Fault positioning method and system of photovoltaic power grid system

Abstract

The embodiment of the invention provides a fault positioning method and system for a photovoltaic power grid system, and belongs to the technical field of smart power grids. The method comprises the steps that control parameters of all inverters in a photovoltaic power grid system are acquired, and a dynamic consistency matrix used for representing the operation difference of the inverters is constructed; power tracking parameters of all inverters are collected, and a power consistency matrix used for representing power response differences is constructed; fusing the dynamic consistency matrix and the power consistency matrix to generate a comprehensive compensation matrix, and performing synchronous correction on the operation signal of each inverter based on the comprehensive compensation matrix to form a synchronous signal set; and establishing an energy convergence degree model based on the synchronization signal set and determining the energy unbalance amount of each branch so as to locate the fault position in the photovoltaic power grid system. According to the scheme of the invention, the synchronous correction of the operation signals of the multi-inverter photovoltaic power grid and the precise recognition of the energy imbalance are realized, and the precision and stability of fault positioning are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and more specifically to a fault location method and system for a photovoltaic power grid system. Background Technology

[0002] In photovoltaic (PV) grid systems, with the continuous increase in the number and installed capacity of grid-connected inverters, parallel operation of multi-inverter clusters has become the mainstream structure for large-scale PV power plants. However, in such systems, each inverter is typically driven by an independent controller. Differences in clock synchronization accuracy, sampling period, communication delay, and phase-locking mechanisms among these controllers make it difficult to maintain strict consistency in time and phase between the inverter output signals. When a system fault or voltage disturbance occurs, these operational differences can be amplified in a short period, manifesting as asynchronous output signal responses from each inverter, waveform misalignment, and inconsistent power transients. This makes it difficult for traditional fault location algorithms that rely on voltage, current amplitude, or phase characteristics to accurately determine the fault location.

[0003] Most existing photovoltaic grid-connected fault detection and location methods are based on the assumption of synchronous inverter operation, using monitoring of current abrupt changes, traveling wave propagation time differences, or impedance changes to locate fault sections. However, in scenarios with multiple inverters in parallel, inconsistent control loop responses and communication link delays make it difficult to guarantee the timing alignment of electrical signals. This leads to false phase and amplitude differences between adjacent branches, causing the location results to deviate from the actual fault location. Furthermore, some methods attempt to compensate by introducing power change characteristics or inverter output waveform fitting parameters, but these still fail to eliminate the impact of differences in the operating states of each inverter at the system level. Especially when there are fluctuations in illumination conditions or MPPT (Maximum Power Point Tracking) tracking state switching, the inconsistency in power response further weakens the reliability of fault determination.

[0004] Therefore, the main problem with existing photovoltaic grid systems is the lack of a unified signal consistency modeling method that can handle the differences in operation of multiple inverters. This makes it impossible for the system to guarantee the synchronization of electrical signals in terms of time and power when inverter outputs are asynchronous, communication is delayed, or power response is asynchronous. This results in unstable fault feature extraction, decreased location accuracy, and in severe cases, even misjudgment of branches or missed fault detection. Summary of the Invention

[0005] The purpose of this invention is to provide a fault location method and system for a photovoltaic power grid system, so as to at least solve the problem of decreased fault location accuracy caused by asynchronous operation signals of multiple inverters.

[0006] To achieve the above objectives, the first aspect of the present invention provides a fault location method for a photovoltaic power grid system. The method includes: acquiring control parameters of each inverter in the photovoltaic power grid system and constructing a dynamic consistency matrix to characterize the differences in inverter operation; collecting power tracking parameters of each inverter and constructing a power consistency matrix to characterize the differences in power response; fusing the dynamic consistency matrix and the power consistency matrix to generate a comprehensive compensation matrix, and performing synchronization correction on the operating signals of each inverter based on the comprehensive compensation matrix to form a synchronization signal set; establishing an energy convergence model based on the synchronization signal set and determining the energy imbalance of each branch to locate the fault location in the photovoltaic power grid system.

[0007] Optionally, the control parameters of each inverter include any one or more of the following: controller clock frequency, sampling period, communication delay, and phase locking parameters; obtain the control parameters of each inverter in the photovoltaic grid system, and construct a dynamic consistency matrix to characterize the differences in inverter operation, including: extracting the corresponding time offset, phase drift angle, and transient response constant based on the collected control parameters of each inverter; combining the time offset, phase drift angle, and transient response constant of each inverter according to the inverter identification number to form a parameter vector, and combining them to construct the corresponding matrix unit; arranging each matrix unit according to the inverter cluster topology order to generate a dynamic consistency matrix.

[0008] Optionally, the acquisition rules for the power point tracking parameters of each inverter are as follows: During the operation cycle of the photovoltaic grid system, the output power, voltage signal and current signal are periodically acquired according to the maximum power point tracking (MPPT) sampling frequency set by the inverter controller, as a real-time dataset; based on the real-time dataset, the power change rate, voltage disturbance amplitude and response delay time are determined in each sampling cycle and combined into power point tracking parameters.

[0009] Optionally, the construction rules for the power consistency matrix used to characterize power response differences are as follows: using the power tracking parameters of each inverter as input, the mean and standard deviation of the power change rate, voltage disturbance amplitude, and response delay time of each inverter are determined to generate a power dynamic feature vector; the power dynamic feature vectors are combined in the order of inverter numbering to form matrix units, and the feature similarity coefficients are calculated between the matrix units; a weighted connection relationship is established between the matrix units based on the feature similarity coefficients to construct the power consistency matrix.

[0010] Optionally, the dynamic consistency matrix and the power consistency matrix are fused to generate a comprehensive compensation matrix, including: performing normalization and weighted balance calculations on the time offset and phase drift angle in the dynamic consistency matrix, and the power change rate and response delay time in the power consistency matrix, to obtain a unified dimension parameter set; extracting time correction coefficients and power correction coefficients based on the unified dimension parameter set, and writing the time correction coefficients and power correction coefficients into the corresponding fusion unit according to the inverter identifier; pairing and superimposing each fusion unit on the inverter identifier dimension according to the preset fusion weight to generate a comprehensive compensation matrix containing time correction coefficients and power correction coefficients.

[0011] Optionally, based on the comprehensive compensation matrix, the operating signals of each inverter are synchronized to form a synchronization signal set, including: applying the time correction coefficient and power correction coefficient in the comprehensive compensation matrix to the voltage signal, current signal and output power signal of each inverter respectively to obtain a synchronized operating signal sequence; resampling each operating signal sequence according to a unified reference time, and extracting the amplitude peak point and phase zero point from the resampling result to complete the amplitude and phase alignment; and combining all operating signal sequences that have undergone time and amplitude synchronization correction in the order of inverter identification to form a synchronization signal set.

[0012] Optionally, an energy convergence model is established based on the synchronization signal set to determine the energy imbalance of each branch in order to locate the fault location in the photovoltaic power grid system. This includes: calculating the instantaneous power of each inverter branch within a preset sampling window using voltage and current signals from the synchronization signal set, and performing an integral operation on the instantaneous power sequence to obtain the cumulative energy value of each branch; establishing an energy convergence model based on the cumulative energy value of each branch and the bus energy conservation constraint, and using the output of the energy convergence model to calculate the energy difference between adjacent branches to generate an energy convergence matrix; identifying branches in the energy convergence matrix whose energy imbalance exceeds a preset imbalance threshold, marking the corresponding branches as faulty branches, and determining the fault location of the photovoltaic power grid system based on the energy imbalance distribution of the faulty branches.

[0013] Optionally, an energy convergence model is established based on the cumulative energy values ​​of each branch and the energy conservation constraints of the bus. This includes: using the cumulative energy values ​​of each branch as input nodes and the total energy of the bus as constraint nodes, establishing an energy transmission mapping matrix according to the system topology; introducing the energy conservation conditions of the bus and the energy balance equations of the nodes into the energy transmission mapping matrix to construct the constraint expression of the energy convergence model; and solving the energy balance coefficient of each node by minimizing the sum of squared energy deviations of the branches to form an energy convergence model for quantifying the consistency of energy transmission.

[0014] A second aspect of the present invention provides a fault location system for a photovoltaic power grid system. The system includes: a first acquisition unit for acquiring control parameters of each inverter in the photovoltaic power grid system and constructing a dynamic consistency matrix to characterize the differences in inverter operation; a second acquisition unit for acquiring power tracking parameters of each inverter and constructing a power consistency matrix to characterize the differences in power response; a synchronization unit for fusing the dynamic consistency matrix and the power consistency matrix to generate a comprehensive compensation matrix, and performing synchronization correction on the operating signals of each inverter based on the comprehensive compensation matrix to form a synchronization signal set; and a fault location unit for establishing an energy convergence model based on the synchronization signal set and determining the energy imbalance of each branch to locate the fault location in the photovoltaic power grid system.

[0015] On the other hand, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described fault location method for a photovoltaic power grid system.

[0016] Through the above technical solution, this invention achieves synchronous correction of multi-inverter operating signals at the time, phase, and power levels by constructing a dynamic consistency matrix and a power consistency matrix and performing fusion compensation. This enables the system to maintain signal consistency even under conditions of controller asynchrony, communication delay, and power response differences. An energy convergence model is established based on the corrected synchronization signal set, accurately reflecting the energy transfer relationships and imbalance states between branches. Therefore, even in complex operating environments with light fluctuations or power disturbances, it can stably identify faulty branches and accurately locate fault positions. This method effectively improves the anti-interference capability and location accuracy of fault detection in photovoltaic power grid systems, enhancing the reliability and safety of system operation.

[0017] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the steps of a fault location method for a photovoltaic power grid system provided by one embodiment of the present invention; Figure 2 This is a schematic diagram of the power consistency matrix construction process provided by one embodiment of the present invention; Figure 3 This is a schematic diagram of the integrated compensation matrix fusion and signal synchronization correction process provided in one embodiment of the present invention; Figure 4This is a system structure diagram of a fault location system for a photovoltaic power grid system provided in one embodiment of the present invention; Figure 5 This is an internal structural diagram of a computer device provided in one embodiment of the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0020] like Figure 1 As shown, an embodiment of the present invention provides a fault location method for a photovoltaic power grid system, the method comprising: Step S10: Obtain the control parameters of each inverter in the photovoltaic grid system and construct a dynamic consistency matrix to characterize the differences in inverter operation.

[0021] Specifically, the control parameters of each inverter include any one or more of the following: controller clock frequency, sampling period, communication delay, and phase locking parameters. The control parameters of each inverter in the photovoltaic grid system are acquired, and a dynamic consistency matrix is ​​constructed to characterize the differences in inverter operation. This includes: extracting the corresponding time offset, phase drift angle, and transient response constant based on the acquired control parameters of each inverter; combining the time offset, phase drift angle, and transient response constant of each inverter according to the inverter identifier to form a parameter vector, and then combining them to construct corresponding matrix units; arranging the matrix units according to the inverter cluster topology to generate the dynamic consistency matrix.

[0022] In this embodiment of the invention, multi-inverter cluster operation has become the norm in photovoltaic grid-connected scenarios. However, due to inherent differences between inverter controllers, their output signals are often not perfectly synchronized. To accurately identify and correct these operational differences, it is necessary to first uniformly model the key parameters of the control layer. Therefore, by collecting the control parameters of each inverter, a dynamic consistency matrix capable of quantifying its dynamic difference characteristics is constructed. Control parameters typically include the controller's clock frequency, sampling period, communication delay, and phase-locking parameters. These quantities exhibit slight deviations under different devices, communication paths, or load conditions, and it is these deviations that cause misalignment of voltage and current signals in time and phase.

[0023] Specifically, by standardizing the control parameters of each inverter, clock frequency offset can be converted into equivalent time offset, phase lock error into phase drift angle, and transient response constant can be calculated based on the controller output response curve. These three parameters together describe the inverter's response characteristics to external disturbances at the control loop level. Subsequently, using the inverter number as an index, the time offset, phase drift angle, and transient response constant corresponding to each inverter are sequentially arranged into a parameter vector. Each parameter vector represents an independent inverter's operating characteristic unit, and these units are combined to construct the matrix units in the dynamic consistency matrix.

[0024] During matrix construction, the inverters must be arranged according to their physical connection order within the cluster topology to ensure that the matrix rows and columns correspond to the actual electrical connections. This allows the matrix to reflect not only the numerical differences in control parameters but also the coupling relationships between different inverters. Each element in the matrix can be considered as the synchronization correlation between inverters. The value of each element is calculated from the differences in the corresponding inverter control parameters (including time offset, phase drift angle, and transient response constant) and then standardized. The closer the value is to zero, the more consistent the control behavior; a larger deviation indicates significant control differences or communication delays. In this way, the dynamic consistency state of the entire inverter cluster can be quantified in matrix form.

[0025] This process transforms previously discrete and difficult-to-compare control parameters into a matrix structure that can intuitively represent overall operational differences, providing a unified mathematical basis for subsequent signal synchronization correction, phase compensation, and power response alignment. Through the dynamic consistency matrix, inconsistencies in the inverter control layer can be identified before fault analysis, providing accurate and quantifiable support for the stable operation and fault location of the photovoltaic grid.

[0026] Step S20: Collect the power tracking parameters of each inverter and construct a power consistency matrix to characterize the differences in power response.

[0027] Specifically, the acquisition rules for the power point tracking parameters of each inverter are as follows: During the operation cycle of the photovoltaic grid system, the output power, voltage signal and current signal are periodically acquired according to the maximum power point tracking (MPPT) sampling frequency set by the inverter controller, as a real-time dataset; based on the real-time dataset, the power change rate, voltage disturbance amplitude and response delay time are determined in each sampling cycle and combined into power point tracking parameters.

[0028] Furthermore, the construction rules for the power consistency matrix used to characterize power response differences are as follows: Using the collected power tracking parameters of each inverter as input, the mean and standard deviation of the power change rate, voltage disturbance amplitude, and response delay time of each inverter are determined to generate a power dynamic feature vector; these feature vectors are combined according to the inverter numbering order to form matrix units, and feature similarity coefficients are calculated between matrix units; based on the feature similarity coefficients, weighted connection relationships are established between matrix units to construct the power consistency matrix. Specifically, such as... Figure 2 This includes the following steps.

[0029] Step S201: Collect the output power, voltage and current signals of each inverter, and calculate the power change rate, voltage disturbance amplitude and response delay time to form power tracking parameters.

[0030] In this embodiment of the invention, in a grid-connected photovoltaic environment with multiple inverters operating in parallel, although the inverters share the same bus, their power regulation behaviors are often not entirely consistent. Especially in the MPPT (Maximum Power Point Tracking) stage, due to differences in controller parameter settings, sampling rates, and internal filtering stages, the dynamic responses of the output power curves of each inverter differ significantly when facing the same changes in illumination or temperature. These differences manifest in the rate of power increase or decrease, the smoothness of the output waveform, and the degree of response lag—characteristics that are precisely the key factors affecting the accuracy of fault diagnosis. To accurately quantify these power response differences, it is necessary to uniformly collect and model the power, voltage, and current signals output by each inverter during the MPPT control process, thereby forming a power consistency matrix that can describe the differences in power dynamic characteristics.

[0031] Specifically, during the photovoltaic grid-connected operation cycle, based on the sampling frequency set by the inverter's own MPPT control circuit, the output power P(t), output voltage U(t), and output current I(t) of each inverter are periodically collected to obtain continuous real-time data sets. MPPT control typically uses a sampling period T... s Power perturbation and tracking updates are performed at intervals of 10 ms or 20 ms, thus obtaining a representative set of power change data within each sampling period. For the acquired data, the power change rate dP / dt, voltage perturbation amplitude ΔU, and power response delay time are calculated within each sampling period. These three elements constitute the power tracking parameter unit for this cycle, which can be represented in vector form as follows:

[0032] Where i represents the sampling period number. Let i be the power point tracking parameter vector of the i-th inverter within a certain sampling period. The power change rate of the i-th inverter Let be the voltage disturbance amplitude of the i-th inverter during this cycle. Let be the response delay time constant of the i-th inverter. By concatenating data from multiple consecutive sampling periods, a power point tracking parameter time series can be formed. This series reflects the dynamic response process of the inverter under different operating conditions, providing a data foundation for subsequent matrix construction.

[0033] Step S202: Perform statistical analysis on the power tracking parameters within the specified time window, extract the mean and standard deviation, and construct the power dynamic feature vector.

[0034] To avoid randomness caused by data fluctuations, statistical processing needs to be performed on the power point tracking parameters of each inverter. Specifically, the mean μ and standard deviation σ of the power change rate, voltage disturbance amplitude, and response delay time for each inverter within a specified time window are calculated to describe the central tendency and fluctuation characteristics of its dynamic behavior. This generates a power dynamic characteristic vector for each inverter.

[0035] Where k is the inverter number. Let K be the power dynamic characteristic vector of the k-th inverter. This represents the average power change rate of the k-th inverter; This represents the average amplitude of the voltage disturbance of the k-th inverter. This represents the average response delay time of the k-th inverter; Let be the standard deviation of the power change rate of the kth inverter; Let be the standard deviation of the voltage disturbance amplitude of the k-th inverter; This represents the standard deviation of the response delay time of the k-th inverter. This eigenvector preserves both the mean and dispersion of each dynamic parameter, allowing for direct quantitative comparison of response differences between different inverters.

[0036] Step S203: Calculate the similarity between the power response characteristics of different inverters based on cosine similarity or Euclidean distance, and establish a weighted connection relationship to form a power consistency matrix.

[0037] After obtaining the power dynamic characteristic vectors of all inverters, these vectors are combined in order of inverter numbering to form power response characteristic matrix units. Next, characteristic similarity coefficients are calculated between the matrix units. Cosine similarity or Euclidean distance can be used to measure the similarity of the power responses between different inverters. For example, the formula for calculating cosine similarity can be expressed as:

[0038] in This indicates the similarity of the power response between inverter m and inverter n. The closer the value is to 1, the more consistent the dynamic response of the two is, and the closer it is to 0, the more obvious the difference is. and These are the power dynamic feature vectors of the m-th and n-th inverters, respectively. Based on the similarity results, weighted connections can be established between matrix units to form a power consistency matrix. The elements of the matrix... It not only represents the power response correlation between inverters, but can also be used to determine the synchronization level of the entire cluster. If the average similarity outside the diagonal of the matrix is ​​low, it indicates that there are inverters with abnormal power responses in the cluster, and their output characteristics may be affected by hardware aging or abnormal MPPT control.

[0039] The constructed power consistency matrix provides a quantitative basis for subsequent integrated compensation and synchronization correction. This matrix clearly identifies the sources of difference among inverters during power point tracking, which can then be used to allocate correction weights for the power signal in the fusion calculation. In other words, this matrix not only reflects the degree of dynamic power consistency but also provides the basic data for correction coefficients in subsequent time and amplitude alignment.

[0040] This process enables effective modeling and quantification of power response differences between inverters, overcoming the problem of traditional methods failing to accurately measure the dynamic inconsistency of power among multiple inverters. Even under complex lighting fluctuations or partial shading conditions, this matrix can still identify inverters with abnormal responses, providing accurate and stable data support for subsequent fault location and improving overall location accuracy and anti-interference capabilities.

[0041] In another possible implementation, the construction of the power consistency matrix does not directly rely on the traditional MPPT power, voltage, and current signals. Instead, it introduces the power disturbance energy spectrum feature method to extract the power response differences of the inverters. This method treats the micro-amplitude power fluctuations generated during MPPT control as a frequency domain signal and analyzes the disturbance energy distribution of each inverter's power signal through short-time Fourier transform (STFT) or discrete wavelet transform (DWT). Specifically, using the continuously sampled output power sequence P(t) as the input signal, its power energy spectral density E(f) within a specific time window Δt is calculated, and three indices—the dominant frequency amplitude, the spectral centroid frequency, and the energy distribution entropy—are extracted to form a power frequency domain feature vector.

[0042]

[0043] in Let be the power frequency domain eigenvector of the k-th inverter. This indicates the magnitude of the power disturbance at the dominant frequency. As the center of gravity of the energy spectrum, This represents the energy distribution entropy. The above indicators effectively reflect the power regulation sensitivity and stability of MPPT control within the disturbance period, and are not directly affected by short-term solar radiation fluctuations or temperature changes.

[0044] When constructing the power consistency matrix, the power frequency domain eigenvectors are used instead of the time-domain power tracking parameters, and a weighted connection matrix is ​​established by calculating the spectral feature similarity between different inverters. Compared with traditional time-domain power change rate or response delay parameters, this frequency domain method can capture the regulation inertia and disturbance damping characteristics inside the MPPT loop, and can maintain high recognition, especially in high-frequency disturbance or low-light operation scenarios.

[0045] This power consistency modeling method based on the energy spectrum characteristics of power disturbance is a frequency domain extension expression of the inverter's dynamic characteristics. It can significantly improve the sensitivity to differences in nonlinear response, providing a new technical path for synchronization correction and fault identification under complex operating conditions, and also opening up a more refined monitoring dimension for the stability diagnosis of photovoltaic power grids.

[0046] Step S30: The dynamic consistency matrix and the power consistency matrix are fused to generate a comprehensive compensation matrix, and the operating signals of each inverter are synchronized based on the comprehensive compensation matrix to form a synchronization signal set.

[0047] Specifically, the dynamic consistency matrix and the power consistency matrix are fused to generate a comprehensive compensation matrix, including: normalizing and weighting the time offset and phase drift angle in the dynamic consistency matrix, and the power change rate and response delay time in the power consistency matrix to obtain a unified dimension parameter set; extracting time correction coefficients and power correction coefficients based on the unified dimension parameter set, and writing the time correction coefficients and power correction coefficients into the corresponding fusion unit according to the inverter identifier; pairing and superimposing each fusion unit on the inverter identifier dimension according to the preset fusion weight to generate a comprehensive compensation matrix containing time correction coefficients and power correction coefficients.

[0048] Optionally, based on the comprehensive compensation matrix, synchronization correction is performed on the operating signals of each inverter to form a synchronization signal set. This includes: applying the time correction coefficient and power correction coefficient in the comprehensive compensation matrix to the voltage signal, current signal, and output power signal of each inverter, respectively, to obtain a synchronized operating signal sequence; resampling each operating signal sequence along the time axis according to a unified reference time, and extracting the amplitude peak and phase zero point from the resampling result to complete amplitude and phase alignment; and combining all operating signal sequences that have undergone time and amplitude synchronization correction according to the inverter identification order to form a synchronization signal set. Specifically, as shown... Figure 3 This includes the following steps.

[0049] Step S301: Extract the key parameters of the dynamic and power consistency matrix, perform normalization and weighted calculation on the time offset, phase drift angle, power change rate and response delay time, and generate a set of unified dimensional parameters that can be fused.

[0050] In this embodiment of the invention, in a photovoltaic grid-connected structure with multiple inverters operating in parallel, the output signals of each inverter are often not synchronized in terms of time and power due to differences in control parameters, sampling delay, communication links, and power regulation characteristics. Simply put, this difference is like each inverter having its own independent rhythm; some react quickly, some slowly, while the grid side requires all outputs to maintain strict time, phase, and power consistency. Once this synchronization is disrupted, subsequent fault feature extraction and location calculations are prone to drift or errors. To solve this type of problem, a unified compensation mechanism that can take into account both control dynamic characteristics and power response characteristics is needed. Therefore, by fusing the aforementioned dynamic consistency matrix and power consistency matrix, a comprehensive compensation matrix with adaptive correction capabilities in the time, phase, and power dimensions is generated. This matrix is ​​then used to synchronize and correct the operating signals, forming a set of synchronization signals that can be used for fault location analysis.

[0051] Specifically, the dynamic consistency matrix provides deviation information at the time domain level, mainly reflecting the differences in time synchronization and control delay among various inverters, such as time offset Δt and phase drift angle. and transient response constant The power consistency matrix, on the other hand, characterizes the dynamic inconsistencies at the power level, primarily composed of the power change rate (dP / dt) and power response delay. The power perturbation amplitude ΔP is used to characterize the data. Since the two types of matrices have different physical dimensions—one primarily based on time and angle, and the other on power and its rate of change—direct matrix operations would lead to inconsistencies in dimensions and imbalanced weights. Therefore, normalization and weighted balancing calculations are first performed on the core parameters of both types of matrices to ensure comparability of different types of data on a unified dimensional scale. The specific approach is as follows:

[0052] Where X is the original parameter value, X n For the normalized result, X min With X max These represent the minimum and maximum values ​​of the parameters, respectively. The normalized time offset, phase drift angle, power change rate, and response delay time are all mapped to the [0,1] interval.

[0053] Step S302: Perform a weighted average based on a unified dimensional parameter set to calculate the time correction coefficient and power correction coefficient for subsequent signal correction.

[0054] Next, based on the unified dimensional parameter set, the time correction coefficient k is determined by weighted averaging. t With power correction factor k p The weighting can be adaptively adjusted according to actual operating conditions. For example, the weight of the time correction coefficient can be increased when communication latency is dominant, while the influence of the power correction coefficient can be increased in scenarios with significant light fluctuations. The following weighting method can be used:

[0055]

[0056] Among them, w t w φ w P w r These are the preset weighting coefficients; This is the average of the time offset within the specified sliding window; This represents the average phase drift angle within a specified sliding window. This is the time derivative of the power mean curve within a specified sliding window, used to reflect the average rate of dynamic change in power. This is the window mean of the response delay time constant. This yields k. t With k p It can dynamically reflect the combined effect of time synchronization error and power response inconsistency.

[0057] Step S303: Write the time correction coefficient and power correction coefficient into the matrix fusion unit, and generate a comprehensive compensation matrix by pairing and superimposing them according to the inverter topology order.

[0058] k t With k p The data are written into the corresponding fusion units of the comprehensive compensation matrix to form matrix units M. i,j = [k t (i,j), k p [i,j)] represents the synchronization compensation relationship between inverter i and inverter j. The matrix must be constructed strictly according to the topological order of the inverter cluster, ensuring that the rows and columns of the matrix correspond to the actual electrical connection locations, thus guaranteeing that the directionality of the compensation is consistent with the physical association. Subsequently, based on the preset fusion weight α, the dynamic consistency matrix and the power consistency matrix are paired and superimposed along the inverter identification dimension:

[0059] Where A is the dynamic consistency matrix, B is the power consistency matrix, and C is the generated comprehensive compensation matrix. By adjusting α, the balance between time and power compensation can be flexibly controlled, ensuring the correction effect under different operating conditions.

[0060] Step S304: Based on the time correction coefficient and power correction coefficient of the integrated compensation matrix, perform synchronous correction on the voltage, current and power signals of each inverter.

[0061] The generated comprehensive compensation matrix is ​​not merely a numerical fusion result; it physically represents the synchronization mapping relationship between multidimensional signals. Next, the time correction coefficient k in the comprehensive compensation matrix will be... t With power correction factor k p The operating signals applied to each inverter, including voltage signal U(t), current signal I(t), and output power P(t), achieve dual synchronous correction in both the time and power domains. The time correction process can be understood as realigning the sampling times, i.e., using k... t Fine-tuning the sampling time series:

[0062] in, The original sampling time point represents the time marker of each inverter signal (voltage, current, or power) during the sampling process; This refers to the corrected unified time point, i.e., the time coordinates of all inverter signals at a unified reference time. Power correction corresponds to the dynamic balance adjustment of power amplitude.

[0063] Among them, P ref For reference power, This is the corrected power signal. The original output power signal after time correction represents the inverter's synchronization time. The actual output power at any given time.

[0064] Step S305: Perform unified time base resampling and amplitude alignment on the operating signal after time and power synchronization correction to form a synchronization signal set.

[0065] The corrected operating signals still need to be resampled at a unified reference time to ensure that the signals from different inverters are comparable at the same time point. During resampling, fixed-step interpolation or phase-aligned interpolation can be used to extract the amplitude peaks and phase zeros within each sampling window, achieving amplitude and phase alignment of the waveform. The aligned inverter signal sequences are then recombined in numerical order to form a synchronization signal set. The synchronization signal set can be considered as a signal library after unified time and power compensation, where the outputs of all inverters are calibrated under the same time base, providing highly consistent input for subsequent energy convergence model calculations and fault location.

[0066] Through the aforementioned fusion and correction process, cross-domain unification of information between the control layer and the power layer is achieved. The timing characteristics provided by the dynamic consistency matrix and the dynamic response characteristics provided by the power consistency matrix are fused in the comprehensive compensation matrix, ensuring that the corrected signal remains consistent across the three dimensions of time phase, power amplitude, and response delay. This implementation overcomes the signal misalignment problem caused by asynchronous operation of multiple inverters, significantly reducing the alignment errors of voltage, current, and power data. Furthermore, it effectively suppresses spurious fluctuations caused by inconsistent power responses, improving the stability of subsequent energy analysis. Finally, it provides stable, highly consistent data input for the fault location algorithm, enabling the energy convergence model to accurately identify energy imbalance sections, fundamentally improving the accuracy and reliability of photovoltaic grid-connected fault detection and location.

[0067] In another possible implementation, the generation of the comprehensive compensation matrix is ​​not limited to the direct weighted fusion of the dynamic consistency matrix and the power consistency matrix, but introduces a nonlinear fusion strategy based on phase-energy coupling. The core idea of ​​this approach is to extract the coupling energy characteristics between the time phase and the power response, making the fusion result more adaptive and thus maintaining high synchronization accuracy under complex operating conditions.

[0068] Specifically, after calculating the dynamic consistency matrix A and the power consistency matrix B, the coupling phase relationship between the control layer and the power layer for each inverter is first extracted. The time drift signal of inverter i is defined as Δt. i The corresponding power disturbance signal is ΔP i The analytic signal Z is obtained by performing a Hilbert transform on both. i (t), and further calculate the phase coupling energy E i :

[0069] Where t0 and t1 are the start and end times of the integration, respectively, and the time window for energy calculation is defined, where the energy E is... i This reflects the coupling strength between the control signal and the power signal of the inverter over a period of time; a larger value indicates a more consistent dynamic response between the control layer and the power layer. Subsequently, E... i Assuming weights, construct an energy weight vector W = [E1, E2, …, E n ], and perform normalization processing on it to make ΣW i = 1.

[0070] During the integration process, a fixed weighting coefficient α is no longer used; instead, the integration ratio is dynamically adjusted based on the energy weight of each inverter. The specific calculation formula is as follows:

[0071] in The elements of the comprehensive compensation matrix represent the coupling correction relationship between inverters i and j. These are elements of the dynamic consistency matrix, reflecting the deviations between inverters i and j in terms of time, phase, and control delay. These are elements of the power consistency matrix, describing the differences between inverters i and j in terms of power dynamic response, such as power change rate, response delay, and disturbance magnitude. Due to the energy weight W... i Calculated under real-time operating conditions, this fusion method can adaptively reflect the differences in the operating states of different inverters. When a certain inverter experiences a significant increase in phase drift due to communication delay or MPPT disturbance, its energy weight will decrease accordingly, thereby automatically reducing its time dimension weight in matrix fusion and enhancing overall power consistency.

[0072] Furthermore, a phase energy constraint term can be introduced when performing synchronization correction. That is, in the time correction coefficient k t and power correction factor k p Based on this, a phase energy correction term is added to give the correction function self-adjusting capability:

[0073] in To correct for the phase difference before and after, k e According to E i The changes are updated in real time. To represent the corrected synchronization signal, The time correction term is used to achieve sampling alignment through time shifting, thereby eliminating time-domain asynchrony caused by controller clock or communication delays. This is a power correction term that achieves dynamic balance of power amplitude. When the inverter output power is lower than the reference power, this term is positive and automatically increases the correction signal amplitude; otherwise, it decreases the output, thereby achieving power coordination among multiple inverters. This is a phase energy correction term, used to introduce phase energy constraints based on dual power and time synchronization, enabling the correction function to have self-adjusting capabilities. e This is the phase energy correction coefficient. This method can maintain a stable transition of the signal waveform in scenarios with severe phase drift or complex power fluctuations, avoiding the over-correction or under-correction problems commonly found in traditional linear compensation.

[0074] In this embodiment, the compensation matrix is ​​transformed from static fusion to dynamic energy-driven nonlinear fusion, enabling time and power features to be self-weighted according to the real-time operating status. The phase energy constraint mechanism improves the robustness of the signal correction process, and can still maintain the phase consistency of the synchronization signal under strong disturbances or weak light conditions.

[0075] Step S40: Based on the synchronization signal set, establish an energy convergence model and determine the energy imbalance of each branch to locate the fault location in the photovoltaic power grid system.

[0076] Specifically, the instantaneous power of each inverter branch within a preset sampling window is calculated using voltage and current signals from the synchronization signal set, and an integral operation is performed on the instantaneous power sequence to obtain the cumulative energy value of each branch; an energy convergence model is established based on the cumulative energy value of each branch and the energy conservation constraint of the bus, and the energy convergence matrix is ​​generated by calculating the energy difference between adjacent branches using the output of the energy convergence model; branches whose energy imbalance exceeds a preset imbalance threshold are identified in the energy convergence matrix, the corresponding branches are marked as faulty branches, and the fault location of the photovoltaic grid system is determined based on the energy imbalance distribution of the faulty branches.

[0077] Optionally, an energy convergence model is established based on the cumulative energy values ​​of each branch and the energy conservation constraints of the bus. This includes: using the cumulative energy values ​​of each branch as input nodes and the total energy of the bus as constraint nodes, establishing an energy transmission mapping matrix according to the system topology; introducing the energy conservation conditions of the bus and the energy balance equations of the nodes into the energy transmission mapping matrix to construct the constraint expression of the energy convergence model; and solving the energy balance coefficient of each node by minimizing the sum of squared energy deviations of the branches to form an energy convergence model for quantifying the consistency of energy transmission.

[0078] In this embodiment of the invention, in a photovoltaic grid structure with multiple inverters operating in parallel, a common problem arises: seemingly all inverters are outputting power normally, but subtle energy transmission anomalies exist in some branches, such as changes in local cable contact resistance, current deviation in combiner box branches, or inverter response lag. These minor issues are easily averaged out and masked by traditional monitoring logic. By the time the fault truly affects the overall power balance of the grid, it often exceeds the controllable range. To identify such energy anomalies at an early stage, a more refined energy consistency discrimination model, namely the energy convergence model, is needed. The core idea of ​​this model is to treat the entire photovoltaic cluster as an energy transmission network. By uniformly modeling the energy output of each inverter branch and the energy convergence state of the bus, the degree of energy imbalance between different branches is quantified, thereby achieving precise location of the faulty branch.

[0079] Specifically, firstly, using the voltage signal U(t) and current signal I(t) from the synchronization signal set, the instantaneous power P(t) of each inverter branch within the preset sampling window Δt is calculated. The calculation method is as follows:

[0080] Here, Δt is typically taken to be from tens to hundreds of milliseconds, balancing sampling resolution and computational stability. The resulting instantaneous power sequence P(t) reflects the continuous output characteristics of the branch energy in the time domain. Subsequently, an integral operation is performed on this sequence to calculate the cumulative energy E of the branch within the sampling window. i :

[0081] Where t0 is the starting time of the current energy integration. E represents the instantaneous output power signal of node i at time t, which can be the inverter output or the bus power component; i The physical meaning of is the total amount of energy transmitted from inverter i to the bus during that time period, which has an intuitive comparative significance.

[0082] After collecting the energy accumulation values ​​of all inverters, the energy accumulation values ​​of each branch are used as input nodes, and the bus energy E is used as the input value. bus As constraint nodes, an energy transmission mapping matrix T is established based on the electrical topology of the photovoltaic power station. The rows of matrix T represent the output nodes of each inverter, and the columns correspond to bus nodes or intermediate combiner nodes, reflecting the energy transmission path and proportion. Subsequently, energy conservation conditions and node energy balance equations are introduced into this mapping matrix to form energy constraint relationships:

[0083] Where ε is the energy loss term, representing the energy transmission deviation between the branch and the bus. Let be the energy transfer coefficient from inverter node i to the common bus. Let be the cumulative energy value of inverter node i within the sampling window, and n be the number of inverters participating in grid connection. Based on this expression, the constraint function of the energy convergence model can be further constructed, with the objective of minimizing the sum of squared branch energy deviations, and the energy balance coefficient λ can be solved. i :

[0084] Where E ref For theoretical equilibrium energy, λ i This reflects the relative balance of energy transfer in the branch circuit. λ is obtained through optimization. i These constitute the core parameters of the energy convergence model, used to characterize the consistency of energy transfer between different branches.

[0085] After the energy convergence model is solved, the output balance coefficients are used to calculate the energy difference ΔE between adjacent branches. i,j Generate the energy convergence matrix C E Each element C of the matrix E(i,j) represents the degree of energy convergence deviation between branch i and branch j, that is, the amount of energy imbalance between branch i and branch j. The smaller the value, the more coordinated the energy transmission; the larger the value, the more obvious the energy imbalance. To facilitate subsequent judgment, the matrix can be normalized to map all deviation values ​​to the interval [0,1], which is convenient for cross-branch comparison.

[0086] In the energy convergence matrix, a preset imbalance threshold δ is set. E Filter the elements of each matrix. When C E (i,j) exceeds δ E If this occurs, it is determined that there is energy non-convergence between the branches. Further, by combining the topology mapping relationship, the concentrated area of ​​the unbalanced node distribution is tracked to determine the location of the corresponding faulty branch. For example, when multiple elements corresponding to a row in the matrix are significantly higher than the threshold, it indicates that the branch corresponding to the inverter has abnormal energy dissipation or reverse energy transmission during energy transfer, and can be directly marked as a faulty branch.

[0087] The advantage of this model lies in its departure from traditional reliance on current surges, waveform zero-crossing points, or harmonic energy characteristics. Instead, it performs a balance analysis from the perspective of energy transmission. Because energy parameters are relatively smooth under short-term fluctuations, it exhibits strong robustness to non-ideal factors such as measurement noise and communication delays. Furthermore, by introducing bus energy conservation constraints, overall energy deviations can be dynamically corrected, ensuring that the judgment results remain consistent with the total output of the power plant.

[0088] In practical applications, to prevent misjudgments caused by minor fluctuations, a moving average filtering process can be introduced after the energy difference calculation. That is, the C values ​​from multiple consecutive sampling windows are filtered. E The matrix performs time-sliding statistics, retaining only energy imbalance characteristics that persist for more than a preset period, while short-term disturbances are automatically smoothed out. This detailed design further improves the stability of fault diagnosis.

[0089] This process achieves a complete closed loop from voltage and current signals to energy difference quantification and fault spatial location. Compared with traditional methods relying on single electrical quantities for judgment, the energy convergence model can detect slight but persistent energy imbalances at an early stage, making it particularly suitable for complex grid-connected scenarios where inverter power output exhibits dynamic differences. Through matrix modeling and differential judgment, false detections caused by local fluctuations can be effectively avoided, improving the spatial resolution of fault location. The final output fault location not only provides the abnormal branch number but also reflects the directionality of the energy imbalance, providing a reliable basis for subsequent maintenance or control strategies.

[0090] Overall, the introduction of the energy convergence model transforms fault identification in photovoltaic power grids from traditional signal mutation detection to multidimensional analysis based on energy transmission consistency. It takes into account time, space, and energy constraints, and uses a mathematical model to dynamically quantify the internal operational consistency of the power grid, thereby significantly improving the fault identification accuracy and operational stability of multi-inverter photovoltaic power plants.

[0091] like Figure 4 As shown, this invention provides a fault location system for a photovoltaic power grid system. The system includes: a first acquisition unit for acquiring control parameters of each inverter in the photovoltaic power grid system and constructing a dynamic consistency matrix to characterize the differences in inverter operation; a second acquisition unit for acquiring power tracking parameters of each inverter and constructing a power consistency matrix to characterize the differences in power response; a synchronization unit for fusing the dynamic consistency matrix and the power consistency matrix to generate a comprehensive compensation matrix, and performing synchronization correction on the operating signals of each inverter based on the comprehensive compensation matrix to form a synchronization signal set; and a fault location unit for establishing an energy convergence model based on the synchronization signal set and determining the energy imbalance of each branch to locate the fault location in the photovoltaic power grid system.

[0092] The present invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described fault location method for a photovoltaic power grid system.

[0093] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor A01, a network interface A02, a memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A04. The network interface A02 is used for communication with external terminals via a network connection. When the computer program B02 is executed by the processor A01, it implements a fault location method for a photovoltaic power grid system.

[0094] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0095] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details described above. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe the various possible combinations.

[0096] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the embodiments of the present invention, they should also be regarded as the content disclosed by the embodiments of the present invention.

Claims

1. A method for fault location in a photovoltaic grid system, characterized by, The method comprises: acquiring control parameters of each inverter in a photovoltaic power grid system, and constructing a dynamic consistency matrix for representing operation differences of the inverters; collecting power tracking parameters of each inverter, and constructing a power consistency matrix for representing power response differences; fusing the dynamic consistency matrix and the power consistency matrix to generate a comprehensive compensation matrix, and performing synchronous correction on operation signals of each inverter based on the comprehensive compensation matrix to form a synchronous signal set; establishing an energy convergence degree model based on the synchronous signal set and determining energy imbalance amounts of each branch to locate a fault position in the photovoltaic power grid system.

2. The method of claim 1, wherein, The control parameters of each inverter include: any one or more of a controller clock frequency, a sampling period, a communication delay, and a phase locking parameter; acquiring control parameters of each inverter in a photovoltaic power grid system, and constructing a dynamic consistency matrix for representing operation differences of the inverters, comprising: based on the collected control parameters of each inverter, respectively extracting corresponding time offsets, phase drift angles, and transient response constants; combining the time offsets, phase drift angles, and transient response constants of each inverter according to inverter identification numbers to form parameter vectors, and combining to construct corresponding matrix units; arranging the matrix units in a topological order of inverter clusters to generate the dynamic consistency matrix.

3. The method of claim 1, wherein, The collection rule of the power tracking parameters of each inverter is: in the operation period of the photovoltaic power grid system, periodically collecting output power, voltage signals, and current signals according to a maximum power point tracking (MPPT) sampling frequency set by the inverter controller as a real-time data set; based on the real-time data set, determining a power change rate, a voltage disturbance amplitude, and a response delay time in each sampling period, and combining them as power tracking parameters.

4. The method of claim 3, wherein, The construction rule of the power consistency matrix for representing power response differences is: taking the collected power tracking parameters of each inverter as input, respectively determining the mean and standard deviation of the power change rate, the voltage disturbance amplitude, and the response delay time of each inverter to generate power dynamic feature vectors; combining the power dynamic feature vectors in the order of inverter numbers to form matrix units, and calculating feature similarity coefficients between the matrix units; establishing a weighted connection relationship between the matrix units according to the feature similarity coefficients to construct the power consistency matrix.

5. The method of claim 1, wherein, Fusing the dynamic consistency matrix and the power consistency matrix to generate a comprehensive compensation matrix comprises: performing normalization and weighted balance calculation on the time offsets and phase drift angles in the dynamic consistency matrix, and the power change rate and response delay time in the power consistency matrix to obtain a unified dimension parameter set; extracting time correction coefficients and power correction coefficients from the unified dimension parameter set, and writing the time correction coefficients and power correction coefficients into corresponding fusion units according to inverter identification; pairing and superimposing each fusion unit in the inverter identification dimension according to a preset fusion weight to generate a comprehensive compensation matrix containing time correction coefficients and power correction coefficients.

6. The method of claim 5, wherein, Performing synchronous correction on operation signals of each inverter based on the comprehensive compensation matrix to form a synchronous signal set comprises: The time correction coefficient and the power correction coefficient in the comprehensive compensation matrix are respectively applied to the voltage signal, the current signal and the output power signal of each inverter to obtain a sequence of operation signals after synchronization adjustment; Each sequence of operation signals is resampled on a time axis according to a unified reference time, and a peak point of an amplitude value and a zero point of a phase are extracted from the resampling result to complete alignment of the amplitude value and the phase; All the sequences of operation signals after time and amplitude synchronization correction are combined in the order of inverter identification to form a set of synchronized signals.

7. The method of claim 1, wherein, An energy convergence degree model is established based on the set of synchronized signals, and an energy imbalance of each branch is determined to locate a fault position in the photovoltaic power grid system, including: The instantaneous power of each inverter branch in a preset sampling window is calculated based on the voltage signal and the current signal in the set of synchronized signals, and an integral operation is performed on the sequence of instantaneous power to obtain an energy accumulation value of each branch; An energy convergence degree model is established based on the energy accumulation value of each branch and a bus energy conservation constraint, and an energy difference between adjacent branches is calculated based on an output of the energy convergence degree model to generate an energy convergence degree matrix; In the energy convergence degree matrix, a branch with an energy imbalance exceeding a preset imbalance threshold is identified, and the corresponding branch is marked as a fault branch, and a fault position of the photovoltaic power grid system is determined according to the energy imbalance distribution of the fault branch.

8. The method of claim 7, wherein, An energy convergence degree model is established based on the energy accumulation value of each branch and a bus energy conservation constraint, including: An energy transmission mapping matrix is established according to a system topology relationship, taking the energy accumulation value of each branch as an input node and taking a total energy of a bus as a constraint node; A bus energy conservation condition and a node energy balance equation are introduced into the energy transmission mapping matrix to construct a constraint expression of the energy convergence degree model; The node energy balance coefficients are solved by minimizing the square sum of branch energy deviations to form an energy convergence degree model for quantifying energy transmission consistency.

9. A fault location system for a photovoltaic grid system, characterized by, The system includes: A first acquisition unit configured to acquire control parameters of each inverter in the photovoltaic power grid system and construct a dynamic consistency matrix for representing operation differences of the inverters; A second acquisition unit configured to acquire power tracking parameters of each inverter and construct a power consistency matrix for representing power response differences; A synchronization unit configured to fuse the dynamic consistency matrix and the power consistency matrix to generate a comprehensive compensation matrix, and perform synchronization correction on operation signals of each inverter based on the comprehensive compensation matrix to form a set of synchronized signals; A fault positioning unit configured to establish an energy convergence degree model based on the set of synchronized signals and determine an energy imbalance of each branch to locate a fault position in the photovoltaic power grid system.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores instructions, which, when executed on a computer, cause the computer to perform the fault positioning method of the photovoltaic power grid system according to any one of claims 1-8.

Images