Risk prediction method and device of main network and electronic equipment

Abstract

The invention discloses a risk prediction method and device of a main network and electronic equipment. The method comprises the following steps: acquiring operation data of a target main network in a preset period, and inputting the operation data into a target model to obtain a group of parameter prediction values of risk factors; determining the weight of each parameter predicted value, and performing weighted summation on the group of parameter predicted values according to the weight of each parameter predicted value to obtain a first predicted value which is corrected for the first time; obtaining a target parameter value of the risk factor, and inputting the target parameter value and the first predicted value into a linear regression model to obtain a second predicted value subjected to secondary correction; and calculating a difference value between the target parameter value and the second predicted value, determining the second predicted value as the target predicted value when the difference value belongs to a preset error interval, and determining that the target main network has a risk when the target predicted value is greater than or equal to a predicted value threshold value. Through the method and the device, the problem of inaccurate prediction of the operation risk of the main network in related technologies is solved.

Description

Technical Field

[0001] The present application relates to the field of main networks, and specifically, to a risk prediction method, device and electronic equipment for a main network. Background Art

[0002] As an important infrastructure of the national economy, the power system plays a decisive role in national security, national defense construction and economic development. As the backbone of the power system, the main grid is the hub and foundation of the power system. Its safe and stable operation is of great significance to ensuring the normal operation of the entire power system and ensuring power supply. In recent years, with the rapid development of my country's economy and society and the rapid growth of the power industry, the risks and challenges faced by the power system have continued to escalate, and the safe operation of the main grid faces increasingly severe challenges.

[0003] The operation status of the main network is affected by many factors. It is necessary to comprehensively consider the main network itself, the equipment in the network, the operating environment, the operation mode, the control measures and other factors to achieve the scientific and reasonable operation of the main network. The dynamic prediction of the operation status of the main network can provide a theoretical basis for the dispatching and commanding personnel and provide decision-making support for the formulation of scientific and reasonable emergency response measures. Accurate prediction of the operation status of the main network can provide accurate decision-making basis for the dispatching and commanding personnel, improve the scientificity and accuracy of dispatching decisions, reduce the cost of emergency response, and improve the efficiency of emergency response.

[0004] In the related technology, the prediction of the main network operation status mainly relies on the prediction model. Since the data set used to train the prediction model cannot represent all possible situations, the prediction results of the prediction model on unseen data will be inaccurate, resulting in deviations. Improper setting of the hyperparameters of the prediction model will affect the performance of the prediction model and lead to prediction deviations. Noise or outliers in the data will mislead the learning of the prediction model, thereby affecting the accuracy of the prediction. There is a certain degree of deviation in the risk factor values ​​predicted by the model, resulting in inaccurate prediction results of the main network's operation risks.

[0005] There is currently no effective solution to the problem of inaccurate prediction of the operating risks of the main network in related technologies. Summary of the invention

[0006] The main purpose of this application is to provide a main network risk prediction method, device and electronic equipment to solve the problem of inaccurate prediction of the main network operation risk in related technologies.

[0007] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a risk prediction method for a main network is provided. The method includes: obtaining the operation data of the target main network within a preset period, and inputting the operation data into the target model to obtain a set of parameter prediction values ​​of the risk factor, wherein a set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operation data at a moment in the preset period, and the target model is trained by multiple sets of training samples, each set of training samples includes historical operation data and a set of historical parameter prediction values ​​of the risk factor; determining the weight of each parameter prediction value in a set of parameter prediction values, and weighted summing a set of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value of the initial correction; obtaining the target parameter value of the risk factor, inputting the target parameter value and the first prediction value into a linear regression model, and obtaining a second prediction value of the secondary correction, wherein the linear regression model is trained by multiple historical first prediction values ​​and target parameter values; calculating the difference between the target parameter value and the second prediction value, and determining the second prediction value as the target prediction value when the difference belongs to the preset error interval, and determining the target main network to be at risk when the target prediction value is greater than or equal to the prediction value threshold.

[0008] Optionally, after calculating the difference between the target parameter value and the second predicted value, the method further includes: when the difference does not fall within a preset error interval, repeating the steps of primary correction and secondary correction until the difference falls within the preset error interval, or the number of repetitions is greater than or equal to a number threshold; when the number of repetitions is greater than or equal to the number threshold, determining the second predicted value as the target predicted value.

[0009] Optionally, the risk factors and target parameter values ​​of the risk factors are obtained by: obtaining historical fault records of the target main network and extracting the fault causes from the historical fault records; determining the risk factors from the operating parameters of the target main network based on the fault causes, and determining the historical parameter values ​​of the risk factors as the target parameter values.

[0010] Optionally, determining the weight of each parameter prediction value in a set of parameter prediction values ​​includes: determining the prediction moment of each parameter prediction value, wherein the prediction moment is a moment within a preset period; when the prediction moment belongs to a first preset time interval, determining the weight of the parameter prediction value as a first weight, wherein the first preset time interval is used to characterize the peak power consumption interval of the target main network; when the prediction moment belongs to a second preset time interval, determining the weight of the parameter prediction value as a second weight, wherein the second preset time interval is used to characterize the low power consumption interval of the target main network; when the prediction moment does not belong to the first preset time interval, and the prediction moment does not belong to the second preset time interval, determining the weight of the parameter prediction value as a third weight, wherein the first weight is greater than the third weight, which is greater than the second weight.

[0011] Optionally, after determining that the target main network is at risk, the method further includes: obtaining historical emergency response strategies for risk factors, and executing the historical emergency response strategies on the target main network; after executing the historical emergency response strategies, obtaining an updated target prediction value; and when the updated target prediction value is greater than or equal to a prediction value threshold, issuing a prompt message, wherein the prompt message is used by operation and maintenance personnel to maintain the target main network.

[0012] Optionally, the target model is trained in the following manner: obtaining historical operating data of the target main network within multiple historical periods, and determining a set of historical parameter values ​​of risk factors during the operation period of the target main network corresponding to the historical operating data; determining the historical operating data within each historical period and a set of historical parameter values ​​as a set of training samples to obtain multiple sets of training samples; training a preset neural network model with multiple sets of training samples to obtain the target model.

[0013] Optionally, the linear regression model is trained in the following manner: a plurality of historical first prediction values ​​are obtained, an initial linear regression model is trained with the target parameter value as the dependent variable, and the plurality of historical first prediction values ​​as the independent variable to obtain the linear regression model.

[0014] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a risk prediction device for a main network is provided. The device comprises: a first acquisition unit, used for acquiring operation data of a target main network within a preset period, and inputting the operation data into a target model to obtain a group of parameter prediction values ​​of risk factors, wherein a group of parameter prediction values ​​comprises multiple parameter prediction values, each parameter prediction value corresponds to the operation data at a moment within the preset period, and the target model is obtained by training multiple groups of training samples, each group of training samples comprises historical operation data and a group of historical parameter prediction values ​​of risk factors; a first determination unit, used for determining the weight of each parameter prediction value in a group of parameter prediction values, and weighted summing a group of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value corrected for the first time; a second acquisition unit, used for acquiring a target parameter value of the risk factor, inputting the target parameter value and the first prediction value into a linear regression model to obtain a second prediction value corrected for the second time, wherein the linear regression model is obtained by training multiple historical first prediction values ​​and target parameter values; a calculation unit, used for calculating the difference between the target parameter value and the second prediction value, and determining the second prediction value as the target prediction value when the difference belongs to a preset error interval, and determining that the target main network has a risk when the target prediction value is greater than or equal to a prediction value threshold.

[0015] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps of the main network risk prediction method described in each embodiment of the present application.

[0016] Through the present application, the following steps are adopted: obtaining the operating data of the target main network within a preset period, and inputting the operating data into the target model to obtain a set of parameter prediction values ​​of the risk factors, wherein a set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operating data at a moment within the preset period, and the target model is trained by multiple sets of training samples, each set of training samples includes historical operating data and a set of historical parameter prediction values ​​of the risk factors; determining the weight of each parameter prediction value in a set of parameter prediction values, and weighted summing a set of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value of the initial correction; obtaining the target parameter value of the risk factor, inputting the target parameter value and the first prediction value into a linear regression model to obtain a second prediction value of the secondary correction, wherein the linear regression model is trained by multiple historical first prediction values ​​and target parameter values; calculating the difference between the target parameter value and the second prediction value, and when the difference is within a preset error interval, determining the second prediction value as the target prediction value, and when the target prediction value is greater than or equal to the prediction value threshold, determining that the target main network has a risk, thereby solving the problem of inaccurate prediction of the operating risk of the main network in the related art. A set of parameter prediction values ​​is obtained through the target model, and then the weighted average method is used to obtain the first predicted value of the initial correction. Then, the linear regression model is used to perform a second correction on the first predicted value of the initial correction. The difference between the target parameter value and the second predicted value is calculated, and a preset error interval is set. When the difference falls within the preset error interval, the second predicted value is determined as the target predicted value. Through the double deviation correction method and the preset error interval determination method, more effective and more accurate risk factor prediction values ​​can be obtained, thereby achieving the effect of improving the accuracy of predicting the operation risk of the main network. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0018] Figure 1 is a flow chart of a risk prediction method for a main network provided in an embodiment of the present application;

[0019] Figure 2 is a schematic diagram of a risk prediction device for a main network provided according to an embodiment of the present application;

[0020] Figure 3 is a schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.

[0025] It should be noted that the information collected is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of the relevant regions, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0026] The present invention is described below in conjunction with preferred implementation steps. Figure 1 is a flow chart of a risk prediction method for a main network provided in an embodiment of the present application, such as Figure 1 As shown, the method comprises the following steps:

[0027] Step S101, obtain the operating data of the target main network within a preset period, and input the operating data into the target model to obtain a set of parameter prediction values ​​of risk factors, wherein a set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operating data at a moment in the preset period, and the target model is trained by multiple groups of training samples, each group of training samples includes historical operating data and a set of historical parameter prediction values ​​of risk factors.

[0028] In step S101, the preset period may be a time length set according to expert experience, such as one day, one month, one week or one year. The operation data may be data reflecting the real-time operation status of the main grid, such as voltage, current, power (active power, reactive power, apparent power), frequency, harmonic content, etc., as well as the oil temperature and load rate of the transformer, the number of opening and closing times and operation time of the switchgear, the temperature and load of the cable, and the operation status of other equipment such as capacitors, circuit breakers, and relay protection devices. Risk factors may include equipment failure, overload, harmonic pollution, short circuit, equipment abnormality, unstable operation status of the power grid, major disturbance of the power grid or generator failure, system overload, leakage current, frequent operation or misoperation, network failure and power quality degradation, etc.

[0029] Among them, the parameter values ​​used to predict whether the risks corresponding to the above risk factors occur include, for example, voltage parameters, including bus voltage, line voltage, and user-end voltage. Current parameters, such as line current and load current, current surges or drops may indicate short circuits, overloads, or equipment abnormalities. Power parameters, such as active power and reactive power monitoring, power imbalance or abnormal fluctuations may indicate unstable grid operation. Equipment operating temperature, such as transformer oil temperature, cable temperature, etc. Abnormal temperature rise may indicate equipment overload or poor heat dissipation. Harmonic parameters, such as harmonic content and distortion rate. Frequency parameters, such as grid frequency, frequency deviation from the normal range may indicate major disturbances in the grid or generator failure. Load parameters, such as real-time load and load forecast of the grid, mismatched or abnormal loads may indicate system overload or demand forecast errors. Insulation parameters, such as insulation resistance and leakage current of equipment, reduced insulation performance may indicate that the equipment is about to fail. Switch operation parameters, such as the number of times the switch is opened and closed, operation time, and misoperation records of the switch device, frequent operation or misoperation may increase the risk of equipment failure. Communication parameters, such as network delay, packet loss rate, etc. Equipment status parameters, such as equipment vibration, noise, oil quality, etc. Abnormal changes in these parameters may indicate internal problems of the equipment.

[0030] The target model can be a neural network model trained by multiple sets of training samples. The neural network model is trained by the historical operating data and risk factors in each set of training samples, so that the neural network model learns how the training samples predict the risk factor parameters from the historical operating data, and finally the target model is trained. By inputting the operating data into the target model, a set of parameter prediction values ​​of the risk factors can be obtained. It should be noted that there can be one or more risk factors, each of which corresponds to a set of parameter prediction values. A set of parameter prediction values ​​includes the predicted parameter value data at each moment within a preset period, and each parameter value data corresponds to the historical operating data at that moment.

[0031] Step S102, determining the weight of each parameter prediction value in a group of parameter prediction values, and weighted summing the group of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value that is initially revised.

[0032] In step S102, the weight of the parameter prediction value is a value used to quantify and adjust the influence of different parameter prediction values ​​on the final risk assessment. The setting of the weight can reflect the relationship between the parameter prediction value and time. For example, during the peak power consumption period (such as afternoon to evening), the operation state of the power grid may be more risky, so the parameter prediction value corresponding to the peak period can be given a higher weight, while the parameter prediction value during the power consumption trough period can be given a lower weight. The weight of the parameter prediction value of the first preset time interval (peak power consumption) is set to the first weight, the parameter prediction value of the second preset time interval (power consumption trough) is set to the second weight, and the parameter prediction value of other times is set to the third weight, and the first weight is greater than the third weight and greater than the second weight.

[0033] In some examples, the weights may also be determined based on the type of parameter prediction values, and different operating parameters may contribute to different degrees to the risk of the power grid. For example, the predicted values ​​of voltage and current may be more critical than the predicted values ​​of temperature parameters in assessing risk, so higher weights may be set for the predicted values ​​of voltage and current.

[0034] In some examples, the weights may also be determined based on historical data analysis. By analyzing historical data, it is determined that some parameter prediction values ​​were most effective for risk prediction in the past, and higher weights are then set for these parameter prediction values.

[0035] After determining the weights of the parameter prediction values, a group of parameter prediction values ​​are weighted and summed to obtain a first prediction value for initial correction for evaluating whether the main network has operational risks.

[0036] Step S103, obtaining a target parameter value of the risk factor, inputting the target parameter value and the first prediction value into a linear regression model to obtain a second prediction value corrected twice, wherein the linear regression model is trained by a plurality of historical first prediction values ​​and target parameter values.

[0037] In step S103, the target parameter value may be a value extracted from the historical operation data to measure or define the actual operation status of a risk factor. The target parameter value is obtained from the actual operation data and is used to compare and verify whether the first prediction value is more in line with the actual operation of the main network. Since the parameter prediction values ​​output by the target model are based on the simulation or hypothesis prediction of the model, there is a certain error between these parameter prediction values ​​and the actual operation data. In order to make the first prediction value more in line with the actual operation of the main network, the first prediction value is corrected twice by the pre-trained linear regression model to obtain a second prediction value that is closer to the actual operation of the main network (target parameter value).

[0038] Step S104, calculate the difference between the target parameter value and the second predicted value, and when the difference is within a preset error range, determine the second predicted value as the target predicted value, and when the target predicted value is greater than or equal to the predicted value threshold, determine that the target main network is at risk.

[0039] In step S104, the preset error interval may be an acceptable deviation range between the target parameter value set according to expert experience and the second predicted value. If the difference falls within the preset error interval, it means that the second predicted value is more consistent with the actual operation of the target main network, and the second predicted value is more accurate. The second predicted value can be directly determined as the target predicted value, and then the target predicted value is compared with the predicted value threshold. If the target predicted value is greater than or equal to the predicted value threshold, it means that the target main network is at risk. If the target predicted value is less than the predicted value threshold, it means that the target main network is not at risk.

[0040] It should be noted that if the difference does not fall within the preset error range, it means that the second predicted value of the second correction does not conform to the actual operating conditions of the target main network and needs to be further corrected. At this time, steps S102 and S103 are repeated until the difference between the corrected second predicted value and the target parameter value falls within the preset error range.

[0041] The risk prediction method for the main network provided in the embodiment of the present application obtains the operating data of the target main network within a preset period, and inputs the operating data into the target model to obtain a set of parameter prediction values ​​of the risk factors, wherein the set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operating data at a moment within the preset period, and the target model is trained by multiple sets of training samples, each set of training samples includes historical operating data and a set of historical parameter prediction values ​​of the risk factors; determines the weight of each parameter prediction value in a set of parameter prediction values, and weighted sums the set of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value for the initial correction; obtains the target parameter value of the risk factor, inputs the target parameter value and the first prediction value into a linear regression model to obtain a second prediction value for the secondary correction, wherein the linear regression model is trained by multiple historical first prediction values ​​and the target parameter value; calculates the difference between the target parameter value and the second prediction value, and when the difference is within a preset error interval, determines the second prediction value as the target prediction value, and when the target prediction value is greater than or equal to the prediction value threshold, determines that the target main network has a risk, thereby solving the problem of inaccurate prediction of the operating risk of the main network in the related art. A set of parameter prediction values ​​is obtained through the target model, and then the weighted average method is used to obtain the first predicted value of the initial correction. Then, the linear regression model is used to perform a second correction on the first predicted value of the initial correction. The difference between the target parameter value and the second predicted value is calculated, and a preset error interval is set. When the difference falls within the preset error interval, the second predicted value is determined as the target predicted value. Through the double deviation correction method and the preset error interval determination method, more effective and more accurate risk factor prediction values ​​can be obtained, thereby achieving the effect of improving the accuracy of predicting the operation risk of the main network.

[0042] In order to improve the accuracy of the second prediction value, when the difference does not fall within the preset error interval, the prediction value is repeatedly corrected. Optionally, in the risk prediction method for the main network provided in the embodiment of the present application, after calculating the difference between the historical parameter value and the second prediction value, the method further includes: when the difference does not fall within the preset error interval, repeating the steps of primary correction and secondary correction until the difference falls within the preset error interval, or the number of repeated executions is greater than or equal to a number threshold; when the number of repeated executions is greater than or equal to a number threshold, determining the second prediction value as the target prediction value.

[0043] In some examples, if the difference does not fall within the preset error interval, it means that the second predicted value does not match the actual operation of the target main network, and the accuracy of the second predicted value is low. At this time, it is necessary to repeat the steps of the first correction and the second correction until the difference between the corrected second predicted value and the target parameter value falls within the preset error interval. After multiple re-executions, if the difference still does not fall within the preset error interval, and the number of repetitions is greater than or equal to the number threshold, the second predicted value before the first correction and the second correction is repeated is used as the target predicted value.

[0044] This embodiment provides a quantitative basis for the accuracy of the predicted value of the risk factor by setting a preset error interval, ensuring that only the predicted value of the risk factor that meets the preset accuracy requirement (that is, the difference belongs to the preset error interval) will be determined as the target predicted value. When the difference does not belong to the preset error interval, the correction process is automatically triggered again until the prediction result meets the accuracy requirement. This closed-loop feedback mechanism ensures the continuous optimization of the predicted value of the risk factor and the continuous improvement of the accuracy.

[0045] In order to evaluate whether there is a risk in the target main network, it is necessary to screen out risk factors and target parameter values ​​from historical operating data. Optionally, in the risk prediction method for the main network provided in an embodiment of the present application, the risk factors and the target parameter values ​​of the risk factors are obtained in the following manner: obtaining historical fault records of the target main network, and extracting the cause of the fault from the historical fault records; determining the risk factors from the operating parameters of the target main network according to the cause of the fault, and determining the historical parameter values ​​of the risk factors as the target parameter values.

[0046] In some examples, the determination of risk factors can be determined based on expert experience or extracted from the fault causes of the target main network, for example, historical fault records are collected from the operation log, fault report, maintenance record and sensor data of the target main network. Extract the fault cause from the historical fault record. For example, if the report mentions "overload causes transformer overheating" many times, "overload" and "transformer overheating" can be identified as frequently occurring fault causes. Based on the classification and frequency statistics of the fault causes, determine the risk factors closely related to the operation of the main network. For example, in a commercial area power grid, "overload" and "voltage anomaly" may be two major risk factors. From the historical operation data, find out the parameters related to the above risk factors and record their historical values. For example, for the "overload" risk factor, the historical values ​​of parameters such as load current, active power and reactive power can be extracted from the historical operation data; for "voltage anomaly", the historical changes of bus voltage and user-end voltage are recorded.

[0047] After determining the risk factors, the historical parameter values ​​of the target main network in actual operation are extracted from the historical operation data according to the risk factors, and the historical parameter values ​​are used as the target parameter values ​​for evaluating whether the predicted values ​​of the risk factors are consistent with the actual operation conditions.

[0048] Through the above embodiments, by analyzing historical fault records, key risk factors affecting the operation of the main network are identified, and then the historical parameter values ​​(target parameter values) related to these risk factors are determined. By comparing the target parameter values ​​with the predicted values ​​of the risk factors, it is possible to more accurately assess whether the main network currently has operational risks, providing timely decision support for operation and maintenance personnel.

[0049] In order to improve the robustness of the predicted value of the risk factor, weights are set for the parameter prediction values ​​at different times. Optionally, in the risk prediction method for the main network provided in the embodiment of the present application, determining the weight of each parameter prediction value in a set of parameter prediction values ​​includes: determining the prediction moment of each parameter prediction value, wherein the prediction moment is a moment within a preset period; when the prediction moment belongs to a first preset time interval, determining the weight of the parameter prediction value as a first weight, wherein the first preset time interval is used to characterize the peak power consumption interval of the target main network; when the prediction moment belongs to a second preset time interval, determining the weight of the parameter prediction value as a second weight, wherein the second preset time interval is used to characterize the low power consumption interval of the target main network; when the prediction moment does not belong to the first preset time interval, and the prediction moment does not belong to the second preset time interval, determining the weight of the parameter prediction value as a third weight, wherein the first weight is greater than the third weight and greater than the second weight.

[0050] In some examples, the preset period can be 24 hours, and 17:00 to 22:00 is defined as the peak power consumption (first preset time interval), and 01:00 to 06:00 is defined as the low power consumption (second preset time interval). The rest of the time is the normal power consumption period. For the parameter prediction value during the peak power consumption period, since the power grid is under the greatest pressure at this time, the accuracy of the parameter prediction value and the identification of abnormal conditions are crucial to risk prediction, so a higher weight (first weight) is set. For example, the weight of the predicted load current or transformer temperature during the period from 17:00 to 22:00 is set to 0.8. For the parameter prediction value during the low power consumption period, the power grid load is small, and the direct risk of parameter abnormality is relatively low, but it still needs to be monitored, so a lower weight (second weight) is set. For example, the weight of the predicted voltage fluctuation during the period from 01:00 to 06:00 is set to 0.3. For the parameter prediction value during the normal power consumption period, its weight is between the peak and low periods (third weight). For example, the weight of the load current or transformer temperature prediction value during these periods can be set to 0.5. Each parameter prediction value is associated with a time interval weight corresponding to the prediction time. For example, if the predicted load current value is generated at 18:30, it will be assigned a first weight (0.8); if the predicted voltage fluctuation value is generated at 04:00, it will be assigned a second weight (0.3).

[0051] This embodiment sets weights based on different time intervals, highlights the importance of parameter prediction values ​​during peak and off-peak periods, and more accurately adjusts the output of the target model in the initial correction of parameter prediction values ​​to reflect the operating status and potential risks of the main grid at different time points. This improves the accuracy of risk prediction and provides more reliable risk warning and decision-making support for operation and maintenance personnel.

[0052] In order to ensure the stable operation of the target main network, after determining that the target main network is at risk, it is necessary to promptly execute an emergency response strategy or notify maintenance personnel. Optionally, in the risk prediction method for the main network provided in the embodiment of the present application, after determining that the target main network is at risk, the method further includes: obtaining historical emergency response strategies for risk factors, and executing the historical emergency response strategies on the target main network; after executing the historical emergency response strategies, obtaining an updated target prediction value; and when the updated target prediction value is greater than or equal to a prediction value threshold, issuing a prompt message, wherein the prompt message is used for operation and maintenance personnel to maintain the target main network.

[0053] In some examples, existing historical emergency response strategies for dealing with risk factors are collected from historical maintenance records and fault reports of the target main network. For example, cooling measures for equipment overheating, voltage regulation operations for voltage abnormalities, load distribution adjustments for overload conditions, etc. After determining that the target main network is at risk, the most relevant emergency measures are selected from the historical emergency response strategy library based on the risk factors. For example, if a transformer overheating is detected, the system will search for all strategies related to dealing with transformer overheating. The operation and maintenance personnel or the automation system executes the selected emergency response strategy. For example, for transformer overheating, emergency treatment includes increasing heat dissipation measures, reducing loads, adjusting the operating status of the transformer, etc.

[0054] After executing the emergency response strategy, the operating status of the target main grid is re-monitored, the latest operating data and risk factor parameter values ​​are re-acquired, and the updated target prediction value is obtained. The system compares the updated target prediction value with the preset prediction value threshold. The prediction value threshold is set according to the grid safety operation standards and historical data statistics, and is used to determine whether the risk has been effectively controlled. If the updated target prediction value is still greater than or equal to the prediction value threshold, that is, the level of the risk factor has not dropped to a safe range, a prompt message is automatically issued to notify the operation and maintenance personnel to take further maintenance actions. The prompt message may include specific risk factors, current parameter values, emergency treatments that have been executed and their effects, and recommended next actions.

[0055] In this embodiment, historical emergency response strategies are notified to respond to current risk factors. By executing relevant strategies and monitoring their effects, risk prediction and maintenance actions are dynamically adjusted, which enhances the main grid's ability to respond to risks. Through real-time warnings and recommended maintenance measures, operation and maintenance personnel can respond quickly to ensure the stability and safety of the power grid. In addition, the system can also continuously learn and optimize emergency strategies to improve the efficiency of risk prediction and processing.

[0056] In order to predict the operation risk of the target main network, it is necessary to train a target model. Optionally, in the risk prediction method of the main network provided in the embodiment of the present application, the target model is trained in the following manner: obtaining historical operation data of the target main network in multiple historical periods, and determining a set of historical parameter values ​​of risk factors during the operation period of the target main network corresponding to the historical operation data; determining the historical operation data in each historical period and a set of historical parameter values ​​as a set of training samples to obtain multiple groups of training samples; training a preset neural network model through multiple groups of training samples to obtain a target model.

[0057] In some examples, data from the past year is collected from the historical operation records of the target main grid. These data may include hourly load current, bus voltage, equipment temperature, humidity, wind speed and other environmental parameters, as well as the operating status of the power grid (such as normal, overloaded, voltage abnormality, etc.). Determine the parameter values ​​related to power grid risks from the historical operation data. For example, the load current values ​​a few hours before and after all overload events are collected as the historical parameter values ​​of the "load current" risk factor; if voltage abnormality is another key risk, the voltage data when the voltage fluctuation exceeds the threshold is used as the historical parameter value of the "voltage abnormality" risk factor.

[0058] For each historical cycle (such as every hour), the historical operating data (load current, voltage, environmental parameters, etc.) in the cycle are combined with the corresponding risk factor parameter values ​​(such as load current when overloaded, voltage value when voltage is abnormal) into a training sample. For example, the operating data from 17:00 to 18:00 on January 1, 2022 (load: 1500A, voltage: 20kV, temperature: 25℃, humidity: 60%, wind speed: 5m / s) and the risk parameter values ​​of "load current" (1500A) and "voltage abnormality" (20kV, beyond the normal range, assuming the normal range is 35kV±0.1kV) in the cycle are combined into a sample.

[0059] All constructed training samples are organized into a structured data set to facilitate input into the neural network model for training. Each row of the data set represents a sample of a historical period, including historical operating data and a set of historical parameter values. The neural network model can be a long short-term memory network model. First, initialize the model parameters, including the number of neurons, the number of layers, the activation function, etc. Then divide the organized data set into a training set, a validation set, and a test set, with a ratio of 70%, 15%, and 15%. The training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the model performance. Use the historical operating data in the training set and a set of historical parameter values ​​to train the neural network model. The goal is to enable the model to learn the correlation between the operating data and the risk parameter values, and then predict the parameter values ​​of future risk factors.

[0060] Based on the feedback from the validation set, adjust the model parameters, such as learning rate, batch size, regularization term, etc., to reduce overfitting and improve the generalization ability of the model. Evaluate the prediction performance of the model on the test set, and use indicators such as mean square error and mean absolute error to measure the gap between the risk parameter value predicted by the model and the actual value. After training is completed, save the optimal neural network model as the target model for risk prediction.

[0061] This embodiment uses the main grid historical operation data and a set of historical parameter values ​​to train a neural network model, and finally obtains a target model that can predict the parameter values ​​of power grid risk factors. The data set division, parameter adjustment and performance evaluation during the model training process improve the prediction accuracy and reliability of the model, so that the model can effectively identify and predict potential risks when facing the actual main grid operation situation.

[0062] In order to make the predicted value of the risk factor more consistent with the actual operation of the power grid, the first predicted value is corrected for the second time by training a linear regression model. Optionally, in the risk prediction method for the main grid provided in an embodiment of the present application, the linear regression model is trained in the following manner: multiple historical first predicted values ​​are obtained, and the target parameter value is used as the dependent variable. The initial linear regression model is trained with the multiple historical first predicted values ​​as the independent variable to obtain a linear regression model.

[0063] In some examples, a linear regression model is initialized using the Scikit-learn library (a machine learning library) in Python (a programming language). Model parameters such as fitting method, loss function, etc. are set. The training data set (i.e., the historical first prediction value and the target parameter value) is standardized to ensure that all independent variables are in a similar numerical range. The sorted data set is input into the initial linear regression model for training. The model will learn the linear relationship between the historical first prediction value and the target parameter value, that is, find a set of optimal weight parameters so that the model has the smallest error when predicting the first prediction value, and finally train a linear regression model.

[0064] This embodiment trains a linear regression model and uses the model to correct the first predicted value to obtain a second corrected value, so that the predicted value of the risk factor is more consistent with the actual operation of the target main network, thereby improving the accuracy of the risk prediction of the main network.

[0065] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0066] The embodiment of the present application also provides a risk prediction device for the main network. It should be noted that the risk prediction device for the main network of the embodiment of the present application can be used to execute the risk prediction method for the main network provided in the embodiment of the present application. The risk prediction device for the main network provided in the embodiment of the present application is introduced below.

[0067] Figure 2 Schematic diagram of a risk prediction device for a main network provided according to an embodiment of the present application. Figure 2 As shown, the device comprises:

[0068] The first acquisition unit 201 is used to acquire the operation data of the target main network within a preset period, and input the operation data into the target model to obtain a set of parameter prediction values ​​of the risk factors, wherein a set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operation data at a moment in the preset period, and the target model is trained by multiple sets of training samples, each set of training samples includes historical operation data and a set of historical parameter prediction values ​​of the risk factors;

[0069] A first determining unit 202 is used to determine the weight of each parameter prediction value in a group of parameter prediction values, and to perform weighted summation on the group of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value that is initially revised;

[0070] The second acquisition unit 203 is used to acquire the target parameter value of the risk factor, input the target parameter value and the first prediction value into a linear regression model, and obtain a second prediction value corrected twice, wherein the linear regression model is trained by a plurality of historical first prediction values ​​and target parameter values;

[0071] The calculation unit 204 is used to calculate the difference between the target parameter value and the second prediction value. When the difference is within a preset error range, the second prediction value is determined to be the target prediction value. When the target prediction value is greater than or equal to the prediction value threshold, it is determined that there is a risk in the target main network.

[0072] The risk prediction device for the main network provided in the embodiment of the present application obtains the operating data of the target main network within a preset period through the first acquisition unit 201, and inputs the operating data into the target model to obtain a set of parameter prediction values ​​of the risk factors, wherein a set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operating data at a moment in the preset period, and the target model is trained by multiple groups of training samples, each group of training samples includes historical operating data and a set of historical parameter prediction values ​​of the risk factors; the first determination unit 202 determines the weight of each parameter prediction value in a set of parameter prediction values, and weighted sums a set of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value corrected for the first time; the second acquisition unit 203 obtains the target parameter value of the risk factor, inputs the target parameter value and the first prediction value into the linear regression model, and obtains a second prediction value corrected for the second time, wherein the linear regression model is composed of multiple historical first prediction values ​​and target parameter values. The numerical training is obtained; the calculation unit 204 calculates the difference between the target parameter value and the second prediction value. When the difference is within the preset error interval, the second prediction value is determined to be the target prediction value. When the target prediction value is greater than or equal to the prediction value threshold, it is determined that there is a risk in the target main network, which solves the problem of inaccurate prediction of the operation risk of the main network in the related technology. A set of parameter prediction values ​​is obtained through the target model, and then the first prediction value of the initial correction is obtained by weighted average. Then, the linear regression model is used to perform a secondary correction on the first prediction value of the initial correction, and then the difference between the target parameter value and the second prediction value is calculated, and the preset error interval is set. When the difference is within the preset error interval, the second prediction value is determined to be the target prediction value. Through the double deviation correction method and the preset error interval determination method, more effective and more accurate prediction values ​​of risk factors can be obtained, thereby achieving the effect of improving the accuracy of predicting the operation risk of the main network.

[0073] Optionally, in the risk prediction device of the main network provided in the embodiment of the present application, the device also includes: an execution unit, which is used to repeat the steps of primary correction and secondary correction when the difference does not fall within the preset error interval, until the difference falls within the preset error interval, or the number of repeated executions is greater than or equal to a number threshold; a second determination unit, which is used to determine that the second prediction value is the target prediction value when the number of repeated executions is greater than or equal to the number threshold.

[0074] Optionally, in the risk prediction device of the main network provided in the embodiment of the present application, the device also includes: an extraction unit, used to obtain historical fault records of the target main network and extract fault causes from the historical fault records; a third determination unit, used to determine risk factors from the operating parameters of the target main network based on the fault causes, and determine the historical parameter values ​​of the risk factors as the target parameter values.

[0075] Optionally, in the risk prediction device of the main network provided in the embodiment of the present application, the first determination unit 202 includes: a first determination module, used to determine the prediction moment of each parameter prediction value, wherein the prediction moment is a moment within a preset period; a second determination module, used to determine the weight of the parameter prediction value as a first weight when the prediction moment belongs to a first preset time interval, wherein the first preset time interval is used to characterize the peak power consumption interval of the target main network; a third determination module, used to determine the weight of the parameter prediction value as a second weight when the prediction moment belongs to a second preset time interval, wherein the second preset time interval is used to characterize the low power consumption interval of the target main network; a fourth determination module, used to determine the weight of the parameter prediction value as a third weight when the prediction moment does not belong to the first preset time interval and the prediction moment does not belong to the second preset time interval, wherein the first weight is greater than the third weight and greater than the second weight.

[0076] Optionally, in the risk prediction device of the main network provided in the embodiment of the present application, the device also includes: a third acquisition unit, used to obtain historical emergency response strategies for risk factors, and execute the historical emergency response strategies on the target main network; a fourth acquisition unit, used to obtain an updated target prediction value after executing the historical emergency response strategy; a prompt unit, used to issue a prompt message when the updated target prediction value is greater than or equal to a prediction value threshold, wherein the prompt information is used for operation and maintenance personnel to maintain the target main network.

[0077] Optionally, in the risk prediction device of the main network provided in the embodiment of the present application, the device also includes: a fifth acquisition unit, used to acquire historical operation data of the target main network within multiple historical periods, and determine a set of historical parameter values ​​of risk factors during the operation period of the target main network corresponding to the historical operation data; a fourth determination unit, used to determine the historical operation data and a set of historical parameter values ​​within each historical period as a set of training samples to obtain multiple groups of training samples; a first training unit, used to train a preset neural network model through multiple groups of training samples to obtain a target model.

[0078] Optionally, in the risk prediction device of the main network provided in the embodiment of the present application, the device also includes: a second training unit, used to obtain multiple historical first prediction values, use the target parameter value as the dependent variable, and use the multiple historical first prediction values ​​as the independent variable to train the initial linear regression model to obtain a linear regression model.

[0079] The risk prediction device of the main network includes a processor and a memory. The above-mentioned first acquisition unit 201, first determination unit 202, second acquisition unit 203 and calculation unit 204 are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0080] The processor contains a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and the accuracy of predicting the operation risk of the main network can be improved by adjusting the kernel parameters.

[0081] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0082] An embodiment of the present invention provides a computer-readable storage medium having a program stored thereon, which implements a risk prediction method for a main network when the program is executed by a processor.

[0083] An embodiment of the present invention provides a processor, which is used to run a program, wherein the risk prediction method of the main network is executed when the program is running.

[0084] Figure 3 is a schematic diagram of an electronic device provided according to an embodiment of the present application. Figure 3 As shown, the electronic device 301 includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining operation data within a preset period of the target main network, and inputting the operation data into a target model to obtain a set of parameter prediction values ​​of the risk factor, wherein a set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operation data at a moment within the preset period, and the target model is trained by multiple sets of training samples, each set of training samples includes historical operation data and a set of historical parameter prediction values ​​of the risk factor; determining the weight of each parameter prediction value in a set of parameter prediction values, and weighted summing a set of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value of the initial correction; obtaining a target parameter value of the risk factor, inputting the target parameter value and the first prediction value into a linear regression model to obtain a second prediction value of the secondary correction, wherein the linear regression model is trained by multiple historical first prediction values ​​and the target parameter value; calculating the difference between the target parameter value and the second prediction value, and determining the second prediction value as the target prediction value when the difference belongs to the preset error interval, and determining that the target main network has a risk when the target prediction value is greater than or equal to the prediction value threshold. The devices in this article can be servers, PCs, PADs, mobile phones, etc.

[0085] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that is initialized with the following method steps: obtaining operating data within a preset period of the target main network, and inputting the operating data into a target model to obtain a set of parameter prediction values ​​of the risk factors, wherein the set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operating data at a moment within the preset period, and the target model is trained by multiple sets of training samples, each set of training samples includes historical operating data and a set of historical parameter prediction values ​​of the risk factors; determining the weight of each parameter prediction value in a set of parameter prediction values, and weighted summing the set of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value for the initial correction; obtaining a target parameter value of the risk factor, inputting the target parameter value and the first prediction value into a linear regression model to obtain a second prediction value for the secondary correction, wherein the linear regression model is trained by multiple historical first prediction values ​​and the target parameter value; calculating the difference between the target parameter value and the second prediction value, and when the difference is within a preset error interval, determining the second prediction value as the target prediction value, and when the target prediction value is greater than or equal to a prediction value threshold, determining that the target main network is at risk.

[0086] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0087] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0088] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0089] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0090] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0091] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0092] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0093] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0094] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0095] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A risk prediction method for a main network, characterized in that: include: Obtaining operation data of a target main network within a preset period, and inputting the operation data into a target model to obtain a set of parameter prediction values ​​of risk factors, wherein the set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to operation data at a time within the preset period, and the target model is trained by multiple sets of training samples, each set of training samples includes historical operation data and a set of historical parameter prediction values ​​of risk factors; Determine the weight of each parameter prediction value in the set of parameter prediction values, and weighted sum the set of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value that is initially revised; Obtaining a target parameter value of the risk factor, inputting the target parameter value and the first predicted value into a linear regression model to obtain a second predicted value corrected twice, wherein the linear regression model is trained by a plurality of historical first predicted values ​​and the target parameter value; Calculate the difference between the target parameter value and the second predicted value. If the difference is within a preset error range, determine that the second predicted value is the target predicted value. If the target predicted value is greater than or equal to a predicted value threshold, determine that there is a risk in the target main network.

2. The method according to claim 1, characterized in that: After calculating the difference between the target parameter value and the second predicted value, the method further includes: If the difference does not fall within the preset error range, repeat the steps of primary correction and secondary correction until the difference falls within the preset error range, or the number of repetitions is greater than or equal to a threshold number; When the number of repeated executions is greater than or equal to the number threshold, the second prediction value is determined to be a target prediction value.

3. The method according to claim 1, characterized in that The risk factors and the target parameter values ​​of the risk factors are obtained in the following manner: Obtaining historical fault records of the target primary network, and extracting fault causes from the historical fault records; The risk factor is determined from the operating parameters of the target main network according to the fault cause, and the historical parameter value of the risk factor is determined as the target parameter value.

4. The method according to claim 1, characterized in that: Determining the weight of each parameter prediction value in the set of parameter prediction values ​​includes: Determine a prediction time of each parameter prediction value, wherein the prediction time is a time within the preset period; In the case where the predicted moment belongs to a first preset time interval, the weight of the parameter prediction value is determined as a first weight, wherein the first preset time interval is used to characterize a peak power consumption interval of the target main grid; In the case where the predicted moment belongs to a second preset time interval, the weight of the parameter prediction value is determined as a second weight, wherein the second preset time interval is used to characterize a low power consumption interval of the target main network; When the predicted moment does not belong to the first preset time interval and the predicted moment does not belong to the second preset time interval, the weight of the parameter prediction value is determined as a third weight, wherein the first weight is greater than the third weight and greater than the second weight.

5. The method according to claim 1, characterized in that After determining that the target main network has risks, the method further includes: Acquire a historical emergency response strategy for the risk factor, and execute the historical emergency response strategy on the target main network; After executing the historical emergency response strategy, obtaining an updated target prediction value; When the updated target prediction value is greater than or equal to the prediction value threshold, a prompt message is issued, wherein the prompt message is used by operation and maintenance personnel to maintain the target main network.

6. The method according to claim 1, characterized in that The target model is trained in the following way: Acquire historical operation data of the target main network in multiple historical periods, and determine a set of historical parameter values ​​of the risk factors during the operation period of the target main network corresponding to the historical operation data; Determine the historical operation data and the set of historical parameter values ​​in each historical period as a set of training samples to obtain multiple sets of training samples; The preset neural network model is trained by the multiple groups of training samples to obtain the target model.

7. The method according to claim 1, characterized in that The linear regression model is trained in the following way: A plurality of historical first prediction values ​​are obtained, and an initial linear regression model is trained with the target parameter value as a dependent variable and the plurality of historical first prediction values ​​as independent variables to obtain the linear regression model.

8. A risk prediction device for a main network, characterized in that: include: A first acquisition unit is used to acquire operation data of a target main network within a preset period, and input the operation data into a target model to obtain a set of parameter prediction values ​​of risk factors, wherein the set of parameter prediction values ​​includes multiple parameter prediction values, each parameter prediction value corresponds to the operation data at a moment in the preset period, and the target model is trained by multiple sets of training samples, each set of training samples includes historical operation data and a set of historical parameter prediction values ​​of risk factors; A first determining unit is used to determine the weight of each parameter prediction value in the set of parameter prediction values, and to perform weighted summation on the set of parameter prediction values ​​according to the weight of each parameter prediction value to obtain a first prediction value that is initially revised; a second acquisition unit, configured to acquire a target parameter value of the risk factor, input the target parameter value and the first prediction value into a linear regression model, and obtain a second prediction value corrected twice, wherein the linear regression model is trained by a plurality of historical first prediction values ​​and the target parameter value; A calculation unit is used to calculate the difference between the target parameter value and the second prediction value, and when the difference is within a preset error range, determine that the second prediction value is a target prediction value, and when the target prediction value is greater than or equal to a prediction value threshold, determine that there is a risk in the target main network.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the risk prediction method for the main network described in any one of claims 1 to 7 is implemented.

10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the main network risk prediction method described in any one of claims 1 to 7.

Images