Box-type substation 5G intelligent information fusion system

Abstract

The invention belongs to the field of box-type substation fault detection, relates to an information fusion technology, is used for solving the problem that the accuracy of a detection result cannot be guaranteed due to the fact that only a single fault diagnosis mode is adopted for fault detection in the prior art, and particularly relates to a box-type substation 5G intelligent information fusion system which comprises an information fusion platform. The information fusion platform is in communication connection with a box transformer fault detection module, a detection verification module and an instant decision module; the box-type substation fault detection module comprises an electrical detection unit and a waveform detection unit, the electrical detection unit is used for performing electrical fault detection on the box-type substation, and the waveform detection unit is used for performing current waveform detection on the box-type substation; the electrical state of the box-type substation is comprehensively monitored, the winding overheating hidden danger can be found in time through dynamic temperature rise calculation, and the problem of insulation degradation or sealing failure can be accurately recognized in combination with partial discharge and gas concentration detection.

Description

Technical Field

[0001] The present invention belongs to the field of box-type substation fault detection and relates to information fusion technology, specifically a 5G intelligent information fusion system for box-type substations. Background Art

[0002] The 5G intelligent information fusion system for box-type substations is a solution that deeply integrates traditional box-type substations with modern 5G communications, the Internet of Things, artificial intelligence, big data analysis and other technologies. It aims to achieve remote monitoring, intelligent operation and maintenance, fault warning and efficient energy management of substations.

[0003] The invention patent with publication number CN116776650B discloses an operating status control system for a box-type substation. Through intelligent system scheduling and control, the workload of manual inspection and evaluation of the operating status of the box-type substation when judging maintenance is reduced. By predicting the probability of failure in the future, it provides supporting data for maintenance decisions, and at the same time quickly realizes the inheritance of the original form of the original box-type substation and maintenance strategy decisions; however, the control system only uses a single fault diagnosis method for fault detection, resulting in the accuracy of the detection results cannot be guaranteed.

[0004] In response to the above technical problems, this application adopts two methods to perform independent detection and mutual verification. When the detection results of the two methods are inconsistent, the accuracy of the results of different detection methods is quickly analyzed, taking into account both response timeliness and the accuracy of detection results. Summary of the Invention

[0005] The purpose of the present invention is to provide a 5G intelligent information fusion system for box-type substations, which is used to solve the problem that the existing technology only uses a single fault diagnosis method for fault detection, resulting in the inability to ensure the accuracy of the detection results; The technical problem to be solved by the present invention is: how to provide a 5G intelligent information fusion system for a box-type substation that can perform independent detection and mutual verification in two ways.

[0006] The purpose of the present invention can be achieved through the following technical solutions: A 5G intelligent information fusion system for a box-type substation includes an information fusion platform, which is communicatively connected to a box-type substation fault detection module, a detection verification module, and an instant decision-making module; The box-type transformer fault detection module includes an electrical detection unit and a waveform detection unit. The electrical detection unit is used to detect electrical faults in the box-type transformer substation, and the waveform detection unit is used to detect current waveforms in the box-type transformer substation. The detection and verification module is used to verify and analyze the electrical detection results and current waveform detection results of the box-type substation: if the electrical detection results and waveform detection results of the box-type substation are both marked as normal, the verification is passed, and it is determined that the operating status of the box-type substation meets the requirements; if the electrical detection results and waveform detection results of the box-type substation are both marked as abnormal, the verification is passed, and it is determined that the operating status of the box-type substation does not meet the requirements, a fault processing signal is generated and sent to the mobile phone terminal of the administrator; otherwise, the verification fails; The instant decision module is used to perform instant decision analysis on the box-type substation when the verification fails.

[0007] Furthermore, the specific process of the electrical detection unit performing electrical fault detection on the box-type substation includes: obtaining the winding temperature value, partial discharge amount and insulating gas concentration value of the box-type substation in real time, marking the difference between the maximum and minimum winding temperature values in the last L1 seconds as the winding temperature rise value, and judging whether the box-type substation has an electrical fault through the winding temperature rise value, partial discharge amount and insulating gas concentration value.

[0008] Furthermore, the specific process of determining whether the box-type substation has an electrical fault includes: comparing the winding temperature rise value, partial discharge amount and insulating gas concentration value with the preset temperature rise threshold, discharge threshold and concentration threshold respectively: if the winding temperature rise value is less than the temperature rise threshold, the partial discharge amount is less than the discharge threshold and the insulating gas concentration value is greater than or equal to the concentration threshold, then it is determined that the box-type substation does not have an electrical fault, and the electrical detection result of the box-type substation is marked as normal; otherwise, it is determined that the box-type substation has an electrical fault, and the electrical detection result of the box-type substation is marked as abnormal.

[0009] Furthermore, the specific process of the waveform detection unit performing current waveform detection on the box-type substation includes: collecting the three-phase current waveform from the high-precision current transformer or smart meter of the box-type substation, using wavelet transform to remove high-frequency noise, dividing the continuous waveform into segments of fixed length, marking the identification label of abnormal waveforms with short-circuit inrush current, harmonic distortion, and arc fault as 1; marking the identification label of the normal waveform of stable load current as 0.

[0010] Furthermore, the specific process of determining whether the box-type substation has a current waveform fault includes: generating and training an LSTM time series model, sliding the real-time streaming data of the box-type substation into the model according to the step size to predict the fault and output the fault probability, and comparing the fault probability with the preset fault threshold: if the fault probability is less than the fault threshold, it is determined that the box-type substation does not have a current waveform fault, and the current waveform detection result of the box-type substation is marked as normal; if the fault probability is greater than or equal to the fault threshold, it is determined that the box-type substation has a current waveform fault, and the current waveform detection result of the box-type substation is marked as abnormal.

[0011] Furthermore, when verification fails before the analysis cycle is generated, a manual verification signal is generated and sent to the administrator's mobile phone terminal. After receiving the manual verification signal, the administrator performs fault verification on the box-type substation, and marks the detection process whose result is the same as the fault verification result as a normal process, and marks the detection process whose result is different from the fault verification result as an abnormal process. At the same time, the value of the operating environment parameter i of the box-type substation during the fault verification is recorded and marked as the verification value YZi. The operating environment parameter i includes air temperature, air humidity and smoke concentration.

[0012] Furthermore, the specific process of the instant decision module performing instant decision analysis on the box-type substation when the verification fails includes: generating an analysis cycle, when the verification fails within the analysis cycle, obtaining the numerical value of the operating environment parameter i of the box-type substation and marking it as the analysis value FXi, performing expansion processing on the analysis value FXi to obtain the expansion range KZi of the operating environment parameter i, marking the verification process in which the verification value YZi in the historical data is within the corresponding expansion range KZi as a matching process, counting the number of times the electrical detection process and the current waveform detection process are marked as normal processes in the matching process and marking them as electrical priority values and waveform priority values respectively, comparing the electrical priority value with the waveform priority value: if the electrical priority value is less than the waveform priority value, marking the current waveform detection result as the instant decision result; if the electrical priority value is greater than the waveform priority value, marking the electrical detection result as the instant decision result; if the electrical priority value is equal to the waveform priority value, marking the box-type substation as having a fault as an instant decision result; and returning the instant decision result to the information fusion platform.

[0013] Furthermore, the specific process of expanding the analysis value FXi includes: obtaining the analysis high value FGi and the analysis low value FDi through the formula FGi=(1+Ki)×FXi and the formula FDi=(1-Ki)×FXi, where Ki is the expansion coefficient of the operating environment parameter i; and the analysis high value FGi and the analysis low value FDi constitute the expansion range KZi of the operating environment parameter i.

[0014] The present invention has the following beneficial effects: 1. This application achieves comprehensive monitoring of the electrical status of box-type substations. Through dynamic temperature rise calculation, potential winding overheating hazards can be discovered in a timely manner. Combined with partial discharge and gas concentration detection, insulation degradation or seal failure problems can be accurately identified, significantly improving the coverage and accuracy of fault detection. 2. This application can effectively eliminate the interference of environmental noise on current waveform detection and improve the accuracy of abnormal waveform identification. By dividing the continuous waveform into standardized segments and performing binary marking, it provides structured input data for subsequent detection and verification modules, ensuring the comparability of results from different detection cycles, and ultimately achieving reliable cross-validation of electrical detection and waveform detection results. 3. This application solves the problem of the lack of credibility verification of test results in the existing technology. When the test results are inconsistent, manual intervention is used to ensure the accuracy of fault diagnosis. At the same time, the correlation between environmental parameters and the reliability of the test process is established, providing a data basis for the subsequent priority selection of high-reliability test methods. While ensuring the response speed, this solution effectively improves the reliability of fault detection in complex working conditions and avoids erroneous operations caused by sensor errors or model misjudgment. 4. This application effectively addresses the issue of decision-making accuracy when inconsistent test results are present. Through an extended environmental parameter matching mechanism, it automatically identifies historical cases similar to the current operating conditions and uses historical verification data to objectively assess the reliability of different test methods, enabling rapid and reliable decision-making without manual intervention. This method is particularly suitable for outdoor substations with high environmental fluctuations, significantly improving the accuracy of abnormal state judgments and decision-making efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0016] Figure 1 This is a system block diagram of Embodiment 1 of the present invention; Figure 2 This is a flow chart of the method of embodiment 2 of the present invention. DETAILED DESCRIPTION

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] Existing technologies for intelligent transformation of box-type substations primarily rely on a single detection method for fault diagnosis, which carries the drawbacks of a high risk of misdiagnosis and insufficiently reliable results. In one substation operation and maintenance scenario, electrical parameter testing revealed equipment anomalies while waveform data was normal. This made it difficult for maintenance personnel to quickly determine the equipment's true condition, leading to delayed maintenance decisions and wasted resources.

[0019] To address these issues, the R&D team discovered that a single detection method, susceptible to environmental interference and prone to false alarms, necessitated the development of a multi-dimensional detection mechanism. By analyzing historical failure cases, they discovered the complementary nature of electrical parameters and current waveform data, leading them to propose a dual-detection, collaborative verification mechanism. When conflicting test results arise, the system must be able to make autonomous decisions to address these situations.

[0020] Example 1: Figure 1 As shown, a 5G intelligent information fusion system for a box-type substation includes an information fusion platform, which is communicatively connected to a box-type substation fault detection module, a detection verification module, and an instant decision-making module.

[0021] The box-type transformer fault detection module includes an electrical detection unit and a waveform detection unit. The electrical detection unit is used to perform electrical fault detection on the box-type substation: the winding temperature value, partial discharge amount and insulating gas concentration value of the box-type substation are obtained in real time, and the difference between the maximum and minimum winding temperature values in the last L1 seconds is marked as the winding temperature rise value. The winding temperature rise value, partial discharge amount and insulating gas concentration value are compared with the preset temperature rise threshold, discharge threshold and concentration threshold respectively: if the winding temperature rise value is less than the temperature rise threshold, the partial discharge amount is less than the discharge threshold and the insulating gas concentration value is greater than or equal to the concentration threshold, it is determined that the box-type substation has no electrical fault and the electrical detection result of the box-type substation is marked as normal; otherwise, it is determined that the box-type substation has an electrical fault and the electrical detection result of the box-type substation is marked as abnormal; the electrical detection result is sent to the detection verification module.

[0022] The winding temperature rise value reflects the degree of temperature fluctuation by calculating the difference between the maximum and minimum winding temperatures within the last L1 seconds. This can be achieved by using a temperature sensor to collect real-time data and a difference algorithm to detect sudden temperature rise anomalies. The partial discharge value refers to the amount of charge generated by partial discharge within the insulating material. This can be achieved using a high-frequency current sensor or an ultrasonic sensor and is used to identify signs of insulation degradation. The insulating gas concentration value monitors the concentration of insulating gas (such as sulfur hexafluoride) in sealed equipment. This can be achieved using a gas concentration sensor and is used to determine the sealing performance of the equipment and the condition of the insulating medium.

[0023] Specifically, during implementation, temperature sensors collect winding temperature data at a frequency of seconds. Using a sliding time window, they calculate the maximum temperature difference within L1 seconds as a temperature rise indicator. A high-frequency pulse current detection device captures the partial discharge signal in real time, filtering it and outputting a quantified value. A gas sensor continuously monitors the insulating gas concentration, triggering an abnormality detection when the concentration falls below a preset threshold. These three parameters are synchronously input into the analysis module. If the temperature rise exceeds the threshold, the discharge is abnormal, or the gas concentration is insufficient, the system will determine that an electrical fault has occurred.

[0024] The waveform detection unit is used to detect the current waveform of the box-type substation. It collects three-phase current waveforms from the high-precision current transformer or smart meter of the box-type substation. After using wavelet transform (such as DB4 wavelet) to remove high-frequency noise, the continuous waveform is divided into segments of fixed length (such as 10 cycles = 200ms) with an overlap rate of 30% to 50% to enhance robustness. Abnormal waveforms with short-circuit inrush current, harmonic distortion, and arc faults are marked with the identification label 1; the identification label of normal waveforms with stable load current is marked with 0. An LSTM time series model is generated and trained. The real-time streaming data of the box-type substation is input into the model in a sliding step size to predict faults and output the fault probability. The fault probability is compared with the preset fault threshold. If the fault probability is less than the fault threshold, the box-type substation is determined to have no current waveform fault and the current waveform detection result of the box-type substation is marked as normal. If the fault probability is greater than or equal to the fault threshold, the box-type substation is determined to have a current waveform fault and the current waveform detection result of the box-type substation is marked as abnormal. The current waveform detection result is sent to the detection verification module.

[0025] Among them, a high-precision current transformer refers to a sensor used to measure current signals, which can be implemented by a Hall effect sensor or a Rogowski coil sensor. Its function is to accurately capture the instantaneous changes of the three-phase current. Wavelet transform refers to a time-frequency domain signal processing method, which can be implemented by a discrete wavelet transform algorithm. Its function is to separate high-frequency noise from valid signals in the current waveform. Fixed-length segments refer to equal-length data segments divided by the time dimension. Specifically, they can be cut using a time window of, for example, 1 second or 0.5 seconds. Its function is to standardize the input data to adapt to subsequent analysis models. Abnormal waveform identification labels refer to binary identifiers used to distinguish fault states. Specifically, they can be assigned using binary coding rules. Their function is to provide clear classification results for the detection and verification module.

[0026] Specifically, the current waveform detection process first collects three-phase current signals in real time through high-precision current transformers or smart meters to form continuous raw waveform data. The raw data is then decomposed using a wavelet transform algorithm to filter out high-frequency noise components generated by electromagnetic interference or equipment vibration. The denoised waveform data is divided into segments of fixed length. For example, each segment can contain data points corresponding to 10 power frequency cycles. After feature extraction for each segment, if abnormal features such as a sudden increase in amplitude, excessive harmonic components, or non-periodic oscillations are detected in the waveform, it is marked as an abnormal waveform and assigned a label of 1; if the waveform is smooth and conforms to the standard sine shape, it is marked as a normal waveform and assigned a label of 0.

[0027] The LSTM time series model is an artificial neural network model with long-short-term memory capabilities. It can be built using the TensorFlow or PyTorch framework. It uses memory cells to capture the time-dependent characteristics of the current waveform and is used to identify dynamic abnormal patterns in the current waveform. Real-time streaming data refers to three-phase current waveform data continuously collected from high-precision current transformers or smart meters. It can be transmitted in milliseconds using a 5G communication module and is used to reflect the real-time operating status of the box-type substation. The fault probability is a numerical value between 0 and 1 output by the model. It can be calculated using the Softmax function and is used to quantify the likelihood of an anomaly in the current waveform. The fault threshold is a pre-set probability judgment boundary value, such as 0.85, which is used to demarcate the critical condition between normal and abnormal conditions.

[0028] Specifically, during current waveform detection, a training set is first constructed based on historical fault data, containing abnormal waveform samples such as short-circuit inrush current and harmonic distortion. The LSTM model extracts features from the input sequence using multiple layers of neurons, with its hidden layer states preserving waveform change information from previous time steps. Real-time current waveform data is segmented into fixed-length segments, for example, each containing 100 sampling points, and fed into the model in a sliding step. When the probability value output by the model exceeds a preset threshold, a current waveform fault is detected, triggering an anomaly flag and initiating subsequent verification procedures. This detection mechanism effectively identifies transient fault characteristics, such as the transient waveform distortion of arc faults.

[0029] The detection and verification module is used to verify and analyze the electrical detection results and current waveform detection results of the box-type substation: if the electrical detection results and waveform detection results of the box-type substation are both marked as normal, the verification is passed, and it is determined that the operating status of the box-type substation meets the requirements; if the electrical detection results and waveform detection results of the box-type substation are both marked as abnormal, the verification is passed, and it is determined that the operating status of the box-type substation does not meet the requirements, and a fault handling signal is generated and sent to the administrator's mobile phone terminal; otherwise, the verification fails, a manual verification signal is generated and sent to the administrator's mobile phone terminal. After receiving the manual verification signal, the administrator performs a fault check on the box-type substation, and marks the detection process with the same result as the fault check result as a normal process, and the detection process with a different result from the fault check result as an abnormal process. At the same time, the value of the operating environment parameter i of the box-type substation during the fault check is recorded and marked as the verification value YZi. The operating environment parameter i includes air temperature, air humidity and smoke concentration.

[0030] The manual verification signal refers to the command signal that triggers manual intervention. This can be implemented through SMS push or mobile app message notifications, and is used to initiate the manual review process when there are contradictions in the system's detection results. Operating environment parameters refer to external condition parameters that affect the operating status of the equipment. They can be collected using IoT devices such as temperature and humidity sensors and gas detectors. For example, a digital temperature sensor with an accuracy of ±0.5°C can be used for air temperature, a capacitive humidity sensor can be used for air humidity, and a photoelectric smoke detector can be used to obtain smoke concentration. The verification value refers to the environmental parameter data actually measured during the manual verification process. It can be stored in a local database or cloud server and used as benchmark data for subsequent decision-making analysis.

[0031] Specifically, when the verification results of electrical testing and waveform testing are inconsistent, the system automatically generates a manual verification instruction containing equipment location information and detection contradictions, and pushes it to the operation and maintenance personnel's mobile terminal through the 5G network. After the operation and maintenance personnel arrive at the site, they use special testing equipment to review key indicators such as winding temperature and current waveform, and enter the measured results into the system. The system compares the manual verification results with the original test data: if the electrical test results are consistent with the manual verification, the test process is marked as a normal process; if there is a deviation between the waveform test results and the manual verification, the test process is marked as an abnormal process. At the same time, the system automatically collects environmental parameters at the time of verification, such as air temperature can be recorded as 25.3℃, air humidity can be recorded as 65%RH, and smoke concentration can be recorded as 0.05mg / m³, and these values are stored as verification values.

[0032] The instant decision module is used to perform instant decision analysis on the box-type substation when the verification fails: generate an analysis cycle, and when the verification fails within the analysis cycle, obtain the value of the operating environment parameter i of the box-type substation and mark it as the analysis value FXi, and perform expansion processing on the analysis value FXi: obtain the analysis high value FGi and analysis low value FDi through the formula FGi=(1+Ki)×FXi and the formula FDi=(1-Ki)×FXi, where Ki is the expansion coefficient of the operating environment parameter i; the analysis high value FGi and analysis low value FDi constitute the expansion range KZi of the operating environment parameter i, and divide the verification value YZi in the historical data into The verification process that is within the corresponding extended range KZi is marked as a matching process. The number of times the electrical detection process and the current waveform detection process are marked as normal processes in the matching process is counted and marked as electrical priority value and waveform priority value respectively. The electrical priority value is compared with the waveform priority value: if the electrical priority value is less than the waveform priority value, the current waveform detection result is marked as an immediate decision result; if the electrical priority value is greater than the waveform priority value, the electrical detection result is marked as an immediate decision result; if the electrical priority value is equal to the waveform priority value, the fault in the box-type substation is marked as an immediate decision result; and the immediate decision result is returned to the information fusion platform.

[0033] Among them, the operating environment parameters include physical quantities such as air temperature, air humidity and smoke concentration that affect the operating status of the equipment. Specifically, they can be implemented using temperature sensors, humidity sensors and gas detection devices to reflect the real-time status of the environment in which the box-type substation is located. Expansion processing refers to a calculation method for expanding the range of the current environmental parameter values. Specifically, it can be implemented through a multiplier formula with a preset expansion coefficient. For example, the expansion coefficient can be a value of 0.1 or 0.2, which is used to consider the normal fluctuation range of the environmental parameters. The matching process refers to screening out verification cases similar to the current environmental conditions from historical records. Specifically, it can be implemented using a database query algorithm to establish a correlation between the current status and historical data. Priority value statistics refers to a quantitative evaluation of the reliability of the two detection methods in the matching case. Specifically, it can be implemented using a counter module to determine the credibility of different detection methods in similar environments.

[0034] Specifically, when the system detects that the electrical test results are inconsistent with the waveform test results, an immediate decision-making process is initiated. The upper and lower limits of the analysis value of each parameter are calculated using the expansion coefficient. For example, when the air temperature analysis value is 30°C and the expansion coefficient is 0.1, the expansion range will cover the temperature range of 27°C to 33°C. The system automatically retrieves all cases in the historical database whose verification values fall within this expansion range, and filters out matching processes with similar environmental conditions. For these matching cases, the number of times the electrical test is marked as a normal process is counted as the electrical priority value, and the number of times the waveform test is marked as a normal process is counted as the waveform priority value. Finally, by comparing the sizes of the two priority values, the test result with higher credibility is selected as the basis for decision-making. If the two are equal, it is directly determined that a fault exists.

[0035] Example 2: Figure 2 As shown, a 5G intelligent information fusion method for a box-type substation includes the following steps: Step 1: Detect electrical faults in the box-type substation: Obtain the winding temperature, partial discharge, and insulating gas concentration of the box-type substation in real time, and determine whether the box-type substation has an electrical fault based on the winding temperature, partial discharge, and insulating gas concentration. Step 2: Detect the current waveform of the box-type substation: Generate and train an LSTM time series model. Slide the real-time streaming data of the box-type substation into the model according to the step size to predict faults and output the fault probability. Use the fault probability to determine whether the box-type substation has a current waveform fault. Step 3: Verify and analyze the electrical test results and current waveform test results of the box-type substation: if the verification fails, mark the normal process and abnormal process, and record the value of the operating environment parameter i of the box-type substation during the fault verification and mark it as the verification value YZi; Step 4: Perform instant decision analysis on the box-type substation when the verification fails: obtain the numerical value of the operating environment parameter i of the box-type substation and mark it as the analysis value FXi, expand the analysis value FXi, expand the range KZi, and mark the verification process in which the verification value YZi in the historical data is within the corresponding expansion range KZi as a matching process, and generate an instant decision result through the matching process.

[0036] A 5G intelligent information fusion system for a box-type substation obtains the winding temperature value, partial discharge amount, and insulating gas concentration value of the box-type substation in real time during operation, and determines whether the box-type substation has an electrical fault based on the winding temperature value, partial discharge amount, and insulating gas concentration value; generates and trains an LSTM time series model, inputs the real-time streaming data of the box-type substation into the model according to the sliding step size, performs fault prediction, and outputs the fault probability, and determines whether the box-type substation has a current waveform fault based on the fault probability; marks the normal process and the abnormal process when the verification fails, and records the value of the operating environment parameter i of the box-type substation during the fault verification and marks it as the verification value YZi; obtains the value of the operating environment parameter i of the box-type substation and marks it as the analysis value FXi, expands the analysis value FXi to an expansion range KZi, and marks the verification process in which the verification value YZi in the historical data is within the corresponding expansion range KZi as a matching process, and generates an instant decision result through the matching process.

[0037] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the scope of protection of the present invention.

[0038] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0039] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A 5G intelligent information fusion system for box-type substations, characterized by: It includes an information fusion platform, which is communicatively connected to a box-type transformer fault detection module, a detection verification module, and an instant decision-making module; The box-type transformer fault detection module includes an electrical detection unit and a waveform detection unit. The electrical detection unit is used to detect electrical faults in the box-type transformer substation, and the waveform detection unit is used to detect current waveforms in the box-type transformer substation. The detection and verification module is used to verify and analyze the electrical detection results and current waveform detection results of the box-type substation: if the electrical detection results and waveform detection results of the box-type substation are both marked as normal, the verification is passed, and it is determined that the operating status of the box-type substation meets the requirements; if the electrical detection results and waveform detection results of the box-type substation are both marked as abnormal, the verification is passed, and it is determined that the operating status of the box-type substation does not meet the requirements, a fault processing signal is generated and sent to the mobile phone terminal of the administrator; otherwise, the verification fails; The instant decision module is used to perform instant decision analysis on the box-type substation when the verification fails.

2. A 5G intelligent information fusion system for box-type substations according to claim 1, characterized in that: The specific process of the electrical detection unit performing electrical fault detection on the box-type substation includes: obtaining the winding temperature value, partial discharge amount and insulating gas concentration value of the box-type substation in real time, marking the difference between the maximum and minimum winding temperature values in the last L1 seconds as the winding temperature rise value, and judging whether the box-type substation has an electrical fault based on the winding temperature rise value, partial discharge amount and insulating gas concentration value.

3. A 5G intelligent information fusion system for box-type substations according to claim 1, characterized in that: The specific process of determining whether a box-type substation has an electrical fault includes: comparing the winding temperature rise value, partial discharge amount and insulating gas concentration value with the preset temperature rise threshold, discharge threshold and concentration threshold respectively: if the winding temperature rise value is less than the temperature rise threshold, the partial discharge amount is less than the discharge threshold and the insulating gas concentration value is greater than or equal to the concentration threshold, then it is determined that the box-type substation does not have an electrical fault, and the electrical detection result of the box-type substation is marked as normal; otherwise, it is determined that the box-type substation has an electrical fault, and the electrical detection result of the box-type substation is marked as abnormal.

4. A 5G intelligent information fusion system for box-type substations according to claim 1, characterized in that: The specific process of the waveform detection unit performing current waveform detection on the box-type substation includes: collecting three-phase current waveforms from the high-precision current transformer or smart meter of the box-type substation, using wavelet transform to remove high-frequency noise, dividing the continuous waveform into segments of fixed length, marking the identification label of abnormal waveforms with short-circuit inrush current, harmonic distortion, and arc fault as 1; marking the identification label of normal waveforms of stable load current as 0.

5. A 5G intelligent information fusion system for box-type substation according to claim 1, characterized in that: The specific process of determining whether a box-type substation has a current waveform fault includes: generating and training an LSTM time series model, sliding the real-time streaming data of the box-type substation into the model according to the step size to predict the fault and output the fault probability, and comparing the fault probability with the preset fault threshold: if the fault probability is less than the fault threshold, it is determined that the box-type substation does not have a current waveform fault, and the current waveform detection result of the box-type substation is marked as normal; if the fault probability is greater than or equal to the fault threshold, it is determined that the box-type substation has a current waveform fault, and the current waveform detection result of the box-type substation is marked as abnormal.

6. A 5G intelligent information fusion system for box-type substation according to claim 1, characterized in that: When verification fails before the analysis cycle is generated, an artificial verification signal is generated and sent to the administrator's mobile terminal. After receiving the artificial verification signal, the administrator performs fault verification on the box-type substation, and marks the detection process with the same result as the fault verification result as a normal process, and marks the detection process with a result different from the fault verification result as an abnormal process. At the same time, the value of the operating environment parameter i of the box-type substation during the fault verification is recorded and marked as the verification value YZi. The operating environment parameter i includes air temperature, air humidity and smoke concentration.

7. A 5G intelligent information fusion system for box-type substations according to claim 1, characterized in that: The specific process of the instant decision module performing instant decision analysis on the box-type substation when the verification fails includes: generating an analysis cycle, when the verification fails within the analysis cycle, obtaining the value of the operating environment parameter i of the box-type substation and marking it as the analysis value FXi, performing expansion processing on the analysis value FXi to obtain the expansion range KZi of the operating environment parameter i, marking the verification process in which the verification value YZi in the historical data is within the corresponding expansion range KZi as a matching process, counting the number of times the electrical detection process and the current waveform detection process are marked as normal processes in the matching process and marked as electrical priority value and waveform priority value respectively, comparing the electrical priority value with the waveform priority value: if the electrical priority value is less than the waveform priority value, the current waveform detection result is marked as the instant decision result; if the electrical priority value is greater than the waveform priority value, the electrical detection result is marked as the instant decision result; if the electrical priority value is equal to the waveform priority value, the fault in the box-type substation is marked as the instant decision result; and returning the instant decision result to the information fusion platform.

8. The 5G intelligent information fusion system for box-type substation according to claim 1 is characterized in that: The specific process of expanding the analysis value FXi includes: obtaining the analysis high value FGi and the analysis low value FDi through the formula FGi=(1+Ki)×FXi and the formula FDi=(1-Ki)×FXi, where Ki is the expansion coefficient of the operating environment parameter i; the analysis high value FGi and the analysis low value FDi constitute the expansion range KZi of the operating environment parameter i.

Images