A method and apparatus for assessing voltage over-limit risks in low-voltage distribution transformer areas.

Abstract

This application discloses a method and device for assessing voltage exceedance risks in low-voltage distribution transformer areas, relating to the field of intelligent monitoring technology. It solves the technical problem that existing technologies often rely on historical or real-time data from data acquisition and monitoring control systems for exceedance alarms, resulting in low efficiency because the voltage has already exceeded the limit by the time the alarm is issued. The method generates a distribution topology map and predicted load data for the transformer area based on multi-source data, and then uses this to generate power flow data. It extracts voltage from the power flow data and determines the voltage drop. When the voltage drop of any phase exceeds the amplitude threshold, it generates a predicted conduction path based on the distribution topology map, and uses this to generate voltage anomaly tracing results and a comprehensive risk index. Based on the comprehensive risk index, it generates an alarm signal, predicts future load in the transformer area, and reflects this prediction back to the current time, performing a risk assessment before the voltage exceeds the limit, thus improving the efficiency of the voltage exceedance risk assessment method.

Description

Technical Field

[0001] This application belongs to the field of intelligent monitoring technology, specifically a method and device for assessing voltage over-limit risks in low-voltage distribution transformer areas. Background Technology

[0002] With the continuous advancement of the construction of new power systems, low-voltage distribution substations, as the "last mile" connecting the end of the power grid to a vast number of power users, directly affect the power supply quality of residents, industrial and commercial operations, and even the overall satisfaction of society with electricity consumption. Voltage exceeding limits, which refers to the voltage of one or more nodes in a substation exceeding the nationally stipulated allowable range, is a common power quality problem in the operation of low-voltage distribution networks. Long-term or frequent voltage exceeding limits can not only damage users' electrical equipment and shorten its lifespan, but may also trigger power safety accidents and affect the reliability and stability of power grid operation.

[0003] Existing technologies often rely on historical or real-time data from data acquisition and monitoring control systems to issue over-limit alarms. By the time the alarm is issued, the voltage may have already exceeded the limit, affecting sensitive electrical equipment. This results in low efficiency of voltage over-limit risk assessment methods. Therefore, further improvements are needed for voltage over-limit risk assessment methods in low-voltage distribution transformer areas. Summary of the Invention

[0004] This application aims to solve at least one of the technical problems existing in the prior art; to this end, this application proposes a voltage over-limit risk assessment method and device for low-voltage distribution transformer areas, which solves the technical problem that the prior art often relies on historical or real-time data from data acquisition and monitoring control systems to issue over-limit alarms, resulting in the voltage already exceeding the limit when the alarm is issued, thus leading to the low efficiency of the voltage over-limit risk assessment method.

[0005] To achieve the above objectives, the first aspect of this application provides a method for assessing voltage over-limit risks in low-voltage distribution transformer areas, comprising:

[0006] Acquire multi-source data of low-voltage distribution transformer areas;

[0007] Constructing a power distribution topology map of the transformer area based on multi-source data of the transformer area;

[0008] Predicted load data is generated based on multi-source data from transformer substations.

[0009] Power flow data is generated based on multi-source data of the transformer area, the distribution topology map of the transformer area, and the predicted load data; the power flow data refers to the result of discounting the predicted load data to the current moment.

[0010] Three-phase total voltage imbalance is generated based on multi-source data from the transformer area;

[0011] Extract voltage from power flow data; determine voltage drop based on voltage and standard voltage; the voltage drop includes three-phase voltage drop.

[0012] When the voltage drop of any phase exceeds the amplitude threshold, a predicted conduction path is generated based on the distribution topology of the transformer area; the amplitude threshold is set empirically.

[0013] Voltage anomaly tracing results are generated based on predicted conduction paths;

[0014] A comprehensive risk index is generated based on trend data;

[0015] An alarm signal is generated based on the three-phase total voltage imbalance and the comprehensive risk index.

[0016] This application achieves scientific prediction of future transformer area load through the above steps, and converts the prediction results into the current operating conditions to generate corresponding power flow data. This method can identify potential risks in advance before the actual voltage over-limit occurs, realizing the transformation from "passive response" to "proactive early warning". Through the pre-risk identification step, sufficient intervention time window is reserved for operation and maintenance personnel, which significantly improves the foresight and effectiveness of voltage over-limit risk assessment, and effectively enhances the overall efficiency and application value of the assessment method.

[0017] Furthermore, the construction of the distribution topology map of the transformer area based on multi-source data of the transformer area includes:

[0018] The multi-source data for the transformer area includes node data and impedance response parameters; the impedance response parameters include node electrical distance and impedance amplitude and phase characteristics; the node data includes node ID and node type;

[0019] Acquire disturbance response data; the disturbance response data refers to the response data obtained after periodically injecting a pseudo-random binary sequence disturbance signal with a rated voltage, including voltage and current;

[0020] The impedance response parameters are dynamically updated based on the disturbance response data.

[0021] A distribution topology for transformer substations, incorporating impedance response parameters, is established by improving the Prim algorithm.

[0022] The distribution topology of a transformer substation is determined based on its distribution topology structure; the distribution topology is represented as G=(V, E, W); where V represents the node ID, E represents the edge connection relationship, and W represents the edge weight matrix; the edge weight matrix satisfies the following formula: Where j and k represent the node IDs, α represents the weight coefficient, and α∈(0,1); Z j，k The impedance value is represented as the impedance value between the j-th node ID and the k-th node ID, and the impedance response parameter includes the impedance value; Cj，k It represents the coupling coefficient between the j-th node ID and the k-th node ID.

[0023] Furthermore, the dynamic updating of impedance response parameters based on disturbance response data includes:

[0024] The impedance spectrum characteristics corresponding to the disturbance response data in a specific frequency band range are extracted using a 1024-point fast Fourier transform.

[0025] The impedance spectrum characteristics are input into the impedance prediction model to obtain the impedance response parameters;

[0026] The impedance prediction model is obtained by training a random forest regression algorithm on an impedance database and is used to update the impedance response parameters; the impedance database refers to a database containing several historical impedance spectrum features and their corresponding impedance response parameters.

[0027] Furthermore, the generation of predicted load data based on multi-source data from transformer substations includes:

[0028] Extract current load data and load impact parameters from multi-source data of the transformer area; the load impact parameters refer to parameters that affect load changes, including ambient temperature, ambient humidity and light intensity.

[0029] The current load data and load impact parameters are combined to form load analysis data;

[0030] Acquire several historical load analysis data points within a sliding time window;

[0031] Several historical load analysis data are spliced ​​together in chronological order to form a load analysis sequence;

[0032] The load analysis sequence is input into the load forecasting model to obtain the predicted load data; the load forecasting model is used to estimate the load status for a future period of time.

[0033] The load forecasting model is constructed using an LSTM model, including:

[0034] Obtain several historical load analysis sequences and their corresponding historical load data;

[0035] Several historical load analysis sequences and their corresponding historical load data are divided into training data, validation data, and test data; data preprocessing is performed on the training data, validation data, and test data to obtain the training set, validation set, and test set;

[0036] The loss function of the LSTM model is set as a hybrid loss function; the hybrid loss function is expressed as: Where γ1, γ2, and γ3 are loss weight coefficients, and γ1, γ2, and γ3 ∈ (0, 1); Mean square error, It is the mean absolute error. It is the mean absolute percentage error;

[0037] The modified LSTM was chosen as the base model.

[0038] The basic model is trained on the training set, and the learning rate and hyperparameters are adjusted on the validation set using the particle swarm optimization algorithm to obtain the pre-trained model.

[0039] By validating the pre-trained model on the test set, a load prediction model is finally obtained, which takes load analysis sequences as input and outputs predicted load data.

[0040] When performing load forecasting for transformer substations, this application acquires historical load data and combines it into a load analysis sequence according to the chronological order. The LSTM model, modified with a loss function, is then trained on the corresponding dataset to obtain a load forecasting model suitable for the current task. This improves the accuracy of the predicted load data and provides accurate data support for subsequent voltage over-limit risk assessment.

[0041] Furthermore, the generation of power flow data based on multi-source data of the transformer area, the transformer area distribution topology map, and the predicted load data includes the following steps:

[0042] Step 1: Extract impedance response parameters, distribution topology diagram, current load data, and predicted load data from the multi-source data of the transformer area;

[0043] Step 2: Construct the admittance matrix containing the three phases A, B, and C and the neutral line; define the initial values ​​of the node voltages and the iteration accuracy threshold;

[0044] Step 3: Reconstruct the admittance matrix in blocks according to the node type, reduce the scale of invalid operations by focusing on non-zero elements, and form a sparse matrix adapted to the GPU parallel architecture; the node types include power nodes and load nodes;

[0045] Step 4: Construct power flow equations based on sparse matrices and the Newton-Raphson method framework; obtain voltage correction by solving the power flow equations; update the voltage value based on the voltage correction and obtain the residual change rate; dynamically adjust the calculation step size based on the residual change rate.

[0046] Step 5: Obtain the iteration residual; when the iteration residual is less than the iteration accuracy threshold, output the power flow data.

[0047] Furthermore, the dynamic adjustment of the calculation step size based on the residual change rate includes:

[0048] When the rate of change of the residual is greater than the change threshold, the calculation step size is reduced to the step size value of one.

[0049] When the rate of change of the residual is less than the change threshold of two, the calculation step size is increased to the step size value of two.

[0050] When the rate of change of the residual is less than or equal to the first threshold and the rate of change of the residual is greater than or equal to the second threshold, the calculation step size remains unchanged; wherein, the first threshold is greater than the second threshold, and the second step size is greater than the first step size.

[0051] This application obtains accurate power flow data through a series of steps when discounting load forecast data to the current time. It also achieves dynamic adjustment of the calculation step size by utilizing the residual change rate, thus achieving a relative balance between computational efficiency and stability and improving the accuracy and efficiency of power flow data generation.

[0052] Furthermore, the generation of three-phase total voltage imbalance based on multi-source data from the transformer substation includes:

[0053] Extract the three-phase voltage drop from the multi-source data of the transformer area, and express it as ΔV. A ΔV B and ΔV C ;

[0054] The three-phase total voltage imbalance (TVUR) is calculated using the formula. adj The formula satisfies:

[0055] Where β represents the difference factor, used to quantify the line impedance deviation; and i represents the three-phase voltage drop ΔV. A ΔV B and ΔV C The corresponding number, It can be expressed as the average value of the three-phase voltage drop, and can be expressed by the formula. The calculation is performed; the difference factor satisfies the following calculation formula: max() and min() represent the operations of finding the maximum and minimum values, respectively; Z A Z B and Z C These are represented as the impedances of phase A, phase B, and phase C lines, respectively.

[0056] Furthermore, the generation of predicted transmission paths based on the distribution topology map of the transformer substation includes:

[0057] The topological spatial features of the distribution topology map of the transformer substation are extracted using the GCN network model;

[0058] Extract the voltage from the multi-source data of the transformer area within the historical Q minutes, and concatenate them into a voltage time series according to the chronological order; where Q is an integer, Q>0;

[0059] Voltage time series and topological spatial features are input into the conduction prediction model to obtain the predicted conduction path; the conduction prediction model is constructed using an LSTM model to predict the optimal conduction path of the fault.

[0060] Furthermore, the generation of voltage anomaly tracing results based on predicted conduction paths includes:

[0061] Extract the voltage corresponding to several node IDs in the predicted conduction path;

[0062] The correlation coefficients of voltage fluctuations at several nodes were calculated using the Pearson correlation coefficient method.

[0063] Select several node IDs whose absolute values ​​of the correlation coefficients are greater than the coefficient threshold as candidate source points;

[0064] Calculate the slope matching degree (SMD) of several candidate source points; the SMD is calculated using a formula that satisfies: , where N total N represents the total number of local time segments into which the time window is divided. match The number of local time segments required to satisfy the slope matching condition;

[0065] Several candidate source points with slope matching degree greater than the matching threshold are selected as candidate source points.

[0066] Extract the voltage fluctuation anomaly start time points of several candidate source points;

[0067] The earliest candidate source point at which the voltage fluctuation anomaly begins is identified as the fault node ID.

[0068] The source of the voltage anomaly was determined based on the fault node ID.

[0069] Furthermore, the generation of a comprehensive risk index based on tidal data includes:

[0070] Extracting the rated transformer capacity P from the multi-source data of the distribution area N And the active power and line impedance Z in the power flow data. line ;

[0071] Define three-dimensional evaluation metrics; these metrics include load fluctuation rate, impedance matching degree, and capacity-to-load ratio.

[0072] Extracting active power P within the sliding window t t represents the time point number within the sliding window;

[0073] The load volatility FBL is calculated using a formula; the formula satisfies:

[0074] ;in, This is expressed as the average active power within the sliding window;

[0075] Through formula Calculate the impedance matching degree ZPD; where Z base Represented as a reference impedance value;

[0076] Through formula Calculate the capacity ratio RZB; P load，j This represents the typical operating load corresponding to the j-th node ID; J represents the total number of node IDs.

[0077] The comprehensive risk index is obtained by weighting and summing the three-dimensional evaluation indicators using indicator weight coefficients; the indicator weight coefficients are obtained through the following methods:

[0078] Each three-dimensional evaluation index is divided into 5 levels and a triangular membership function is defined to quantify the membership degree;

[0079] 100,000 sets of random samples covering historical operating scenarios were generated through Monte Carlo simulation, and the initial index weights were calculated by combining the entropy weight method and correlation analysis.

[0080] The objective function is optimized using the gradient descent algorithm to ultimately determine the index weight coefficients; the objective function satisfies: Where LYS is the number of historical samples, CZQ is the initial indicator weight to be optimized, and y s A comprehensive risk index is set for the s-th historical sample. Calculate the comprehensive risk index for the s-th historical sample.

[0081] This application, upon detecting a voltage drop exceeding a preset threshold in any phase, immediately triggers a pre-trained voltage propagation path prediction model. This model rapidly deduces the potential propagation path of the abnormal voltage. By comprehensively analyzing the electrical location, topological association, and response timing of multiple key node IDs in the predicted propagation path, it accurately locates the source region of the voltage anomaly. This achieves a closed-loop analysis from "phenomenon monitoring" to "root cause tracing," significantly improving the interpretability and operational guidance value of voltage over-limit events. It not only enables real-time assessment of voltage over-limit risks but also proactive tracing of abnormal events, making the risk assessment system more comprehensive and intelligent. Simultaneously, in calculating the comprehensive risk index, the weight coefficients of the three-dimensional assessment indicators are adaptively adjusted, effectively overcoming the assessment bias caused by traditional fixed weights. This allows the comprehensive risk index to more accurately reflect the current risk level of the transformer area. The resulting high-precision risk index provides a scientific and reliable decision-making basis for the generation of subsequent graded early warning signals, significantly improving the accuracy and practicality of voltage over-limit risk early warning.

[0082] Furthermore, the generation of alarm signals based on the three-phase total voltage imbalance and comprehensive risk index includes:

[0083] When the three-phase total voltage imbalance exceeds the imbalance threshold, a three-phase total voltage abnormality alarm signal is generated; the imbalance threshold is set based on experience.

[0084] When the comprehensive risk index is within the normal range, a normal voltage signal is generated;

[0085] When the comprehensive risk index is within the attention index range, a voltage anomaly attention signal is generated;

[0086] When the comprehensive risk index is within the alarm index range, a voltage abnormality alarm signal is generated;

[0087] When the comprehensive risk index is within the emergency index range, a voltage abnormality emergency signal is generated; the normal index range, attention index range, alarm index range, and emergency index range are set based on experience.

[0088] Another aspect of the present invention provides a voltage over-limit risk assessment device for a low-voltage distribution transformer area, comprising: a data acquisition unit, a data analysis unit, and an early warning unit; the data acquisition unit and the data analysis unit are connected; the data analysis unit and the early warning unit are connected.

[0089] The data acquisition unit acquires multi-source data of the low-voltage distribution transformer area through data acquisition equipment.

[0090] The data analysis unit: constructs a distribution topology map of the transformer area based on multi-source data of the transformer area; generates predicted load data based on the multi-source data of the transformer area; generates power flow data based on the multi-source data of the transformer area, the distribution topology map of the transformer area, and the predicted load data; the power flow data refers to the result of discounting the predicted load data to the current moment; generates the three-phase total voltage imbalance based on the multi-source data of the transformer area; extracts the voltage from the power flow data; and determines the voltage drop based on the voltage and the standard voltage; the voltage drop includes the three-phase voltage drop.

[0091] When the voltage drop of any phase exceeds the magnitude threshold, a predicted conduction path is generated based on the distribution topology of the transformer area; voltage anomaly tracing results are generated based on the predicted conduction path; and a comprehensive risk index is generated based on power flow data.

[0092] An alarm signal is generated based on the three-phase total voltage imbalance and the comprehensive risk index;

[0093] The early warning unit: provides corresponding prompts based on the alarm signal.

[0094] Compared with the prior art, the beneficial effects of this application are:

[0095] 1. This application constructs a distribution topology map of a transformer substation based on multi-source data; generates predicted load data based on the multi-source data; generates power flow data based on the multi-source data, distribution topology map, and predicted load data; generates three-phase total voltage imbalance based on the multi-source data; extracts voltage from the power flow data; determines voltage drop based on voltage and standard voltage; generates a predicted conduction path based on the distribution topology map when the voltage drop of any phase exceeds the amplitude threshold; generates voltage anomaly tracing results based on the predicted conduction path; generates a comprehensive risk index based on the power flow data; and generates an alarm signal based on the three-phase total voltage imbalance and comprehensive risk index. It predicts future load in the transformer substation and reflects this prediction back to the current time point to obtain power flow data. This allows for risk assessment before voltage exceeds limits, achieving a leap from passive response to proactive early warning, providing maintenance personnel with a valuable intervention window, and improving the efficiency of voltage exceedance risk assessment methods.

[0096] 2. This application achieves dynamic updating of impedance response parameters by periodically injecting pseudo-random binary sequence perturbation signals of rated voltage into the response data. This enables the dynamic updating of the distribution topology of the transformer substation by establishing a distribution topology containing impedance response parameters through the improved Prim algorithm. Consequently, the final distribution topology map has dynamic updating capability, improving the accuracy of the distribution topology map and providing accurate data support for subsequent voltage over-limit risk assessment.

[0097] 3. When the voltage drop of any phase exceeds the amplitude threshold, this application obtains the predicted conduction path through a pre-trained conduction prediction model, and performs comprehensive analysis on several node IDs in the predicted conduction path to determine the source of the voltage anomaly. This not only realizes the real-time assessment of voltage over-limit but also the source tracing of the abnormal result, making the voltage over-limit risk assessment method more comprehensive. At the same time, when calculating the comprehensive risk index, the index weight coefficients corresponding to the three-dimensional assessment indicators are dynamically adjusted, which improves the accuracy of the comprehensive risk index and provides accurate data support for the generation of subsequent alarm signals. Attached Figure Description

[0098] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0099] Figure 1 This is a flowchart of a voltage over-limit risk assessment method for a low-voltage distribution transformer area according to this application;

[0100] Figure 2 This is a schematic diagram illustrating the principle of a voltage over-limit risk assessment system for a low-voltage distribution transformer area according to this application. Detailed Implementation

[0101] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0102] Please see Figure 1 The first aspect of this application provides a method for assessing the voltage over-limit risk of a low-voltage distribution transformer area, including:

[0103] Acquire multi-source data of low-voltage distribution transformer areas;

[0104] Constructing a power distribution topology map of the transformer area based on multi-source data of the transformer area;

[0105] Predicted load data is generated based on multi-source data from transformer substations; predicted load data refers to the prediction results of load at future times.

[0106] Power flow data is generated based on multi-source data of transformer substations, distribution topology diagrams of transformer substations, and predicted load data; power flow data refers to the result of discounting predicted load data to the current moment.

[0107] The three-phase total voltage imbalance is generated based on multi-source data from the transformer area; the three-phase total voltage imbalance refers to the degree of abnormality in voltage imbalance.

[0108] Extract voltage from power flow data; determine voltage drop based on voltage and standard voltage; voltage drop includes three-phase voltage drop; in this embodiment, voltage drop is expressed as the difference between voltage and standard voltage;

[0109] When the voltage drop of any phase exceeds the amplitude threshold, a predicted conduction path is generated based on the distribution topology of the transformer area; the amplitude threshold is set empirically, and in this embodiment, the amplitude threshold is set to 0.9pu;

[0110] Voltage anomaly tracing results are generated based on predicted conduction paths;

[0111] A comprehensive risk index is generated based on power flow data; the comprehensive risk index measures the overall degree of anomaly in a low-voltage distribution transformer area.

[0112] An alarm signal is generated based on the three-phase total voltage imbalance and the comprehensive risk index.

[0113] The method for constructing a distribution topology map of a transformer area based on multi-source data in this embodiment includes:

[0114] The multi-source data for the transformer area includes node data and impedance response parameters; the impedance response parameters include node electrical distance and impedance amplitude and phase characteristics; the node data includes node ID and node type.

[0115] Obtain disturbance response data; disturbance response data refers to the response data obtained after periodically injecting a pseudo-random binary sequence disturbance signal with a rated voltage, including voltage and current; in this embodiment, the rated voltage range is between 0.5% and 2%;

[0116] The impedance response parameters are dynamically updated based on the disturbance response data.

[0117] A distribution topology for transformer substations, incorporating impedance response parameters, is established by improving the Prim algorithm.

[0118] The distribution topology of the transformer substation is determined based on the distribution topology of the substation area; the distribution topology of the substation area is represented as G=(V,E,W); where V represents the node ID, E represents the edge connection relationship, and W represents the edge weight matrix; the edge weight matrix satisfies the following formula: Where j and k represent the node IDs, α represents the weight coefficient, α∈(0,1); the specific value is set based on experience, and in this embodiment, α is set to 0.7; Z j，k The impedance value is represented as the impedance value between the j-th node ID and the k-th node ID. The impedance response parameter includes the impedance value; C j，k It represents the coupling coefficient between the j-th node ID and the k-th node ID.

[0119] The dynamic updating of impedance response parameters based on disturbance response data in this embodiment includes:

[0120] The impedance spectrum features corresponding to the disturbance response data in a specific frequency band range are extracted using a 1024-point fast Fourier transform. In this embodiment, the specific frequency band range is between 0.1 kHz and 1 kHz. The impedance spectrum features in this embodiment have 16 dimensions, including line length, conductor type, and ambient temperature.

[0121] The impedance spectrum characteristics are input into the impedance prediction model to obtain the impedance response parameters;

[0122] The impedance prediction model is obtained by training a random forest regression algorithm on an impedance database and is used to update the impedance response parameters. The impedance database refers to a database containing several historical impedance spectrum features and their corresponding impedance response parameters.

[0123] This embodiment achieves dynamic identification and real-time updating of transformer area impedance response parameters by periodically injecting pseudo-random binary sequence disturbance signals under rated voltage and collecting their response data. Based on the updated impedance parameters, a transformer area distribution topology containing electrical characteristic information is constructed using an improved Prim algorithm. This allows the topology to be dynamically adjusted according to changes in the actual operating status of the transformer area, significantly improving the timeliness and accuracy of the transformer area distribution topology diagram. It effectively overcomes the limitation of traditional static topology modeling, which is difficult to reflect actual wiring changes. The generated dynamic topology results provide high-precision basic data support for subsequent voltage over-limit risk assessment, enhance the assessment model's ability to perceive and predict the operating status of the transformer area, and further improve the accuracy and practicality of risk warning.

[0124] In this embodiment, the generation of predicted load data based on multi-source data from transformer substations includes:

[0125] Extract current load data and load impact parameters from multi-source data of the transformer area; load impact parameters refer to parameters that affect load changes, including ambient temperature, ambient humidity and light intensity.

[0126] The current load data and load impact parameters are combined to form load analysis data;

[0127] Acquire several historical load analysis data within a sliding time window; the sliding time window refers to the time range between a past time point and the current time point; the specific value is set based on experience, and in this embodiment, the sliding time window is set to 15 minutes.

[0128] Several historical load analysis data are spliced ​​together in chronological order to form a load analysis sequence;

[0129] The load analysis sequence is input into the load forecasting model to obtain the forecast load data; the load forecasting model is used to estimate the load status in the future.

[0130] The load forecasting model is constructed using an LSTM model, including:

[0131] Obtain several historical load analysis sequences and their corresponding historical load data;

[0132] Several historical load analysis sequences and their corresponding historical load data are divided into training data, validation data, and test data; data preprocessing is performed on the training data, validation data, and test data to obtain training sets, validation sets, and test sets; the ratio between the training set, test set, and validation set is 6:2:2.

[0133] Set the loss function of the LSTM model to a hybrid loss function; the hybrid loss function is expressed as: Wherein, γ1, γ2, and γ3 are loss weight coefficients, γ1, γ2, and γ3 ∈ (0, 1); the specific values ​​are set according to experience, and in this embodiment, γ1, γ2, and γ3 are set to 0.7, 0.2, and 0.1, respectively; Mean square error, It is the mean absolute error. It is the mean absolute percentage error;

[0134] The modified LSTM was chosen as the base model.

[0135] The basic model is trained on the training set, and the learning rate and hyperparameters are adjusted on the validation set using the particle swarm optimization algorithm to obtain the pre-trained model.

[0136] By validating the pre-trained model on the test set, a load prediction model is finally obtained, which takes load analysis sequences as input and outputs predicted load data.

[0137] In this embodiment, generating power flow data based on multi-source data of the transformer area, the transformer area distribution topology map, and predicted load data includes the following steps:

[0138] Step 1: Extract impedance response parameters, distribution topology diagram, current load data, and predicted load data from the multi-source data of the transformer area;

[0139] Step 2: Construct the admittance matrix containing the three phases A, B, and C and the neutral line; define the initial values ​​of the node voltages and the iteration accuracy threshold;

[0140] Step 3: Reconstruct the admittance matrix in blocks according to node type, reduce the scale of invalid operations by focusing on non-zero elements, and form a sparse matrix adapted to the GPU parallel architecture; node types include power nodes and load nodes;

[0141] Step 4: Construct power flow equations based on sparse matrices and the Newton-Raphson method framework, and obtain voltage correction values ​​by solving the power flow equations; update the voltage values ​​based on the voltage correction values ​​and obtain the residual change rate; dynamically adjust the calculation step size based on the residual change rate; in this embodiment, the residual refers to the difference between the node power calculated by the current voltage and the actual power; the voltage value update in this embodiment can be divided into two stages: the first stage is during the initial iteration, where the updated voltage value is the initial voltage value during the current iteration; the second stage is during subsequent iterations, where the updated voltage value is the voltage value at the beginning of the current iteration; the initial voltage value during subsequent iterations is determined by the voltage value updated in the previous iteration and the calculation step size.

[0142] Step 5: Obtain the iteration residual. When the iteration residual is less than the iteration accuracy threshold, output the power flow data.

[0143] The dynamic adjustment of the calculation step size based on the residual change rate in this embodiment includes:

[0144] When the rate of change of the residual is greater than the change threshold, the calculation step size is reduced to the step size value of one.

[0145] When the rate of change of the residual is less than the change threshold of two, the calculation step size is increased to the step size value of two.

[0146] When the residual change rate is less than or equal to the change threshold one and the residual change rate is greater than or equal to the change threshold two, the calculation step size remains unchanged; wherein, the change threshold one is greater than the change threshold two; the step size two is greater than the step size one; the specific values ​​are set according to experience. In this embodiment, the change threshold one and the change threshold two are set to 5% and 1% respectively; the step size one and the step size two are set to 0.3 and 1.2 respectively.

[0147] This embodiment achieves a reasonable mapping of future operating conditions through a series of refined processing steps in the process of discounting load forecast data to the current operating time, thereby generating power flow data for voltage risk assessment. To improve the accuracy and adaptability of the calculation process, the residual change rate is introduced as a dynamic adjustment basis to monitor the convergence trend in the power flow iteration process in real time and adaptively adjust the calculation step size accordingly. While ensuring the stability of the solution, it effectively accelerates the convergence speed and avoids the computational redundancy or oscillation risk caused by the traditional fixed step size. Thus, a good balance is achieved between computational efficiency and numerical stability. This method significantly improves the accuracy and timeliness of power flow data generation and provides a high-quality and highly reliable data foundation for the accurate assessment of subsequent voltage over-limit risks.

[0148] In this embodiment, the generation of three-phase total voltage imbalance based on multi-source data from the transformer substation includes:

[0149] Extract the three-phase voltage drop from the multi-source data of the transformer area, and express it as ΔV. A ΔV B and ΔV C In this embodiment, the three-phase voltage drop is the result after a 15-point moving average filtering operation. The averaging filtering operation eliminates high-frequency noise to improve data accuracy.

[0150] The three-phase total voltage imbalance (TVUR) is calculated using the formula. adj The formula satisfies:

[0151] Where β represents the difference factor, used to quantify the line impedance deviation; and i represents the three-phase voltage drop ΔV. A ΔV B and ΔV C The corresponding number, It can be expressed as the average value of the three-phase voltage drop, and can be expressed by the formula. The calculation is performed; the difference factor satisfies the following calculation formula: max() and min() represent the operations of finding the maximum and minimum values, respectively; Z A Z B and Z C These are represented as the impedances of phase A, phase B, and phase C lines, respectively.

[0152] In this embodiment, the generation of predicted conduction paths based on the distribution topology map of the transformer substation includes:

[0153] The topological spatial features of the distribution topology map of the transformer substation are extracted using the GCN network model;

[0154] Voltage data from multiple sources within a historical Q-minute period is extracted and concatenated into a voltage time series based on chronological order. Here, Q is an integer, Q>0; the specific value is set based on experience, and in this embodiment, Q is set to 10.

[0155] Voltage time series and topological spatial features are input into the conduction prediction model to obtain the predicted conduction path; the conduction prediction model is constructed using an LSTM model to predict the optimal conduction path of the fault; in this embodiment, the conduction prediction model is obtained by training the LSTM model on the corresponding historical data, and the training method is consistent with the training method of general network models in the prior art.

[0156] The voltage anomaly tracing results generated based on predicted conduction paths in this embodiment include:

[0157] Extract the voltage corresponding to several node IDs in the predicted conduction path;

[0158] The correlation coefficients of voltage fluctuations at several nodes were calculated using the Pearson correlation coefficient method.

[0159] Several node IDs whose absolute values ​​of correlation coefficients are greater than the coefficient threshold are selected as candidate source points; the coefficient threshold is set based on experience, and in this embodiment, the coefficient threshold is set to 0.7.

[0160] Calculate the slope matching degree (SMD) of several candidate source points; the SMD is calculated using the formula, which satisfies: , where N total N represents the total number of local time segments into which the time window is divided. match To determine the number of local time segments that meet the slope matching criteria, the slope matching criteria in this embodiment include consistent trends and similar magnitudes of change; consistent trends are expressed as: , and Let represent the slopes of the m-th node ID and the n-th node ID in the h-th local time segment, respectively, and let sign() be the sign function; similar changes are represented as follows: , where ε is a constant, ε>0; the specific value is set according to experience, in this embodiment ε is set to 0.0001 to avoid the denominator being 0, δ is the threshold of the change range, δ∈(0,1); the specific value is set according to experience, in this embodiment δ is set to 0.1;

[0161] Several candidate source points with slope matching degrees greater than the matching threshold are selected as candidate source points; the matching threshold is set based on experience, and in this embodiment, the matching threshold is set to 85%;

[0162] Extract the voltage fluctuation anomaly start time points of several candidate source points;

[0163] The earliest candidate source point at which the voltage fluctuation anomaly begins is identified as the fault node ID.

[0164] The source of the voltage anomaly was determined based on the fault node ID.

[0165] The generation of a comprehensive risk index based on tidal flow data in this embodiment includes:

[0166] Extracting the rated transformer capacity P from the multi-source data of the distribution area N And the active power and line impedance Z in the power flow data. line ;

[0167] Define three-dimensional evaluation metrics; the three-dimensional evaluation metrics include load fluctuation rate, impedance matching degree, and capacity-to-load ratio;

[0168] Extracting active power P within the sliding window t The sliding window size is set based on experience; in this embodiment, the sliding window size is set to 15 minutes; t represents the time point number within the sliding window.

[0169] The load volatility FBL is calculated using the formula; the formula satisfies:

[0170] ;in, This is expressed as the average active power within the sliding window;

[0171] Through formula Calculate the impedance matching degree ZPD; where Z base Represented as a reference impedance value;

[0172] Through formula Calculate the capacity ratio RZB; P load，j This represents the typical operating load corresponding to the j-th node ID; J represents the total number of node IDs.

[0173] The comprehensive risk index is obtained by weighting and summing the three-dimensional assessment indicators using indicator weight coefficients. The indicator weight coefficients are obtained through the following methods:

[0174] Each three-dimensional evaluation index is divided into 5 levels, and a triangular membership function is defined to quantify the membership degree. In this embodiment, the three-dimensional evaluation indexes are divided into low risk, lower risk, medium risk, higher risk, and high risk. For example, load volatility is divided into: [0, 10], (10, 20], (20, 30], (30, 40], (40, 50]; impedance matching is divided into: [0, 1.2], (1.2, 1.4], (1.4, 1.6], (1.6, 1.8], (1.8, 2]; and capacity ratio is divided into: ( [1.8, 2], (1.6, 1.8], (1.4, 1.6], (1.2, 1.4], [1, 1.2]; Construct a triangle membership function: For a certain level interval [b, c), define the triangle coordinate vertices (a, 0), (b, 1), (c, 0), where a = b - ω × (bc), indicating that adjacent levels have overlap, ω is the overlap ratio, ω∈(0, 1), the specific value is set according to experience, in this embodiment ω is set to 0.05, then the membership degree calculation formula of the three-dimensional evaluation index value at this level is: Where x represents the three-dimensional evaluation index value;

[0175] 100,000 random samples covering historical operating scenarios were generated using Monte Carlo simulation. Initial indicator weights were calculated using the entropy weight method and correlation analysis. In this embodiment, the entropy weight SQ corresponding to the three-dimensional evaluation indicators was first determined using the entropy weight method. Then, the Pearson correlation coefficients between each three-dimensional evaluation indicator were calculated to obtain the average correlation between the three-dimensional evaluation indicators and other three-dimensional evaluation indicators. , where r d，e This represents the correlation coefficient between the d-th three-dimensional evaluation index and the e-th three-dimensional evaluation index;

[0176] The initial index weights corresponding to the three-dimensional evaluation indicators are calculated using the formula; the formula satisfies: , where tx is the adjustment coefficient, tx∈(0,1), and the specific value is set according to experience. In this embodiment, tx is set to 0.2;

[0177] The gradient descent algorithm is used to optimize the objective function, and the index weight coefficients are finally determined; the objective function satisfies: Where LYS is the number of historical samples, CZQ is the initial indicator weight to be optimized, and y s A comprehensive risk index is set for the s-th historical sample. Calculate the comprehensive risk index for the s-th historical sample. Among them, scored，s Let be the membership degree of the d-th three-dimensional evaluation index for the s-th sample, represented as a 1-dimensional matrix containing 5 levels of triangular membership functions [μ1, μ2, μ3, μ4, μ5].

[0178] In this embodiment, the alarm signal is generated based on the three-phase total voltage imbalance and the comprehensive risk index, including:

[0179] When the three-phase total voltage imbalance exceeds the imbalance threshold, a three-phase total voltage abnormality alarm signal is generated; the imbalance threshold is set based on experience, and in this embodiment, the imbalance threshold is set to 5%;

[0180] When the comprehensive risk index is within the normal range, a normal voltage signal is generated;

[0181] When the comprehensive risk index is within the attention index range, a voltage anomaly attention signal is generated;

[0182] When the comprehensive risk index is within the alarm index range, a voltage abnormality alarm signal is generated;

[0183] When the comprehensive risk index is within the emergency index range, a voltage abnormality emergency signal is generated. The normal index range, attention index range, alarm index range and emergency index range are set according to experience. In this embodiment, the normal index range, attention index range, alarm index range and emergency index range are set to [0, 1.25], (1.25, 2.5], (2.5, 3.75], (3.75, 5) respectively.

[0184] Please see Figure 2 Another embodiment of this application provides a voltage over-limit risk assessment device for a low-voltage distribution transformer area, including: a data acquisition unit, a data analysis unit, and an early warning unit; the data acquisition unit and the data analysis unit are electrically and / or communicatively connected; the data analysis unit and the early warning unit are electrically and / or communicatively connected.

[0185] Data acquisition unit: Acquires multi-source data from low-voltage distribution transformer areas through data acquisition equipment;

[0186] Data Analysis Unit: Constructs a distribution topology map of the transformer substation based on multi-source data; generates predicted load data based on multi-source data of the transformer substation; generates power flow data based on multi-source data of the transformer substation, distribution topology map of the transformer substation, and predicted load data; power flow data refers to the result of discounting the predicted load data to the current moment; generates the three-phase total voltage imbalance based on multi-source data of the transformer substation; extracts voltage from the power flow data; determines voltage drop based on voltage and standard voltage; voltage drop includes three-phase voltage drop.

[0187] When the voltage drop of any phase exceeds the magnitude threshold, a predicted conduction path is generated based on the distribution topology of the transformer area; voltage anomaly tracing results are generated based on the predicted conduction path; and a comprehensive risk index is generated based on power flow data.

[0188] An alarm signal is generated based on the three-phase total voltage imbalance and the comprehensive risk index;

[0189] Early warning unit: Provides corresponding prompts based on alarm signals.

[0190] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0191] The working principle of this application is as follows: Multi-source data of low-voltage distribution transformer areas is acquired; a distribution topology map of the transformer area is constructed based on the multi-source data; predicted load data is generated based on the multi-source data; power flow data is generated based on the multi-source data, the distribution topology map, and the predicted load data; three-phase total voltage imbalance is generated based on the multi-source data; voltage is extracted from the power flow data; voltage drop is determined based on the voltage and standard voltage; when the voltage drop of any phase exceeds the amplitude threshold, a predicted conduction path is generated based on the distribution topology map; voltage anomaly tracing results are generated based on the predicted conduction path; and a comprehensive wind power distribution system is generated based on the power flow data. The risk index generates alarm signals based on the three-phase total voltage imbalance and comprehensive risk index, predicts the future load of the transformer area, and reflects it to the current time point to obtain power flow data. It can conduct risk assessment before the voltage actually exceeds the limit, realizing a leap from passive response to proactive early warning. It provides valuable intervention time windows for operation and maintenance personnel, improves the efficiency of voltage limit exceedance risk assessment methods, and avoids the problem that existing technologies often rely on historical or real-time data from data acquisition and monitoring control systems to issue limit exceedance alarms, which means that the voltage has already exceeded the limit when the alarm is issued, resulting in low efficiency of voltage limit exceedance risk assessment methods.

[0192] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.

Claims

1. A method for assessing voltage over-limit risk in a low-voltage distribution transformer area, characterized in that, include: Acquire multi-source data of low-voltage distribution transformer areas; Constructing a power distribution topology map of the transformer area based on multi-source data of the transformer area; Predicted load data is generated based on multi-source data from transformer substations. Power flow data is generated based on multi-source data of the transformer area, the distribution topology of the transformer area, and the predicted load data; the power flow data refers to the result of discounting the predicted load data to the current moment; the three-phase total voltage imbalance is generated based on the multi-source data of the transformer area; the voltage is extracted from the power flow data; the voltage drop is determined based on the voltage and the standard voltage; the voltage drop includes the three-phase voltage drop. When the voltage drop of any phase exceeds the amplitude threshold, a predicted conduction path is generated based on the distribution topology of the transformer area; and a voltage anomaly source tracing result is generated based on the predicted conduction path. A comprehensive risk index is generated based on trend data; An alarm signal is generated based on the three-phase total voltage imbalance and the comprehensive risk index; The construction of the distribution topology map of the transformer area based on multi-source data of the transformer area includes: The multi-source data for the transformer area includes node data and impedance response parameters; the impedance response parameters include node electrical distance and impedance amplitude and phase characteristics; the node data includes node ID and node type; Acquire disturbance response data; the disturbance response data refers to the response data obtained after periodically injecting a pseudo-random binary sequence disturbance signal with a rated voltage, including voltage and current; The impedance response parameters are dynamically updated based on the disturbance response data. An improved Prim algorithm was used to establish a distribution topology for transformer substations that includes impedance response parameters. The distribution topology of a transformer substation is determined based on its distribution topology structure; the distribution topology is represented as G=(V, E, W); where V represents the node ID, E represents the edge connection relationship, and W represents the edge weight matrix; the edge weight matrix satisfies the following formula: Where j and k represent the node IDs, α represents the weight coefficient, and α∈(0,1); Z j，k The impedance value is represented as the impedance value between the j-th node ID and the k-th node ID, and the impedance response parameter includes the impedance value; C j，k This is represented as the coupling coefficient between the j-th node ID and the k-th node ID; The generation of power flow data based on multi-source data of the transformer area, the distribution topology map of the transformer area, and the predicted load data includes the following steps: Step 1: Extract impedance response parameters, distribution topology diagram, current load data, and predicted load data from the multi-source data of the transformer area; Step 2: Construct the admittance matrix containing the three phases A, B, and C and the neutral line; define the initial values ​​of the node voltages and the iteration accuracy threshold; Step 3: Reconstruct the admittance matrix in blocks according to the node type, reduce the scale of invalid operations by focusing on non-zero elements, and form a sparse matrix adapted to the GPU parallel architecture; the node types include power nodes and load nodes; Step 4: Construct power flow equations based on sparse matrices and the Newton-Raphson method framework; obtain voltage correction by solving the power flow equations; update the voltage value based on the voltage correction and obtain the residual change rate; dynamically adjust the calculation step size based on the residual change rate. Step 5: Obtain the iteration residual; when the iteration residual is less than the iteration accuracy threshold, output the power flow data; The generation of three-phase total voltage imbalance based on multi-source data from the transformer substation includes: Extract the three-phase voltage drop from the multi-source data of the transformer area, and express it as ΔV. A ΔV B and ΔV C ; The three-phase total voltage imbalance (TVUR) is calculated using the formula. adj The formula satisfies: Where β represents the difference factor; i represents the three-phase voltage drop ΔV A ΔV B and ΔV C The corresponding number, It can be expressed as the average value of the three-phase voltage drop, and can be expressed by the formula. The calculation is performed; the difference factor satisfies the following calculation formula: max() and min() represent the operations of finding the maximum and minimum values, respectively; Z A Z B and Z C These are represented as the impedances of phase A, phase B, and phase C lines, respectively.

2. The method for assessing voltage over-limit risk in a low-voltage distribution transformer area according to claim 1, characterized in that, The dynamic updating of impedance response parameters based on disturbance response data includes: The impedance spectrum characteristics corresponding to the disturbance response data in a specific frequency band range are extracted using a 1024-point fast Fourier transform. The impedance spectrum characteristics are input into the impedance prediction model to obtain the impedance response parameters; The impedance prediction model is obtained by training a random forest regression algorithm on an impedance database and is used to update the impedance response parameters; the impedance database refers to a database containing several historical impedance spectrum features and their corresponding impedance response parameters.

3. The method for assessing voltage over-limit risk in a low-voltage distribution transformer area according to claim 1, characterized in that, The generation of predicted load data based on multi-source data from transformer substations includes: Extract current load data and load impact parameters from multi-source data of the transformer area; the load impact parameters refer to parameters that affect load changes, including ambient temperature, ambient humidity and light intensity; The current load data and load impact parameters are combined to form load analysis data; Acquire several historical load analysis data points within a sliding time window; Several historical load analysis data are spliced ​​together in chronological order to form a load analysis sequence; The load analysis sequence is input into the load forecasting model to obtain the predicted load data; the load forecasting model is used to estimate the load status for a future period of time. The load forecasting model is constructed using an LSTM model, including: Obtain several historical load analysis sequences and their corresponding historical load data; Several historical load analysis sequences and their corresponding historical load data are divided into training data, validation data, and test data; data preprocessing is performed on the training data, validation data, and test data to obtain the training set, validation set, and test set; The loss function of the LSTM model is set as a hybrid loss function; the hybrid loss function is expressed as: Where γ1, γ2, and γ3 are loss weight coefficients, and γ1, γ2, and γ3 ∈ (0, 1); Mean squared error, It is the mean absolute error. It is the mean absolute percentage error; The modified LSTM was chosen as the base model. The basic model is trained on the training set, and the learning rate and hyperparameters are adjusted on the validation set using the particle swarm optimization algorithm to obtain the pre-trained model. By validating the pre-trained model on the test set, a load prediction model is finally obtained, which takes load analysis sequences as input and outputs predicted load data.

4. The method for assessing voltage over-limit risk in a low-voltage distribution transformer area according to claim 1, characterized in that, The dynamic adjustment of the calculation step size based on the residual change rate includes: When the rate of change of the residual is greater than the change threshold, the calculation step size is reduced to the step size value of one. When the rate of change of the residual is less than the change threshold of two, the calculation step size is increased to the step size value of two. When the rate of change of the residual is less than or equal to the first threshold and the rate of change of the residual is greater than or equal to the second threshold, the calculation step size remains unchanged; wherein, the first threshold is greater than the second threshold, and the second step size is greater than the first step size.

5. The method for assessing voltage over-limit risk in a low-voltage distribution transformer area according to claim 1, characterized in that, The generation of predicted transmission paths based on the distribution topology map of the transformer substation includes: The topological spatial features of the distribution topology map of the transformer substation are extracted using the GCN network model; Extract the voltage from the multi-source data of the transformer area within the historical Q minutes, and concatenate them into a voltage time series according to the chronological order; where Q is an integer, Q>0; Voltage time series and topological spatial features are input into the conduction prediction model to obtain the predicted conduction path; the conduction prediction model is constructed using an LSTM model to predict the optimal conduction path of the fault.

6. The method for assessing voltage over-limit risk in a low-voltage distribution transformer area according to claim 1, characterized in that, The voltage anomaly tracing results generated based on the predicted conduction path include: Extract the voltage corresponding to several node IDs in the predicted conduction path; The correlation coefficients of voltage fluctuations at several nodes were calculated using the Pearson correlation coefficient method. Select several node IDs whose absolute values ​​of the correlation coefficients are greater than the coefficient threshold as candidate source points; Calculate the slope matching degree (SMD) of several candidate source points; the SMD is calculated using a formula that satisfies: , where N total N represents the total number of local time segments into which the time window is divided. match The number of local time segments required to satisfy the slope matching condition; Several candidate source points with slope matching degree greater than the matching threshold are selected as candidate source points. Extract the voltage fluctuation anomaly start time points of several candidate source points; The earliest candidate source point at which the voltage fluctuation anomaly begins is identified as the fault node ID. The source of the voltage anomaly was determined based on the fault node ID.

7. The method for assessing voltage over-limit risk in a low-voltage distribution transformer area according to claim 1, characterized in that, The comprehensive risk index generated based on tidal flow data includes: Extracting the rated transformer capacity P from the multi-source data of the distribution area N And the active power and line impedance Z in the power flow data. line ; Define three-dimensional evaluation metrics; these metrics include load fluctuation rate, impedance matching degree, and capacity-to-load ratio. Extracting active power P within the sliding window t t represents the time point number within the sliding window; The load volatility FBL is calculated using a formula; the formula satisfies: ;in, This is expressed as the average active power within the sliding window; Through formula Calculate the impedance matching degree ZPD; where Z base Represented as a reference impedance value; Through formula Calculate the capacity ratio RZB; P load，j This represents the typical operating load corresponding to the j-th node ID; J represents the total number of node IDs. The comprehensive risk index is obtained by weighting and summing the three-dimensional evaluation indicators using indicator weight coefficients; the indicator weight coefficients are obtained through the following methods: Each three-dimensional evaluation index is divided into 5 levels and a triangular membership function is defined to quantify the membership degree; 100,000 sets of random samples covering historical operating scenarios were generated through Monte Carlo simulation, and the initial index weights were calculated by combining the entropy weight method and correlation analysis. The objective function is optimized using the gradient descent algorithm to ultimately determine the index weight coefficients; the objective function satisfies: Where LYS is the number of historical samples, CZQ is the initial indicator weight to be optimized, and y s A comprehensive risk index is set for the s-th historical sample. Calculate the comprehensive risk index for the s-th historical sample.

8. The method for assessing voltage over-limit risk in a low-voltage distribution transformer area according to claim 1, characterized in that, The alarm signal generated based on the three-phase total voltage imbalance and comprehensive risk index includes: When the three-phase total voltage imbalance exceeds the imbalance threshold, a three-phase total voltage abnormality alarm signal is generated; the imbalance threshold is set based on experience. When the comprehensive risk index is within the normal range, a normal voltage signal is generated; When the comprehensive risk index is within the attention index range, a voltage anomaly attention signal is generated; When the comprehensive risk index is within the alarm index range, a voltage abnormality alarm signal is generated; When the comprehensive risk index is within the emergency index range, a voltage abnormality emergency signal is generated; the normal index range, attention index range, alarm index range, and emergency index range are set based on experience.

9. A voltage over-limit risk assessment device for a low-voltage distribution transformer area, applied to the voltage over-limit risk assessment method for a low-voltage distribution transformer area as described in claim 1, characterized in that, include: Data acquisition unit, data analysis unit, and early warning unit; The data acquisition unit is connected to the data analysis unit; the data analysis unit is connected to the early warning unit. The data acquisition unit acquires multi-source data of the low-voltage distribution transformer area through data acquisition equipment. The data analysis unit: constructs a power distribution topology map of the transformer area based on multi-source data of the transformer area; Predicted load data is generated based on multi-source data from the transformer area; power flow data is generated based on multi-source data from the transformer area, the transformer area distribution topology map, and the predicted load data; the power flow data refers to the result of reflecting the predicted load data to the current moment; three-phase total voltage imbalance is generated based on multi-source data from the transformer area; voltage is extracted from the power flow data; voltage drop is determined based on voltage and standard voltage; the voltage drop includes the three-phase voltage drop. When the voltage drop of any phase exceeds the amplitude threshold, a predicted conduction path is generated based on the distribution topology of the transformer area; and a voltage anomaly source tracing result is generated based on the predicted conduction path. A comprehensive risk index is generated based on trend data; An alarm signal is generated based on the three-phase total voltage imbalance and the comprehensive risk index; The early warning unit: provides corresponding prompts based on the alarm signal.

Images