Automatic calculation method for constant value accounting of power supply and distribution system based on deep learning

Abstract

The invention discloses an automatic calculation method for constant value accounting of a power supply and distribution system based on deep learning. The method comprises the following steps: establishing an equipment database; in combination with historical operation and maintenance data and electrical parameters, constructing a deep neural network model to learn and predict protection constant value calculation of the power supply and distribution equipment; according to the power grid topological structure and the equipment parameters, a deep learning short-circuit current calculation model is constructed and trained; inputting equipment parameters into the short-circuit current calculation model, automatically calculating a constant value of the equipment, and outputting a setting calculation book according to a setting calculation principle in combination with power grid operation conditions and a short-circuit current calculation result; outputting a constant value matching graph by combining a calculation result of the deep neural network model with a constant value parameter of a setting calculation book; generating a short-circuit current calculation book and a setting calculation book report according to the results of the automatic constant value accounting and the protection cooperation calculation; and a single line diagram, a short circuit impedance diagram and a constant value matching diagram interface are realized through a configuration interface, and related calculation results are displayed in corresponding diagrams.

Description

Technical Field

[0001] This invention relates to the field of information and intelligent technology, and in particular to an automatic calculation method for setting value calculation of power supply and distribution systems based on deep learning. Background Technology

[0002] The setting calculation of the power supply and distribution system is an important part of the operation of the power system. It is directly related to the safety and stability of the power system. Traditional setting calculation methods mainly rely on manual calculation and judgment, which has problems such as large amount of calculation, long time consumption, and susceptibility to human factors.

[0003] Existing power supply and distribution system setting calculations mostly rely on traditional methods such as lookup tables and empirical formulas. While these methods are effective in certain situations, they may not achieve ideal accuracy in complex electrical environments and are difficult to dynamically adjust according to changes in grid operating conditions. Furthermore, existing power supply and distribution systems typically only perform setting calculations during the initial configuration phase. Once the system is operational, real-time adjustments and optimizations are difficult, and the lack of an effective feedback mechanism prevents dynamic correction of settings based on actual operating data. This can lead to settings becoming disconnected from actual needs, affecting system stability and reliability. Therefore, this paper proposes an automatic setting calculation method for power supply and distribution systems based on deep learning. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide an automatic calculation method for setting values ​​of power supply and distribution systems based on deep learning. This method overcomes the defects of traditional setting value calculation, and through a deep neural network model and continuous adjustment of model parameters, the model can accurately predict the system protection setting values, thereby ensuring the stability and reliability of the power supply and distribution system.

[0005] To address the aforementioned technical problems, the present invention provides an automatic calculation method for power supply and distribution system setpoints based on deep learning, comprising the following steps:

[0006] Step 1: Automatically extract equipment nameplate information using image recognition technology, establish an equipment database, collect power grid topology, load information, and line parameter data, and perform preprocessing.

[0007] Step 2: Combining historical operation and maintenance data and electrical parameters, construct a deep neural network model to learn and predict the protection settings of power supply and distribution equipment.

[0008] Step 3: Based on the power grid topology and equipment parameters, construct and train a deep learning-based short-circuit current calculation model;

[0009] Step 4: Input the equipment parameters into the short-circuit current calculation model, automatically calculate the equipment settings, and output the setting calculation report based on the setting calculation principles, combined with the power grid operating conditions and the short-circuit current calculation results.

[0010] Step 5: By combining the calculation results of the deep neural network model with the setpoint parameters in the tuning calculation sheet, output the setpoint matching diagram;

[0011] Step 6: Based on the results of automatic setting calculation and protection coordination calculation, generate short-circuit current calculation report and setting calculation report;

[0012] Step 7: Implement the single-line diagram, short-circuit impedance diagram, and setting coordination diagram through the configuration interface, and display the relevant calculation results in the corresponding diagrams.

[0013] Furthermore, in step one, the equipment nameplates in the power supply and distribution system are scanned using image recognition technology to obtain nameplate images. The acquired nameplate images are preprocessed by denoising, grayscale conversion, and binarization. The preprocessed nameplate images are then used for text recognition using OCR. The identified nameplate data is parsed to extract key information such as equipment name, model, rated voltage, and rated current. Based on the type and characteristics of the equipment information, a database table structure is designed, and the identified equipment information is entered into the database according to the designed database table structure to form an equipment database.

[0014] Furthermore, in step two, historical operation and maintenance data and electrical parameters are integrated into an input feature vector, which is then input into a deep neural network model. The deep neural network model outputs protection settings.

[0015] Let the input feature vector be , Indicates the rated current of the equipment. Indicates the rated voltage of the equipment. Indicates the rated power of the equipment. This indicates the temperature parameter of the equipment. The load data of the equipment is used to calculate the protection settings through a deep neural network model.

[0016]

[0017] In the formula, This represents the protection setpoint of the output, and f represents the prediction function of the deep neural network model. Indicates model parameters;

[0018] The overcurrent protection setting is calculated by inputting the current current and historical fault data. :

[0019]

[0020] In the formula, This indicates the overcurrent protection setting. This represents the custom calculation function for overcurrent protection, where I represents the current current and ts represents historical fault data.

[0021] The protection action time setting is calculated by inputting the overcurrent characteristics and current waveform of the device. :

[0022]

[0023] In the formula, This indicates the timing setpoint for the protective action. This represents the time setpoint calculation function, P represents the current waveform, and hy represents the load data.

[0024] During model training, the model output values ​​are calibrated using historical operation and maintenance data and equipment rated parameters.

[0025]

[0026] In the formula, L represents the loss function value, which measures the overall error between the model's predicted value and the actual value. The smaller the value, the more accurate the model's prediction. N represents the total sample size. This represents a constant value in the model output. This indicates the actual protection setting.

[0027] Furthermore, in step three, assuming that the impedance of each transformer consists of equivalent resistance and reactance, the transformer impedance is expressed as:

[0028]

[0029] In the formula, Indicates the transformer impedance. Indicates the equivalent resistance. This represents reactance, and j represents the imaginary unit;

[0030] The total impedance is obtained by adding the impedances of all transformers in the system.

[0031]

[0032] In the formula, This represents the total impedance, and n represents the nth transformer.

[0033] The short-circuit current calculation model calculates the short-circuit current based on the short-circuit type, assuming the power supply voltage is... The total impedance is Short circuit current for:

[0034]

[0035] Let the short-circuit current prediction model be y. The error between the predicted and actual values ​​is measured using this model.

[0036]

[0037] In the formula, This represents the short-circuit current predicted by the short-circuit current prediction model for the i-th sample. Let N represent the actual short-circuit current of the i-th sample, and N represent the total number of samples.

[0038] Furthermore, in step four, the equipment parameters and the operating conditions of the power grid are input into the short-circuit current calculation model. The model calculates appropriate protection settings based on the input equipment parameters, power grid status, and short-circuit current. According to the setting calculation principle, the analysis results of the protection settings and setting calculation principle, the evaluation results of the power grid operating conditions, and the calculation results of the short-circuit current are sorted and summarized. Based on the sorted and summarized results, a setting calculation report is generated and output.

[0039] Furthermore, in step five, the device setpoint results output by the deep neural network model are matched with the collected setpoint parameters. The model prediction results, setpoint parameters, and power grid operating conditions are combined into a feature vector. Based on the setpoints predicted by the model and the power grid operating conditions, the selectivity, sensitivity, and reliability of protection devices at all levels are calculated. The results of the selectivity, sensitivity, and reliability calculations are organized to form a structured data table. The layout and style of the setpoint coordination diagram are designed and verified, and the generated setpoint coordination diagram is output.

[0040] Furthermore, in step six, based on the results of automatic setting calculation and protection coordination calculation, the short-circuit current calculation results are analyzed to determine the maximum short-circuit current value of each node and its influence range. The protection coordination calculation results are analyzed to check whether the selectivity, sensitivity and reliability of protection devices at all levels meet the requirements. The short-circuit current calculation report and the setting calculation report are then combined into a complete report.

[0041] Furthermore, in step seven, the configuration interface uses a circuit configuration tool as a carrier to realize single-line diagrams, short-circuit impedance diagrams, and setting coordination diagrams. The circuit configuration tool includes configuration functions, setting functions, auxiliary functions, and preview functions.

[0042] Furthermore, the configuration function performs manual configuration of the loop diagram. The settings function includes position height and width, legend color, and associated device selection. The auxiliary functions include selection, zoom in / out, full screen, and undo. The preview function includes loop diagram preview, device details, device parameters, device history, device documentation, historical status, device comparison, and historical loops.

[0043] This invention, based on deep learning, employs the aforementioned technical solution for automatic setting value calculation in power supply and distribution systems. Specifically, it establishes an equipment database; combines historical operation and maintenance data and electrical parameters to construct a deep neural network model that learns and predicts the protection setting values ​​of power supply and distribution equipment; constructs and trains a deep learning-based short-circuit current calculation model based on the power grid topology and equipment parameters; inputs equipment parameters into the short-circuit current calculation model to automatically calculate equipment settings; outputs a setting calculation report based on setting calculation principles, combined with power grid operating conditions and short-circuit current calculation results; outputs a setting coordination diagram by combining the calculation results of the deep neural network model with the setting parameters in the setting calculation report; generates short-circuit current calculation reports and setting calculation reports based on the results of automatic setting value calculation and protection coordination calculation; and displays single-line diagrams, short-circuit impedance diagrams, and setting coordination diagrams through a configuration interface, showing the relevant calculation results in the corresponding diagrams. This method overcomes the shortcomings of traditional setting value calculation methods by using a deep neural network model and continuously adjusting model parameters to enable the model to accurately predict system protection settings, ensuring the stability and reliability of the power supply and distribution system. Attached Figure Description

[0044] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments:

[0045] Figure 1 This is a flowchart of the automatic calculation method for setpoint accounting of power supply and distribution system based on deep learning, as described in this invention.

[0046] Figure 2 Here is a flowchart illustrating the learning and prediction process of the deep neural network model in this method;

[0047] Figure 3 This is a flowchart of the deep learning-based short-circuit current calculation model in this method. Detailed Implementation

[0048] Implementation, for example Figure 1 As shown, the automatic calculation method for power supply and distribution system setpoints based on deep learning of the present invention includes the following steps:

[0049] Step 1: Automatically extract equipment nameplate information using image recognition technology, establish an equipment database, collect power grid topology, load information, and line parameter data, and perform preprocessing.

[0050] Step 2: Combining historical operation and maintenance data and electrical parameters, construct a deep neural network model to learn and predict the protection settings of power supply and distribution equipment.

[0051] Step 3: Based on the power grid topology and equipment parameters, construct and train a deep learning-based short-circuit current calculation model;

[0052] Step 4: Input the equipment parameters into the short-circuit current calculation model, automatically calculate the equipment settings, and output the setting calculation report based on the setting calculation principles, combined with the power grid operating conditions and the short-circuit current calculation results.

[0053] Step 5: By combining the calculation results of the deep neural network model with the setpoint parameters in the tuning calculation sheet, output the setpoint matching diagram;

[0054] Step 6: Based on the results of automatic setting calculation and protection coordination calculation, generate short-circuit current calculation report and setting calculation report;

[0055] Step 7: Implement the single-line diagram, short-circuit impedance diagram, and setting coordination diagram through the configuration interface, and display the relevant calculation results in the corresponding diagrams.

[0056] Preferably, in step one, the equipment nameplates in the power supply and distribution system are scanned using image recognition technology to obtain nameplate images. The acquired nameplate images are preprocessed by denoising, grayscale conversion, and binarization. The preprocessed nameplate images are then used for text recognition using OCR. The identified nameplate data is parsed to extract key information such as equipment name, model, rated voltage, and rated current. Based on the type and characteristics of the equipment information, a database table structure is designed, and the identified equipment information is entered into the database according to the designed database table structure to form an equipment database.

[0057] Preferred, such as Figure 2 As shown, in step two, historical operation and maintenance data and electrical parameters are integrated into an input feature vector, which is then input into a deep neural network model. The deep neural network model outputs protection settings.

[0058] Let the input feature vector be , Indicates the rated current of the equipment. Indicates the rated voltage of the equipment. Indicates the rated power of the equipment. This indicates the temperature parameter of the equipment. The load data of the equipment is used to calculate the protection settings through a deep neural network model.

[0059]

[0060] In the formula, This represents the protection setpoint of the output, and f represents the prediction function of the deep neural network model. Indicates model parameters;

[0061] The overcurrent protection setting is calculated by inputting the current current and historical fault data. :

[0062]

[0063] In the formula, This indicates the overcurrent protection setting. This represents the custom calculation function for overcurrent protection, where I represents the current current and ts represents historical fault data.

[0064] The protection action time setting is calculated by inputting the overcurrent characteristics and current waveform of the device. :

[0065]

[0066] In the formula, This indicates the timing setpoint for the protective action. This represents the time setpoint calculation function, P represents the current waveform, and hy represents the load data.

[0067] During model training, the model output values ​​are calibrated using historical operation and maintenance data and equipment rated parameters.

[0068]

[0069] In the formula, L represents the loss function value, which measures the overall error between the model's predicted value and the actual value. The smaller the value, the more accurate the model's prediction. N represents the total sample size. This represents a constant value in the model output. This indicates the actual protection setting.

[0070] Preferably, in step three, as follows: Figure 3 As shown, assuming that the impedance of each transformer consists of equivalent resistance and reactance, the transformer impedance can be expressed as:

[0071]

[0072] In the formula, Indicates the transformer impedance. Indicates the equivalent resistance. This represents reactance, and j represents the imaginary unit;

[0073] The total impedance is obtained by adding the impedances of all transformers in the system.

[0074]

[0075] In the formula, This represents the total impedance, and n represents the nth transformer.

[0076] The short-circuit current calculation model calculates the short-circuit current based on the short-circuit type, assuming the power supply voltage is... The total impedance is Short circuit current for:

[0077]

[0078] Let the short-circuit current prediction model be y. The error between the predicted and actual values ​​is measured using this model.

[0079]

[0080] In the formula, This represents the short-circuit current predicted by the short-circuit current prediction model for the i-th sample. Let N represent the actual short-circuit current of the i-th sample, and N represent the total number of samples.

[0081] Preferably, in step four, the equipment parameters and the operating conditions of the power grid are input into the short-circuit current calculation model. The model calculates appropriate protection settings based on the input equipment parameters, power grid status, and short-circuit current. According to the setting calculation principle, the analysis results of the protection settings and setting calculation principle, the evaluation results of the power grid operating conditions, and the calculation results of the short-circuit current are sorted and summarized. Based on the sorted and summarized results, a setting calculation report is generated and output.

[0082] Preferably, in step five, the device setpoint results output by the deep neural network model are matched with the collected setpoint parameters. The model prediction results, setpoint parameters, and power grid operating conditions are combined into a feature vector. Based on the model-predicted setpoints and power grid operating conditions, the selectivity, sensitivity, and reliability of protection devices at all levels are calculated. The results of the selectivity, sensitivity, and reliability calculations are organized to form a structured data table. The layout and style of the setpoint coordination diagram are designed and verified, and the generated setpoint coordination diagram is output.

[0083] Preferably, in step six, based on the results of automatic value calculation and protection coordination calculation, the short-circuit current calculation results are analyzed to determine the maximum short-circuit current value and its influence range at each node. The protection coordination calculation results are analyzed to check whether the selectivity, sensitivity and reliability of protection devices at all levels meet the requirements. The short-circuit current calculation report and the setting calculation report are then combined into a complete report.

[0084] Preferably, in step seven, the configuration interface uses a circuit configuration tool as a carrier to realize single-line diagrams, short-circuit impedance diagrams, and setting coordination diagrams. The circuit configuration tool includes configuration functions, setting functions, auxiliary functions, and preview functions.

[0085] Preferably, the configuration function performs manual configuration of the loop diagram. The setting function includes position height and width, legend color, and associated device selection. The auxiliary functions include selection, zoom in / out, full screen, and undo. The preview function includes loop diagram preview, device details, device parameters, device history, device documentation, historical status, device comparison, and historical loops.

[0086] This method uses image recognition technology to automatically extract key information from the equipment nameplate, preprocesses this information including denoising, grayscale conversion, and binarization, and then records the identified equipment information into the database to form an equipment database.

[0087] A deep neural network model is constructed by combining historical operation and maintenance data and electrical parameters. This data and parameters are integrated into a single input feature vector, which is then fed into the model. The model outputs protection settings, and by continuously adjusting the model parameters, it can accurately predict these settings. Based on the system's circuit topology and equipment parameters, a deep learning-based short-circuit current calculation model is built. This model calculates the short-circuit current at various points in the system and assesses its impact. Equipment parameters and grid operating conditions are input into the short-circuit current calculation model, which automatically calculates the equipment settings. Based on the setting calculation principles, combined with the grid operating conditions and the short-circuit current calculation results, the model outputs a setting calculation report. This report details the equipment settings and the setting calculation principles. The analysis results, the assessment results of power grid operating conditions, and the calculation results of short-circuit current are used. The equipment setting results output by the deep neural network model are matched with the collected setting parameters to calculate the selectivity, flexibility, and reliability of protection devices at all levels. The results are then organized into structured data tables, and setting coordination diagrams are designed and output. Based on the results of automatic setting calculation and protection coordination calculation, short-circuit current calculation reports and setting calculation reports are generated. These reports provide a detailed analysis of the short-circuit current calculation results and protection setting calculation results, ensuring that the selectivity, sensitivity, and reliability of protection devices at all levels meet the requirements. Using circuit configuration tools as a carrier, single-line diagrams, short-circuit impedance diagrams, and setting coordination diagrams are implemented through the configuration interface, displaying the relevant calculation results in the corresponding diagrams.

Claims

1. An automatic calculation method for setting value calculation in power supply and distribution systems based on deep learning, characterized in that... Includes the following steps: Step 1: Automatically extract equipment nameplate information using image recognition technology, establish an equipment database, collect power grid topology, load information, and line parameter data, and perform preprocessing. Step 2: Combining historical operation and maintenance data and electrical parameters, construct a deep neural network model to learn and predict the protection settings of power supply and distribution equipment. Step 3: Based on the power grid topology and equipment parameters, construct and train a deep learning-based short-circuit current calculation model; Step 4: Input the equipment parameters into the short-circuit current calculation model, automatically calculate the equipment settings, and output the setting calculation report based on the setting calculation principles, combined with the power grid operating conditions and the short-circuit current calculation results. Step 5: By combining the calculation results of the deep neural network model with the setpoint parameters in the tuning calculation sheet, output the setpoint matching diagram; Step 6: Based on the results of automatic setting calculation and protection coordination calculation, generate short-circuit current calculation report and setting calculation report; Step 7: Implement the single-line diagram, short-circuit impedance diagram, and setting coordination diagram through the configuration interface, and display the relevant calculation results in the corresponding diagrams.

2. The automatic calculation method for setpoint accounting of power supply and distribution systems based on deep learning according to claim 1, characterized in that: In step one, image recognition technology is used to scan the nameplates of equipment in the power supply and distribution system to obtain nameplate images. The acquired nameplate images are preprocessed by denoising, grayscale conversion, and binarization. The preprocessed nameplate images are then used for text recognition by OCR. The identified nameplate data is parsed to extract key information such as equipment name, model, rated voltage, and rated current. Based on the type and characteristics of the equipment information, a database table structure is designed, and the identified equipment information is entered into the database according to the designed database table structure to form an equipment database.

3. The automatic calculation method for setpoint accounting of power supply and distribution systems based on deep learning according to claim 1, characterized in that: In step two, historical operation and maintenance data and electrical parameters are integrated into an input feature vector, which is then input into a deep neural network model. The deep neural network model outputs protection settings. Let the input feature vector be , Indicates the rated current of the equipment. Indicates the rated voltage of the equipment. Indicates the rated power of the equipment. This indicates the temperature parameter of the equipment. The load data of the equipment is used to calculate the protection settings through a deep neural network model. ， In the formula, This represents the protection setpoint of the output, and f represents the prediction function of the deep neural network model. Indicates model parameters; The overcurrent protection setting is calculated by inputting the current current and historical fault data. : ，， In the formula, This indicates the overcurrent protection setting. This represents the custom calculation function for overcurrent protection, where I represents the current current and ts represents historical fault data. The protection action time setting is calculated by inputting the overcurrent characteristics and current waveform of the device. : ， In the formula, This indicates the timing setpoint for the protective action. This represents the time setpoint calculation function, P represents the current waveform, and hy represents the load data. During model training, the model output values ​​are calibrated using historical operation and maintenance data and equipment rated parameters. ， In the formula, L represents the loss function value, which measures the overall error between the model's predicted value and the actual value. The smaller the value, the more accurate the model's prediction. N represents the total sample size. This represents a constant value in the model output. This indicates the actual protection setting.

4. The automatic calculation method for setpoint accounting of power supply and distribution systems based on deep learning according to claim 1, characterized in that: In step three, assuming that the impedance of each transformer consists of equivalent resistance and reactance, the transformer impedance is expressed as: ， In the formula, Indicates the transformer impedance. Indicates the equivalent resistance. This represents reactance, and j represents the imaginary unit; The total impedance is obtained by adding the impedances of all transformers in the system. ， In the formula, This represents the total impedance, and n represents the nth transformer. The short-circuit current calculation model calculates the short-circuit current based on the short-circuit type, assuming the power supply voltage is... The total impedance is Short circuit current for: ， Let the short-circuit current prediction model be y. The error between the predicted and actual values ​​is measured using this model. ， In the formula, This represents the short-circuit current predicted by the short-circuit current prediction model for the i-th sample. Let N represent the actual short-circuit current of the i-th sample, and N represent the total number of samples.

5. The automatic calculation method for setpoint accounting of power supply and distribution systems based on deep learning according to claim 1, characterized in that: In step four, the equipment parameters and the operating conditions of the power grid are input into the short-circuit current calculation model. The model calculates appropriate protection settings based on the input equipment parameters, power grid status, and short-circuit current. According to the setting calculation principle, the analysis results of the protection settings and setting calculation principle, the evaluation results of the power grid operating conditions, and the calculation results of the short-circuit current are sorted and summarized. Based on the sorted and summarized results, the setting calculation report is generated and output.

6. The automatic calculation method for setpoint accounting of power supply and distribution system based on deep learning according to claim 1, characterized in that: In step five, the device setpoint results output by the deep neural network model are matched with the collected setpoint parameters. The model prediction results, setpoint parameters, and power grid operating conditions are combined into a feature vector. Based on the model-predicted setpoints and power grid operating conditions, the selectivity, sensitivity, and reliability of protection devices at all levels are calculated. The results of the selectivity, sensitivity, and reliability calculations are organized into a structured data table. The layout and style of the setpoint coordination diagram are designed and verified, and the generated setpoint coordination diagram is output.

7. The automatic calculation method for setpoint accounting of power supply and distribution system based on deep learning according to claim 1, characterized in that: In step six, based on the results of automatic setting calculation and protection coordination calculation, the short-circuit current calculation results are analyzed to determine the maximum short-circuit current value of each node and its influence range. The protection coordination calculation results are analyzed to check whether the selectivity, sensitivity and reliability of protection equipment at all levels meet the requirements. The short-circuit current calculation report and the setting calculation report are combined into a complete report.

8. The automatic calculation method for setpoint accounting of power supply and distribution system based on deep learning according to claim 1, characterized in that: In step seven, the configuration interface uses the circuit configuration tool as a carrier to realize single-line diagrams, short-circuit impedance diagrams, and setting coordination diagrams. The circuit configuration tool includes configuration functions, setting functions, auxiliary functions, and preview functions.

9. The automatic calculation method for setpoint accounting of power supply and distribution system based on deep learning according to claim 8, characterized in that: The configuration function performs manual configuration of the loop diagram. The settings function includes position height and width, legend color, and associated device selection. The auxiliary functions include selection, zoom in / out, full screen, and undo. The preview function includes loop diagram preview, device details, device parameters, device history, device documentation, historical status, device comparison, and historical loops.

Images