Ring main unit fault rapid early warning method and system

Abstract

The invention relates to the field of ring main unit engineering, in particular to a ring main unit fault rapid early warning method and system, and the method comprises the steps: obtaining the operation environment data collected by each preset sensor in a ring main unit housing, and generating an environment disturbance factor vector according to the operation environment data; and screening the environmental disturbance factor vectors based on a preset threshold to obtain an abnormal disturbance factor vector. According to the invention, multi-dimensional comprehensive monitoring of the operation environment of the ring main unit is realized, and potential fault factors such as gas leakage, humidity abnormity and mechanical vibration can be identified in time; compared with a single monitoring means, a monitoring blind area is effectively eliminated, and failure report omission is avoided; noise interference is filtered through signal processing, effective abnormal features are extracted, high-reliability data support is provided for fault early warning, and the system sensitivity is remarkably improved.

Description

TECHNICAL FIELD

[0001] The application relates to the field of ring main unit engineering, in particular to a ring main unit fault rapid early warning method and system. BACKGROUND

[0002] In a power distribution system, as a key control and protection device, the running state of a ring main unit directly affects the stability and power supply reliability of the power grid. When the ring main unit fails, it is easy to cause regional power outage accidents, causing significant economic losses and social impact. Therefore, an effective fault early warning mechanism needs to be established to significantly reduce the incidence of power outage accidents by early detection and handling of potential fault hazards, thereby improving the reliability and continuity indicators of the power supply system.

[0003] In the prior art, ring main unit fault early warning mainly adopts a monitoring scheme based on a single sensor to judge faults by collecting local state parameters of the equipment. However, this method has obvious technical defects: due to the lack of comprehensive consideration of environmental disturbance factors, it is difficult to accurately assess the combined influence on the running state of the equipment, thereby often causing early warning response lag or fault omission, seriously affecting the practical value of the early warning system.

[0004] Therefore, there is an urgent need for a ring main unit fault rapid early warning method and system. SUMMARY

[0005] Therefore, it is necessary to provide a ring main unit fault rapid early warning method and system which is reliable, timely and accurate in early warning.

[0006] The technical scheme of the application is as follows: A ring main unit fault rapid early warning method, the method comprising: acquiring running environment data collected by each preset sensor in a ring main unit shell, and generating an environmental disturbance factor vector according to the running environment data; screening the environmental disturbance factor vector based on a preset threshold to obtain an abnormal disturbance factor vector; acquiring equipment response data collected by each preset sensor in the ring main unit shell, obtaining an equipment response vector by combination, and generating a time delay input vector according to the equipment response vector and the abnormal disturbance factor vector; processing the time delay input vector based on a preset neural network ring main unit response prediction model to output a predicted response vector, and generating an equipment response amplitude according to the predicted response vector; calculating the coupling strength score of the equipment response vector and the environmental disturbance factor vector, constructing a response risk map based on the coupling strength score, and obtaining an active path edge weight sum according to the response risk map; Output a risk level and an associated early warning level strategy according to the device response amplitude and the sum of the active path edge weights.

[0007] Specifically, the operation environment data collected by each preset sensor in the ring main unit shell is acquired, and an environment disturbance factor vector is generated according to the operation environment data, including: Operation environment data collected by each preset sensor in the ring main unit shell is acquired; The operation environment data is combined to construct and generate an environment disturbance factor vector.

[0008] Specifically, the environment disturbance factor vector is screened based on a preset threshold to obtain an abnormal disturbance factor vector, including: The environment disturbance factor vector is screened based on a preset threshold to obtain an abnormal disturbance feature; The abnormal disturbance features are integrated to generate an abnormal disturbance factor vector.

[0009] Specifically, device response data collected by each preset sensor in the ring main unit shell is acquired, a device response vector is obtained by combination, a time delay input vector is generated according to the device response vector and the abnormal disturbance factor vector, including: Device response data collected by each preset sensor in the ring main unit shell is acquired, and a device response vector is obtained by combination; Based on the correlation analysis of the device response vector and the abnormal disturbance factor vector, a cross-correlation function is constructed; The peak position of the cross-correlation function is calculated and determined, the corresponding time delay parameter is extracted, and a time delay input vector is constructed according to the time delay parameter.

[0010] Specifically, the time delay input vector is processed based on a preset neural network ring main unit response prediction model to output a predicted response vector, and a device response amplitude is generated according to the predicted response vector, including: The time delay input vector is processed based on a preset neural network ring main unit response prediction model to output a predicted response vector; Based on the difference between the predicted value in the predicted response vector and the preset device response reference value, a device response amplitude is calculated and generated.

[0011] Specifically, the coupling strength score of the device response vector and the environment disturbance factor vector is calculated, and a response risk map is constructed based on the coupling strength score, an active path edge weight sum is obtained according to the response risk map, including: The coupling strength score of the device response vector and the environment disturbance factor vector is calculated; obtain historical ring main unit fault data and simulation ring main unit experimental data, and construct a response risk map based on the coupling strength score, the historical ring main unit fault data, and the simulation ring main unit experimental data; According to the response risk map, the total edge weight of the active path is calculated.

[0012] Specifically, the device response amplitude and the total edge weight of the active path are used to output a risk level and an associated early warning level strategy, including: According to the device response amplitude and the total edge weight of the active path, a ring main unit fault comprehensive score is calculated; Based on the ring main unit fault comprehensive score, a risk level is generated; According to the risk level, an associated early warning level strategy is matched.

[0013] Specifically, a ring main unit fault rapid early warning system is also provided, and the system comprises: An environmental disturbance vector generation module is configured to obtain operation environment data collected by each preset sensor in a ring main unit shell, and generate an environmental disturbance factor vector based on the operation environment data; An abnormal disturbance vector acquisition module is configured to filter the environmental disturbance factor vector based on a preset threshold, and obtain an abnormal disturbance factor vector; A time delay input vector generation module is configured to obtain device response data collected by each preset sensor in a ring main unit shell, obtain a device response vector by combination, and generate a time delay input vector based on the device response vector and the abnormal disturbance factor vector; A device response amplitude generation module is configured to process the time delay input vector based on a preset neural network ring main unit response prediction model, output a predicted response vector, and generate a device response amplitude based on the predicted response vector; An active path weight acquisition module is configured to calculate a coupling strength score of the device response vector and the environmental disturbance factor vector, construct a response risk map based on the coupling strength score, and obtain a total edge weight of an active path based on the response risk map; An early warning level strategy output module is configured to output a risk level and an associated early warning level strategy based on the device response amplitude and the total edge weight of the active path.

[0014] Specifically, the environmental disturbance vector generation module is further configured to obtain operation environment data collected by each preset sensor in a ring main unit shell, and combine the operation environment data to construct and generate an environmental disturbance factor vector.

[0015] Specifically, the abnormal disturbance vector acquisition module is further configured to: filter the environmental disturbance factor vector based on a preset threshold to obtain an abnormal disturbance feature; and integrate the abnormal disturbance feature to generate an abnormal disturbance factor vector.

[0016] Specifically, the time delay input vector generation module is further configured to: obtain equipment response data collected by each preset sensor in the ring main unit shell to obtain an equipment response vector by combination; construct a cross-correlation function based on correlation analysis of the equipment response vector and the abnormal disturbance factor vector; determine a peak position of the cross-correlation function, extract a corresponding time delay parameter, and construct a time delay input vector according to the time delay parameter.

[0017] Specifically, the equipment response amplitude generation module is further configured to: process the time delay input vector based on a preset neural network ring main unit response prediction model to output a predicted response vector; and calculate and generate an equipment response amplitude based on a difference between a predicted value in the predicted response vector and a preset equipment response reference value.

[0018] Specifically, the active path weight acquisition module is further configured to: calculate a coupling strength score of the equipment response vector and the environmental disturbance factor vector; obtain historical ring main unit fault data and simulated ring main unit experiment data, and construct a response risk map based on the coupling strength score, the historical ring main unit fault data and the simulated ring main unit experiment data; and calculate and obtain an active path edge weight sum according to the response risk map.

[0019] Specifically, the early warning level strategy output module is further configured to: calculate and obtain a ring main unit fault comprehensive score according to the equipment response amplitude and the active path edge weight sum; generate a risk level based on the ring main unit fault comprehensive score; and match an associated early warning level strategy according to the risk level.

[0020] Optionally, a computer device is also provided, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the ring main unit fault rapid early warning method when executing the computer program.

[0021] Optionally, a computer readable storage medium is also provided, which stores a computer program, and the computer program implements the steps of the ring main unit fault rapid early warning method when executed by a processor.

[0022] The present application relates to machine learning and deep learning technology, and the technical effects achieved are as follows: (1) The ring main unit fault rapid early warning method and system successively acquires operation environment data collected by each preset sensor in the ring main unit shell, generates an environment disturbance factor vector according to the operation environment data, screens the environment disturbance factor vector based on a preset threshold, acquires device response data collected by each preset sensor in the ring main unit shell, obtains a device response vector by combination, generates a time delay input vector according to the device response vector and the abnormal disturbance factor vector, processes the time delay input vector based on a preset neural network ring main unit response prediction model, outputs a predicted response vector, and generates a device response amplitude according to the predicted response vector. The coupling strength score of the device response vector and the environment disturbance factor vector is calculated, a response risk map is constructed based on the coupling strength score, an active path edge weight sum is acquired according to the response risk map, the risk level and the associated early warning level strategy are output according to the device response amplitude and the active path edge weight sum, multi-dimensional comprehensive monitoring of the ring main unit operation environment is realized, potential fault factors such as gas leakage, humidity anomaly and mechanical vibration can be identified in time, compared with a single monitoring means, the monitoring blind area is effectively eliminated, and fault omission is avoided. Noise interference is filtered through signal processing, effective abnormal features are extracted, high-reliability data support is provided for fault early warning, and the system sensitivity is significantly improved. (2) The time delay characteristics of the environment disturbance factor and the device response are calculated by constructing a cross-correlation function, a time delay input vector is formed and input into an LSTM neural network model, and the time sequence prediction of the device dynamic response is realized. This technical scheme innovatively breaks through the dependence of traditional prediction models on real-time monitoring data, and can identify the fault development trend in advance. In specific implementation, mechanical voiceprint features are extracted through short-time Fourier transform, and the time delay input vector can effectively predict potential fault risks such as mechanical component wear; (3) The edge weight sum of the active path in the response risk map and the device response amplitude are combined to obtain a ring main unit fault comprehensive score, and the risk level and the matching early warning level strategy are evaluated accordingly. The comprehensive quantitative evaluation and graded early warning of the overall fault risk of the ring main unit are realized, and the accuracy of risk identification and the timeliness of early warning are significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 It is a flowchart of the ring main unit fault rapid early warning method in one embodiment; Figure 2 It is a structural block diagram of the ring main unit fault rapid early warning system in one embodiment; Figure 3 It is a structural block diagram of the computer device in one embodiment. DETAILED DESCRIPTION

[0024] In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, circuits, and

[0025] It will be understood that the term "includes," "including," "has," "having," "comprises," "comprising," "contains" or "containing," when used in this specification and in the following claims, specifies the presence of the stated features, integers, steps, operations, elements, and / or components but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0026] It will be understood that the term "and / or," when used in this specification and in the following claims, can encompass the meaning of "and" and / or the meaning of "or." Similarly, the term "and / or" when used in the context of the "and / or" of a list of items, covers all of the following interpretations of the list of items: alternative (at least one of the items in the list is present and none of the items in the list are present), and / or conjunctive (all of the items in the list are present).

[0027] As used in this specification and claims, the terms "if" and "when" can be interpreted to mean "upon" or "in response to a determination" or "in response to a detection" depending on the context. Similarly, the phrase "if it is determined" or "if [a described condition or event] is detected" can be interpreted to mean "upon determining" or "in response to determining" or "upon detecting [the described condition or event]" or "in response to detecting [the described condition or event]" depending on the context.

[0028] In addition, the terms "first," "second," "third," etc. are used herein only to describe different instances of an element, and do not imply a relative importance of the elements.

[0029] The terms "one embodiment," "an embodiment,” "some embodiments,” or the like, as used in the specification, mean that the particular feature, structure, or characteristic being described is included in at least one embodiment of the application. Thus, the appearances of the phrase "in one embodiment,” "in some embodiments,” "in other embodiments,” "in additional embodiments,” or the like, in various places in the specification are not necessarily referring to the same embodiment, unless otherwise noted. The terms "including,” "containing,” "comprising,” "having,” and the like are meant to be inclusive and mean that there can be additional such features, elements, steps, operations, articles, or components in addition to those specifically recited.

[0030] In one embodiment, a terminal is provided, which is configured to: acquire operation environment data collected by each preset sensor in a ring main unit shell, and generate an environment disturbance factor vector according to the operation environment data; filter the environment disturbance factor vector based on a preset threshold to obtain an abnormal disturbance factor vector; acquire device response data collected by each preset sensor in the ring main unit shell, obtain a device response vector by combination, and generate a time delay input vector according to the device response vector and the abnormal disturbance factor vector; process the time delay input vector based on a preset neural network ring main unit response prediction model, output a predicted response vector, and generate a device response amplitude according to the predicted response vector; calculate a coupling strength score of the device response vector and the environment disturbance factor vector, construct a response risk graph based on the coupling strength score, acquire an active path edge weight sum according to the response risk graph; and output a risk level and an associated early warning level strategy according to the device response amplitude and the active path edge weight sum.

[0031] The terminal can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers and portable wearable devices.

[0032] In one embodiment, as shown in Figure 1 A ring main unit fault rapid early warning method is provided, which includes: Step S100: acquiring operation environment data collected by each preset sensor in a ring main unit shell, and generating an environment disturbance factor vector according to the operation environment data; Step S200: filtering the environment disturbance factor vector based on a preset threshold to obtain an abnormal disturbance factor vector; Step S300: acquiring device response data collected by each preset sensor in the ring main unit shell, obtaining a device response vector by combination, and generating a time delay input vector according to the device response vector and the abnormal disturbance factor vector; Step S400: processing the time delay input vector based on a preset neural network ring main unit response prediction model, outputting a predicted response vector, and generating a device response amplitude according to the predicted response vector; Step S500: calculating a coupling strength score of the device response vector and the environment disturbance factor vector, constructing a response risk graph based on the coupling strength score, and acquiring an active path edge weight sum according to the response risk graph; Step S600: outputting a risk level and an associated early warning level strategy according to the device response amplitude and the active path edge weight sum.

[0033] The application is based on a ring net cabinet fault rapid early warning method and system. The operation environment data collected by each preset sensor in the ring net cabinet shell is obtained in turn, and an environment disturbance factor vector is generated according to the operation environment data. The environment disturbance factor vector is screened based on a preset threshold to obtain an abnormal disturbance factor vector. The equipment response data collected by each preset sensor in the ring net cabinet shell is obtained, and an equipment response vector is obtained by combination. A time delay input vector is generated according to the equipment response vector and the abnormal disturbance factor vector. The time delay input vector is processed based on a preset neural network ring net cabinet response prediction model to output a predicted response vector, and an equipment response amplitude is generated according to the predicted response vector. The coupling strength score of the equipment response vector and the environment disturbance factor vector is calculated, and a response risk map is constructed based on the coupling strength score. The total weight of the active path edge is obtained according to the response risk map. The risk level and the associated early warning level strategy are output according to the equipment response amplitude and the total weight of the active path edge, realizing multi-dimensional comprehensive monitoring of the operation environment of the ring net cabinet, which can timely identify potential fault factors such as gas leakage, humidity anomaly and mechanical vibration. Compared with a single monitoring method, the monitoring blind area is effectively eliminated, and fault omission is avoided. Through signal processing, noise interference is filtered out, effective abnormal features are extracted, high reliability data support is provided for fault early warning, and the system sensitivity is significantly improved. The time delay characteristics of the environment disturbance factor and the equipment response are calculated by constructing a cross-correlation function, a time delay input vector is formed and input into an LSTM neural network model, and the time sequence prediction of the equipment dynamic response is realized. The technical scheme innovatively breaks through the dependence of the traditional prediction model on real-time monitoring data, and can identify the fault development trend in advance. In specific implementation, the mechanical soundprint features are extracted by short-time Fourier transform, and the time delay input vector can effectively predict potential fault risks such as mechanical component wear; by combining the total weight of the edge weight of the active path in the response risk map and the equipment response amplitude, a ring net cabinet fault comprehensive score is obtained, and the risk level and the matching early warning level strategy are evaluated. The comprehensive quantitative evaluation and graded early warning of the overall fault risk of the ring net cabinet are realized, and the accuracy of risk identification and the timeliness of early warning are significantly improved.

[0034] In one embodiment, step S100: obtaining operation environment data collected by each preset sensor in the ring net cabinet shell, and generating an environment disturbance factor vector according to the operation environment data, comprises: Step S100: obtaining operation environment data collected by each preset sensor in the ring net cabinet shell, and generating an environment disturbance factor vector according to the operation environment data; Step S110: obtaining operation environment data collected by each preset sensor in the ring net cabinet shell; Step S120: combining the operation environment data to construct and generate an environment disturbance factor vector.

[0035] In this embodiment, the collection of operating environment data is mainly through the installation of sensors on the surface and internal key areas (such as switch contacts, bus connection, mechanical transmission parts) of the ring net cabinet shell. Specifically, the sensor types include: gas leakage sensor, distributed humidity probe, low-frequency microseismic detection device.

[0036] Further, the sensor setting method is as follows: Gas leakage sensor: Install a miniature sensor at the SF6 gas leakage point (such as a sealing ring, a pipe joint), collect distributed layout (install one every 5 cm), collect SF6 concentration, and record it as (unit: ppm, t is time). This kind of sensor has high precision and high sensitivity, and can detect trace amounts of SF6 leakage. At the same time, in order to ensure the accuracy of the data, the sensor should be calibrated regularly, and the calibration period can be set according to the frequency of use of the equipment and environmental conditions, for example, calibrated once every 3-6 months.

[0037] Distributed humidity probe: Install a patch-type humidity sensor on the surface of the cabinet metal parts (such as bus support), use magnetic attraction or support to fix, collect relative humidity, and record it as (unit: %RH). The humidity probe should have fast response and anti-interference ability, and can accurately measure the humidity in a complex electromagnetic environment. In addition, the probe should have a certain protection level to prevent water vapor condensation from damaging the probe itself. In order to improve the reliability of humidity monitoring, a redundant design can be used, that is, multiple humidity probes are installed in key areas, and the collected data is processed through a data fusion algorithm to obtain more accurate humidity values.

[0038] Low-frequency microseismic detection device: Install vibration sensors on both ends of the cabinet transmission shaft, connecting rod hinge, and bolt head to capture abnormal stress changes and collect vibration acceleration, and record it as . The microseismic detection device should have wideband and high resolution, and can capture small vibration signals. At the same time, the device should have good anti-noise performance, and can distinguish between normal vibration and abnormal vibration generated by the equipment. For the low-frequency microseismic detection device, the sampling frequency is reasonably set, and it is generally recommended that the sampling frequency is not less than 10 times the vibration frequency to ensure that the characteristics of the vibration signal can be completely collected. At the same time, real-time analysis is performed on the collected vibration data, and the wavelet transform signal processing method is used to extract the characteristic parameters (vibration acceleration) of the vibration signal for subsequent fault diagnosis.

[0039] In this embodiment, the environmental disturbance factor is defined as a time series vector: where each component corresponds to a type of environmental disturbance data.

[0040] In one embodiment, step S200: filtering the environmental disturbance factor vector based on a preset threshold value to obtain an abnormal disturbance factor vector, comprising: Step S210: filtering the environmental disturbance factor vector based on a preset threshold value to obtain an abnormal disturbance feature; Step S220: integrating the abnormal disturbance feature to generate an abnormal disturbance factor vector.

[0041] In this embodiment, the preset threshold value is set as follows:

[0042] wherein, is the statistical mean of historical environmental data, is the standard deviation (j=1, 2, 3, corresponding to the gas concentration, humidity, and vibration acceleration in the environmental disturbance factor vector), is the preset threshold value of the jth signal at time t, is an adaptive coefficient that changes with time t and is used to adjust the looseness of the preset threshold value The preset threshold value formula is used to determine a dynamically changing threshold value for determining whether the currently collected signal belongs to an abnormal situation.

[0043] Further, the adaptive coefficient is calculated by the sliding window mean method, and the calculation method is as follows:

[0044] wherein, α is an adjustment coefficient (such as 0.1), is the window length (such as 1 hour), and are the mean values at the current time and at a certain time in the past (the current time minus the window length), respectively, is the standard deviation at the current time.

[0045] This calculation method combines the statistical mean of the historical environmental data and the standard deviation and introduces the adaptive coefficient , so that the preset threshold value can be adjusted according to the changes in the environment and the fluctuations in the data. Specifically, at T1, the historical data is t-1, t-2, t-n, and the dynamic preset threshold value corresponding to T1 is obtained at this time; at T2, the historical data is T1, t-1, t-2, t-n, and the dynamic preset threshold value corresponding to T2 is obtained at this time, thereby dynamically adjusting the preset threshold value with real-time environmental data and improving the accuracy of abnormal situation judgment.

[0046] Further, the preset threshold value is normalized as a dimensionless parameter as follows: The method of normalizing as a dimensionless parameter is as follows:

[0047] wherein, is the normalized preset threshold value of the jth signal at time t.

[0048] Further, the normalized value of each signal component is calculated as follows:

[0049] wherein, is the normalized value of the jth signal at time t. is the actual acquisition value of the jth signal at time t (for example, is the gas concentration, is the humidity, is the vibration acceleration). The formula is used to normalize the actual signal value acquired, so as to convert it into a dimensionless value. The normalized value can be more conveniently compared with the preset threshold value to determine whether the signal is abnormal.

[0050] Further, for each signal component, the following is compared: and If > it is determined as an abnormal disturbance feature and extracted, and finally integrated into an abnormal disturbance factor vector, which is set as follows:

[0051] wherein, is the abnormal disturbance factor vector at time t, is an indicator function, which takes 1 when the condition is met, and 0 otherwise (when > 1 when the condition is met, and 0 otherwise). According to the comparison result of the normalized value and the preset threshold value, the abnormal feature is extracted from the original acquisition signal in the environmental disturbance factor vector. Only when the normalized value exceeds the dynamic threshold value, the corresponding original signal value will be retained in the abnormal feature vector for subsequent analysis and processing.

[0052] In one embodiment, step S300: obtaining the equipment response data acquired by each preset sensor in the ring network cabinet shell, obtaining the equipment response vector by combination, generating the time delay input vector according to the equipment response vector and the abnormal disturbance factor vector, comprising: Step S310: obtaining the equipment response data acquired by each preset sensor in the ring network cabinet shell, obtaining the equipment response vector by combination; Step S320: Based on the correlation analysis of the device response vector and the abnormal disturbance factor vector, a cross-correlation function is constructed; Step S330: Calculate the peak position of the cross-correlation function, extract the corresponding time delay parameter, and construct a time delay input vector according to the time delay parameter.

[0053] In this embodiment, the collected device response data includes the load switch contact temperature: (unit: ℃), bus node hot spot temperature: (unit: ℃), mechanical transmission noise characteristics: extract the frequency domain feature vector through short-time Fourier transform The data collection process is as follows: Load switch contact temperature: Install high-precision temperature sensors such as thermocouples or thermistors at the load switch contacts. These sensors should have the characteristics of fast response and high accuracy, and can measure the temperature changes of the contacts in real time. The sensors are connected to the data acquisition system through wired or wireless methods, and the data acquisition system collects the contact temperature at a certain sampling frequency (for example, sampling once every 10 seconds). In order to ensure the reliability of the data, when installing the sensor, attention should be paid to the good contact between the sensor and the contact, and the measurement error caused by poor contact should be avoided. At the same time, the sensor should be calibrated regularly, and the calibration period can be determined according to the use frequency and importance of the device, for example, calibrated once every 36 months.

[0054] Bus node hot spot temperature: For bus nodes, especially key parts prone to hot spots, install infrared temperature sensors or optical fiber temperature sensors. Infrared temperature sensors can measure temperature non-contact, and optical fiber temperature sensors have the advantage of strong anti-electromagnetic interference ability. These sensors will transmit the real-time collected bus node temperature data to the data acquisition system, and the sampling frequency can be set according to the actual situation, for example, sampling once every 15 seconds. In the data acquisition process, the influence of the bus operating environment should be considered, such as the interference of the surrounding electromagnetic field. For infrared temperature sensors, ensure that the measurement line of sight is not blocked to obtain accurate temperature data.

[0055] Mechanical transmission noise feature: high-precision acoustic sensors such as piezoelectric or capacitive microphones are installed on key moving parts such as gearboxes, bearings, and transmission shafts to capture sound signals generated during mechanical transmission. The acoustic sensors transmit the collected sound signals to the data acquisition device, which collects the sound signals at a high sampling frequency (e.g., tens of thousands of samples per second) to ensure that the details of the sound signals can be captured. The collected sound signals are time-domain signals, and in order to extract the voiceprint features, the short-time Fourier transform (STFT) is used to process the sound signals. STFT converts time-domain signals into frequency-domain signals by performing Fourier transform on the sound signals in different time windows to obtain time-varying frequency-domain feature vectors When performing STFT, the size and moving step of the time window should be reasonably selected to balance the time resolution and frequency resolution. For example, the time window size can be selected as 50 milliseconds, and the moving step is 10 milliseconds, so that the changes in the sound signals can be effectively captured while sufficient frequency resolution is obtained. All the above sensors use IEEE1588 precision clock protocol to avoid data timestamp errors.

[0056] In this embodiment, the device response vector The above three device response parameters are integrated to comprehensively reflect the running state of the key devices inside the ring network cabinet. Through real-time monitoring and analysis of the device response vector, abnormal operation of the devices can be found in time, providing strong support for fault warning and maintenance.

[0057] Further, in order to limit the search range in the subsequent extraction of the time delay parameter process, reduce unnecessary calculation, improve analysis efficiency, and focus on the time range where there may be a causal relationship between the disturbance and the device response, the time window , wherein is the time when the disturbance occurs, is the maximum time delay search range set, used to search for the possible time delay relationship between the disturbance and the device response on the time axis, and is set to 0.01-0.5 seconds, since the duration of the vibration signal caused by bearing failure impact is usually <0.5 seconds, the transient process of the fault impact needs to be covered.

[0058] Further, by calculating the cross-correlation function, the correlation between the environmental disturbance factor vector and the device response vector at different time delays can be found. The peak position of the cross-correlation function corresponds to the time delay variable that may exist between the two, which helps to determine the lag time of the device response after the occurrence of the abnormal disturbance, wherein the cross-correlation function formula is set as follows:

[0059] wherein, is the cross-correlation function of the jth component of the environmental disturbance factor vector and the kth component of the equipment response at a time lag of is the mathematical expectation, is the value of the jth component of the environmental disturbance factor vector at time t, is the mean value of is the time lag variable, representing the time delay of the equipment response vector relative to the environmental disturbance factor vector, is the value of the kth component of the equipment response vector at time is the mean value of the kth component of the equipment response vector at time

[0060] Further, to determine the optimal time delay estimate between the jth component of the environmental disturbance factor and the kth component of the equipment response vector, in order to determine the residence time of the equipment response after the occurrence of the abnormal disturbance, the time lag variable corresponding to the peak position of the cross-correlation function is set as follows:

[0061] wherein, is the optimal time delay estimate between the jth component of the environmental disturbance factor and the kth component of the equipment response, is the time lag variable value when the function takes the maximum value.

[0062] Further, in order to incorporate the time delay relationship between the environmental disturbance factor and the equipment response vector into the model input, so as to more accurately model the dynamic relationship between the two, and then used for subsequent neural network model prediction, the formula for constructing the input vector with time delay is set as follows: Let be a column vector, which can be written as:

[0063] wherein, is the input vector with time delay at time t, is the value of the first disturbance factor at time t minus the optimal time delay between it and a certain equipment response vector, and the same applies subsequently.

[0064] In one embodiment, step S400: processing the time delay input vector based on the preset neural network ring network cabinet response prediction model, outputting a predicted response vector, and generating an equipment response amplitude according to the predicted response vector, comprises: ​​​Step S410: processing the time-delay input vector based on the preset neural network ring network cabinet response prediction model, and outputting a predicted response vector; Step S420: calculating the generated device response amplitude based on the difference between the predicted value in the predicted response vector and the preset device response reference value.

[0065] In this embodiment, the preset neural network ring network cabinet response prediction model includes a training set, an input layer, an output layer and a hidden layer, which are set as follows: Training set: the input features of the training set are the time-delay input vectors, i.e. the calculated , containing the values of historical environmental disturbance factors at different time delays, reflecting the dynamic time correlation between historical disturbances and device responses. The label of the training set is the corresponding historical device response vector , as the "target value" of model learning, used to train the neural network to fit the mapping relationship of "time-delay input vector → device response vector", so that the model can learn how disturbances affect device responses through historical data.

[0066] Input layer: the input vector is , the number of neurons in the input layer is the same as the dimension of , i.e. 9 neurons (assuming the above 9 elements).

[0067] Hidden layer: multiple layers of structure can be collected, including: First layer of hidden layer: using Long Short-Term Memory (LSTM) as part of the non-linear mapping function . The LSTM layer helps to process long-term dependencies in time series data; the number of neurons can be adjusted according to actual conditions, for example, set to 32 neurons.

[0068] Second layer of hidden layer: add a fully connected layer (Dense layer) to further process the data output from the LSTM layer. The number of neurons is set to 16; the activation function can choose the ReLU (Rectified Linear Unit) function, i.e. , which helps to speed up model convergence.

[0069]

[0070] wherein, is the predicted vector of device response at time t, is the LSTM network non-linear mapping function, is the parameter of the neural network model.

[0071] Output layer: the output vector is , whose elements are the predicted values of device response. The response includes the temperature of the load switch contact busbar node hot spot temperature and mechanical drive voiceprint features then The number of neurons in the output layer is the same as the dimension of the device response vector, for example, 3 neurons.

[0072] Further, it is necessary to determine the normal range of the device response. For example, for the load switch contact temperature , Assuming its normal operating temperature range is where is the minimum normal temperature, is the maximum normal temperature, and Similarly. These normal range values are determined by the specifications of the device manufacturer or long-term operating experience data.

[0073] Further, the device response amplitude calculation method is as follows:

[0074] where, is the device response amplitude, is the predicted vector of the device response at time t, is the normal range of the device response reference value (which can be taken ), is the upper limit of the normal range, is the lower limit of the normal range.

[0075] In one embodiment, step S500: calculate the coupling strength score of the device response vector and the environmental disturbance factor vector, and construct a response risk map based on the coupling strength score, obtain the active path edge weight sum according to the response risk map, comprising: Step S510: calculate the coupling strength score of the device response vector and the environmental disturbance factor vector; Step S520: obtain historical ring network cabinet failure data and simulated ring network cabinet experimental data, and construct a response risk map based on the coupling strength score, the historical ring network cabinet failure data and the simulated ring network cabinet experimental data; Step S530: calculate and obtain the active path edge weight sum according to the response risk map.

[0076] In this embodiment, the coupling strength score is calculated by the Pearson correlation coefficient formula, which is set as follows:

[0077] where, Let be the Pearson correlation coefficient between the j-th component of the environmental disturbance factor vector and the k-th component of the equipment response vector. for covariance, for The standard deviation is used to measure the dispersion of a set of data.

[0078] Furthermore, the coupling strength score is calculated as follows:

[0079] in, To score the coupling strength of the ring main unit, the Pearson correlation coefficient is converted into a score of 0-100. exp() represents the Pearson correlation coefficient. The exponent is taken, and the exponential function can convert the correlation coefficient between (-1) and 1 into a positive number, which facilitates subsequent calculations; m and n are index variables used to traverse all disturbance factor components and equipment response components; the denominator is the sum of the exponents of all possible Pearson correlation coefficients. This formula converts the Pearson correlation coefficient into a score of 0-100 through the softmax function, which is used to intuitively evaluate the correlation strength between disturbance factors and equipment responses.

[0080] Furthermore, the steps for constructing the response risk map are as follows: (1) Data collection: ①Historical Data Collection: Collect historical warning events and actual fault case data from sources such as equipment operation records and maintenance logs. This data includes information such as the type of equipment failure, the time of failure, and abnormal phenomena before the failure.

[0081] ②Simulated data collection: Conduct simulated anomaly experiments to simulate various disturbances that may cause equipment failure, such as gas leakage, humidity changes, mechanical vibration, etc., and record the equipment's response data under these simulated anomalies, including temperature changes, sound characteristics, electrical signal fluctuations, etc.

[0082] (2) Graph initialization: ① Disturbance factor nodes: including “SF6 leakage”, “internal wall condensation”, and “structural loosening”.

[0083] ② Equipment response nodes: "Contact temperature rise", "Busbar hot spot", "Abnormal mechanical noise".

[0084] ③ Risk nodes: These are the final results of the fault evolution, such as "insulation failure", "short circuit", "mechanical failure", etc.

[0085] (3) Edge weight determination: Based on historical fault cases, simulated ring network cabinet experimental data and ring network cabinet coupling strength score, initial weights are assigned to the edges between nodes.

[0086] ①Calculate the risk propagation probability, the method is as follows:

[0087] wherein, is the edge weight of node j to node k, representing the risk propagation probability, is the frequency of node j to node k from historical failure cases and simulation ring network cabinet experimental data, is the out-degree of node k, that is, the number of edges from node k, is the sum of all possible edge weights from node k, the purpose is to normalize the sum of all edge weights from node j to 1 (or between 0 and 1).

[0088] ②The calculated risk propagation probability is directly used as the edge weight , that is, = .

[0089] ③According to the current active disturbance factor and the device response node (such as "SF6 leakage + contact temperature rise"), search all possible paths from the disturbance factor node to the device response node in the response risk graph, and the active path is defined as the path from the starting node to the target node.

[0090] ④For each path, calculate the sum of its edge weights. For example, path a passes through nodes (where is the starting node, is the target node), then the sum of the edge weights of path a is:

[0091] wherein, represents the sum of the edge weights of the active path a. also represents the edge weight from node n1 to node n i+1 . The summation symbol in the formula represents the sum of the edge weights between all adjacent nodes on the path .

[0092] In one embodiment, step S600: according to the device response amplitude and the sum of the active path edge weights, output the risk level and the associated warning level strategy, including: Step S610: according to the device response amplitude and the sum of the active path edge weights, calculate the ring network cabinet failure comprehensive score; Step S620: based on the ring network cabinet failure comprehensive score, evaluate and generate a risk level; Step S630: According to the risk level, the associated early warning level strategy is matched.

[0093] In this embodiment, the edge weight sum of the active path in the response risk graph and the device response amplitude are combined to obtain the comprehensive score of the ring netowrk cabinet fault, which is set as follows:

[0094] wherein, is the edge weight sum of the normalized path a, is the normalized i-th device response amplitude, and a and β are weight coefficients, mainly relying on the in-depth understanding of experts in the relevant field on the device failure mechanism and risk factors. For example, for a (weight coefficient of the edge weight sum of the active path), the expert will consider the importance of the risk propagation path in the device failure process. If the expert believes that the risk propagation path plays a key role in the occurrence of the fault, the value of a may be larger. For β (weight coefficient of the device response amplitude value), the expert will determine according to the importance of the device response amplitude to the fault evaluation. If the device response amplitude (such as temperature rise amplitude, vibration amplitude, etc.) has an important indication on the occurrence and development of the fault, the value of β may be larger. Through this weight coefficient determination method based on expert experience, the risk propagation path and the actual response of the device can be considered comprehensively to give a reasonable comprehensive score of the device fault.

[0095] In a specific embodiment, it is assumed that there are currently two active paths, and the edge weight sums are and In terms of device response amplitude, the “contact temperature rise” amplitude value A 触头温升 = 20°C, and the “bus bar hot spot” amplitude value A 母线热斑 = 15°C. It is assumed that , .

[0096] Then S = 0.6 × (0.4 + 0.3) + 0.4 × (20 + 15) = 14.42 Further, the risk level evaluation step is as follows: (1) Evaluate the risk level: divide the risk level interval: according to historical data and experience, different risk level intervals are divided. For example: low risk: S < 10; medium risk: 10 ≤ S < 20; high risk: S ≥ 20.

[0097] (2) Determine the current risk level: according to the calculated comprehensive score S of the fault, determine the risk level of the current operating state. For example, in the above embodiment, S = 14.42 is calculated, so the current is in the medium risk level.

[0098] (3) Automatically match the early warning level strategy: ① Define early warning level strategy: Low risk early warning strategy: may only need to record the current state, routine monitoring, no special measures need to be taken; Medium risk early warning strategy: send early warning notice, increase monitoring frequency, arrange personnel to carry out preliminary inspection.

[0099] High risk early warning strategy: immediately issue an emergency alert, take emergency measures such as shutdown or switch to backup equipment, organize professional personnel to carry out comprehensive inspection and maintenance; According to the risk level obtained by evaluation, automatically match the corresponding early warning level strategy. For example, for medium risk level, the system automatically sends early warning notice and increases monitoring frequency, etc.

[0100] ② Self-optimization mechanism: The early warning disposal results (such as "no failure occurred after emergency warning") are fed back to the response risk graph. According to the feedback results, the edge weights between related nodes are adjusted. For example, if no failure occurs after maintenance measures are taken on a high-risk path, the weights between related nodes on this path can be appropriately reduced to reflect the reduced risk.

[0101] In this embodiment, through the steps of multi-dimensional data collection, abnormality identification, time delay analysis, neural network prediction, risk graph construction and comprehensive risk assessment, comprehensive monitoring of the running state of the ring net cabinet and rapid early warning of faults are realized. Through dynamic adjustment and self-learning mechanism, the accuracy and adaptability of early warning are continuously improved, ensuring the reliable operation of the ring net cabinet in complex environment.

[0102] In one embodiment, as shown in Figure 2 , a ring net cabinet fault rapid early warning system is also provided, which comprises: An environmental disturbance vector generation module is configured to obtain running environment data collected by each preset sensor in the ring net cabinet shell, and generate an environmental disturbance factor vector based on the running environment data; An abnormal disturbance vector acquisition module is configured to filter the environmental disturbance factor vector based on a preset threshold, and obtain an abnormal disturbance factor vector; A time delay input vector generation module is configured to obtain device response data collected by each preset sensor in the ring net cabinet shell, obtain a device response vector by combination, and generate a time delay input vector based on the device response vector and the abnormal disturbance factor vector; A device response amplitude generation module is configured to process the time delay input vector based on a preset neural network ring net cabinet response prediction model, output a predicted response vector, and generate a device response amplitude based on the predicted response vector; an active path weight obtaining module, configured to calculate a coupling strength score of the device response vector and the environmental disturbance factor vector, and construct a response risk atlas based on the coupling strength score, and obtain an active path edge weight sum according to the response risk atlas; a warning level strategy output module, configured to output a risk level and an associated warning level strategy according to the device response amplitude and the active path edge weight sum.

[0103] In another embodiment, the environmental disturbance vector generation module is further configured to: obtain operation environment data collected by each preset sensor in the ring main unit shell; and combine the operation environment data to construct and generate the environmental disturbance factor vector.

[0104] In another embodiment, the abnormal disturbance vector obtaining module is further configured to: filter the environmental disturbance factor vector based on a preset threshold to obtain an abnormal disturbance feature; and integrate the abnormal disturbance feature to generate an abnormal disturbance factor vector.

[0105] In another embodiment, the time delay input vector generation module is further configured to: obtain device response data collected by each preset sensor in the ring main unit shell to obtain a device response vector by combination; construct a cross-correlation function based on a correlation analysis of the device response vector and the abnormal disturbance factor vector; calculate a peak position of the cross-correlation function to extract a corresponding time delay parameter, and construct a time delay input vector according to the time delay parameter.

[0106] In another embodiment, the device response amplitude generation module is further configured to: process the time delay input vector based on a preset neural network ring main unit response prediction model to output a predicted response vector; and calculate a device response amplitude based on a difference between a predicted value in the predicted response vector and a preset device response reference value.

[0107] In another embodiment, the active path weight obtaining module is further configured to: calculate a coupling strength score of the device response vector and the environmental disturbance factor vector; obtain historical ring main unit failure data and simulated ring main unit experimental data, and construct a response risk atlas based on the coupling strength score, the historical ring main unit failure data and the simulated ring main unit experimental data; and calculate an active path edge weight sum according to the response risk atlas.

[0108] In another embodiment, the warning level strategy output module is further configured to: calculate a ring main unit failure comprehensive score based on the device response amplitude and the active path edge weight sum; generate a risk level based on the ring main unit failure comprehensive score; and match an associated warning level strategy according to the risk level.

[0109] In one embodiment, as Figure 3As shown, a computer device is also provided, comprising a memory and a processor, the memory storing a computer program and an operating system, and the processor implementing the steps of the machine vision-based separation detection method of the separation screen surface of the separator described above when executing the computer program. The computer device further comprises a system bus, an internal memory, a network structure, a display screen, an input device, and the like.

[0110] In one embodiment, a computer-readable storage medium is also provided, which stores a computer program, and the computer program is executed by a processor to implement the steps of the ring main unit fault rapid early warning method described above.

[0111] It should be noted that the information interaction, execution process and the like between the above modules, since based on the same concept as the method embodiments of the present application, the specific functions and the technical effects brought about, specific reference can be made to the method embodiments part, and will not be repeated here.

[0112] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit, and the integrated unit can be realized in the form of hardware or software. In addition, the specific name of each functional unit and module is only for easy distinction, and does not limit the protection scope of the present application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0113] It should be noted that the information interaction, execution process and the like between the above modules, since based on the same concept as the method embodiments of the present application, the specific functions and the technical effects brought about, specific reference can be made to the method embodiments part, and will not be repeated here.

[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the above described functions. Each functional unit or module in the embodiment can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit. In addition, the specific name of each functional unit or module is only for convenient distinction, and does not limit the protection scope of the present application. The specific working process of the unit or module in the system can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0115] The embodiments of the present application further provide a network device, which comprises at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps in any of the method embodiments described above when executing the computer program.

[0116] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program, wherein the computer program is executable on a processor to implement the steps in any of the method embodiments described above.

[0117] The embodiments of the present application provide a computer program product, which, when running on a mobile terminal, enables the mobile terminal to implement the steps in any of the method embodiments described above.

[0118] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the present application can implement all or part of the processes in the above-mentioned embodiment methods through a computer program to instruct relevant hardware to complete, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of each method embodiment described above can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms. The computer readable medium can at least include any entity or device capable of carrying the computer program code to the photographing device / terminal equipment, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium. For example, U disk, mobile hard disk, magnetic disk or optical disk, etc. In some jurisdictions, according to legislation and patent practice, the computer readable medium can not be an electrical carrier signal and a telecommunication signal.

[0119] In the above embodiments, the description of each embodiment has its own focus, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0120] Those skilled in the art can appreciate that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0121] In the embodiments provided in the present application, it should be understood that the disclosed apparatus / network device and method can be implemented in other ways. For example, the apparatus / network device embodiments described above are only schematic. The division of the modules or units is only a logical function division, and there can be another division in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual coupling or direct coupling or communication connection can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0122] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e. may be located in one place, or may be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0123] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

[0124] An embodiment of the present application further provides a computer device, which comprises at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to implement the steps in any of the above-mentioned embodiments.

[0125] The computer device can include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above description is an example of the computer device, and does not constitute a limitation on the computer device, and can include more or fewer components than the above description, or combine certain components, or different components, for example, can also include input / output devices, network access devices, etc.

[0126] The processor can be a central processing unit (CPU), and the processor can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0127] The memory can be an internal storage unit of the computer device in some embodiments, such as a hard disk or a memory of the computer device. The memory can also be an external storage device of the computer device in other embodiments, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, and the like. Further, the memory can include both an internal storage unit and an external storage device of the computer device. The memory is used to store an operating system, an application program, a boot loader, data, and other programs, such as program codes of the computer program, and the like. The memory can also be used to temporarily store data that has been output or is to be output.

[0128] The technical features of the above embodiments can be combined in any manner. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described, but it should be understood that any combination of the technical features is within the scope of the present disclosure as long as the combination does not result in a contradiction.

[0129] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the present application. It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these are within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for rapid early warning of ring main unit faults, characterized in that, The method includes: The system acquires operating environment data collected by each preset sensor inside the ring main unit housing, and generates an environmental disturbance factor vector based on the operating environment data. The environmental disturbance factor vector is filtered based on a preset threshold to obtain the abnormal disturbance factor vector; The device response data collected by each preset sensor inside the ring main unit is obtained, and the device response vector is obtained by combining them. Based on the device response vector and the abnormal disturbance factor vector, a time delay input vector is generated. The time delay input vector is processed based on a preset neural network ring main unit response prediction model, and a predicted response vector is output. The device response amplitude is then generated based on the predicted response vector. Calculate the coupling strength score between the device response vector and the environmental disturbance factor vector, construct a response risk map based on the coupling strength score, and obtain the sum of active path edge weights based on the response risk map; Based on the device response magnitude and the sum of the active path edge weights, the risk level and associated early warning level strategy are output.

2. The rapid early warning method for ring main unit faults according to claim 1, characterized in that, The step of acquiring operating environment data collected by each preset sensor inside the ring main unit housing, and generating an environmental disturbance factor vector based on the operating environment data, includes: Acquire operating environment data collected by various preset sensors inside the ring main unit housing; The operating environment data is combined to construct an environmental disturbance factor vector.

3. The rapid early warning method for ring main unit faults according to claim 1, characterized in that, The step of filtering the environmental disturbance factor vector based on a preset threshold to obtain the abnormal disturbance factor vector includes: The environmental disturbance factor vector is filtered based on a preset threshold to obtain abnormal disturbance characteristics; The aforementioned abnormal perturbation features are integrated to generate an abnormal perturbation factor vector.

4. The rapid early warning method for ring main unit faults according to claim 1, characterized in that, The process of acquiring equipment response data collected by preset sensors within the ring main unit housing, combining this data to obtain an equipment response vector, and generating a time delay input vector based on the equipment response vector and the abnormal disturbance factor vector includes: Acquire the device response data collected by each preset sensor inside the ring main unit housing, and obtain the device response vector by combining them; Based on the correlation analysis of the device response vector and the abnormal disturbance factor vector, a cross-correlation function is constructed; The peak position of the cross-correlation function is calculated and determined, the corresponding time delay parameters are extracted, and a time delay input vector is constructed based on the time delay parameters.

5. The rapid early warning method for ring main unit faults according to claim 1, characterized in that, The preset neural network ring main unit response prediction model processes the time delay input vector, outputs a predicted response vector, and generates the device response amplitude based on the predicted response vector, including: The time delay input vector is processed based on a preset neural network ring main unit response prediction model, and the predicted response vector is output. The device response amplitude is calculated based on the difference between the predicted value in the predicted response vector and the preset device response benchmark value.

6. The rapid early warning method for ring main unit faults according to claim 1, characterized in that, The calculation of the coupling strength score between the device response vector and the environmental disturbance factor vector, the construction of a response risk map based on the coupling strength score, and the acquisition of the sum of active path edge weights based on the response risk map include: Calculate the coupling strength score between the device response vector and the environmental disturbance factor vector; Historical ring main unit fault data and simulated ring main unit experimental data are obtained, and a response risk map is constructed based on the coupling strength score, the historical ring main unit fault data and the simulated ring main unit experimental data; The sum of active path edge weights is calculated based on the response risk graph.

7. The rapid early warning method for ring main unit faults according to claim 1, characterized in that, The strategy for outputting a risk level and associated early warning level based on the sum of the device response magnitude and the active path edge weights includes: The comprehensive fault score of the ring main unit is calculated based on the sum of the device response amplitude and the active path edge weights. Based on the comprehensive fault score of the ring main unit, a risk level is assessed and generated. Match the associated early warning level strategy according to the risk level.

8. A rapid early warning system for ring main unit faults, characterized in that, The system includes: The environmental disturbance vector generation module is used to acquire the operating environment data collected by each preset sensor inside the ring main unit housing, and generate an environmental disturbance factor vector based on the operating environment data. An abnormal disturbance vector acquisition module is used to filter the environmental disturbance factor vector based on a preset threshold to obtain the abnormal disturbance factor vector. The time delay input vector generation module is used to acquire the equipment response data collected by each preset sensor inside the ring main unit housing, obtain the equipment response vector by combining them, and generate the time delay input vector based on the equipment response vector and the abnormal disturbance factor vector. The device response amplitude generation module is used to process the time delay input vector based on a preset neural network ring network cabinet response prediction model, output a predicted response vector, and generate the device response amplitude based on the predicted response vector. The active path weight acquisition module is used to calculate the coupling strength score between the device response vector and the environmental disturbance factor vector, construct a response risk map based on the coupling strength score, and obtain the sum of active path edge weights based on the response risk map; The early warning level strategy output module is used to output the risk level and associated early warning level strategy based on the device response magnitude and the sum of the active path edge weights.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images