A method and system for processing monitoring data of a box-type substation

Abstract

The present invention discloses a method and system for processing monitoring data of a box-type substation, which relates to the technical field of monitoring data processing, and includes: clarifying the types of monitoring data, constructing a monitoring data set and an abnormal fault set, and establishing the association between the two; calculating the correlation index between the monitoring data and the abnormal types, as well as the weights of each associated monitoring data in the abnormal types; adopting a periodic monitoring strategy to capture data changes in real time, screen out abnormal values, and eliminate data fluctuation interference through analysis; for the quantity of the initial monitoring data, adjust and execute the abnormal detection strategy: when there is single data, comprehensively consider all abnormal types associated with it, and use the buffer time to predict future abnormal risks; when there are multiple data, first perform an abnormal judgment of complete matching, and then judge whether it is necessary to perform an abnormal judgment on a single data; realizing comprehensive and real-time monitoring of the operation status of the box-type substation and efficient abnormal early warning, and improving the safe and stable operation ability of the substation.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring data processing, and particularly to a method and system for processing monitoring data of a box-type substation. Background Art

[0002] A box-type substation (hereinafter referred to as a box substation) is a compact substation that integrates high-voltage switchgear, a distribution transformer, and low-voltage distribution equipment in a moisture-proof, rust-proof, dust-proof, mouse-proof, fire-proof, anti-theft, heat-insulating, fully enclosed, and movable steel structure box; it has the characteristics of compact structure, convenient installation, reliable performance, and beautiful appearance, and is especially suitable for urban network construction and renovation, as well as places such as mines, factories and enterprises, oil and gas fields, and wind power stations, becoming a new type of complete power distribution device.

[0003] As an important part of the power system, the safe and stable operation of a box-type substation is crucial for ensuring the reliability of the power grid and the power quality of users; however, during the operation of a box-type substation, it will be affected by various internal and external factors, such as electrical parameter fluctuations, environmental temperature and humidity changes, equipment aging, etc., which may all lead to faults or abnormal states in the substation; therefore, real-time monitoring and data processing of the box-type substation, timely discovery and early warning of potential faults, are of great significance for improving the operation reliability of the substation and reducing maintenance costs.

[0004] Most traditional methods for processing monitoring data of box-type substations only rely on limited sensors and cannot comprehensively capture various key data during the operation of box-type substations, which may lead to an incomplete understanding of the overall operation state of the substation; in addition, most traditional methods for processing monitoring data of box-type substations do not make a comprehensive judgment based on the weights of associated monitoring data, but view each monitoring data in isolation, which may lead to misjudgment of abnormal situations because some abnormalities may be caused by the combined small changes of multiple parameters; finally, traditional methods for processing monitoring data of box-type substations often cannot predict the possible types of abnormalities based on the initial monitoring data and can only be discovered when the abnormality actually occurs, which makes the management personnel in a passive position when dealing with faults and unable to take preventive measures in advance. Summary of the Invention

[0005] (I) Technical Problems to be Solved

[0006] In view of the technical problems in the background art, the present invention proposes a method and system for processing monitoring data of a box-type substation. By selecting multiple types of monitoring data, constructing a monitoring data set and an abnormal fault set, calculating the correlation index between each monitoring data and the corresponding abnormal type, and making a comprehensive judgment based on the weights of the associated monitoring data; predicting the possible abnormal types based on the prediction and weight calculation of the associated monitoring data; and sending out a warning signal in advance by calculating the buffer time, thereby solving the technical problems recorded in the background art.

[0007] (II) Technical Solution

[0008] To achieve the above objectives, the present invention is realized through the following technical solutions:

[0009] A method for processing monitoring data of a box-type substation, comprising:

[0010] Determine multiple types of data that need to be monitored during the operation of the box-type substation; construct a monitoring data set and an abnormal fault set, and establish a corresponding relationship between the data in the monitoring data set and the abnormal fault set;

[0011] Calculate the correlation index between each monitoring data and the corresponding abnormal type; based on multiple associated monitoring data within each abnormal type, calculate the weight of each associated monitoring data;

[0012] Periodically monitor the determined types of data, and at the end of each period, screen out all the monitoring data whose monitoring values exceed the corresponding first threshold, and respectively judge whether the reason for exceeding the first threshold is data fluctuation to screen out the initial monitoring data;

[0013] Perform corresponding abnormal detection operations based on the number of initial monitoring data: when the number of initial monitoring data is single, screen out all the corresponding abnormal types, and based on the associated monitoring data of each abnormal type, sequentially predict whether the corresponding abnormality will occur after the buffer time calculated from the initial monitoring data;

[0014] When the number of initial monitoring data is multiple, first consider the abnormal types that all or some of the initial monitoring data completely correspond to, and then sequentially perform the abnormal type detection of the remaining single initial monitoring data.

[0015] Specifically, put the determined types of multiple monitoring data into the monitoring data set, and denote the monitoring data set as {A1, A2, …, A n} where A i represents the i-th data in the monitoring data set;

[0016] Put all the abnormal types that occurred during the historical operation of the box-type substation into the abnormal fault set, and denote the abnormal fault set as {B1, B2, …, B m}, where B j represents the j-th data in the abnormal fault concentration.

[0017] Specifically, obtain the monitoring data records when all abnormalities occur, and calculate the correlation index between each monitoring data in the monitoring data set and the corresponding abnormal type based on the first threshold of each type of data set. The expression is:

[0018] where represents the total number of times the j-th type of abnormality in the abnormal fault concentration appears in the historical operation records of the box-type substation. represents the total number of times the i-th type of data in the monitoring data set exceeds the corresponding first threshold in all operation records of the j-th type of abnormality in the historical box-type substation with abnormal faults.

[0019] Further, the monitoring data with a non-zero correlation index with each type of abnormality is recorded as the associated monitoring data of the corresponding abnormal type, and each associated monitoring data will record these abnormalities as their respective associated abnormal types;

[0020] Calculate the weight of each type of associated monitoring data based on the correlation index The expression is:

[0021] where represents the sum of the correlation indices of all associated monitoring data of the j-th type of abnormality in the abnormal fault concentration.

[0022] Specifically, at the end of each monitoring cycle, compare the recorded data values of each type with the corresponding first threshold. When it is detected that there is a monitoring data value exceeding or lower than the corresponding first threshold, obtain the monitoring values of this monitoring data at the end of the previous several monitoring cycles, and based on the linear regression algorithm, judge whether a linear function that fits the change of this monitoring data over time can be obtained; if a linear function can be fitted, execute the abnormal judgment strategy; if a linear function cannot be fitted, judge whether the value detected in the next monitoring cycle has returned to the normal value range. If it has returned to the normal value range, do not record this monitoring data as the initial monitoring data.

[0023] Further, if the value detected in the next monitoring cycle still has not returned to the normal value range, but the difference between the monitoring value and the first threshold is less than the difference between the monitoring value and the first threshold in the current monitoring cycle, further judge whether there is data fluctuation according to the monitoring value of this monitoring data in the next monitoring cycle;

[0024] If the value detected in the next monitoring period still does not return to the normal value range, and the difference between the monitoring value and the first threshold is not less than the difference between the monitoring value in the current monitoring period and the first threshold, then record the monitoring data as the initial monitoring data, and based on the monitoring value in the current monitoring period and the monitoring values in the next several monitoring periods, fit a linear function based on the linear regression algorithm.

[0025] Specifically, at the end of a monitoring period, if there is initial monitoring data, execute the abnormal judgment strategy. When there is only one piece of initial monitoring data, sequentially retrieve all associated abnormal types of this initial monitoring data, and sequentially retrieve the remaining associated monitoring data corresponding to each abnormal type;

[0026] Obtain the first monitoring value of each associated monitoring data when the corresponding abnormal type appears, and take the maximum or minimum value among the multiple first monitoring values as the second threshold for the initial monitoring data and each associated monitoring data;

[0027] Obtain the second threshold of the initial monitoring data, and calculate the difference between the monitoring value of the current monitoring data and the second threshold; Based on the fitted linear function, calculate the buffer time required for the value of the initial monitoring data to change to the second threshold.

[0028] Furthermore, fit a linear function based on the monitoring values of each remaining associated monitoring data at the end of the previous several monitoring periods; Based on the fitted linear function, calculate the fitted values of each remaining associated monitoring data after the buffer time, and compare them with the corresponding second thresholds, record the associated monitoring data that exceeds or is lower than the corresponding second threshold, and add the sum of the weights of these associated monitoring data to the weight of the initial monitoring data to obtain the total predicted weight of the corresponding abnormal type in the current monitoring period If it means that based on the data in the current monitoring period, it is predicted that the corresponding abnormal type will not appear after the buffer time;

[0029] If it means that based on the data in the current monitoring period, it is predicted that the corresponding abnormality may appear after the buffer time, and further judgment is required. Specifically, mark the current monitoring period with the corresponding abnormality. If the total predicted weight of this abnormal type analyzed in the next monitoring period it means that based on the data in the next monitoring period, it is predicted that the corresponding abnormal type will appear after the buffer time;

[0030] If it means that based on the data in the current monitoring period, it is predicted that the corresponding abnormal type will appear after the buffer time.

[0031] Specifically, when multiple initial monitoring data appear simultaneously, screen out the abnormal types that exactly correspond to all or some of the initial monitoring data; if all the initial monitoring data exactly correspond to one abnormal type, then based on the fitting function of each initial monitoring data, calculate the buffer time required for the numerical change of each initial monitoring data to reach the corresponding second threshold respectively;

[0032] Calculate the difference t between the maximum and minimum values of multiple buffer times. If t < t0, it is predicted that the corresponding abnormal type will appear after the maximum buffer time; where t0 represents the buffer time difference threshold;

[0033] If t ≥ t0, determine whether there are some initial monitoring data that exactly correspond to one abnormal type. If so, calculate multiple buffer times in the same way for abnormal judgment; if not, perform abnormal prediction under each initial monitoring data.

[0034] A monitoring data processing system for a box-type substation, comprising:

[0035] A monitoring data determination module, used to determine multiple types of data that need to be monitored during the operation of the box-type substation;

[0036] A corresponding relationship establishment module, used to construct a monitoring data set and an abnormal fault set, establish the corresponding relationship between the data in the monitoring data set and the abnormal fault set; calculate the correlation index between each monitoring data and the corresponding abnormal type; based on multiple associated monitoring data within each abnormal type, calculate the weight of each associated monitoring data;

[0037] An abnormal analysis module, including an initial monitoring data screening unit and an associated monitoring data prediction unit. The initial monitoring data screening unit is used to periodically monitor the determined various types of data, screen out all the monitoring data whose monitoring values exceed the corresponding first threshold at the end of each cycle, and respectively judge whether the reason for exceeding the first threshold is data fluctuation to screen out the initial monitoring data;

[0038] The associated monitoring data prediction unit performs corresponding abnormal detection operations based on the number of initial monitoring data. When the number of initial monitoring data is single, screen out all the abnormal types corresponding to it, and based on the associated monitoring data of each abnormal type, predict in turn whether the corresponding abnormality will occur after the buffer time calculated from the initial monitoring data; when the number of initial monitoring data is multiple, first consider the abnormal types that exactly correspond to all or some of the initial monitoring data, and then perform the abnormal type detection of the remaining single initial monitoring data in turn.

[0039] (III) Beneficial effects

[0040] The present invention provides a method and system for processing monitoring data of a box-type substation, which has the following beneficial effects:

[0041] 1. By carefully planning and placing various types of sensors, comprehensive collection of various key data during the operation of the box-type substation is ensured. This avoids measurement errors that may be brought by a single sensor, thereby improving the overall monitoring accuracy and providing a comprehensive data basis for subsequent anomaly analysis of the box-type substation;

[0042] 2. By constructing a monitoring data set and an abnormal fault set and establishing the corresponding relationship between them, in-depth analysis and intelligent judgment of the operation status of the box-type substation are realized; by calculating the correlation index between the monitoring data and the corresponding abnormal type and making a comprehensive judgment based on the weights of the associated monitoring data, the actual operation status of the substation can be more accurately reflected, and the probabilities of false alarms and missed alarms are reduced;

[0043] 3. By setting a monitoring period and systematically collecting, preprocessing, and comparing and analyzing data, not only the accuracy and effectiveness of the data are ensured, but also the monitoring data is dynamically analyzed through a linear regression algorithm, effectively distinguishing data fluctuations from real anomalies, thereby improving the accuracy and timeliness of anomaly detection; this can timely discover potential problems in the operation of the substation, provide a reliable decision-making basis for management personnel, and contribute to ensuring the safe and stable operation of the box-type substation;

[0044] 4. Through a refined anomaly judgment strategy, not only can accurate prediction of single or multiple initial monitoring data be performed, but also the possibility and urgency of anomalies can be effectively evaluated in combination with historical operation data and linear regression analysis; by calculating the buffer time and the total prediction weight, a scientific early warning basis is provided for management personnel, which helps to timely discover and handle potential faults, thereby significantly improving the operation safety and stability of the box-type substation. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flowchart of the steps of a method for processing monitoring data of a box-type substation provided by the present invention;

[0046] Figure 2 is a schematic structural diagram of a system for processing monitoring data of a box-type substation provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] Reference Figure 1 , the present invention provides a method for processing monitoring data of a box-type substation, including:

[0049] Step 1: Determine multiple types of data that need to be monitored during the operation of the box-type substation; construct a monitoring data set and an abnormal fault set, and establish the corresponding relationship between each data in the monitoring data set and the abnormal fault set; calculate the correlation index between each monitoring data and the corresponding abnormal type; based on multiple associated monitoring data within each abnormal type, calculate the weight of each associated monitoring data;

[0050] The first step includes the following steps:

[0051] Step 101: Place multiple types of sensors in the box-type substation to obtain various data during the operation of the box-type substation;

[0052] The various data during the operation of the box-type substation include electrical parameter data, internal and external environment data, oil level and oil quality data, insulation performance data, and other data, etc. Among them, the electrical parameter data includes high-voltage side voltage, current and power, low-voltage side voltage, current and power, bus voltage, phase voltage, line voltage, zero-sequence voltage and current, current harmonic content, etc.; the internal and external environment data includes the temperature and humidity inside the box-type substation, transformer winding temperature, oil temperature, wire contact point temperature, etc.; the oil level and oil quality data includes the oil level in the transformer and the gas content in the oil, etc.; the insulation performance data includes the insulation resistance and breakdown voltage of transformer oil, the insulation resistance and dielectric loss of each device in the box-type substation, etc.; other data includes the discharge times of lightning arresters and the rotation speed of internal fans, etc.; the devices in the box-type substation include transformers, instrument transformers, lightning arresters, capacitors, etc.;

[0053] For the above various data during the operation of the box-type substation, different types of sensors are used to monitor the corresponding data, and the sensors are placed at designated positions. Specifically: Hall current sensors, electromagnetic voltage transformers, and power factor meters are respectively used to monitor the high-voltage side voltage, current and power, and the low-voltage side voltage, current and power in real time; the sensors corresponding to the high-voltage side voltage, current and power are installed in the high-voltage switchgear and are close to the high-voltage bus or cable connection; the sensors corresponding to the low-voltage side voltage, current and power are installed in the low-voltage switchgear and are close to the low-voltage bus or cable connection;

[0054] The bus voltage, phase voltage, and line voltage are respectively monitored using voltage sensors in the bus voltage detection device and are installed near the bus;

[0055] The zero-sequence voltage and current are respectively monitored in real time using a zero-sequence voltage sensor and a zero-sequence current transformer and are installed near the grounding system;

[0056] The amount of current harmonics is monitored in real time by a harmonic current sensor, which is installed near all sensors used to detect current;

[0057] The temperature and humidity inside the box-type substation are monitored by a temperature sensor and a humidity sensor respectively, and one temperature sensor and one humidity sensor are installed on the inner walls of the high-voltage chamber, low-voltage chamber, transformer chamber and the box of the box-type substation respectively;

[0058] The temperature of the transformer winding and the oil temperature are monitored by a fluorescence optical fiber temperature sensor, which are installed on the transformer winding wall and the oil duct wall respectively; the temperature of the wire contact point is monitored by an infrared temperature sensor, which is installed near the wire contact point where it is difficult to directly install a sensor;

[0059] The oil level of the transformer is monitored by a magnetic flap oil level gauge, which is installed on the oil tank of the transformer; the gas content in the oil is monitored by an on-line gas monitoring device, which is installed near the transformer oil tank and connected to the oil tank through an oil pipe to detect the gas content in the oil in real time;

[0060] The insulation resistance and breakdown voltage of the transformer oil are monitored by an insulating oil withstand voltage tester, which is placed near the selected sampling point of the transformer oil; the insulation resistance and dielectric loss of the equipment are monitored by a megohmmeter, which is placed near the equipment to be tested;

[0061] The discharge times of the lightning rod are monitored by monitoring the discharge situation to perform real-time cumulative update of the data; the rotation speed of the internal fan is monitored by a fan speed sensor;

[0062] Step 102: Put the above-obtained various types of monitoring data into the monitoring data set. The temperature data at the above different positions are different types of data, and the insulation resistance, dielectric loss and operation duration data of different equipment are all different types of data; perform standard deviation operations on various types of voltage data and current data, and record the results as the fluctuation values of the corresponding voltage and current, and also put these data into the monitoring data set; record the monitoring data set as {A1, A2, …, A n}, where A i represents the i-th data in the monitoring data set;

[0063] Step 103: Establish an abnormal fault set, which stores all possible abnormal situations that may occur in the box-type substation, including overheating of the box-type substation, failure of the on-load switch of the box substation, tripping fault of the low-voltage switch of the box substation, busbar fault, overheating of the transformer winding, deterioration of the insulation performance of the transformer oil, poor ventilation and heat dissipation, abnormal high (low) voltage measurement data, etc.; record the abnormal fault set as {B1, B2, …, B m}, where B j represents the j-th data in the abnormal fault set;

[0064] Step 104: Establish the correspondence between the monitoring data set and each data in the abnormal fault set. Obtain all the monitoring data records when abnormalities occur from the historical operation data of the box-type substation. Based on the first threshold of each type of data set, calculate the correlation index between each monitoring data in the monitoring data set and the corresponding abnormal type. The expression is:

[0065] Among them, represents the total number of occurrences of the j-th type of abnormality in the abnormal fault set in the historical operation records of the box-type substation. represents the total number of times that the i-th type of data in the monitoring data set exceeds the corresponding first threshold in all the operation records of the historical box-type substation when the j-th type of abnormality occurs in the abnormal fault set; the first threshold of each type of data is set by the substation staff themselves and is used to determine whether further judgment of the abnormality is required.

[0066] Step 105: Record the monitoring data with a non-zero correlation index for each type of abnormality as the associated monitoring data for the corresponding abnormal type. Similarly, these associated monitoring data will record these abnormalities as their respective associated abnormal types. One type of abnormality can have multiple associated monitoring data at the same time, and one type of monitoring data can also have multiple associated abnormal types at the same time;

[0067] Calculate the weight of each type of associated monitoring data based on the correlation index. The weight is used for subsequent comprehensive judgment of whether the corresponding abnormality occurs. The weight of the associated monitoring data The expression is:

[0068] Among them, represents the sum of the correlation indices of all the associated monitoring data of the j-th type of abnormality in the abnormal fault set.

[0069] When in use, combine the content in Steps 101 to 105:

[0070] By carefully planning and placing various types of sensors, the comprehensive collection of various key data during the operation of the box-type substation is ensured. This avoids the measurement errors that may be brought by a single sensor, thereby improving the overall monitoring accuracy and providing a comprehensive data basis for the subsequent abnormality analysis of the box-type substation;

[0071] By constructing the monitoring data set and the abnormal fault set and establishing the correspondence between them, the in-depth analysis and intelligent judgment of the operation state of the box-type substation are realized; by calculating the correlation index between the monitoring data and the corresponding abnormal type and making a comprehensive judgment based on the weight of the associated monitoring data, the actual operation state of the substation can be more accurately reflected, and the probability of false alarms and missed alarms can be reduced.

[0072] Step 2: Periodically monitor the determined various types of data. At the end of each period, screen out all the monitored data whose monitored values exceed the corresponding first threshold, and respectively determine whether the reason for exceeding the first threshold is data fluctuation, so as to screen out the initial monitored data; perform corresponding anomaly detection operations based on the number of the initial monitored data. When the number of the initial monitored data is single, screen out all the corresponding anomaly types, and based on the associated monitored data of each anomaly type, predict in turn whether the corresponding anomaly will occur after the buffer time calculated from the initial monitored data; when the number of the initial monitored data is multiple, first consider the anomaly types that all the initial monitored data or some of the initial monitored data completely correspond to, and then detect the anomaly types of the remaining single initial monitored data in turn.

[0073] The Step 2 includes the following steps:

[0074] Step 201: Set the monitoring period, monitor the various types of data in Step 101 at the end of each monitoring period and record the specific values; for the type of data that needs to be calculated, such as the voltage and current fluctuation values, perform the corresponding calculations and record the calculation results.

[0075] Preprocess the collected raw data, including data cleaning (removing incorrect or invalid data), data calibration (performing necessary corrections according to the sensor characteristics), and data normalization (converting data with different dimensions to a unified standard), to ensure the accuracy and effectiveness of data analysis.

[0076] Step 202: At the end of each monitoring period, compare the recorded values of various types of data with the corresponding first threshold. When it is detected that there is a monitored data value exceeding the corresponding first threshold, determine whether the monitored data exceeding the corresponding first threshold is caused by fluctuation, specifically:

[0077] Obtain the monitored values of this monitored data at the end of the previous several monitoring periods, and based on the linear regression algorithm, fit the linear function of the change of this monitored data over time. If a linear function cannot be fitted, then based on the monitored value of this monitored data in the next monitoring period, determine whether it is data fluctuation. If the value detected in the next monitoring period has returned to the normal value range, it indicates that this monitored data has data fluctuation; the normal value range represents the value range that does not exceed the corresponding first threshold, that is, the monitored data value range when the corresponding device is operating normally.

[0078] If the value detected in the next monitoring period still has not returned to the normal value range, but the difference between the monitored value and the first threshold is less than the difference between the monitored value and the first threshold in the current monitoring period, then use the same method to further determine whether there is data fluctuation according to the monitored value of this monitored data in the next next monitoring period.

[0079] If the value detected in the next monitoring period still does not return to the normal value range, and the difference between the monitoring value and the first threshold is not less than the difference between the monitoring value in the current monitoring period and the first threshold, it is determined that the box-type substation is abnormal, and the abnormal judgment strategy is executed;

[0080] After determining to execute the abnormal judgment strategy, based on the monitoring value in the current monitoring period and the monitoring values in the next several monitoring periods, a linear function is fitted based on the linear regression algorithm;

[0081] If a linear function can be fitted, it is determined that the box-type substation is abnormal, and the abnormal judgment strategy is executed;

[0082] Step 203, execute the abnormal judgment strategy: Mark the monitoring data that exceeds the corresponding first threshold and is determined not to have data fluctuations as the initial monitoring data. If at the end of a monitoring period, only one initial monitoring data appears, directly predict the remaining specified monitoring data. Sequentially take out each type of abnormal associated with the abnormal fault concentration in order, and sequentially take out the associated monitoring data corresponding to each type of abnormal, and perform the prediction of each associated monitoring data, specifically including:

[0083] Step 2031, for several associated monitoring data corresponding to the same type of abnormal, obtain the initial monitoring data from the historical operation data of the box-type substation and the first monitoring value, that is, the first monitoring value, of each associated monitoring data when the corresponding abnormal type appears, and take the maximum or minimum value among the multiple first monitoring values as the second threshold of the initial monitoring data and each associated monitoring data; if the monitoring data is abnormal when it exceeds the first threshold, take the minimum value among the multiple first monitoring values as the second threshold of the corresponding monitoring data; if the monitoring data is abnormal when it is lower than the first threshold, take the maximum value among the multiple first monitoring values as the second threshold of the corresponding monitoring data;

[0084] Step 2032, obtain the second threshold of the initial monitoring data, calculate the difference between the current monitoring value of this monitoring data and the second threshold, and based on the linear function fitted in step 202, calculate the time required for the value of the initial monitoring data to change to the second threshold, that is, the buffer time;

[0085] Step 2033, similarly obtain the monitoring values of each associated monitoring data at the end of the previous several monitoring periods, and respectively perform the fitting of the linear function, ensuring that the monitoring values under the obtained number of monitoring periods can be used to fit a linear function. For example, for a certain associated monitoring data, the monitoring values under the previous 4 monitoring periods cannot be used to obtain a linear function, but the monitoring values under the previous 3 monitoring periods can be used to obtain a linear function, then the linear function is fitted based on the monitoring values under the previous 3 monitoring periods;

[0086] Based on the buffering time calculated in step 2032, calculate the fitting value of each associated monitoring data after the buffering time, and compare it with the corresponding second threshold. Record the associated monitoring data that exceeds or is lower than the corresponding second threshold, and add the sum of the weights of these associated monitoring data to the weight of the initial monitoring data to obtain the total predicted weight of the corresponding abnormal type in the current monitoring cycle. If it indicates that based on the data of the current monitoring cycle, it is predicted that the corresponding abnormal type will not occur after the buffering time;

[0087] If it indicates that based on the data of the current monitoring cycle, it is predicted that the corresponding abnormality may occur after the buffering time, and further judgment is required. Specifically, mark the current monitoring cycle for the corresponding abnormality. If the total predicted weight of this abnormal type obtained from the analysis of the next monitoring cycle it indicates that based on the data of the next monitoring cycle, it is predicted that the corresponding abnormal type will occur after the buffering time;

[0088] If it indicates that based on the data of the current monitoring cycle, it is predicted that the corresponding abnormal type will occur after the buffering time;

[0089] Step 2034: Perform the operations in steps 2031 to 2033 for each type of abnormality in the associated abnormal types in turn to predict whether the box-type substation will have the corresponding abnormal type;

[0090] If multiple initial monitoring data appear simultaneously at the end of a monitoring cycle, first screen the abnormal types corresponding to these initial monitoring data, and then judge whether other monitoring data values need to be predicted. The specific steps are as follows:

[0091] Step 2131: Screen out the abnormal types that exactly correspond to all or some of the initial monitoring data, that is, the associated monitoring data corresponding to the screened abnormal types are all initial monitoring data; if all the initial monitoring data exactly correspond to one abnormal type, then based on the fitting function of each initial monitoring data, calculate the buffer time required for the numerical change of each initial monitoring data to reach the corresponding second threshold respectively; if some of the initial monitoring data exactly correspond to one abnormal type, then extract these initial monitoring data for analysis of the corresponding abnormal type, and also based on the fitting function of each initial monitoring data, calculate the buffer time required for the numerical change of each initial monitoring data to reach the corresponding second threshold respectively, and the remaining initial monitoring data are respectively subjected to abnormal prediction under each initial monitoring data according to steps 2031 to 2033; for example, if at the end of a certain monitoring period, 3 initial monitoring data are identified, and after further analysis, it is found that these 3 initial monitoring data exactly correspond to the associated monitoring data of an abnormal type, then based on the fitting function of each initial monitoring data, calculate the buffer time required for the numerical change of each initial monitoring data to reach the corresponding second threshold respectively, and further determine whether the corresponding abnormal type has occurred; if after further analysis, it is found that 2 of these 3 initial monitoring data exactly correspond to the associated monitoring data of an abnormal type, then extract these two initial monitoring data, and again based on the fitting function of these two initial monitoring data, calculate the buffer time required for the numerical change of each initial monitoring data to reach the corresponding second threshold respectively, and further determine whether the corresponding abnormal type has occurred, and the remaining one initial monitoring data is subjected to abnormal identification according to the operations in steps 2031 to 2033.

[0092] Step 2132: Calculate the difference t between the maximum value and the minimum value among multiple buffer times. If t < t0, it means that the change trend of these initial monitoring data is the same as the change trend of the monitoring data under the corresponding abnormal type, and it is predicted that the corresponding abnormal type will occur after the maximum buffer time; where t0 represents the buffer time difference threshold, which is set by the substation management personnel themselves.

[0093] If t ≥ t0, it means that the change trend of these initial monitoring data is different from the change trend of the monitoring data under the corresponding abnormal type, and these initial monitoring data need to be separated for abnormal judgment to determine whether there are some initial monitoring data that exactly correspond to one abnormal type. If so, abnormal judgment is carried out based on the operations in steps 2131 to 2132; if not, abnormal prediction is carried out for each initial monitoring data based on steps 2031 to 2033.

[0094] Step 204: Based on the prediction results of each abnormal type, send early warnings for the corresponding abnormal types.

[0095] During use, in combination with the content in Steps 201 to 204:

[0096] By setting the monitoring period and systematically collecting, preprocessing, and comparing and analyzing data, not only the accuracy and effectiveness of the data are ensured, but also the monitoring data is dynamically analyzed through the linear regression algorithm to effectively distinguish data fluctuations from real anomalies, thereby improving the accuracy and timeliness of anomaly detection; this can timely discover potential problems in the operation of the substation, provide reliable decision-making basis for management personnel, and contribute to ensuring the safe and stable operation of the box-type substation;

[0097] Through the refined anomaly judgment strategy, not only can accurate predictions be made for single or multiple initial monitoring data, but also the possibility and urgency of anomalies can be effectively evaluated by combining historical operation data and linear regression analysis; by calculating the buffer time and the total prediction weight, a scientific early warning basis is provided for management personnel, which helps to timely discover and handle potential faults, thereby significantly improving the operation safety and stability of the box-type substation.

[0098] Reference Figure 2 , the present invention also provides a box-type substation monitoring data processing system, including:

[0099] A monitoring data determination module for determining various types of data that need to be monitored during the operation of the box-type substation;

[0100] A corresponding relationship establishment module for constructing a monitoring data set and an abnormal fault set, establishing the corresponding relationship between the data in the monitoring data set and the abnormal fault set; calculating the correlation index between each monitoring data and the corresponding abnormal type; calculating the weight of each associated monitoring data based on multiple associated monitoring data within each abnormal type;

[0101] An anomaly analysis module includes an initial monitoring data screening unit and an associated monitoring data prediction unit. The initial monitoring data screening unit is used to periodically monitor the determined various types of data, and at the end of each cycle, screen out all monitoring data whose monitoring values exceed the corresponding first threshold, and respectively judge whether the reason for exceeding the first threshold is data fluctuation to screen out the initial monitoring data;

[0102] The associated monitoring data prediction unit performs corresponding anomaly detection operations based on the number of initial monitoring data. When the number of initial monitoring data is single, screen out all the corresponding abnormal types, and based on the associated monitoring data of each abnormal type, predict in turn whether the corresponding anomaly will occur after the buffer time calculated from the initial monitoring data; when the number of initial monitoring data is multiple, first consider the abnormal types that all initial monitoring data or some initial monitoring data completely correspond to, and then perform the anomaly type detection of the remaining single initial monitoring data in turn.

[0103] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer storage medium or transmitted through a computer storage medium.

[0104] The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired manner (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or a wireless manner (e.g., infrared, wireless, microwave, etc.). The computer storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more integrated available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)), etc.

[0105] As described above, the above are only specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for processing monitoring data of a box-type substation, characterized in that: Including: Determine multiple types of data that need to be monitored during the operation of the box-type substation; Construct a monitoring data set and an abnormal fault set, and establish the corresponding relationship between the data in the monitoring data set and the abnormal fault set; Calculate the correlation index between each monitoring data and the corresponding abnormal type; Based on multiple associated monitoring data within each abnormal type, calculate the weight of each associated monitoring data; Periodically monitor the determined various types of data. At the end of each period, screen out all the monitoring data whose monitoring values exceed the corresponding first threshold, and respectively judge whether the reason for exceeding the first threshold is data fluctuation to screen out the initial monitoring data; Perform corresponding anomaly detection operations based on the number of initial monitoring data: when the number of initial monitoring data is single, screen out all the corresponding abnormal types, and based on the associated monitoring data of each abnormal type, predict in turn whether the corresponding anomaly will occur after the buffer time calculated from the initial monitoring data; At the end of a monitoring period, if there is initial monitoring data, execute the anomaly judgment strategy. When only one initial monitoring data appears, sequentially take out all the associated abnormal types of this initial monitoring data, and sequentially take out the remaining associated monitoring data corresponding to each abnormal type; Obtain the first monitoring value of each associated monitoring data when the corresponding abnormal type appears, and take the maximum or minimum value among the multiple first monitoring values as the second threshold of the initial monitoring data and each associated monitoring data; Obtain the second threshold of the initial monitoring data, and calculate the difference between the monitoring value of this monitoring data currently and the second threshold; Based on the fitted linear function, calculate the buffer time required for the numerical change of the initial monitoring data to reach the second threshold; Fitting a linear function based on the monitored values of each associated monitoring data remaining at the end of the previous several monitoring cycles; calculating the fitted values of each remaining associated monitoring data after a buffer time based on the fitted linear function, comparing with the corresponding second threshold, recording the associated monitoring data that exceeds or is lower than the corresponding second threshold, and adding the sum of the weights of these associated monitoring data to the weight of the initial monitoring data to obtain the total predicted weight of the corresponding abnormal type in the current monitoring cycle If it indicates that based on the data of the current monitoring cycle, it is predicted that the corresponding abnormal type will not occur after the buffer time; If it means that based on the data of the current monitoring period, it is predicted that the corresponding abnormality may occur after the buffer time, and further judgment is required. Specifically, the current monitoring period is marked with the corresponding abnormality type. If the total predicted weight of this abnormality type obtained from the analysis of the next monitoring period it means that based on the data of the next monitoring period, it is predicted that the corresponding abnormality type will occur after the buffer time; If it means that based on the data of the current monitoring period, it is predicted that the corresponding abnormal type will occur after the buffer time; When the number of initial monitoring data is multiple, first consider the abnormal types that all the initial monitoring data or some of the initial monitoring data completely correspond to, and then sequentially perform the anomaly type detection of the remaining single initial monitoring data.

2. The method for processing monitoring data of a box-type substation according to claim 1, characterized in that: Put the determined types of multi-monitoring data into the monitoring data set, and denote the monitoring data set as {A1, A2, …, A i , …, A n}, where A i represents the i-th data in the monitoring data set; Put all the abnormal types that occurred during the historical operation of the box-type substation into the abnormal fault set, and denote the abnormal fault set as {B1, B2, …, B j , …, B m}, where B j represents the j-th data in the abnormal fault set.

3. The method for processing monitoring data of a box-type substation according to claim 2, characterized in that: Obtain the monitoring data records when all exceptions occur, and calculate the correlation index between each monitoring data in the monitoring dataset and the corresponding exception type based on the first threshold of each type of data set The expression is: Among them, represents the total number of occurrences of the j-th type of abnormality in the abnormal fault concentration in the historical operation records of the box-type substation. represents the total number of times that the i-th type of data in the monitoring data set exceeds the corresponding first threshold among all operation records of the j-th type of abnormality in the abnormal fault concentration of the historical box-type substation.

4. The method for processing monitoring data of a box-type substation according to claim 3, characterized in that: Record the monitoring data with a non-zero correlation index with each type of anomaly as the associated monitoring data of the corresponding abnormal type, and each associated monitoring data will record these anomalies as its respective associated abnormal types; Calculate the weight of each type of associated monitoring data based on the relevance index The expression is as follows: Among them, represents the sum of the relevance indices of all associated monitoring data of the j-th type of anomaly in the abnormal fault concentration.

5. The method for processing monitoring data of a box-type substation according to claim 4, characterized in that: At the end of each monitoring period, compare the recorded data values of various types with the corresponding first threshold. When it is detected that there is a monitoring data value that exceeds or is lower than the corresponding first threshold, obtain the monitoring values of this monitoring data at the end of the previous several monitoring periods, and based on the linear regression algorithm, judge whether it is possible to fit the linear function of the change of this monitoring data over time; if a linear function can be fitted, then execute the anomaly judgment strategy; If a linear function cannot be fitted, it is determined whether the value detected in the next monitoring period has returned to the normal value range. If it has returned to the normal value range, the monitoring data is not recorded as the initial monitoring data.

6. A method for processing monitoring data of a box-type substation according to claim 5, characterized in that: If the value detected in the next monitoring period still has not returned to the normal value range, but the difference between the monitoring value and the first threshold is less than the difference between the monitoring value in the current monitoring period and the first threshold, it is further determined whether there is data fluctuation according to the monitoring value of the monitoring data in the next next monitoring period; If the value detected in the next monitoring period still has not returned to the normal value range, and the difference between the monitoring value and the first threshold is not less than the difference between the monitoring value in the current monitoring period and the first threshold, the monitoring data is recorded as the initial monitoring data, and a linear function is fitted based on the monitoring value in the current monitoring period and the monitoring values in the next several monitoring periods by using the linear regression algorithm.

7. A method for processing monitoring data of a box-type substation according to claim 1, characterized in that: When multiple initial monitoring data appear simultaneously, the abnormal types that completely correspond to all or some of the initial monitoring data are screened out; if all the initial monitoring data completely correspond to one abnormal type, then based on the fitting functions of each initial monitoring data, the buffer time required for the value of each initial monitoring data to change to the corresponding second threshold is calculated respectively; Calculate the difference t between the maximum value and the minimum value of the multiple buffer times. If t < t0, it is predicted that the corresponding abnormal type will occur after the maximum buffer time; where t0 represents the buffer time difference threshold; If t ≥ t0, it is determined whether there are some initial monitoring data that completely correspond to one abnormal type. If so, multiple buffer times are also calculated for abnormal judgment; if not, abnormal prediction is performed for each initial monitoring data.

8. A monitoring data processing system for a box-type substation, which is used to implement the method described in any one of claims 1 to 7, and is characterized in that, Including: A monitoring data determination module, used to determine multiple types of data that need to be monitored during the operation of the box-type substation; A corresponding relationship establishment module, used to construct a monitoring data set and an abnormal fault set, and establish the corresponding relationship between the data in the monitoring data set and the abnormal fault set; Calculate the correlation index between each monitoring data and the corresponding abnormal type; Based on multiple associated monitoring data within each abnormal type, calculate the weight of each associated monitoring data; An abnormal analysis module, including an initial monitoring data screening unit and an associated monitoring data prediction unit; Among them, the initial monitoring data screening unit is used to periodically monitor each type of data determined, and at the end of each period, screen out all monitoring data whose monitoring values exceed the corresponding first threshold, and respectively determine whether the reason for exceeding the first threshold is data fluctuation, so as to screen out the initial monitoring data; The associated monitoring data prediction unit performs corresponding anomaly detection operations based on the number of initial monitoring data. When the number of initial monitoring data is single, it filters out all corresponding anomaly types, and based on the associated monitoring data of each anomaly type, predicts in turn whether the corresponding anomaly will occur after the buffer time calculated from the initial monitoring data. When the number of initial monitoring data is multiple, it first considers the anomaly types that all or some of the initial monitoring data exactly correspond to, and then sequentially performs the anomaly type detection of the remaining single initial monitoring data.

Images