A method and device for predicting active load of a low-voltage distribution area

Abstract

The application provides an active load prediction method and device for low-voltage distribution area, relates to the field of area load prediction, and solves the technical problem that the prediction error will increase sharply when facing the nonlinear response of temperature mutation influence. The method comprises the following steps: collecting and cleaning the multi-source data of the area to obtain electrical measurement data; obtaining meteorological forecast data of the area and performing tensor conversion to obtain a three-dimensional tensor; obtaining an electricity regulation policy, aligning the effective time of the electricity regulation policy with the three-dimensional tensor in time sequence to form a bimodal input matrix, inputting the bimodal input matrix and the electrical measurement data into a multi-factor coupling probability prediction model, and outputting a load prediction value. The application is used in the active load prediction process of the low-voltage distribution area. Unstructured data such as meteorological parameters and policy texts are converted into quantifiable features, and a multi-factor coupling prediction model is used to realize high-precision load prediction within a time period of 48, 72 hours, etc.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of transformer area load forecasting, in particular to an active load forecasting method and device for low-voltage distribution transformer area. BACKGROUND

[0002] With the accelerated construction of new power systems, low-voltage distribution transformer areas are facing unprecedented operating pressure. Under the background of rapid urbanization and electrification, the proportion of impact loads such as central air conditioning, energy storage devices in commercial complexes, data centers and intelligent building groups continues to rise. In some coastal cities, the daily active load peak-valley difference rate of low-voltage transformer areas can reach 3-5 times the rated capacity. Especially during the summer peak power consumption period, regional voltage sag events occur frequently, directly causing major economic losses such as precision equipment downtime and production line shutdown. More seriously, the large-scale access of new loads such as distributed photovoltaic and electric vehicle charging piles has changed the traditional one-way power supply-based distribution network structure to a multi-source two-way interactive mode. The three-phase imbalance of active power presents a complex spatial and temporal distribution trend, and the negative sequence voltage imbalance of typical transformer areas can reach more than 4.2%. Under the condition of continuous overload of 20%, the additional loss of distribution transformers increases by 12%-25%, and the average service life of equipment is shortened by 3-5 years.

[0003] Although mainstream load forecasting methods such as ARIMA and support vector regression perform well in stable load scenarios, the prediction error increases dramatically to more than 12% when facing the nonlinear response of air conditioning load affected by sudden temperature changes. In terms of state assessment, static power flow-based analysis methods cannot capture the impact of millisecond-level load fluctuations. SUMMARY

[0004] The present application provides an active load forecasting method and device for low-voltage distribution transformer area, which solves the technical problem of rapid increase of prediction error when facing the nonlinear response of sudden temperature changes.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0006] In a first aspect, an active load forecasting method for low-voltage distribution transformer area is provided, comprising:

[0007] Collecting and cleaning the multi-source data of the transformer area to obtain electrical measurement data;

[0008] Obtaining meteorological forecast data of the location of the transformer area and performing tensor conversion to obtain a three-dimensional tensor; wherein the meteorological forecast data includes temperature, humidity and wind speed;

[0009] An electric regulation policy is acquired, semantic vectors and the effective time of the electric regulation policy are extracted through a BERT model, the effective time of the electric regulation policy is time-aligned with a three-dimensional tensor, and a double-modal input matrix is formed;

[0010] The double-modal input matrix and the electrical measurement data are input into a multi-factor coupling probability prediction model to output a load prediction value and a confidence interval boundary, wherein the multi-factor coupling probability prediction model is constructed based on a physical information neural network (PINN) and an LSTM-CRF model.

[0011] Based on the above technical solution, in the active load prediction method for a low-voltage distribution area provided in the application, unstructured data such as meteorological parameters and policy texts are converted into quantifiable features, and then a multi-factor coupling prediction framework is used to realize high-precision load prediction within a future time n.

[0012] In combination with the first aspect, in a possible implementation manner, the acquisition manner of the electrical measurement data includes:

[0013] Multi-source heterogeneous data are acquired through integrated smart meters, wherein the multi-source heterogeneous data include voltage, current and active power;

[0014] The multi-source heterogeneous data are aligned and calibrated in time and space dimensions through an adaptive Kalman filtering algorithm to obtain calibrated data;

[0015] The calibrated data are processed through a hybrid data cleaning pipeline to obtain electrical measurement data, wherein the hybrid data cleaning pipeline includes an isolation forest algorithm, a K-nearest neighbor algorithm, a wavelet transform, a hash time-locked contract and a knowledge graph.

[0016] It should be noted that the electrical measurement data are a dynamic device association graph of data-topology-confidence, which provides a high-quality input basis for subsequent prediction.

[0017] In combination with the first aspect, in a possible implementation manner, the processing of the calibrated data through the hybrid data cleaning pipeline includes:

[0018] The calibrated data are detected and abnormal data are removed through an isolation forest algorithm, missing points in the calibrated data from which the abnormal data are removed are filled through a K-nearest neighbor algorithm, and the filled calibrated data are filtered through a wavelet transform to obtain modified data;

[0019] The modified data are processed through a hash time-locked contract to obtain data blocks with confidence labels, a device association model is constructed based on a knowledge graph, a distribution area physical topology is obtained by analyzing a distribution area topology relationship through a graph attention network, and the data blocks with confidence labels and the distribution area physical topology are bound to obtain electrical measurement data.

[0020] In conjunction with the first aspect above, in one possible implementation, the method for obtaining the three-dimensional tensor includes:

[0021] The Z-score standardization method was used to eliminate the dimensional differences in meteorological forecast data, and the meteorological forecast data were sorted according to the collection time to obtain the forecast sequence;

[0022] pass The predicted sequence is processed to obtain a three-dimensional tensor; where, , , They represent the first Standardized temperature, humidity, and wind speed for hours, where n is the future time of the weather forecast data, t∈n.

[0023] In conjunction with the first aspect above, in one possible implementation, the construction method of the multi-factor coupled probability prediction model includes:

[0024] A spatiotemporal feature extractor is constructed based on the Transformer-XL architecture and by introducing relative position encoding and gated attention mechanisms.

[0025] The thermodynamic equations are transformed into differentiable constraint terms using a physical information neural network to construct a residual loss function.

[0026] A load change inflection point layer is constructed based on the LSTM-CRF model and by introducing the historical feature difference of the lag time window.

[0027] A joint probability distribution layer of meteorological, social, and load is constructed based on the Gaussian Copula function;

[0028] The computational output layer is constructed based on Monte Carlo sampling.

[0029] It should be noted that long-term temporal dependencies can be captured by introducing a positional encoding formula.

[0030] The gating attention mechanism enables the model to adaptively adjust feature importance according to the characteristics of the prediction period; the temporal correlation features captured by the LSTM-CRF model are used as "additional constraints" and incorporated into subsequent multi-factor coupling modeling to ensure the temporal logic consistency of the coupled prediction.

[0031] In conjunction with the first aspect above, in one possible implementation, the relative position is encoded as follows: , where A i,j q represents the attention score from position i to j; i and j are the indices of the i-th and j-th positions in the sequence; i Let k be the query vector at the i-th position; jLet R be the key vector at the j-th position; i-j This is a relative position encoding matrix, where w represents the relative distance between encoding positions i and j. R It is a learnable relative position weight vector.

[0032] In conjunction with the first aspect above, in one possible implementation, the gate-controlled attention mechanism is as follows: ; The weight matrix is ​​a learnable matrix. This is a vector concatenated from meteorological and policy characteristics. It is the sigmoid activation function.

[0033] It should be noted that, This is a hidden state related to meteorological characteristics. This refers to a hidden state related to policy characteristics.

[0034] In conjunction with the first aspect above, in one possible implementation, the residual loss function is: , Let be the partial derivative of the load with respect to temperature, and k be the thermal conductivity coefficient. For the outside temperature, To set the temperature, This is the regularization coefficient, which is the weight that balances prediction error and physical constraints.

[0035] In conjunction with the first aspect mentioned above, one possible implementation also includes: obtaining a load change risk warning signal based on the load forecast value.

[0036] In conjunction with the first aspect above, in one possible implementation, obtaining the load change risk warning signal based on the load forecast value includes:

[0037] The load current signal is decomposed into three levels using wavelet packet transform to obtain time-frequency energy sequences of several sub-bands. Constructing a multi-scale detection index based on different frequency bands—energy gradient change rate. The normalized rate of change of energy gradient is calculated by weighted summation. Among them, the rate of change of energy gradient Through calculation formula Calculated; For time intervals; For the k-th sub-band at t- Energy at time t; Let Ek,t be the variance of the energy sequence of the kth sub-band over the entire time range [1,T]. Let k be the rate of change of the energy gradient of the k-th sub-band. The energy gradient change rate index for node i;

[0038] Constructing a double-input fuzzy inference system: gradient mean with x-second window length and variance as inputs, defining a dynamic threshold fuzzy rule base, and solving the fuzzy output by the barycenter method to update the dynamic threshold at the minute level , defining a risk level fuzzy rule base according to the dynamic threshold and the energy gradient change rate to match the risk level L;

[0039] The risk source node is obtained by establishing a risk propagation equation to quantify the spatiotemporal diffusion law of the mutation signal in the topology of the transformer area;

[0040] The risk level L, the propagation speed v, and the influence range S are set to the hue , the saturation , and the transparency respectively to obtain the dynamic color code of the warning; wherein the propagation speed v is obtained by the calculation formula , dij is the electrical distance between nodes i and j, Δtij is the time difference between the energy gradient change rate of node i and node j exceeding the set threshold for the first time, and N is the number of adjacent nodes of node j; the influence range S is the number of nodes with risk influence score greater than the set threshold.

[0041] In combination with the first aspect, in a possible implementation manner, the risk source node is a node corresponding to the maximum risk influence score; wherein the risk influence score is calculated by the calculation formula ; wherein Rj(t) is the risk influence score of node j at time t, N(j) is the adjacent node set of node j, is the electrical coupling weight from node i to j, tonset is the mutation start time, and a is the time decay factor. It should be noted that the second-level visualization is realized by superimposing the SCADA system on the geographical wiring diagram, which provides a solution for micro-signal perception to macro-risk warning for high-density commercial load scenarios.

[0042] In combination with the first aspect, in a possible implementation manner, the risk source node is a node corresponding to the maximum risk influence score; wherein the risk influence score is calculated by the calculation formula ; wherein Rj(t) is the risk influence score of node j at time t, N(j) is the adjacent node set of node j, is the electrical coupling weight from node i to j, tonset is the mutation start time, and a is the time decay factor.

[0043] In a second aspect, an electronic device is provided, comprising: a communication unit and a processing unit; the communication unit is configured to collect multi-source data of a transformer area; obtain meteorological forecast data of the location of the transformer area; obtain an electric control policy;

[0044] The processing unit is configured to clean the multi-source data of the transformer area to obtain electrical measurement data, and perform tensor conversion on the weather forecast data to obtain a three-dimensional tensor.

[0045] The semantic vector in the electricity regulation policy and the effective time of the electricity regulation policy are extracted by the BERT model, the effective time of the electricity regulation policy is time-aligned with the three-dimensional tensor, and a dual-modal input matrix is formed.

[0046] The dual-modal input matrix and the electrical measurement data are input into a multi-factor coupling probability prediction model to output a load prediction value and a confidence interval boundary; wherein the multi-factor coupling probability prediction model is constructed based on a physical information neural network (PINN) and an LSTM-CRF model.

[0047] In a third aspect, the present application provides a processing device, comprising a processor and a storage medium; the storage medium comprises instructions, and the processor is configured to execute the instructions to implement the method described in the first aspect and any possible implementation manner of the first aspect. The processing device can be an electronic device, or a chip in an electronic device.

[0048] In a fourth aspect, the present application provides an active load prediction system for a low-voltage distribution transformer area, comprising an information acquisition module, a data processing module and a prediction module; wherein the information acquisition module is configured to acquire multi-source data of the transformer area.

[0049] In a fifth aspect, the present application provides a computer readable storage medium, which stores instructions, and when the instructions are executed on an electronic device, the electronic device executes the method described in the first aspect and any possible implementation manner of the first aspect.

[0050] In a sixth aspect, the present application provides a computer program product comprising instructions, and when the computer program product is executed on an electronic device, the electronic device executes the method described in the first aspect and any possible implementation manner of the first aspect.

[0051] The application provides an active load prediction method and device for a low-voltage distribution area, which can break through the traditional data barrier through a multi-source heterogeneous data fusion perception system, deploy intelligent electric meters, environmental sensors and other multi-source terminals at edge computing nodes, realize spatio-temporal data alignment by using an adaptive Kalman filtering algorithm, ensure data credibility in combination with a blockchain hash time lock contract, analyze the topology relationship of the area by using a graph attention network (GAT), and construct a dynamic device association graph with a credibility label, so as to provide a high-precision data base for an upper-layer model. On this basis, a meteorological-social multi-dimensional load prediction model innovatively fuses a Transformer-XL architecture and a physical information neural network (PINN), converts unstructured data such as meteorological parameters and policy texts into quantifiable features, captures a load lagging effect by using an LSTM-CRF model, constructs a multi-factor coupling prediction model based on a Copula theory, and realizes high-precision load prediction in a time period of 48, 72 hours and the like, and the prediction dimension is expanded from single power consumption to comprehensive environmental and social influence analysis.

[0052] It should be understood that the description of technical features, technical solutions, beneficial effects or similar language in the present application does not imply that all features and advantages can be realized in any single embodiment. On the contrary, it can be understood that the description of a feature or beneficial effect means that the specific technical feature, technical solution or beneficial effect is included in at least one embodiment. Therefore, the description of technical features, technical solutions or beneficial effects in the specification does not necessarily refer to the same embodiment. Further, the technical features, technical solutions and beneficial effects described in the embodiments can be combined in any appropriate manner. Those skilled in the art will understand that the embodiments can be implemented without one or more specific technical features, technical solutions or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects can be identified in specific embodiments that do not embody all embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 A system architecture diagram of an active load prediction system for a low-voltage distribution area is provided for embodiments of the present application;

[0054] Figure 2 A flowchart of an active load prediction method for a low-voltage distribution area is provided for embodiments of the present application;

[0055] Figure 3 A flowchart of obtaining electrical measurement data is provided for embodiments of the present application;

[0056] Figure 4 A load probability prediction effect comparison diagram is provided for embodiments of the present application;

[0057] Figure 5This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application;

[0058] Figure 6 This is a schematic diagram of the hardware structure of a processing device provided in an embodiment of this application. Detailed Implementation

[0059] In the description of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. The "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" means one or more, and "multiple" means two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0060] It should be noted that, in this application, the terms "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0061] The active load prediction method for low-voltage distribution substations provided in this application can be applied to, for example... Figure 1 In the active load prediction system 100 of the low-voltage distribution substation shown, such as Figure 1 As shown, the communication system includes: an integrated smart meter 101, an environmental monitoring device 102, and a cloud computer 103.

[0062] Among them, the integrated smart meter 101 is used to collect multi-source data of the transformer area and send it to the cloud computer 103;

[0063] Environmental monitoring equipment 102 is used to acquire meteorological forecast data of the location of the station area and send it to cloud computer 103;

[0064] The cloud computer 103 is used to clean multi-source data from the transformer substation to obtain electrical measurement data; and to perform tensor transformation on meteorological forecast data to obtain three-dimensional tensors.

[0065] The semantic vectors and effective time of the electricity regulation policy are extracted by the BERT model. The effective time of the electricity regulation policy is then temporally aligned with the three-dimensional tensor to form a bimodal input matrix.

[0066] The bimodal input matrix and the electrical measurement data are input into a multi-factor coupling probability prediction model, and a load prediction value and a confidence interval boundary are output.

[0067] To solve the technical problem that the prediction error increases sharply when facing the nonlinear response of temperature mutation in the prior art, an active load prediction method for a low-voltage distribution area is provided, which includes: collecting and cleaning multi-source data of the area to obtain electrical measurement data;

[0068] Meteorological forecast data of the area are obtained and tensor conversion is performed to obtain a three-dimensional tensor; the meteorological forecast data include temperature, humidity and wind speed;

[0069] An electrical regulation policy is obtained, and a semantic vector and an effective time of the electrical regulation policy are extracted through a BERT model; the effective time of the electrical regulation policy is time-aligned with the three-dimensional tensor to form a bimodal input matrix;

[0070] The bimodal input matrix and the electrical measurement data are input into a multi-factor coupling probability prediction model, and a load prediction value and a confidence interval boundary are output. Based on this, high-precision load prediction in future time n is realized.

[0071] As shown in Figure 2 A kind of active load prediction method for low-voltage distribution area provided in the present application embodiment, including:

[0072] S201, the multi-source data of the area are collected and cleaned to obtain electrical measurement data.

[0073] The multi-source heterogeneous data include voltage, current and active power.

[0074] In some implementations, the multi-source heterogeneous data are obtained by integrating smart meters; the multi-source heterogeneous data include voltage, current and active power.

[0075] The multi-source heterogeneous data are aligned and calibrated in time and space dimensions by an adaptive Kalman filtering algorithm to obtain calibrated data.

[0076] The calibrated data are processed by a hybrid data cleaning pipeline to obtain electrical measurement data; the hybrid data cleaning pipeline includes isolated forest algorithm, K nearest neighbor, wavelet transform, hash time lock contract and knowledge graph.

[0077] Further, the processing of the calibrated data by the hybrid data cleaning pipeline includes:

[0078] The isolated forest algorithm is used to detect and remove outliers from the calibration data. The K-nearest neighbor algorithm is used to fill in the missing points in the calibration data after removing outliers. Finally, wavelet transform is used to filter the filled calibration data to obtain the corrected data.

[0079] The corrected data is processed by a hash time-locked contract to obtain data blocks with trust labels; a device association model is constructed based on a knowledge graph, and the transformer area topology relationship is analyzed by a graph attention network to obtain the transformer area physical topology; the data blocks with trust labels are bound to the transformer area physical topology to obtain electrical measurement data.

[0080] It should be noted that electrical measurement data is a dynamic device correlation map integrating data, topology, and reliability, providing a high-quality input basis for subsequent predictions.

[0081] S202. Obtain meteorological forecast data for the location of the substation area and perform tensor transformation to obtain a three-dimensional tensor.

[0082] The meteorological forecast data includes temperature, humidity, and wind speed.

[0083] In some implementations, the Z-score normalization method is used to eliminate the dimensional differences in meteorological forecast data, and the meteorological forecast data is sorted according to the collection time to obtain the forecast sequence;

[0084] pass The predicted sequence is processed to obtain a three-dimensional tensor; where, , , They represent the first Standardized temperature, humidity, and wind speed for hours, where n is the future time of the weather forecast data, t∈n.

[0085] S203. Obtain the electricity regulation policy, and extract the semantic vector and the effective time of the electricity regulation policy through the BERT model. Align the effective time of the electricity regulation policy with the three-dimensional tensor in time to form a dual-modal input matrix.

[0086] It should be noted that, based on the policy effective time window, the semantic vector is time-series aligned with the standardized meteorological data three-dimensional tensor to ensure that policy features and meteorological features match in the time dimension, forming a "meteorological-social" dual-modal input matrix.

[0087] S204. Input the dual-modal input matrix and electrical measurement data into the multi-factor coupled probability prediction model, and output the load prediction value and confidence interval boundary.

[0088] The multi-factor coupled probability prediction model is constructed based on the physical information neural network PINN and the LSTM-CRF model.

[0089] In some implementations, the multi-factor coupling probability prediction model is constructed in the following manner:

[0090] A space-time feature extractor is constructed based on the Transformer-XL architecture and by introducing relative position encoding and a gate-controlled attention mechanism;

[0091] A residual loss function is constructed by converting a thermodynamic equation into a differentiable constraint term through a physical information neural network;

[0092] A load change inflection point layer is constructed based on an LSTM-CRF model and by introducing a historical feature difference of a lag time window;

[0093] A joint probability distribution layer of meteorology-social load is constructed based on a Gaussian Copula function;

[0094] A calculation output layer is constructed based on Monte Carlo sampling.

[0095] It should be noted that the relative position encoding formula is introduced to solve the problem of insufficient dependence relationship capture of long time series load by traditional Transformer; and is a learnable relative position matrix, and is a learnable relative position weight vector.

[0096] The gate-controlled attention mechanism enables the model to adaptively adjust the feature importance according to the characteristics of the prediction period; and in view of the lag effect of load change (such as air conditioning refrigeration which needs 1-2 hours to reach stable load), a long short-term memory network space-time conditional random field (LSTM-CRF) model is adopted, a historical load feature difference of 6 hours is introduced, a load change inflection point is accurately positioned through sequence probability calculation of the LSTM-CRF layer, and the problem of ignoring load response delay by traditional models is solved. The time series correlation features captured by the LSTM-CRF model are taken as “additional constraints” and are integrated into subsequent multi-factor coupling modeling to ensure the time series logical consistency of coupling prediction.

[0097] The relative position encoding is where A i,j is an attention score of position i to j; i and j are indexes of the i-th and j-th positions in the sequence; q i is a query vector of the i-th position; k j is a key vector of the j-th position; R i-j is a relative position encoding matrix, encoding the relative distance between positions i and j; w R is a learnable relative position weight vector.

[0098] Further, the gate-controlled attention mechanism is ; is a learnable weight matrix, is a meteorological and policy feature concatenation vector, is a sigmoid activation function.

[0099] Further, the residual loss function is , is the partial derivative of load with respect to temperature, k is the thermal conductivity, is the outside temperature, is the set temperature, is a regularization coefficient, through which the model is forced to“learn physical logic” during training – to avoid the pure data-driven model outputting results that violate physical common sense due to historical data noise (such as abnormal electricity meter readings, short-term load fluctuations), ensuring that the prediction results are consistent with the physical logic of the actual electricity consumption scenario.

[0100] Based on the above technical solutions, the active load prediction method for a low-voltage distribution area provided by the application further comprises: obtaining a load mutation risk early warning signal according to the load prediction value.

[0101] In some implementations, the load mutation risk early warning signal is obtained according to the load prediction value, comprising:

[0102] The load current signal is decomposed into three layers by wavelet packet transform to obtain time-frequency energy sequences of a plurality of sub-bands ; a multi-scale detection index, energy gradient change rate , is constructed based on different frequency bands, and the normalized energy gradient change rate is calculated by weighted summation; is calculated by the calculation formula ; is a time interval; is the energy of the kth sub-band at t- t moment; is the variance of the energy sequence Ek,t of the kth sub-band in the entire time range [1, T]; is the energy gradient change rate of the kth sub-band, is the energy gradient change rate index of node i;

[0103] A double-input fuzzy inference engine is constructed: the gradient mean and variance of the x-second window are taken as inputs, and a dynamic threshold fuzzy rule base is defined (for example: when the gradient mean and variance When all values ​​are greater than 3, the dynamic alarm threshold is 9. The dynamic threshold is updated minute-by-minute by defuzzifying the output using the centroid method. Define a fuzzy rule base for risk levels (e.g.: >8 and When the value is greater than 4, the risk level is 8, based on the dynamic threshold and the rate of change of the energy gradient. Matching risk level L;

[0104] By establishing a risk propagation equation, the spatiotemporal diffusion pattern of mutation signals in the transformer area topology is quantified to obtain the risk source node;

[0105] The risk level L, the propagation speed v, and the impact range S should each be assigned a color tone. saturation and transparency The dynamic color code for early warning is obtained; wherein, the propagation speed v is calculated using the formula... Get, d ij Let Δt be the electrical distance between nodes i and j. ij The time difference between the first time the rate of change of the energy gradient of nodes i and j exceeds a set threshold is defined as N, where N is the number of neighboring nodes of node j; and the influence range S is the number of nodes whose risk influence score is greater than the set threshold.

[0106] It should be noted that by overlaying the SCADA system onto the geographic wiring diagram to achieve second-level visualization, a solution from micro-signal perception to macro-risk early warning is provided for high-density commercial load scenarios.

[0107] Furthermore, the risk source node is the node corresponding to the maximum risk impact score; wherein, the risk impact score is calculated using the formula... The calculation yields Rj(t), where Rj(t) is the risk influence score of node j at time t, and N(j) is the set of neighboring nodes of node j. Let be the electrical coupling weight from node i to j, tonset be the mutation start time, and α be the time decay factor.

[0108] To address power grid safety risks, the load mutation risk early warning system innovatively integrates wavelet packet transform and mutation detection theories, constructs multi-scale gradient monitoring indicators, combines a digital twin platform to perform risk propagation simulation, and develops an HSV color space early warning coding scheme to achieve a 10-second anomaly detection response.

[0109] For example, please refer to Figure 4As shown, the actual load curve in the figure shows a significant peak fluctuation from 18:00 on August 1 to 7:00 the next day (annotated in the gray area) due to the passage of a typhoon, with the maximum load value reaching 693 MW, an increase of 38.6% over the base load. The dynamic gate control attention mechanism of the Transformer-XL architecture automatically increases the weight of meteorological features to 0.67 (policy feature weight 0.33) at the typhoon warning release time, and the predicted mean curve (red dashed line) accurately tracks the load trend, showing an upward trend 5 hours before the mutation (13:00 on August 1), with a peak prediction error of only 1.8%, and perfectly reproducing the gradient decay characteristics caused by policy regulation during the load decline phase. It is worth noting that the 95% confidence interval (orange area) generated by the Copula probability prediction of the model intelligently broadens during the core period of the typhoon, with a maximum width of ±60 MW, accurately quantifying the double uncertainty caused by the sudden change in temperature and the lag in policy implementation. In contrast, the error band of traditional deterministic prediction methods is fixed at ±20 MW in this scenario, and cannot reflect the dynamic changes in risk.

[0110] The above describes the scheme of the embodiments of the present application mainly from the perspective of device implementation. It can be understood that each device, for example, an electronic device, includes at least one of a corresponding hardware structure and a software module for executing each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of the examples described in the embodiments disclosed herein, the present application can be realized in the form of hardware or a combination of hardware and computer software. Whether a certain function is driven by hardware or computer software to drive hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods for each specific application to implement the described functions, but such implementation should not be considered beyond the scope of the present application.

[0111] The embodiments of the present application can divide the functional units of the electronic device according to the above method examples, for example, each functional unit can be divided according to each function, or two or more functions can be integrated in one processing unit. The integrated unit can be realized in the form of hardware or software functional unit. It should be noted that the division of units in the embodiments of the present application is illustrative, and is only a logical functional division. Actual implementation can have another division method.

[0112] In the case of integrated units, Figure 5 A possible structure schematic diagram of the electronic device (denoted as electronic device 50) involved in the above embodiments is shown, which includes a processing unit 501 and a communication unit 502, and can also include a storage unit 503. Figure 5 The structure schematic diagram shown can be used to illustrate the structure of the electronic device involved in the above embodiments.

[0113] When Figure 5 The structure diagram shown is used to illustrate the structure of the electronic device involved in the above embodiment, the processing unit 501 is used to control and manage the action of the electronic device, the communication unit 502 is used for communication between the electronic device and other devices, and the storage unit 503 is used to store the program code and data of the electronic device.

[0114] For example, the communication unit 502 is used to collect multi-source data of the transformer area; obtain meteorological forecast data of the location of the transformer area; and obtain electric regulation and control policy.

[0115] The processing unit 501 is used to clean the multi-source data of the transformer area to obtain electrical measurement data; and perform tensor conversion on the meteorological forecast data to obtain a three-dimensional tensor.

[0116] The semantic vector in the electric regulation and control policy and the effective time of the electric regulation and control policy are extracted by the BERT model, the effective time of the electric regulation and control policy is time-aligned with the three-dimensional tensor, and a dual-modal input matrix is formed.

[0117] The dual-modal input matrix and the electrical measurement data are input into a multi-factor coupling probability prediction model to output a load prediction value and a confidence interval boundary; wherein the multi-factor coupling probability prediction model is constructed based on a physical information neural network PINN and an LSTM-CRF model.

[0118] The processing unit 501 can be a processor or a controller, and the communication unit 502 can be a communication interface, a transceiver, a transceiver, a transceiver circuit, a transceiver device, etc. The communication interface is a general term and can include one or more interfaces. The storage unit 503 can be a memory. When the electronic device 50 is a chip, the processing unit 501 can be a processor or a controller, and the communication unit 502 can be an input interface and / or an output interface, a pin or a circuit, etc. The storage unit 503 can be a storage unit (e.g., a register, a cache, etc.) within the chip, or can be a storage unit (e.g., a read-only memory (ROM), a random access memory (RAM), etc.) located outside the chip.

[0119] The communication unit can also be referred to as a transceiver unit. The antenna and the control circuit with transceiving functions in the electronic device 50 can be regarded as a communication unit 502 of the electronic device 50, and the processor with processing functions can be regarded as a processing unit 501 of the electronic device 50. Optionally, the device for realizing the receiving function in the communication unit 502 can be regarded as a communication unit, which is used to perform the receiving steps in the embodiments of the present application, and the communication unit can be a receiver, a receiver, a receiving circuit, etc. The device for realizing the sending function in the communication unit 502 can be regarded as a sending unit, which is used to perform the sending steps in the embodiments of the present application, and the sending unit can be a transmitter, a sender, a sending circuit, etc.

[0120] Figure 5 The integrated units in the above embodiments can be stored in a computer readable storage medium if they are realized in the form of software function modules and sold or used as independent products. Based on such understanding, the technical solutions of the embodiments of the present application can be embodied in the form of software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods described in the embodiments of the present application. The storage medium storing the computer software product includes a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and various media that can store program codes.

[0121] Figure 5 The units in the above embodiments can also be referred to as modules, for example, the processing unit can be referred to as a processing module.

[0122] The embodiments of the present application also provide a hardware structure diagram of a processing device (denoted as processing device 60), which is shown in Figure 6 The processing device 60 includes a processor 601, and optionally further includes a memory 602 connected with the processor 601.

[0123] In a first possible implementation, referring to Figure 6 The processing device 60 further includes a transceiver 603. The processor 601, the memory 602 and the transceiver 603 are connected through a bus. The transceiver 603 is used for communicating with other devices or communication networks. Optionally, the transceiver 603 can include a transmitter and a receiver. The device for realizing the receiving function in the transceiver 603 can be regarded as a receiver, which is used to perform the receiving steps in the embodiments of the present application. The device for realizing the sending function in the transceiver 603 can be regarded as a transmitter, which is used to perform the sending steps in the embodiments of the present application.

[0124] Based on the first possible implementation, Figure 6 The structural schematic diagram shown can be used to show the structure of the processing apparatus involved in the above embodiments.

[0125] Among them, Figure 6 The system chip in the processing apparatus can also be shown. In this case, the actions performed by the processing apparatus described above can be implemented by the system chip, and the specific actions performed can be referred to above and will not be described here again.

[0126] In the implementation process, each step in the method provided by the embodiment can be completed by the integrated logic circuit of hardware in the processor or the instruction in the form of software. The steps of the method disclosed by the embodiment of the present application can be directly embodied as hardware processor execution completion, or executed by the combination of hardware and software modules in the processor.

[0127] The processor in the present application can include but is not limited to at least one of the following: central processing unit (CPU), microprocessor, digital signal processor (DSP), microcontroller unit (MCU), or artificial intelligence processor and various types of computing devices running software, each of which can include one or more cores for executing software instructions to perform operations or processing. The processor can be a separate semiconductor chip, or can be integrated with other circuits as a semiconductor chip, for example, it can constitute a SoC (system on chip) with other circuits (such as coding and decoding circuits, hardware acceleration circuits or various bus and interface circuits), or it can be integrated as an internal processor in the ASIC, and the ASIC integrated with the processor can be packaged separately or packaged together with other circuits. In addition to including cores for executing software instructions to perform operations or processing, the processor can further include necessary hardware accelerators, such as field programmable gate array (FPGA), PLD (programmable logic device), or logic circuits implementing special logic operations.

[0128] The memory in the embodiments of the present application can include at least one of the following types: read-only memory (ROM) or other types of static storage devices that can store static information and instructions, random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, and can also be electrically erasable programmable read-only memory (EEPROM). In some scenarios, the memory can also be a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage medium or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and capable of being accessed by a computer, but not limited to this.

[0129] The embodiments of the present application also provide a computer readable storage medium including instructions, which, when executed on a computer, cause the computer to perform any of the above methods.

[0130] The embodiments of the present application also provide a computer program product including instructions, which, when executed on a computer, cause the computer to perform any of the above methods.

[0131] The embodiments of the present application also provide a chip, which includes a processor and an interface circuit, the interface circuit is coupled with the processor, the processor is used to run a computer program or instructions to implement the above method, and the interface circuit is used to communicate with other modules outside the chip.

[0132] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present application are generated. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer readable storage medium can be any available medium that can be accessed by a computer or data storage device including one or more servers, data centers, etc. integrated with the medium. The available medium can be magnetic medium (such as floppy disk, hard disk, magnetic tape), optical medium (such as DVD), or semiconductor medium (such as solid state disk (solid state disk, SSD)) and the like.

[0133] Although the present application is described herein in conjunction with various embodiments, other variations and modifications of the disclosed embodiments are possible in light of this disclosure, and it is intended that the scope of the application encompass such variations and modifications as can come within the scope of the claims. In the claims, the term "comprising" does not exclude other elements or steps, the term "one" or "a" does not exclude a plurality, and the term "one" or "a" means "one or more but can include only one". Individual features or units can implement several functions of several claim items. Measures described in mutually different dependent claims can be combined and implemented to produce desirable results.

[0134] Although the present application is described herein in conjunction with various embodiments, other variations and modifications of the disclosed embodiments are possible in light of this disclosure, and it is intended that the scope of the application encompass such variations and modifications as can come within the scope of the claims. In the claims, the term "comprising" does not exclude other elements or steps, the term "one" or "a" does not exclude a plurality, and the term "one" or "a" means "one or more but can include only one". Individual features or units can implement several functions of several claim items. Measures described in mutually different dependent claims can be combined and implemented to produce desirable results.

[0134] Although the present application is described herein in conjunction with various embodiments, other variations and modifications of the disclosed embodiments are possible in light of this disclosure, and it is intended that the scope of the application encompass such variations and modifications as can come within the scope of the claims. In the claims, the term "comprising" does not exclude other elements or steps, the term "one" or "a" does not exclude a plurality, and the term "one" or "a" means "one or more but can include only one". Individual features or units can implement several functions of several claim items. Measures described in mutually different dependent claims can be combined and implemented to produce desirable results.

Claims

1. A method for predicting active load of a low-voltage distribution area, characterized in that, The method comprises the following steps: Collecting and cleaning multi-source data of a transformer area to obtain electrical measurement data; Obtaining meteorological forecast data of the location of the transformer area and performing tensor conversion to obtain a three-dimensional tensor; Obtaining electric regulation and control policies, extracting semantic vectors and effective time of the electric regulation and control policies through a BERT model, time-aligning the effective time of the electric regulation and control policies with the three-dimensional tensor to form a double-modal input matrix; Inputting the double-modal input matrix and the electrical measurement data into a multi-factor coupled probability prediction model to output a load prediction value; wherein the multi-factor coupled probability prediction model is constructed based on a physical information neural network (PINN) and an LSTM-CRF model; The construction method of the multi-factor coupled probability prediction model comprises: Constructing a spatio-temporal feature extractor based on a Transformer-XL architecture and introducing relative position encoding and gate-controlled attention mechanism; Converting a thermodynamic equation into a differentiable constraint term through a physical information neural network to construct a residual loss function; Constructing a load change inflection point layer based on an LSTM-CRF model and introducing a historical feature difference value of a lag time window; Constructing a joint probability distribution layer of meteorology-society-load based on a Gaussian Copula function; Constructing a calculation output layer based on Monte Carlo sampling; The relative position encoding is ; wherein A i,j is an attention score of positions i to j; i, j are indices of the i-th and j-th positions in the sequence; q i is a query vector of the i-th position; k j is a key vector of the j-th position; R i-j is a relative position encoding matrix, encoding the relative distance of positions i and j; w R is a learnable relative position weight vector; The gate control attention mechanism is ; is a learnable weight matrix, is a weather and policy feature concatenation vector, is a sigmoid activation function; The residual loss function is , is the partial derivative of the load to the temperature, k is the thermal conductivity, is the ambient temperature, is the set temperature, is the regularization coefficient.

2. The method for predicting the active load of a low-voltage distribution area according to claim 1, characterized in that, The method for obtaining the electrical measurement data comprises: Obtaining multi-source heterogeneous data through an integrated smart meter; Calibrating the multi-source heterogeneous data in time and space dimensions through an adaptive Kalman filtering algorithm to obtain calibrated data; Processing the calibrated data through a hybrid data cleaning pipeline to obtain the electrical measurement data; wherein the hybrid data cleaning pipeline comprises an isolation forest algorithm, a K-nearest neighbor algorithm, a wavelet transform, a hash time-locked contract, and a knowledge graph.

3. A method of predicting active load for low voltage distribution area according to claim 2, characterized in that, The method for processing the calibrated data through the hybrid data cleaning pipeline comprises: Detecting and removing abnormal data from the calibrated data through an isolation forest algorithm, filling missing points in the calibrated data after removing abnormal data through a K-nearest neighbor algorithm, and filtering the filled calibrated data through a wavelet transform to obtain modified data; Processing the modified data through a hash time-locked contract to obtain data blocks with credibility labels; constructing a device association model based on a knowledge graph, analyzing the topology of the transformer area through a graph attention network to obtain a physical topology of the transformer area, and binding the data blocks with credibility labels and the physical topology of the transformer area to obtain the electrical measurement data.

4. The method for predicting the active load of a low-voltage distribution area according to claim 1, characterized in that, The method for obtaining the three-dimensional tensor comprises: Eliminating the dimensional differences of the meteorological forecast data by using a Z-score standardization method, sorting the meteorological forecast data according to the collection time to obtain a prediction sequence; wherein the meteorological forecast data includes temperature, humidity, and wind speed. By processing the forecast sequence to obtain a three-dimensional tensor; wherein, , , respectively represent the normalized temperature, humidity, wind speed of the tth hour, n is the future time of the meteorological forecast data, t ∈ n.

5. The method for predicting the active load of a low-voltage distribution area according to claim 1, characterized in that, The method further comprises: Obtaining a load mutation risk early warning signal according to the load prediction value.

6. A method of predicting active load for low voltage distribution area according to claim 5, characterized in that, The method for obtaining the load mutation risk early warning signal according to the load prediction value comprises: The load current signal is decomposed into three layers by wavelet packet transform to obtain time-frequency energy sequences of a plurality of sub-bands ; a multi-scale detection index, energy gradient change rate, is constructed based on different frequency bands , and a normalized energy gradient change rate is calculated by weighted summation ; wherein the energy gradient change rate is calculated by the calculation formula ; is a time interval ; Ek,t is the energy of the kth sub-band at time t ; and is the variance of the energy sequence Ek,t of the kth sub-band in the entire time range [1, T] ; and is the energy gradient change rate of the node i Constructing double-input fuzzy inference engine: with gradient mean of x seconds window length and variance as input, define dynamic threshold fuzzy rule base, solve fuzzy output by barycenter method minute-level updated dynamic threshold , define risk level fuzzy rule base, match risk level L according to dynamic threshold and energy gradient change rate ; Obtaining a risk source node by establishing a risk propagation equation to quantify the spatio-temporal diffusion law of the mutation signal in the topology of the transformer area. The risk level L, the propagation speed v and the influence range S are respectively set as the hue, the saturation and the transparency to obtain a dynamic color code of the early warning; wherein, the propagation speed v is obtained by the calculation formula , the saturation and the transparency , and the influence range S is the number of nodes whose risk influence scores are greater than a set threshold. , d ij is the electrical distance between nodes i and j, Δt ij is the time difference when the energy gradient change rate of nodes i and j exceeds a set threshold for the first time, and N is the number of adjacent nodes of node j.

7. A method of predicting active load for low voltage distribution area according to claim 6, characterized in that, The risk source node is a node corresponding to a maximum value of a risk influence score; wherein the risk influence score is calculated by a formula R (t) = R (t - 1) + (t - t j (t) is a risk influence score of the node j at the time t, N(j) is a set of adjacent nodes of the node j, is an electrical coupling weight of the node i to j, t onset is a mutation starting time, and a is a time attenuation factor.

8. An electronic device for applying to the active load forecasting method for low-voltage distribution areas as claimed in claim 1, characterized by, The device comprises a communication unit and a processing unit. The communication unit is configured to collect multi-source data of a transformer area, acquire meteorological forecast data of a location of the transformer area, and acquire an electricity regulation policy; The processing unit is configured to clean the multi-source data of the transformer area to obtain electrical measurement data, and convert the meteorological forecast data into a three-dimensional tensor. The semantic vector in the electricity regulation policy and the effective time of the electricity regulation policy are extracted through a BERT model, the effective time of the electricity regulation policy is time-aligned with the three-dimensional tensor, and a double-modal input matrix is formed. The double-modal input matrix and the electrical measurement data are input into a multi-factor coupling probability prediction model to output a load prediction value and a confidence interval boundary, wherein the multi-factor coupling probability prediction model is constructed based on a physical information neural network (PINN) and an LSTM-CRF model.

Images