimension, time dimension, and space dimension of the time-frequency domain impedance characteristic matrix, and a point multiplication and accumulation operation is performed at each sliding position to calculate the spatiotemporal coupling characteristic response of the local time-frequency region. A fixed step size is maintained during the sliding process of the convolution kernel to ensure that the dimensions of the output spatiotemporal coupling feature tensor match. By performing parallel calculations on multiple groups of three-dimensional convolution kernels with different parameters, the spatiotemporal correlation characteristics of different modes in the time-frequency domain impedance characteristic matrix are extracted respectively. The output feature maps of each group of convolution kernels are stacked and combined according to the channel dimension to form a multidimensional feature tensor with spatiotemporal correlation.

[0054] Based on the multi-dimensional feature tensor, a frequency-time domain joint feature vector is generated through a cross-channel attention fusion mechanism;

[0055] It should be noted that the global average pooling value of each channel feature map is calculated along the channel dimension of the multidimensional feature tensor, which serves as the basis for the channel importance score. The pooled value of each channel is input into an attention network consisting of two fully connected layers to generate a channel attention weight vector. After normalization, the attention weight vector is multiplied with the original multidimensional feature tensor in a channel-weighted manner to highlight the sensitivity feature expression of the key channels. The weighted multidimensional feature tensor is globally average pooled along the spatial dimension, preserving the energy distribution characteristics of the frequency and time dimensions. The pooling result is flattened into a one-dimensional vector to form a joint feature vector that integrates the frequency and time domain characteristics.

[0056] Based on the frequency-domain and time-domain joint feature vector, the fault detection result is generated through a multi-layer perceptron classifier. The formula is as follows:

[0057]

[0058] where y is the fault detection score, σ is the Sigmoid activation function, β k is the contribution weight of the k-th sub-band, K is the total number of frequency band blocks, F (k) is the k-th time-frequency feature block, W k is the fault-sensitive filter for the k-th frequency band, k is the frequency band index, and b is the control score benchmark;

[0059] It should be noted that first, the frequency-domain and time-domain joint feature vector F is divided into k sub-feature blocks F(k) according to the frequency band. Each feature block is element-wise multiplied by the fault-sensitive filter W k for the corresponding frequency band to obtain the weighted feature block. The fault-sensitive filter W k is obtained by training with historical fault data and can highlight the characteristic response of the fault-sensitive frequency band. Then, the L2 norm is calculated for each weighted feature block to quantify the significance intensity of the frequency band feature. Next, the norm values of each frequency band are multiplied by the pre-trained contribution weight β k and accumulated, and the bias term b is added to obtain the linear combination value. Finally, the linear combination value is non-linearly mapped to the interval (0,1) through the Sigmoid function, and the fault detection score y is output. The closer the score is to 1, the higher the fault probability, and the closer it is to 0, the more normal the operating state. The entire calculation process effectively integrates multi-band feature information and obtains an intuitive probabilistic fault score through non-linear transformation.

[0060] Based on historical data, a fault threshold A is defined;

[0061] When y < A, it is considered that the device is operating normally, and the current monitoring mode is maintained;

[0062] When y ≥ A, it is considered that the device is in an abnormal state, and the operation and maintenance personnel need to be notified immediately for repair.

[0063] It should be noted that when the fault detection score y is less than threshold A, for example, a fault detection score of 0.35 is monitored and threshold A is set to 0.5, the amplitude fluctuations of the time-frequency domain impedance characteristic matrix in each frequency band are within the historical normal range, the values of each channel of the multidimensional feature tensor output by the lightweight convolutional neural network remain stable, and no control command adjustment or alarm notification is triggered. The DC distribution network equipment is determined to be in normal operation, and the current monitoring mode is maintained. When the fault detection score y is greater than or equal to threshold A, for example, a fault detection score of 0.65 is monitored and threshold A is set to 0.5, the exception handling process is immediately executed. At this time, the time-frequency domain impedance characteristic matrix experiences a sudden increase in amplitude in a specific frequency band, and the values of key channels of the multidimensional feature tensor output by the lightweight convolutional neural network fluctuate abnormally, and the DC distribution network equipment is determined to be in an abnormal operating state.

[0064] S3. Based on the fault detection score, the node detection results are synchronized through a distributed network protocol. The federated learning framework is used to perform weighted aggregation on the weight parameters of the frequency-time domain joint feature extraction to generate a grid health report including the regional impedance deviation index.

[0065] Through the distributed consistency protocol, the node detection data is time-aligned and quality-checked to generate a synchronized data set;

[0066] It should be explained that the distributed consensus algorithm based on the Byzantine fault tolerance mechanism aligns the timestamps of the fault detection results reported by each monitoring node to ensure that all node data has a unified time base; verifies the integrity of the node data through the Merkle tree structure, calculates the hash value of each data block and compares it with the root hash stored in the master node, and eliminates abnormal data that fails the hash verification; performs sliding window mean filtering on the data that passes the verification to eliminate the timing jitter caused by network delay; establishes a data quality weight matrix based on signal-to-noise ratio evaluation, and downgrades the detection results of nodes with low signal-to-noise ratios; finally, the multi-node data that has undergone time alignment and quality weighting are integrated in time series to form a synchronized data set with consistency and reliability.

[0067] Based on the synchronized data set, the federated learning framework aggregates the neural network parameter updates of each monitoring terminal and uses a weighted average algorithm to generate global shared parameters.

[0068] It should be noted that the lightweight convolutional neural network parameter update amount is calculated based on the local data in the synchronized data set, and the update amount is encrypted and transmitted to the aggregation server; the aggregation server calculates the aggregation weight according to the data quality and historical accuracy of each terminal, and uses the anomaly detection method based on Euclidean distance to eliminate abnormal parameter updates that deviate from the group distribution; the parameter update amounts that pass the verification are weighted and summed according to the corresponding weights to calculate the global parameter update amount; the global parameter update amount is added to the global shared parameters of the previous round to generate a new round of global shared parameters.

[0069] Based on global shared parameters, a grid health report is generated through impedance deviation analysis and health score mapping.

[0070] It should be noted that the lightweight convolutional neural network of each monitoring terminal is initialized with global shared parameters, and the time-frequency domain impedance characteristics in the synchronized data set are recalculated to obtain a standardized impedance characteristic distribution; by comparing the Euclidean distance between the impedance characteristics of each node and the global reference impedance, the regional impedance deviation index is calculated to quantify the impedance matching degree between the local power grid and the overall network; a health score mapping function based on Mahalanobis distance is established to project the multidimensional impedance characteristics into the scalar health score space, where a lower score indicates a healthier power grid state; the regional impedance deviation index and health score results are integrated to generate a structured power grid health report.

[0071] S4. Based on the grid health report, the support vector machine algorithm is used to perform feature space mapping and classification decisions on the time-frequency domain impedance characteristics and regional impedance deviation index to generate a broadband impedance map and graded warning information;

[0072] Based on the power grid health report, the time-frequency domain impedance characteristics and regional impedance deviation index are mapped to a high-dimensional space through kernel space mapping to generate a high-dimensional feature representation.

[0073] It should be noted that the time-frequency domain impedance characteristics and regional impedance deviation index in the power grid health report are nonlinearly transformed using the radial basis kernel function to generate the original feature vector, which is then projected into the reproducing kernel Hilbert space. For each element in the time-frequency domain impedance feature matrix, its distance from the center of the preset kernel function is calculated, and a high-dimensional representation is obtained through Gaussian kernel transformation; the regional impedance deviation index is transformed with a polynomial kernel function to enhance its separability in high-dimensional space. The transformed impedance characteristics and deviation index characteristics are spliced and combined in the kernel space to form a unified high-dimensional feature representation. This high-dimensional feature representation retains the spatiotemporal correlation characteristics of the original time-frequency domain impedance characteristics, while amplifying the discriminative characteristics of the regional impedance deviation through kernel techniques. The resulting high-dimensional feature representation has stronger linear separability than the original features and can be directly used in the support vector machine optimization algorithm to solve the optimal classification decision boundary.

[0074] Based on high-dimensional feature representation, the optimal classification decision boundary is generated through the support vector machine optimization algorithm;

[0075] It should be noted that the sample points in the high-dimensional feature space are mapped to the reproducing kernel Hilbert space using the kernel technique. Based on the principle of structural risk minimization, a separating hyperplane with maximum geometric margin is constructed. This optimization problem is formulated as a quadratic programming problem with slack variables, whose objective function simultaneously maximizes margin and minimizes classification error. A Gaussian kernel function is used to calculate the similarity matrix between the sample points, and the corresponding Lagrangian dual problem is established. This problem is then iteratively solved using a sequential minimum optimization algorithm. In each iteration, the sample pairs that most severely violate the KKT condition are selected to update the Lagrangian multipliers, while the regularization parameters are dynamically adjusted to balance model complexity and classification accuracy. The optimization process continues until all sample points satisfy the KKT condition. At this point, the decision boundary is uniquely determined by support vectors with nonzero Lagrangian multipliers. The sample points corresponding to these support vectors lie on the maximum margin boundary, and their high-dimensional feature representations contain key information for distinguishing different operating modes. The resulting nonlinear decision boundary effectively separates the normal operating condition feature clusters from the various failure mode feature clusters, constructing a classification region with maximum fault tolerance in the kernel space. The classifier is robust to abnormal points in the feature space and can adapt to the dynamic changes of the operating conditions of the DC distribution network.

[0076] Based on the classification decision boundary, the spectrum is reconstructed by clustering impedance features and locating abnormal areas to generate a wide-band impedance spectrum and graded warning information.

[0077] It should be noted that in the high-dimensional feature space determined by the classification decision boundary, a density clustering algorithm is used to automatically group the impedance feature points and identify feature clusters with similar impedance characteristics; based on the clustering results, the high-dimensional features are back-projected into the original time-frequency domain, and the impedance distribution pattern corresponding to each feature cluster is reconstructed to form a wide-band impedance spectrum reflecting the characteristics of different working conditions; at the same time, the relative distance from each cluster center to the classification decision boundary is calculated, the degree of anomaly is quantified, and the abnormal area is accurately located in combination with the regional impedance deviation index; the warning level is divided into three levels according to the anomaly degree threshold: the first-level warning corresponds to a significant abnormal cluster that directly crosses the classification decision boundary, the second-level warning corresponds to a potential abnormal cluster adjacent to the boundary, and the third-level warning is for a cluster to be observed whose impedance characteristics deviate from the typical distribution although it has not exceeded the limit; the final output wide-band impedance spectrum presents the impedance amplitude and phase characteristics of each frequency band in the form of a time-frequency matrix, and the graded warning information marks the abnormal frequency band range and recommended disposal measures, which together constitute a complete power grid status assessment result.

[0078] S5. Based on the wideband impedance spectrum and graded warning information, the converter PWM control parameters are adjusted in real time through the adaptive control algorithm, and the equipment maintenance strategy is generated in combination with the grid health report.

[0079] Based on the broadband impedance spectrum and graded warning information, the dominant resonant frequency band characteristics are generated through Hilbert spectrum analysis;

[0080] It should be noted that the Hilbert-Huang transform method is used to perform time-frequency analysis on the frequency band area marked as the warning state in the wide-band impedance spectrum. First, the impedance signal is decomposed into several intrinsic mode function components through empirical mode decomposition; each component is subjected to Hilbert transform to calculate the instantaneous frequency and amplitude, and the Hilbert energy spectrum is constructed; in the energy spectrum, the frequency bands whose energy concentration exceeds the preset threshold are identified as candidate resonant frequency bands; the quality factor and stability index of each candidate frequency band are calculated, and the dominant resonant frequency band that meets both the high Q value and low frequency fluctuation conditions is screened; the dominant resonant frequency band characteristics finally output include center frequency, resonance intensity, quality factor and attenuation coefficient.

[0081] Based on the dominant resonant frequency band characteristics, the PWM carrier frequency correction is generated through the PID regulator;

[0082] It should be noted that the center frequency of the dominant resonant frequency band characteristic is compared with the current PWM carrier frequency, and the frequency deviation is calculated as the input signal of the PID regulator; the PID regulator generates a carrier frequency adjustment amount according to the combined effect of the three links of proportion, integration and differentiation, among which the proportional link responds to the real-time frequency deviation, the integral link eliminates the steady-state error, and the differential link suppresses frequency oscillation; the regulator output generates a PWM carrier frequency correction amount after limiting processing, and the correction amount is superimposed on the basic carrier frequency to form a new carrier frequency setting value; the amplitude and direction of the correction amount are jointly determined by the strength index and stability index of the resonant frequency band characteristic, ensuring the stable operation of the converter while suppressing the resonance; the final output PWM carrier frequency correction amount directly acts on the converter control loop to achieve active avoidance of the resonant frequency point.

[0083] Based on the graded warning information, a dynamic limiting strategy is used to generate a safe adjustment range for the PWM modulation ratio;

[0084] It should be explained that based on the abnormality levels divided in the graded warning information and according to the limit threshold boundary, the first-level warning corresponds to the most stringent modulation ratio constraint range, the second-level warning adopts moderately relaxed boundary conditions, and the third-level warning maintains the standard operating limit; a constraint condition calculation model based on Lyapunov stability theory is established, and the upper limit of the maximum allowable modulation ratio is calculated in combination with the current DC bus voltage fluctuation amplitude and the AC side harmonic distortion rate; at the same time, considering the junction temperature change trend and heat dissipation conditions of the power device, the lower limit of the minimum modulation ratio is derived through the thermal stress algorithm; the upper and lower limit parameters are input into the hysteresis comparator for smoothing to eliminate boundary mutations; the final generated PWM modulation ratio safe adjustment range ensures that the converter always operates in a safe and stable area under various warning states.

[0085] Furthermore, when establishing a constraint calculation model based on Lyapunov stability theory, a state-space expression is first constructed based on the equivalent circuit equation of the DC distribution network. The converter output voltage and current are used as state variables, and an energy function is constructed using the Lyapunov function to describe the dynamic characteristics. Solving the Lyapunov matrix inequality in the time domain yields a sufficient condition for asymptotic stability, which is directly converted into a constraint on the upper bound of the PWM modulation ratio. Combining the current DC bus voltage fluctuation amplitude, the voltage deviation is substituted into the Lyapunov stability criterion to derive the upper bound of the maximum allowable modulation ratio required to maintain system stability. Furthermore, based on the AC side harmonic distortion rate indicator, a quadratic constraint correction is applied to the upper bound calculated using the predefined harmonic distortion rate-modulation ratio relationship curve. Ultimately, the upper bound parameter for the modulation ratio is output, satisfying both the Lyapunov stability theory and the harmonic distortion rate constraints.

[0086] Based on the PWM carrier frequency correction value and the safe adjustment range of the PWM modulation ratio, the remaining service life of the DC link capacitor is estimated using a life prediction algorithm.

[0087] It should be noted that the PWM carrier frequency correction is converted into harmonic current components through Fourier series expansion, and the effective value of the ripple current generated by each harmonic on the capacitor is calculated. The DC bus voltage fluctuation amplitude is determined based on the safe adjustment range of the PWM modulation ratio. Combined with real-time monitoring of capacitor surface temperature data, a three-dimensional stress space of current, voltage, and temperature is established. By continuously collecting parameters such as the capacitor's equivalent series resistance, capacitance, and loss tangent value, a parameter change curve is plotted over time. The ripple current, voltage fluctuation, and temperature parameter input curves under the current operating conditions are extrapolated to obtain two evaluation results: the lower limit of the remaining service life considering the maximum stress condition and the expected value of the remaining service life based on typical operating conditions.

[0088] Based on the remaining usage time evaluation results, a preventive maintenance strategy with current control parameters is generated through the maintenance decision optimization method.

[0089] It should be noted that the guaranteed predicted value and the reference predicted value in the remaining usage time assessment result are matched and processed according to the pre-set maintenance rules, and the corresponding maintenance recommendation is triggered when the predicted value is lower than a specific threshold. For the PWM carrier frequency correction amount, a carrier frequency adjustment plan is generated based on the correspondence between the ripple current stress level and the lifespan; for the PWM modulation ratio safety adjustment range, a modulation ratio limit optimization recommendation is formed based on the correlation characteristics between voltage stress and lifespan. Combined with the real-time monitoring trend of the capacitor's equivalent series resistance and capacitance, a preventive maintenance strategy including maintenance timing, priority and specific measures is generated according to the preset maintenance level standards. The final output maintenance strategy directly corresponds to the current control parameters, achieving the optimal configuration of maintenance resources while ensuring the reliable operation of the equipment.

[0090] This embodiment further provides a computer device applicable to a wideband impedance analysis method based on a network structure in a DC distribution network, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the wideband impedance analysis method based on a network structure in a DC distribution network as proposed in the above embodiment.

[0091] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0092] This embodiment also provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the program implements the wideband impedance analysis method based on the network structure in the DC distribution network proposed in the above embodiment. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0093] In summary, this invention utilizes a distributed network protocol and a federated learning framework to synchronize and aggregate data from each node, generating a grid health report containing a regional impedance deviation index. This process significantly enhances the interoperability of grid systems across different locations. This method not only accurately assesses the overall health of the grid but also dynamically generates a regional impedance deviation index, providing a key basis for subsequent maintenance decisions.

[0094] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A broadband impedance analysis method based on the network structure in a DC distribution network, characterized by: including, real-time collecting voltage and current data and performing preprocessing, and at the same time performing time-frequency analysis through short-time Fourier transform to generate time-frequency domain impedance characteristics; based on the time-frequency domain impedance characteristics, extracting frequency-domain and time-domain joint characteristics through a lightweight convolutional neural network, and generating a fault detection result through a multi-layer perceptron classifier; based on the fault detection score, synchronizing node detection results through a distributed network protocol, and using a federated learning framework to weighted aggregate the weight parameters of the frequency-domain and time-domain joint feature extraction to generate a power grid health report including the regional impedance deviation index; based on the power grid health report, performing feature space mapping and classification decision on the time-frequency domain impedance characteristics and the regional impedance deviation index through a support vector machine algorithm to generate a broadband impedance map and a hierarchical warning message; based on the broadband impedance map and the hierarchical warning message, adjusting the converter PWM control parameters in real time through an adaptive control algorithm, and generating a device maintenance strategy in combination with the power grid health report.

2. The broadband impedance analysis method based on the network structure in a DC distribution network according to claim 1, characterized in that: The preprocessing includes eliminating power frequency harmonics and random noise through a Gaussian filter and performing signal normalization processing.

3. The broadband impedance analysis method based on the network structure in a DC distribution network according to claim 2, characterized in that: The time-frequency analysis through short-time Fourier transform to generate time-frequency domain impedance characteristics is specifically as follows. Performing time-frequency decomposition on the preprocessed voltage and current data through short-time Fourier transform to generate a time-frequency spectrum matrix; Based on the time-frequency spectrum matrix, performing impedance characteristic conversion through the frequency-domain Ohm's law to generate time-frequency domain impedance characteristics.

4. The broadband impedance analysis method based on the network structure in a DC distribution network according to claim 3, characterized in that: The extracting frequency-domain and time-domain joint characteristics through a lightweight convolutional neural network based on the time-frequency domain impedance characteristics and generating a fault detection result through a multi-layer perceptron classifier is specifically as follows. Based on the time-frequency domain impedance characteristic matrix, generating a multi-dimensional feature tensor with spatio-temporal correlation through sliding calculation of a three-dimensional convolution kernel; Based on the multi-dimensional feature tensor, generating a frequency-domain and time-domain joint feature vector through a cross-channel attention fusion mechanism; Based on the frequency-domain and time-domain joint feature vector, generating a fault detection result y through a multi-layer perceptron classifier; Defining a fault threshold A based on historical data; When y < A, it is considered that the device is operating normally, and the current monitoring mode is maintained; When y ≥ A, it is considered that the device is in an abnormal state, and the operation and maintenance personnel need to be notified immediately for repair.

5. The broadband impedance analysis method based on the network structure in a DC distribution network according to claim 4, characterized in that: The synchronizing node data through a distributed protocol based on the fault detection result and using a federated learning framework to extract model parameters for the frequency-domain and time-domain joint features to generate a power grid health report including the regional impedance deviation index is specifically as follows. Through a distributed consistency protocol, performing time alignment and quality verification on the node detection data to generate a synchronized data set; Based on the synchronized data set, aggregating the neural network parameter update amounts of each monitoring terminal through a federated learning framework, and using a weighted average algorithm to generate globally shared parameters; Based on the globally shared parameters, generating a power grid health report through impedance deviation analysis and health score mapping.

6. The broadband impedance analysis method based on the network structure in a DC distribution network according to claim 5, characterized in that: The performing feature space mapping and classification decision on the time-frequency domain impedance characteristics and the regional impedance deviation index through a support vector machine algorithm based on the power grid health report to generate a broadband impedance map and a hierarchical warning message is specifically as follows. Based on the power grid health report, the time-frequency domain impedance characteristics and regional impedance deviation index are mapped to a high-dimensional space through kernel space mapping to generate a high-dimensional feature representation. Based on high-dimensional feature representation, the optimal classification decision boundary is generated through the support vector machine optimization algorithm; Based on the classification decision boundary, the spectrum is reconstructed by clustering impedance features and locating abnormal areas to generate a wide-band impedance spectrum and graded warning information.

7. The broadband impedance analysis method based on the network structure in a DC distribution network according to claim 6, characterized in that: Based on the broadband impedance spectrum and graded warning information, the PWM control parameters of the converter are adjusted in real time through the adaptive control algorithm. The specific steps are as follows Based on the broadband impedance spectrum and graded warning information, the dominant resonant frequency band characteristics are generated through Hilbert spectrum analysis; Based on the dominant resonant frequency band characteristics, the PWM carrier frequency correction is generated through the PID regulator; Based on the graded warning information, a safe adjustment range of the PWM modulation ratio is generated through a dynamic limiting strategy.

8. The broadband impedance analysis method based on the network structure in a DC distribution network according to claim 7, characterized in that: The specific steps of generating equipment maintenance strategy based on power grid health report are as follows: Based on the PWM carrier frequency correction value and the safe adjustment range of the PWM modulation ratio, the remaining service life of the DC link capacitor is estimated using a life prediction algorithm. Based on the remaining usage time evaluation results, a preventive maintenance strategy with current control parameters is generated through the maintenance decision optimization method.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the wide-band impedance analysis method based on the network structure in the DC distribution network according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the wide-band impedance analysis method based on the network structure in a DC distribution network according to any one of claims 1 to 8 are implemented.

Images