Fault prediction method and system for smart electricity room

Abstract

The invention discloses a fault prediction method and system for an intelligent electricity room, and the method comprises the steps: collecting multi-dimensional operation data, such as electrical parameters, temperature parameters, vibration parameters and operation duration parameters, comprehensively monitoring the operation state of equipment, enabling the electrical parameters to reflect the electrical performance, the temperature parameters to reflect the heating condition, and the vibration parameters to reflect the mechanical state, compared with single parameter acquisition, the aging evaluation assisted by the operation duration parameter can obtain richer information and lay a foundation for accurate judgment; meanwhile, a fault prediction comprehensive index is calculated based on multi-dimensional data, parameter relations and cooperative influences are comprehensively considered, complex relations are quantified, the fault possibility is evaluated more accurately, and misjudgment caused by neglecting of the parameter relations in the prior art is avoided; in addition, the comprehensive index is compared with the preset threshold value, compared with single parameter judgment, the fault risk can be recognized more accurately, the problem that in the prior art, fault early warning accuracy is insufficient is effectively solved, and the operation and maintenance management level of intelligent electricity room equipment is improved.

Description

Technical Field

[0001] This invention relates to the field of power room fault prediction technology, and in particular to a fault prediction method and system for smart power rooms. Background Technology

[0002] With the continuous development of the power system and the advancement of intelligent construction, smart substations, as key nodes for power distribution and conversion, are crucial for ensuring power supply through stable and reliable operation. Smart substations are equipped with a large number of electrical devices, such as transformers, circuit breakers, and switchgear. During long-term operation, these devices are affected by various factors such as electrical stress, temperature changes, and mechanical vibration, which inevitably lead to aging and wear, thereby increasing the risk of failure.

[0003] To more effectively ensure the reliable operation of electrical equipment in smart power substations and reduce the probability of failures, fault prediction technology has emerged. However, most existing fault prediction methods for electrical equipment are based on only a single parameter or a limited number of parameters, making it difficult to comprehensively and accurately reflect the actual operating status of the equipment. For example, relying solely on electrical parameters for fault prediction may fail to detect potential faults caused by abnormal temperatures or mechanical vibrations; while considering only temperature parameters may ignore the impact of changes in electrical parameters on the equipment. Summary of the Invention

[0004] In view of this, the present invention proposes a fault prediction method and system for smart power substations, which can effectively solve the shortcomings of existing technologies that are unable to fully and accurately reflect the actual operating status of equipment.

[0005] The technical solution of this invention is implemented as follows:

[0006] A fault prediction method for smart power substations includes:

[0007] Collect multi-dimensional operating data of various electrical devices in the smart power room, including electrical parameters, temperature parameters, vibration parameters, and operating time parameters;

[0008] Based on the collected multi-dimensional operational data, a comprehensive fault prediction index for each electrical device is calculated.

[0009] The calculated fault prediction index is compared with the preset fault threshold. When the fault prediction index is greater than or equal to the preset fault threshold, the electrical equipment is determined to have a fault risk and a fault warning signal is issued. When the fault prediction index is less than the preset fault threshold, the electrical equipment is determined to be operating normally.

[0010] As a further optional solution to the fault prediction method for smart substations, the collection of multi-dimensional operational data of various electrical devices in the smart substation specifically includes:

[0011] Inside the smart power room, various types of sensors are installed to match different types of electrical equipment;

[0012] Based on the operating characteristics and importance of electrical equipment, different data acquisition frequencies are set for different equipment and their corresponding sensors;

[0013] The data collected by the sensors is transmitted to the data aggregation node, which preprocesses the received sensor data to obtain multi-dimensional operational data.

[0014] As a further optional solution to the fault prediction method for smart power substations, the data aggregation node preprocesses the received sensor data to obtain multi-dimensional operational data, specifically including:

[0015] The data aggregation node performs preliminary verification on the received sensor data to check whether the data meets the preset basic format requirements;

[0016] For data that passes the initial verification, the anomaly identification index calculation formula is used to determine whether the data is anomaly. If so, a correction method based on historical data trend prediction is used for correction.

[0017] Check if there are any missing data after the abnormal data correction. If there are missing data, use the data missing imputation index calculation formula to determine the method of imputing the data.

[0018] After correcting for abnormal data and handling missing data, the data is normalized to obtain multi-dimensional operational data.

[0019] As a further optional solution to the fault prediction method for smart substations, the formula for calculating the abnormal data identification index is as follows:

[0020] ;

[0021] in, This is represented as an anomaly data identification index. The current data value to be detected. This represents the average data value of the sensor over the most recent n normal data periods. Let n be the standard deviation of the sensor's data over the most recent n normal data periods. This refers to the data value collected by the sensor in the previous data cycle. These are the data values ​​collected by the sensor over two data cycles. It is a very small positive number, used to avoid the case where the denominator is zero. and These are the weighting coefficients, and .

[0022] As a further optional solution to the fault prediction method for smart substations, the formula for calculating the data missing information index is as follows:

[0023] ;

[0024] in, This is represented by the data missing imputation index. The number of historical data time periods similar to the current missing data time period. This represents the total number of historical data time periods. For data values ​​within a similar historical data time period, This represents the average of data within a similar historical time period. Let k be the median of data within a similar historical data period, and k be the number of data points within that same historical data period. and These are the weighting coefficients, and .

[0025] As a further optional solution to the fault prediction method for smart substations, the fault prediction comprehensive index of each electrical device is calculated based on the fault prediction comprehensive index calculation formula, specifically as follows:

[0026] ;

[0027] in, This is represented as a comprehensive index for fault prediction. These are normalized values ​​for electrical parameters. These are reference values ​​after normalization of the historical maximum values ​​of electrical parameters. This is the normalized value of the temperature parameter. This is a reference value after normalizing the historical maximum value of the temperature parameter. These are the normalized values ​​of the vibration parameters. This is a reference value after normalizing the historical maximum value of the vibration parameter. This is the normalized value of the runtime parameter. This is a reference value after normalizing the historical maximum value of the runtime parameter. , , , These are the weighting coefficients corresponding to electrical parameters, temperature parameters, vibration parameters, and running time parameters, respectively. .

[0028] As a further optional embodiment of the fault prediction method for smart power rooms, the electrical parameters include voltage, current and power factor, the temperature parameters include the temperature of key parts of the equipment and the ambient temperature, and the vibration parameters include the equipment vibration frequency and vibration amplitude.

[0029] A fault prediction system for smart power substations, comprising:

[0030] The data acquisition module is used to collect multi-dimensional operating data of various electrical devices in the smart power room. The multi-dimensional operating data includes electrical parameters, temperature parameters, vibration parameters, and operating time parameters.

[0031] The comprehensive index calculation module is used to calculate the comprehensive fault prediction index of each electrical device based on the collected multi-dimensional operational data;

[0032] The threshold comparison module is used to compare the calculated fault prediction comprehensive index with the preset fault threshold. The preset fault threshold is set based on the historical fault data of the electrical equipment, equipment characteristics and actual operating experience. Different types of electrical equipment have corresponding fault thresholds.

[0033] The fault determination and early warning module is used to determine that the electrical equipment has a fault risk and issue a fault early warning signal when the fault prediction comprehensive index is greater than or equal to the preset fault threshold; when the fault prediction comprehensive index is less than the preset fault threshold, the electrical equipment is determined to be operating normally.

[0034] A computing device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the steps of the fault prediction method for smart substations described above.

[0035] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the above-described fault prediction methods for smart power substations.

[0036] The beneficial effects of this invention are as follows: By collecting multi-dimensional operational data such as electrical parameters, temperature parameters, vibration parameters, and operating time parameters of various electrical devices in a smart power room, this method can monitor the operating status of the equipment from multiple key aspects. Electrical parameters directly reflect the electrical performance of the equipment; for example, abnormal changes in voltage and current may indicate electrical faults. Temperature parameters reflect the heat generation of the equipment; excessively high temperatures are often a sign of equipment overload or internal faults. Vibration parameters reflect the mechanical operating status of the equipment; abnormal vibration may indicate wear or loosening of mechanical parts. Operating time parameters provide an important basis for assessing the aging degree of the equipment. Compared with existing technologies based on only a single parameter or a limited number of parameters, this multi-dimensional data acquisition method can obtain more comprehensive and richer equipment operating information, laying a solid foundation for accurately judging the equipment status. Simultaneously, based on the collected multi-dimensional operational data, calculations can be performed on each electrical device... The comprehensive index for predicting equipment failures integrates and analyzes multiple data types. During the calculation process, it fully considers the interrelationships between various parameters and their synergistic impact on equipment failures. For example, abnormal electrical parameters may lead to temperature increases, and temperature changes may further affect the mechanical performance of the equipment. The comprehensive index quantifies these complex relationships, thus more accurately assessing the likelihood of equipment failure. In contrast, existing technologies, considering only a single or few parameters, fail to capture the inherent connections between these parameters, easily leading to misjudgments of equipment status. Furthermore, comparing the calculated comprehensive index with a preset failure threshold, this judgment method based on the comprehensive index and preset thresholds, compared to the single-parameter judgment of existing technologies, can more accurately identify equipment failure risks, effectively solving the problem of insufficient accuracy in fault early warning in existing technologies. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a flowchart of a fault prediction method for a smart power substation according to the present invention;

[0039] Figure 2 This is a schematic diagram of the composition of a fault prediction system for a smart power substation according to the present invention.

[0040] Figure 3 This is a schematic diagram of the composition of a computing device according to the present invention. Detailed Implementation

[0041] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] refer to Figures 1 to 3 A fault prediction method for smart power substations, comprising:

[0043] Collecting multi-dimensional operational data of various electrical devices in the smart power substation, including electrical parameters, temperature parameters, vibration parameters, and operating time parameters; in some embodiments, the collection of multi-dimensional operational data of various electrical devices in the smart power substation specifically includes:

[0044] Inside the smart power room, various types of sensors are installed to match different types of electrical equipment. For transformers, temperature sensors are installed to monitor winding and oil temperatures in real time, oil level sensors monitor oil level, and partial discharge sensors detect partial discharge. For circuit breakers, contact temperature sensors monitor contact temperature, operating mechanism displacement sensors monitor operating mechanism displacement, and gas pressure sensors (for sulfur hexafluoride circuit breakers) monitor gas pressure. For power cables, distributed fiber optic temperature sensors monitor temperature distribution along the entire cable length, and grounding current sensors monitor cable grounding current. Simultaneously, environmental sensors are strategically placed throughout the power room, including temperature and humidity sensors to monitor ambient temperature and humidity, smoke sensors to detect smoke concentration, and water immersion sensors to detect water accumulation.

[0045] Based on the operating characteristics and importance of electrical equipment, different data acquisition frequencies are set for different equipment and their corresponding sensors. For critical equipment such as main transformers, the acquisition frequency of key parameters such as temperature and oil level is set to once per minute, and the acquisition frequency of partial discharge data is set to once every 10 minutes. For general equipment such as low-voltage distribution cabinets, the acquisition frequency of parameters such as voltage and current is set to once every 5 minutes. Among environmental parameters, the acquisition frequency of temperature and humidity is set to once every 10 minutes, and smoke and water immersion data adopt a real-time monitoring mode, and the data is uploaded immediately once an abnormality is detected.

[0046] The sensor-collected data is transmitted to the data aggregation node using a combination of wired and wireless communication methods. The data aggregation node preprocesses the received sensor data to obtain multi-dimensional operational data.

[0047] Specifically, within the smart power substation, various types of sensors are installed to match different types of electrical equipment. These different devices have different working principles and operating characteristics. For example, transformers require monitoring winding temperature and oil gas content, while circuit breakers need to monitor contact temperature and operating mechanism status. By installing suitable sensors for each type of equipment, the unique operating information of each device can be accurately captured, ensuring that the collected data comprehensively covers all key operating parameters of the equipment and avoiding data omissions or inaccuracies caused by sensor mismatch. Based on the operating characteristics and importance of the electrical equipment, differentiated data acquisition frequencies are set for different equipment and their corresponding sensors. For critical equipment, such as the main transformer, whose operating status is crucial to the power supply of the entire substation, a higher acquisition frequency is set to obtain its operating data in real time and in detail, promptly detecting even minor anomalies. For some less important equipment, the acquisition frequency is appropriately reduced to minimize data redundancy and storage pressure while ensuring data validity. This differentiated approach makes data acquisition more targeted and efficient, improving data accuracy and usability.

[0048] The data collected by the sensors is transmitted to the data aggregation node. This node acts as a hub for data transmission, centrally receiving data from various sensors. By rationally designing the communication architecture and transmission protocol of the data aggregation node, the stability and reliability of the data during transmission are ensured, reducing data loss and errors. The data aggregation node preprocesses the received sensor data to obtain multi-dimensional operational data. The preprocessed multi-dimensional operational data is more standardized and accurate, facilitating subsequent in-depth data analysis and fault prediction, thereby improving the efficiency of the entire data processing workflow.

[0049] In some embodiments, the data aggregation node preprocesses the received sensor data to obtain multi-dimensional operational data, specifically including:

[0050] The data aggregation node performs preliminary verification on the received sensor data, checking whether the data meets the preset basic format requirements, including data length, data type, timestamp format, etc. If the data does not meet the basic format requirements, the data is marked as abnormal data and the abnormal information is recorded. At the same time, the data that meets the format requirements is processed in the next step.

[0051] For data that passes the initial verification, the anomaly identification index calculation formula is used to determine whether the data is anomaly. If so, a correction method based on historical data trend prediction is used for correction.

[0052] Check if there are any missing data after the abnormal data correction. If there are missing data, use the data missing imputation index calculation formula to determine the method of imputing the data.

[0053] After correcting for abnormal data and handling missing data, the data is normalized to obtain multi-dimensional operational data.

[0054] Specifically, the data aggregation node first performs a preliminary verification on the received sensor data to check whether the data conforms to the preset basic format requirements. This step can effectively intercept data with format errors caused by sensor failure, communication interference, or human error. For example, it prevents the data from containing non-compliant data types or incorrect timestamp formats. Through preliminary verification, it ensures that the data entering the subsequent processing flow has a unified and standardized format, laying the foundation for accurate data processing and analysis, and avoiding processing errors and analysis biases caused by chaotic data formats.

[0055] For data that passes the initial verification, an anomaly identification index calculation formula is used to determine whether the data is abnormal. This formula is calculated based on the historical statistical characteristics of the data and the relationship between relevant parameters, which can more scientifically and accurately identify data points that deviate from the normal range. Once the data is determined to be abnormal, a correction method based on historical data trend prediction is used. This method makes full use of the historical patterns of equipment operation and predicts the reasonable value of the current abnormal data based on the trend of past normal data, so that the corrected data is more consistent with the actual operation of the equipment, greatly enhancing the reliability and authenticity of the data and reducing the interference of abnormal data on subsequent analysis.

[0056] Check whether there are any missing data in the data after anomaly correction. If there are missing data, use the data missing imputation index calculation formula to determine the method of imputation. This calculation formula takes into account factors such as data similarity, historical patterns and stability, and can select the most appropriate imputation method according to different data characteristics, such as imputation based on the mean of similar historical data or imputation based on time series interpolation. By reasonably imputing missing data, the integrity of the data is guaranteed, so that subsequent analysis and modeling can be based on the complete dataset, avoiding the problems of biased analysis results or inaccurate models caused by missing data.

[0057] In some embodiments, the formula for calculating the abnormal data identification index is specifically as follows:

[0058] ;

[0059] in, This is represented as an anomaly data identification index. The current data value to be detected. This represents the average data value of the sensor over the most recent n normal data periods. Let n be the standard deviation of the sensor's data over the most recent n normal data periods. This refers to the data value collected by the sensor in the previous data cycle. These are the data values ​​collected by the sensor over two data cycles. It is a very small positive number, used to avoid the case where the denominator is zero. and These are the weighting coefficients, and Its value is adjusted based on the stability and correlation of the sensor data. For sensors with small data fluctuations, Take the larger value; for sensors with strong data correlation, Take the larger value;

[0060] When calculated greater than the preset abnormal threshold If the data is deemed abnormal, it is corrected using a correction method based on historical data trend prediction. Specifically, based on data from the sensor's past m normal data periods, a linear regression model is used to predict the normal data value for the current data period. and use Replace the current data value to be detected .

[0061] Specifically, the abnormal data identification index is calculated by combining the current data value to be detected with the mean and standard deviation of the most recent normal data period, as well as the relationship with the values ​​collected in the previous two data periods. This multi-factor comprehensive calculation method can fully consider the statistical characteristics and time series correlation of the data. For example, it not only focuses on the degree of deviation of the data from the mean ( (Partial), and also considered the changes in the data over time ( (Partially), thereby more accurately capturing abnormal data and avoiding misjudgments or omissions that may be caused by judging a single factor;

[0062] Weighting coefficient and The settings allow the formula to be adjusted based on the stability and correlation of the sensor data. For sensors with small data fluctuations, Taking a larger value, the formula focuses more on the deviation of the data from the mean to judge anomalies, because the data from these types of sensors is relatively stable, and a large deviation from the mean is likely to indicate an anomaly; for sensors with strong data correlation, Taking a larger value focuses more on the trend of data changes over time, which can better identify anomalies caused by time correlation and improve the accuracy of anomaly data identification.

[0063] Introducing a minimal positive number into the formula This effectively avoids the calculation The fact that the denominator may sometimes be zero ensures the stability and reliability of the calculation process. This allows the formula to operate correctly under various data conditions, preventing calculation interruptions or incorrect results due to special data values, thus providing a stable foundation for subsequent data processing and analysis. The average of the data over the most recent n normal data periods is used. and standard deviation As the basis for the calculation, it is ensured that the statistical quantities used are based on data under normal equipment operating conditions. This allows for a more accurate reflection of the data characteristics under normal equipment conditions. By using normal data as a benchmark to determine whether the current data is abnormal, the rationality and reliability of abnormal data identification are improved. It should be noted that extremely small positive numbers... The range of values ​​may be arrive The value is within a certain range, but the specific value needs to be determined based on the actual data and accuracy requirements of the smart power room sensors.

[0064] In some embodiments, the formula for calculating the data missing information index is specifically as follows:

[0065] ;

[0066] in, This is represented by the data missing imputation index. The number of historical data time periods similar to the current missing data time period. This represents the total number of historical data time periods. For data values ​​within a similar historical data time period, This represents the average of data within a similar historical time period. Let k be the median of data within a similar historical data period, and k be the number of data points within that same historical data period. and These are the weighting coefficients, and Its value is adjusted based on the historical regularity and stability of the data. For data with strong historical regularity, Take the larger value; for data with good stability, Take the larger value;

[0067] when Greater than the preset fill threshold When this is the case, a mean-based imputation method based on similar historical data is used, that is, the mean of the data within the same historical data time period is taken as the imputation value; when Less than or equal to When the missing data is missing, a time series-based interpolation imputation method, such as linear interpolation or spline interpolation, is used to calculate the imputation value by interpolating the adjacent data values ​​before and after the missing data.

[0068] Specifically, this is achieved by combining the proportion of historical data time periods similar to the current missing data time period. and the degree of dispersion of data distribution To calculate the data missing imputation index This calculation method comprehensively considers the similarity of historical data and the distribution characteristics of the data itself. The proportion of similar historical data time periods reflects the richness of historical data that can be used for reference, while the dispersion of data distribution reflects the stability and regularity of the data. By comprehensively considering these two factors, the data filling method can be determined more scientifically, avoiding unreasonable filling that may be caused by a single factor judgment.

[0069] Weighting coefficient and This setting allows the formula to be adjusted based on the historical regularity and stability of the data. For data with strong historical regularity, Taking a larger value, the formula focuses more on using the quantity of similar historical data to select the imputation method, because this type of data has obvious patterns, and the more similar historical data there is, the more accurate the imputation based on the historical data mean may be; for data with good stability, Taking larger values ​​shows greater attention to the data's distribution and dispersion. If the data dispersion is small, specific methods such as mean imputation or interpolation imputation may be more appropriate, thus improving the scientificity and rationality of the data imputation method selection.

[0070] Using historical data similar to the current missing data period as a reference is beneficial because similar historical data is comparable to the current data segment in terms of operating status, environmental conditions, etc. Filling in the missing data based on this data can make the filled data more consistent with the actual operating conditions of the equipment. For example, if the current missing data segment corresponds to the operating data of the equipment under a specific load, and the similar historical data is also the operating data of the equipment under a similar load, then filling in the missing data based on this similar historical data can more accurately restore the data and improve the accuracy of the filled data.

[0071] The formula takes into account the distribution characteristics of data such as the mean and median within similar historical data time periods. The mean reflects the average level of the data, while the median reflects the middle position of the data. By calculating the degree of deviation of the data from the mean and median, we can better grasp the distribution characteristics of the data. When determining the imputation method, combining these distribution characteristics can make the imputed data more in line with the overall distribution pattern of the data, and further improve the accuracy of the imputed data.

[0072] Based on the collected multi-dimensional operational data, a comprehensive fault prediction index for each electrical device is calculated. In some embodiments, the comprehensive fault prediction index for each electrical device is calculated based on the formula for calculating the comprehensive fault prediction index, specifically as follows:

[0073] ;

[0074] in, This is represented as a comprehensive index for fault prediction. These are normalized values ​​for electrical parameters. These are reference values ​​after normalization of the historical maximum values ​​of electrical parameters. This is the normalized value of the temperature parameter. This is a reference value after normalizing the historical maximum value of the temperature parameter. These are the normalized values ​​of the vibration parameters. This is a reference value after normalizing the historical maximum value of the vibration parameter. This is the normalized value of the runtime parameter. This is a reference value after normalizing the historical maximum value of the runtime parameter. , , , These are the weighting coefficients corresponding to electrical parameters, temperature parameters, vibration parameters, and running time parameters, respectively. The weighting coefficients are determined using the analytic hierarchy process (AHP) based on the degree of influence of each parameter on electrical equipment failures.

[0075] Specifically, multi-dimensional operational data, including electrical parameters, temperature parameters, vibration parameters, and operating time parameters, were comprehensively calculated. Electrical parameters directly reflect the electrical performance status of electrical equipment; for example, abnormal voltage and current may indicate electrical faults. Temperature parameters reflect the equipment's heat generation; excessively high temperatures are often a sign of overload or internal faults. Vibration parameters reflect the mechanical operating status of the equipment; abnormal vibration may indicate wear or loosening of mechanical components. Operating time parameters provide a basis for assessing the aging degree of the equipment. By integrating these different types of data into a comprehensive fault prediction index... In the calculation, it can comprehensively consider various factors affecting equipment failure, avoiding the limitations of single parameter judgment; it normalizes various parameters to eliminate the influence of differences in dimensions and orders of magnitude between different parameters. For example, electrical parameters may differ greatly from temperature parameters in numerical range. After normalization, each parameter has an equal status in the failure prediction calculation, which can more reasonably integrate the influence of each parameter on equipment failure, making the calculation results more comparable and reliable.

[0076] Weighting coefficient , , , Based on the degree of influence of each parameter on electrical equipment failure, the Analytic Hierarchy Process (AHP) is used to determine the relative importance of each parameter. AHP can systematically analyze the relative importance of each parameter, making the weight allocation more scientific and reasonable. For example, if the analysis finds that an electrical parameter has the greatest impact on equipment failure, then... The electrical parameters will be assigned larger values, thus having a greater impact on the calculation of the comprehensive fault prediction index and more accurately reflecting the actual fault risk of the equipment. Different types of electrical equipment may have different degrees of influence on faults for each parameter. The formula can flexibly adapt to the characteristics of various equipment through adjustable weight coefficients. For example, for equipment that is mainly based on mechanical motion, the weight of vibration parameters may be relatively high; while for equipment that is based on electrical performance, the weight of electrical parameters will be greater. This flexibility allows the formula to be widely used in fault prediction of various electrical equipment in smart power rooms.

[0077] The calculated fault prediction index is compared with the preset fault threshold. When the fault prediction index is greater than or equal to the preset fault threshold, the electrical equipment is determined to have a fault risk and a fault warning signal is issued. When the fault prediction index is less than the preset fault threshold, the electrical equipment is determined to be operating normally.

[0078] Specifically, when the comprehensive fault prediction index is greater than or equal to the preset fault threshold, the system promptly determines that the equipment has a fault risk and issues a fault warning signal. This warning mechanism can detect potential fault hazards in advance before the equipment shows obvious fault symptoms, giving maintenance personnel sufficient time to respond. For example, equipment may gradually accumulate fault risks due to long-term minor abnormal operation. This technical solution can detect this in advance and avoid the sudden occurrence of faults. Timely fault warnings help reduce the impact of equipment failures on the operation of smart power substations. Measures can be taken in advance to avoid power supply interruptions caused by equipment failures, reducing the impact on industrial production, commercial operations, and residents' lives. At the same time, it can also reduce the maintenance costs and losses caused by the expansion of faults.

[0079] When the comprehensive fault prediction index is less than the preset fault threshold, the electrical equipment is judged to be operating normally. This accurate determination of operating status can avoid excessive intervention in normally operating equipment, reduce unnecessary inspection and maintenance work, and lower operation and maintenance costs. Based on accurate status determination, operation and maintenance personnel can rationally allocate operation and maintenance resources, concentrate more energy and resources on equipment with fault risks, improve resource utilization efficiency, and optimize the operation and maintenance management of smart power substations.

[0080] In some embodiments, the electrical parameters include voltage, current, and power factor; the temperature parameters include the temperature of critical parts of the equipment and the ambient temperature; and the vibration parameters include the equipment vibration frequency and vibration amplitude.

[0081] In some embodiments, the preset fault threshold is set based on historical fault data of each electrical device in the smart power room and actual operating experience, and different fault thresholds are set for different types of electrical devices.

[0082] In some embodiments, the methods for issuing fault warning signals include, but are not limited to, audible and visual alarms, SMS notifications, email pushes, and displaying warning information on a designated interface of the smart power room monitoring platform.

[0083] A fault prediction system for smart power substations, comprising:

[0084] The data acquisition module is used to collect multi-dimensional operating data of various electrical devices in the smart power room. The multi-dimensional operating data includes electrical parameters, temperature parameters, vibration parameters, and operating time parameters.

[0085] The comprehensive index calculation module is used to calculate the comprehensive fault prediction index of each electrical device based on the collected multi-dimensional operational data;

[0086] The threshold comparison module is used to compare the calculated fault prediction comprehensive index with the preset fault threshold. The preset fault threshold is set based on the historical fault data of the electrical equipment, equipment characteristics and actual operating experience. Different types of electrical equipment have corresponding fault thresholds.

[0087] The fault determination and early warning module is used to determine that the electrical equipment has a fault risk and issue a fault early warning signal when the fault prediction comprehensive index is greater than or equal to the preset fault threshold; when the fault prediction comprehensive index is less than the preset fault threshold, the electrical equipment is determined to be operating normally.

[0088] A computing device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the steps of the fault prediction method for smart substations described above.

[0089] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the above-described fault prediction methods for smart power substations.

[0090] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A fault prediction method for smart power substations, characterized in that, include: Collect multi-dimensional operating data of various electrical devices in the smart power room, including electrical parameters, temperature parameters, vibration parameters, and operating time parameters; Based on the collected multi-dimensional operational data, a comprehensive fault prediction index for each electrical device is calculated. The calculated fault prediction index is compared with the preset fault threshold. When the fault prediction index is greater than or equal to the preset fault threshold, the electrical equipment is determined to have a fault risk and a fault warning signal is issued. When the comprehensive fault prediction index is less than the preset fault threshold, the electrical equipment is considered to be operating normally.

2. The fault prediction method for smart substations according to claim 1, characterized in that, The collection of multi-dimensional operational data from various electrical devices in the smart power substation specifically includes: Inside the smart power room, various types of sensors are installed to match different types of electrical equipment; Based on the operating characteristics and importance of electrical equipment, different data acquisition frequencies are set for different equipment and their corresponding sensors; The data collected by the sensors is transmitted to the data aggregation node, which preprocesses the received sensor data to obtain multi-dimensional operational data.

3. The fault prediction method for smart power substations according to claim 2, characterized in that, The data aggregation node preprocesses the received sensor data to obtain multi-dimensional operational data, specifically including: The data aggregation node performs preliminary verification on the received sensor data to check whether the data meets the preset basic format requirements; For data that passes the initial verification, the anomaly identification index calculation formula is used to determine whether the data is anomaly. If so, a correction method based on historical data trend prediction is used for correction. Check if there are any missing data after the abnormal data correction. If there are missing data, use the data missing imputation index calculation formula to determine the method of imputing the data. After correcting for abnormal data and handling missing data, the data is normalized to obtain multi-dimensional operational data.

4. The fault prediction method for smart substations according to claim 3, characterized in that, The formula for calculating the abnormal data identification index is as follows: ； in, This is represented as an anomaly data identification index. The current data value to be detected. This represents the average data value of the sensor over the most recent n normal data periods. Let n be the standard deviation of the sensor's data over the most recent n normal data periods. This refers to the data value collected by the sensor in the previous data cycle. These are the data values ​​collected by the sensor over two data cycles. It is a very small positive number, used to avoid the case where the denominator is zero. and These are the weighting coefficients, and .

5. The fault prediction method for smart power substations according to claim 4, characterized in that, The formula for calculating the data missing information imputation index is as follows: ； in, This is represented by the data missing imputation index. The number of historical data time periods similar to the current missing data time period. This represents the total number of historical data time periods. For data values ​​within a similar historical data time period, This represents the average of data within a similar historical time period. Let k be the median of data within a similar historical data period, and k be the number of data points within that same historical data period. and These are the weighting coefficients, and .

6. The fault prediction method for smart power substations according to claim 5, characterized in that, The comprehensive fault prediction index for each electrical device is calculated based on the formula for calculating the comprehensive fault prediction index, specifically as follows: ； in, This is represented as a comprehensive index for fault prediction. These are normalized values ​​for electrical parameters. These are reference values ​​after normalization of the historical maximum values ​​of electrical parameters. This is the normalized value of the temperature parameter. This is a reference value after normalizing the historical maximum value of the temperature parameter. These are the normalized values ​​of the vibration parameters. This is a reference value after normalizing the historical maximum value of the vibration parameter. This is the normalized value of the runtime parameter. This is a reference value after normalizing the historical maximum value of the runtime parameter. , , , These are the weighting coefficients corresponding to electrical parameters, temperature parameters, vibration parameters, and running time parameters, respectively. .

7. The fault prediction method for smart power substations according to claim 6, characterized in that, The electrical parameters include voltage, current, and power factor; the temperature parameters include the temperature of key parts of the equipment and the ambient temperature; and the vibration parameters include the equipment vibration frequency and vibration amplitude.

8. A fault prediction system for smart power substations, characterized in that, include: The data acquisition module is used to collect multi-dimensional operating data of various electrical devices in the smart power room. The multi-dimensional operating data includes electrical parameters, temperature parameters, vibration parameters, and operating time parameters. The comprehensive index calculation module is used to calculate the comprehensive fault prediction index of each electrical device based on the collected multi-dimensional operational data; The threshold comparison module is used to compare the calculated fault prediction comprehensive index with the preset fault threshold. The preset fault threshold is set based on the historical fault data of the electrical equipment, equipment characteristics and actual operating experience. Different types of electrical equipment have corresponding fault thresholds. The fault determination and early warning module is used to determine that the electrical equipment has a fault risk and issue a fault early warning signal when the fault prediction comprehensive index is greater than or equal to the preset fault threshold; when the fault prediction comprehensive index is less than the preset fault threshold, the electrical equipment is determined to be operating normally.

9. A computing device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the fault prediction method for a smart power substation as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the fault prediction method for smart power substations as described in any one of claims 1-7.

Images