Power equipment multi-mode fault pre-judgment system based on knowledge base

Abstract

The invention belongs to the technical field of power equipment fault processing, and relates to a power equipment multi-mode fault pre-judgment system based on a knowledge base. According to the method, the parameter value range corresponding to each related parameter of each fault type of each type of power equipment is obtained, and the more scientific parameter value range of each related parameter when the power equipment fails is formulated, so that the fault can be accurately positioned, the power failure time is greatly shortened, and huge economic loss caused by long-time power failure is avoided; by obtaining normal operation, potential faults, all types of power equipment corresponding to the faults and all faults of all types of power equipment corresponding to the faults, state changes of recent power equipment are quickly focused, the troubleshooting and repairing period is shortened, timeliness and continuity of power supply are guaranteed, and the power supply efficiency is improved. By analyzing the matching degree of the power equipment fault clustering and the manual verification result, the diagnosis deviation can be corrected in time, and the fault judgment is more accurate.

Description

Technical Field

[0001] The present invention belongs to the technical field of power equipment fault data processing, and relates to a multi-modal fault prediction system for power equipment based on a knowledge base. Background Art

[0002] With the continuous expansion of the scale of the power system and technological progress, power equipment presents the dual characteristics of a sharp increase in quantity and a leap in complexity. The frequent occurrence of equipment failures not only threatens the stable operation of the power grid but also has a significant impact on production and life. As the material basis of the power system, equipment reliability directly determines the robustness of the system. Therefore, it is urgent to build an intelligent fault management system for power equipment. The knowledge base systematically integrates multi-source data and domain knowledge, providing core cognitive support for fault prediction, diagnosis, and disposal, and becoming an intelligent infrastructure to ensure the safe operation of the power system.

[0003] In the prior art, there are also some related solutions for power equipment data processing. For example, an invention patent application for a method and system for processing power equipment fault data based on power big data with the Chinese patent publication number CN118035924B includes: collecting fault data of electric energy meters in real time, preprocessing the collected fault data, and performing qualified detection to provide comprehensive and accurate data support for the fault judgment of electric energy meters, calculating the fault characterization coefficient of electric energy meters, determining the abnormal degree of electric energy meter faults, which is beneficial to accurately diagnosing electric energy meter fault problems, calculating the fault characterization difference of electric energy meters based on the fault data of electric energy meters, calculating the influence value of data deviation on the fault characterization coefficient of electric energy meters based on the obtained fault characterization difference of electric energy meters and the t-statistic of two groups of data, and analyzing the influence degree of data deviation on electric energy meter faults according to the influence value, which is beneficial to reducing the misjudgment rate of the fault degree of electric energy meters and is also beneficial to performing corresponding optimization and adjustment on electric energy meters to improve their metering accuracy and operation stability.

[0004] Another Chinese patent with the publication number CN118673199A discloses a data processing method, device, storage medium, and electronic device for power equipment, which includes: receiving a target query message uploaded by a setting value setting system, where the target query message includes the device identifier of the relay protection device to be queried and the substation identifier to which the relay protection device belongs; sending the target query message to the protection information master station, determining the target protection information substation according to the substation identifier, where the target protection information substation corresponds to the substation indicated by the substation identifier; sending the target query message to the target protection information substation, and extracting the target device parameters of the relay protection device. Through this application, the technical problem of low processing efficiency of device parameters is solved.

[0005] Although the above solutions propose some solutions for the data processing of power equipment, there are still certain limitations: (1) Some existing technologies collect and analyze the fault data of electric energy meters to obtain the fault degree of electric energy meters, and some existing technologies detect and analyze the data of relay protection devices to finally solve the technical problem of their low processing efficiency. However, the existing technologies all analyze the data of a certain power equipment, so it is impossible to achieve a wide range of power equipment fault location, which is not conducive to ensuring the continuity of industrial production and residential electricity consumption.

[0006] (2) The existing technologies lack the screening and analysis of the historical data of power equipment, so as to formulate a more scientific parameter value range of each relevant parameter when the power equipment fails, and then it is impossible to quickly compare and refer to it, which is not conducive to accurately locating the fault and cannot greatly shorten the power outage time.

[0007] (3) The existing technologies also ignore verifying the fault conditions of power equipment through manual inspections, which is not conducive to timely correcting the diagnostic deviation and making the fault judgment more accurate. Summary of the Invention

[0008] In view of this, to solve the problems raised in the above background technology, a multi-modal fault prediction system for power equipment based on a knowledge base is proposed.

[0009] The object of the present invention can be achieved by the following technical solutions: The present invention provides a multi-modal fault prediction system for power equipment based on a knowledge base, including: the parameter acquisition module, the fault induction module, the parameter screening module, the fault clustering analysis module and the verification module, where:

[0010] The parameter acquisition module obtains the historical various types of data of each type of power equipment from the operation log, including historical electrical data, historical operation data and historical maintenance data.

[0011] The fault induction module analyzes the correlation coefficients between the historical various types of data parameters of each type of power equipment and each parameter and each fault type, and accordingly obtains the parameter value ranges corresponding to each relevant parameter of each fault type.

[0012] The parameter screening module screens the various types of data parameters of each type of power equipment within a set period.

[0013] The fault clustering analysis module analyzes the probability evaluation index and fault risk index of each fault type of each type of power equipment within a set period, and accordingly conducts clustering division to obtain the normal operation, potential faults and faults corresponding to each type of power equipment and the faults corresponding to each type of power equipment.

[0014] The verification module verifies the normal operation, potential faults, and faults corresponding to each power device and each fault through manual inspections, analyzes the matching degree between the power device fault clustering and the manual verification results, and feeds them back.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. By analyzing, the present invention obtains the corresponding parameter value ranges of the relevant parameters of each fault type of each power device belonging to each type. Through the screening and analysis of historical data, the present invention formulates more scientific parameter value ranges of the relevant parameters when a power device fails, and can quickly compare and reference, accurately locate the fault, which is beneficial to greatly shorten the power outage time, ensure the continuity of industrial production and residential electricity consumption, avoid huge economic losses caused by long-term power outages, and improve power supply reliability.

[0016] 2. By analyzing, the present invention obtains the normal operation, potential faults, and faults corresponding to each power device belonging to each type, and the faults corresponding to each power device belonging to each type. By accurately screening the power device data, it is beneficial to quickly focus on the recent state changes of the power devices, thereby shortening the fault troubleshooting and repair cycle, improving the overall operation and maintenance efficiency, ensuring the timeliness and continuity of power supply, reducing the power outage duration caused by equipment failures, and improving power supply reliability.

[0017] 3. The verification module verifies the normal operation, potential faults, and faults corresponding to each power device belonging to each type, and the faults corresponding to each power device belonging to each type through manual inspections, analyzes the matching degree between the power device fault clustering and the manual verification results, and compares the fault clustering results with the actual manual inspection results, which helps to promptly correct the diagnostic deviation and make the fault judgment more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 It is a schematic diagram of the system composition of the present invention.

[0020] Figure 2 It is a schematic diagram of the implementation process of the method corresponding to the system of the present invention.

[0021] Figure 3 It is a schematic diagram of the screening of power devices corresponding to normal operation, potential faults, and fault types of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0023] Please refer to Figure 1 As shown, the present invention provides a multimodal fault prediction system for power equipment based on a knowledge base, specifically including a parameter acquisition module that obtains historical various types of data of each type of power equipment from the operation log, which includes historical electrical data, historical operation data, and historical maintenance data.

[0024] It should be further noted that the schematic diagram of the implementation process of the corresponding method of this system is as Figure 2 shown.

[0025] It should be further noted that the operation log includes equipment operation monitoring data, power grid dispatching data, power user data, meteorological and geographical data, etc.

[0026] As a preferred feasible embodiment, the historical electrical data parameters include the voltage, current, power, frequency, and power factor detected in each historical test.

[0027] The historical operation data parameters include the operation duration, start-stop times, load rate, temperature, vibration frequency, and rotational speed detected in each historical test.

[0028] The historical maintenance data parameters include the maintenance times, maintenance frequencies, maintenance contents, and maintenance costs detected in each historical test.

[0029] A fault induction module analyzes the correlation coefficients between the historical various types of data parameters of each type of power equipment corresponding to each parameter and each fault type, and accordingly obtains the value ranges of the corresponding parameters of each fault type.

[0030] As a preferred feasible embodiment, the specific analysis method for the correlation coefficients between the historical various types of data parameters of each type of power equipment corresponding to each parameter and each fault type includes: respectively extracting the voltage, current, power, frequency, and power factor detected in each historical test from the historical electrical data parameters of each type of power equipment, the operation duration, start-stop times, load rate, temperature, vibration frequency, and rotational speed detected in each historical test from the historical operation data parameters, and the maintenance times, maintenance frequencies, maintenance contents, and maintenance costs detected in each historical test from the historical maintenance data parameters, and collectively referring to them as the values of the corresponding parameters detected in each historical test of the historical various types of data parameters of each type of power equipment.

[0031] Extract the classification labels of each fault type for each historical detection from the operation log, and determine the corresponding coding values of the classification labels of each fault type for each historical detection.

[0032] Furthermore, calculate the correlation coefficients between each parameter corresponding to each type of historical data of each power equipment belonging to each type and each fault type through the standard calculation formula of the Pearson correlation coefficient.

[0033] It should be further explained that the specific calculation method of the correlation coefficients between each parameter corresponding to each type of historical data of each power equipment belonging to each type and each fault type is: according to the standard calculation formula of the Pearson correlation coefficient Obtain the correlation coefficient between a certain parameter corresponding to a certain type of historical data of a certain power equipment belonging to a certain type and a certain fault type , where is the value of this parameter for the historical data of this type of this power equipment corresponding to the th detection, is the coding value corresponding to the classification label of the fault type of this type for the th detection, and , , is the number of each historical detection, is the number of historical detections.

[0034] It should be further explained that if the coding value corresponding to the classification label of the fault type of this type for a certain historical detection is 1, then the classification label of the fault type of this type for this historical detection is recorded as a fault. If the coding value corresponding to the classification label of the fault type of this type for a certain historical detection is 0, then the classification label of the fault type of this type for this historical detection is recorded as normal.

[0035] Furthermore, obtain the correlation coefficients between each parameter corresponding to each type of historical data of each power equipment belonging to each type and each fault type , where , is the number of each type, is the number of types, , is the number of each power equipment, is the number of power equipment, , is the number of each type of historical data parameter, is the number of types of historical data parameters, , is the number of each parameter, is the number of parameters, , is the number for each fault type, is the number of fault types.

[0036] As a preferred feasible embodiment, the specific obtaining method of the parameter value ranges corresponding to the relevant parameters of each fault type of each power device belonging to each type includes: comparing the historical data parameters of each power device belonging to each type corresponding to each parameter with the correlation coefficients of each fault type respectively with a set correlation coefficient threshold, screening out the parameters corresponding to the historical data parameters of each power device belonging to each type whose correlation coefficients with each fault type are greater than the correlation coefficient threshold, and collectively referring to them as the relevant parameters of each power device belonging to each type related to each fault type, and further denoting them as the relevant parameters of each fault type of each power device belonging to each type.

[0037] Screen the parameter values corresponding to the historical detections of the relevant parameters of each fault type of each power device belonging to each type from the parameter values of the historical detections of the historical data parameters of each power device belonging to each type, and accordingly determine the parameter value ranges corresponding to the relevant parameters of each fault type of each power device belonging to each type.

[0038] It should be further noted that the specific method for determining the parameter value ranges corresponding to the relevant parameters of each fault type of each power device belonging to each type is: sorting the parameter values corresponding to the historical detections of the relevant parameters of each fault type of each power device belonging to each type in ascending order, and then taking the minimum value and the maximum value among them as the minimum boundary value and the maximum boundary value of the parameter value ranges corresponding to the relevant parameters of each fault type of each power device belonging to each type, that is, determining the parameter value ranges corresponding to the relevant parameters of each fault type of each power device belonging to each type.

[0039] The present invention obtains the parameter value ranges corresponding to the relevant parameters of each fault type of each power device belonging to each type through analysis, and formulates more scientific parameter value ranges of the relevant parameters during power device failures through screening and analysis of historical data, so as to be able to quickly make a comparison and reference, accurately locate the fault, which is beneficial to greatly shorten the power outage time, ensure the continuity of industrial production and residential electricity consumption, avoid huge economic losses caused by long-term power outages, and improve power supply reliability.

[0040] A parameter screening module that screens the data parameters of each type of each power device belonging to each type within a set period.

[0041] As a preferred feasible embodiment, the electrical data parameters include the voltage, current, power, frequency, and power factor of each detection.

[0042] The operation data parameters include the operation duration, start-stop times, load rate, temperature, vibration frequency, and rotational speed for each detection.

[0043] The maintenance data parameters include the maintenance times, maintenance frequency, maintenance content, and maintenance cost for each detection.

[0044] The fault clustering analysis module analyzes the probability evaluation index and fault risk index of each fault type of each power equipment belonging to each type within a set period, and performs clustering division based on this to obtain the normal operation, potential faults, and faults corresponding to each power equipment belonging to each type, as well as the faults of each power equipment belonging to each type corresponding to the faults.

[0045] As a preferred feasible embodiment, the specific analysis method of the probability evaluation index of each fault type of each power equipment belonging to each type within the set period is as follows: in the same way as the historical data parameters of each type corresponding to the parameter values of each detection in the history of each power equipment belonging to each type, extract the parameter values of each type of data parameter corresponding to each detection of each power equipment belonging to each type within the set period, screen out the parameter values of each relevant parameter of each fault type of each power equipment belonging to each type corresponding to each detection within the set period, match them with the parameter value ranges of each relevant parameter of each fault type of each power equipment belonging to each type respectively, and obtain the parameter values of each relevant parameter of each fault type of each power equipment belonging to each type corresponding to each detection within the range of its relevant parameter corresponding parameter values, and record them as the parameter values of each fault parameter of each fault type of each power equipment belonging to each type corresponding to each detection within the set period , where , is the number of each type, is the number of types, , is the number of each power equipment, is the number of power equipment, , is the number of each fault type, is the number of fault types, , is the number of each fault parameter, is the number of fault parameters, , is the number of each detection, is the number of detections.

[0046] It should be further noted that the specific method for obtaining the parameter values corresponding to each detection of each relevant parameter of each failure type of each power device belonging to each type within the set period belonging to its relevant parameter corresponding parameter value range is as follows: If the parameter value corresponding to a certain detection of a certain relevant parameter of a certain failure type of a certain power device belonging to a certain type within the set period is within the range of the relevant parameter corresponding parameter value of the failure type of the power device belonging to the type, then the parameter value corresponding to the detection of the relevant parameter of the failure type of the power device belonging to the type within the set period is recorded as the parameter value corresponding to the detection of the relevant parameter of the failure type of the power device belonging to the type within the set period belonging to its relevant parameter corresponding parameter value range, and then the parameter values corresponding to each detection of each relevant parameter of each failure type of each power device belonging to each type within the set period belonging to its relevant parameter corresponding parameter value range are obtained.

[0047] Analyze the probability evaluation indexes of each failure type of each power device belonging to each type within the set period , where is the critical parameter value corresponding to the non-occurrence of a failure of the th failure parameter of the th power device belonging to the th failure type of the th type extracted from the database.

[0048] As a preferred feasible embodiment, the specific analysis method for the failure risk index of each power device belonging to each type within the set period is as follows: Extract the maintenance times, maintenance frequencies, each maintenance content, and each maintenance cost of each detection of each power device belonging to each type within the set period, and accordingly obtain the failure risk index corresponding to the maintenance times, the failure risk index corresponding to the maintenance frequencies, the failure risk index corresponding to each maintenance content, and the failure risk index corresponding to each maintenance cost of each detection of each power device belonging to each type within the set period, which are respectively denoted as , where , is the number of each maintenance, is the number of maintenance times.

[0049] It should be further noted that the specific methods for obtaining the failure risk indices corresponding to the maintenance times of each inspection of each power equipment of each type within the set period, the failure risk indices corresponding to the maintenance frequencies, the failure risk indices corresponding to each maintenance content, and the failure risk indices corresponding to each maintenance cost are as follows: The maintenance times, maintenance frequencies, each maintenance content, and each maintenance cost of each inspection of each power equipment of each type within the set period are respectively matched with the failure risk indices corresponding to each maintenance time range, each maintenance frequency range, each maintenance content, and each maintenance cost range stored in the database to obtain the failure risk indices corresponding to the maintenance times of each inspection of each power equipment of each type within the set period, the failure risk indices corresponding to the maintenance frequencies, the failure risk indices corresponding to each maintenance content, and the failure risk indices corresponding to each maintenance cost.

[0050] In a specific example, each of the maintenance time ranges includes, but is not limited to, 0 - 5, 6 - 15, 16 - 25, and above 25, etc. Among them, the failure risk indices corresponding to the maintenance time ranges of 0 - 5, 6 - 15, 16 - 25, and above 25 can be 0.1, 0.3, 0.5, and 0.9 respectively.

[0051] Each of the maintenance frequency ranges includes, but is not limited to, once a year, once every six months, once every three months, once a month, and once every ten days, etc. Among them, the failure risk indices corresponding to the maintenance frequency ranges of once a year, once every six months, once every three months, once a month, and once every ten days can be 0.1, 0.3, 0.5, 0.7, and 0.9 respectively.

[0052] Each of the maintenance contents includes, but is not limited to, cleaning and maintenance, tightening and adjustment, and replacement of parts, etc. Among them, the failure risk indices corresponding to cleaning and maintenance, tightening and adjustment, and replacement of parts can be 0.1, 0.2, and 0.5 respectively.

[0053] Each of the maintenance cost ranges includes, but is not limited to, 0 - 100 yuan, 100 - 1000 yuan, 1000 - 5000 yuan, and above 5000 yuan, etc. Among them, the failure risk indices corresponding to the maintenance cost ranges of 0 - 100 yuan, 100 - 1000 yuan, 1000 - 5000 yuan, and above 5000 yuan can be 0.1, 0.2, 0.4, and 0.8 respectively.

[0054] Analyze the failure risk indicators of each power equipment of each type within the set period , where are respectively the weight factors of the failure risk indicators corresponding to the set maintenance times, maintenance frequencies, maintenance contents, and maintenance costs.

[0055] A specific embodiment .

[0056] The maintenance content directly determines the quality and effect of maintenance work. Correct maintenance content can detect and repair potential faults in a timely manner, thereby reducing the probability of faults occurring. If the maintenance content is missing or inappropriate, it may lead to hidden dangers in the equipment even after maintenance, increasing the risk of faults. Therefore, the weight factor of the maintenance content corresponding to the fault risk index is assigned a value of 0.4.

[0057] The maintenance frequency reflects the timeliness of equipment maintenance. High-frequency maintenance can detect equipment problems earlier and reduce the occurrence of sudden faults. However, too high a maintenance frequency may also bring additional maintenance costs and workloads, and a balance needs to be found. Therefore, the weight factor of the maintenance frequency corresponding to the fault risk index is assigned a value of 0.3.

[0058] The number of maintenance times is related to the maintenance frequency, but it focuses more on the total number within the entire maintenance cycle. Appropriate maintenance times can ensure that the equipment receives sufficient attention and maintenance, but too many maintenance times may not always be beneficial, especially when the maintenance content or methods are inappropriate. Therefore, the weight factor of the number of maintenance times corresponding to the fault risk index is assigned a value of 0.2.

[0059] The maintenance cost is an important indicator to measure the economy of maintenance work. Although low-cost maintenance may attract some organizations, too low a cost may lead to a decline in maintenance quality, thereby increasing the risk of faults. On the other hand, too high a maintenance cost may also affect the overall efficiency of the organization. Therefore, the weight factor of the maintenance cost corresponding to the fault risk index is assigned a value of 0.1.

[0060] As a preferred feasible embodiment, the specific acquisition methods for the normal operation, potential faults, and faults corresponding to each type of each power equipment are as follows: Compare the probability evaluation indicators of each fault type of each type of each power equipment within a set period with the preset probability evaluation index threshold respectively, screen out each type of each power equipment whose probability evaluation indicators of each fault type within the set period are greater than the probability evaluation index threshold, and record them as the power equipment corresponding to each type of fault.

[0061] Compare the fault risk indicators of each type of each power equipment within a set period with the preset fault risk index threshold respectively, screen out each type of each power equipment whose fault risk indicators within the set period are greater than the fault risk index threshold, and record them as the power equipment corresponding to each type of potential fault.

[0062] Furthermore, record the power equipment of each type after excluding the power equipment corresponding to each type of potential fault and fault as the power equipment corresponding to the normal operation of each type.

[0063] It should be further noted that the screening schematic diagrams of the power equipment corresponding to the normal operation, potential faults, and faults of each type are as Figure 3as shown

[0064] As a preferred feasible embodiment, the specific acquisition method for each fault corresponding to each type of each power equipment is as follows: Screen the probability evaluation indexes of each fault type corresponding to each type of each power equipment from the probability evaluation indexes of each fault type corresponding to each type of each power equipment within a set period, compare them with the probability evaluation index threshold respectively, screen out each fault type corresponding to each type of each power equipment whose probability evaluation index is greater than the probability evaluation index threshold, and record them as each fault corresponding to each type of each power equipment.

[0065] Through analysis, the present invention obtains normal operation, potential faults, each type of each power equipment corresponding to faults, and each fault corresponding to each type of each power equipment. By accurately screening power equipment data, it is conducive to quickly focusing on the recent state changes of power equipment, thereby shortening the fault troubleshooting and repair cycle, improving the overall operation and maintenance efficiency, ensuring the timeliness and continuity of power supply, reducing the power outage duration caused by equipment failures, and improving power supply reliability.

[0066] The verification module verifies normal operation, potential faults, each power equipment corresponding to faults, and each fault through manual inspection, analyzes the matching degree between the power equipment fault clustering and the manual verification results, and feeds it back.

[0067] It should be further noted that after feeding back the matching degree between the power equipment fault clustering and the manual verification results, if the matching degree between the power equipment fault clustering and the manual verification results is less than the preset matching degree threshold, then S1 is further executed.

[0068] As a preferred feasible embodiment, the specific analysis method for the matching degree between the power equipment fault clustering and the manual verification results is as follows: Conduct manual inspection on normal operation, potential faults, and each type of each power equipment corresponding to faults to obtain the consistent quantity of normal operation, potential faults, and each type of power equipment corresponding to faults, which are respectively recorded as Conduct manual inspection on each fault corresponding to each type of each power equipment to obtain the fault consistent quantity of each fault corresponding to each type of each power equipment where is the number of each power equipment corresponding to the fault corresponding type is the quantity of power equipment corresponding to the fault corresponding type.

[0069] Analyze the matching degree between the power equipment fault clustering and the manual verification results where are respectively normal operation, potential faults, and the corresponding first ​The number of power equipment belonging to each type and the faults corresponding to the number of the power equipment belonging to each type.

[0070] It should be further explained that the specific acquisition methods for the number of power equipment belonging to each type corresponding to normal operation, potential faults and faults, and the number of faults of each power equipment belonging to each type corresponding to faults are as follows: count the power equipment belonging to each type corresponding to normal operation, potential faults and faults to obtain the number of power equipment belonging to each type corresponding to normal operation, potential faults and faults, and count the faults of each power equipment belonging to each type corresponding to faults to obtain the number of faults of each power equipment belonging to each type corresponding to faults.

[0071] As a preferred feasible embodiment, a database is used in the execution process of the system to store the critical parameter values when no faults occur corresponding to the fault parameters of each fault type of each power equipment belonging to each type, and to store the fault risk index corresponding to each maintenance frequency, each maintenance content and each maintenance cost.

[0072] The present invention verifies the power equipment belonging to each type corresponding to normal operation, potential faults and faults, and the faults of each power equipment belonging to each type corresponding to faults through manual inspection, analyzes the matching degree between the power equipment fault clustering and the manual verification results, and compares the fault clustering results with the actual manual inspection results, which helps to correct the diagnosis deviation in time and make the fault judgment more accurate.

[0073] The above content is only an example and explanation of the concept of the present invention. Those skilled in the art of the present technology make various modifications or supplements to the described specific embodiments or use similar methods for substitution. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should fall within the protection scope of the present invention.

Claims

1. A multi-modal fault prediction system for power equipment based on a knowledge base, characterized in that: Including: A parameter acquisition module that obtains historical various types of data of each type of power equipment from the operation log, including historical electrical data, historical operation data, and historical maintenance data; A fault induction module that analyzes the correlation coefficients between the historical various types of data parameters of each type of power equipment and each parameter and each fault type, and thereby obtains the corresponding parameter value ranges of each relevant parameter of each fault type; A parameter screening module that screens the various types of data parameters of each type of power equipment within a set period; A fault clustering analysis module that analyzes the probability evaluation indicators and fault risk indicators of each fault type of each type of power equipment within a set period, and thereby conducts clustering division to obtain normal operation, potential faults, and faults corresponding to each type of power equipment and each fault corresponding to each type of power equipment; A verification module that verifies the normal operation, potential faults, and faults corresponding to each power equipment and each fault through manual inspection, analyzes the matching degree between the power equipment fault clustering and the manual verification results, and feeds it back; 2. The multimodal fault prediction system for power equipment based on a knowledge base according to claim 1, wherein: The historical electrical data includes the voltage, current, power, frequency, and power factor detected in each historical measurement; The historical operation data includes the operation duration, start-stop times, load rate, temperature, vibration frequency, and rotational speed detected in each historical measurement; The historical maintenance data includes the maintenance times, maintenance frequencies, maintenance contents, and maintenance costs detected in each historical measurement; 3. The multi-modal fault prediction system for power equipment based on a knowledge base according to claim 2, wherein: The specific analysis method for the correlation coefficients between the historical various types of data parameters of each type of power equipment and each parameter and each fault type includes: Respectively extract the voltage, current, power, frequency, and power factor detected in each historical measurement from the historical electrical data of each type of power equipment, the operation duration, start-stop times, load rate, temperature, vibration frequency, and rotational speed detected in each historical measurement from the historical operation data, and the maintenance times, maintenance frequencies, maintenance contents, and maintenance costs detected in each historical measurement from the historical maintenance data, and collectively call them the corresponding parameter values of each parameter detected in each historical measurement of the historical various types of data parameters of each type of power equipment; Extract the classification labels of each fault type detected in each historical measurement from the power big database, and determine the corresponding coded values of the classification labels of each fault type detected in each historical measurement; Calculate the correlation coefficients between the historical various types of data parameters of each type of power equipment and each parameter and each fault type through the standard calculation formula of the Pearson correlation coefficient; 4. The multi-modal fault prediction system for power equipment based on a knowledge base according to claim 3, characterized in that: The specific obtaining method for the corresponding parameter value ranges of each relevant parameter of each fault type of each type of power equipment includes: Compare the correlation coefficients between the historical various types of data parameters of each type of power equipment and each parameter and each fault type with the set correlation coefficient threshold respectively, screen out the corresponding parameters of the historical various types of data parameters of each type of power equipment whose correlation coefficients with each fault type are greater than the correlation coefficient threshold, and collectively call them the respective relevant parameters of each type of power equipment related to each fault type, and further record them as the respective relevant parameters of each fault type of each type of power equipment; From the parameter values of each historical detection corresponding to the historical data parameters of each type of power equipment belonging to each type, screen the parameter values of each detection corresponding to the relevant parameters of each fault type of each type of power equipment belonging to each type, and accordingly determine the range of parameter values corresponding to the relevant parameters of each fault type of each type of power equipment belonging to each type.

5. The multi-modal fault prediction system for power equipment based on a knowledge base according to claim 3, wherein: The specific analysis method for the probability evaluation index of each fault type of each type of power equipment belonging to each type within the set period is as follows: In the same way as the historical data parameters of each type corresponding to the parameter values of each historical detection of each power equipment belonging to each type, extract the parameter values of each type of data parameter of each power equipment belonging to each type corresponding to each detection within a set period. Screen from them the parameter values of each relevant parameter of each fault type of each power equipment belonging to each type within the set period. Match them respectively with the parameter value ranges of each relevant parameter of each fault type of each power equipment belonging to each type to obtain the parameter values of each relevant parameter of each fault type of each power equipment belonging to each type within the set period that fall within the relevant parameter value ranges, and record them as the parameter values of each fault parameter of each fault type of each power equipment belonging to each type corresponding to each detection within the set period , where , is the number of each type, is the quantity of types, , is the number of each power equipment, is the quantity of power equipment, , is the number of each fault type, is the number of fault types, , is the number of each fault parameter, is the quantity of fault parameters, , is the number of each detection, is the number of detections; According to and Comprehensively analyze the probability evaluation indexes of each fault type of each power equipment belonging to each type within the set period based on the relative deviation, Is the critical parameter value corresponding to the th type belonging to the th power equipment, the th fault type, and the th fault parameter when the fault has not occurred.

6. The multi-modal fault prediction system for power equipment based on a knowledge base according to claim 5, characterized in that: The specific analysis method for the fault risk index of each type of power equipment belonging to each type within the set period is as follows: Extract the maintenance times, maintenance frequencies, maintenance contents, and maintenance costs for each inspection of each power equipment belonging to each type within the set period. Based on this, obtain the failure risk indices corresponding to the maintenance times, the failure risk indices corresponding to the maintenance frequencies, the failure risk indices corresponding to each maintenance content, and the failure risk indices corresponding to each maintenance cost for each inspection of each power equipment belonging to each type within the set period, and denote them respectively as , where , is the number of each maintenance, is the number of maintenance times; According to and using multi-factor weight fusion calculation and analysis to set the fault risk indicators of each type of power equipment within a certain period.

7. The multi-modal fault prediction system for power equipment based on a knowledge base according to claim 6, characterized in that: The specific acquisition method for each type of power equipment belonging to each type corresponding to normal operation, potential fault, and fault is as follows: Compare the probability evaluation indexes of each fault type of each type of power equipment belonging to each type within the set period with the preset probability evaluation index threshold respectively, screen out each type of power equipment belonging to each type in which the probability evaluation index of each fault type within the set period is greater than the probability evaluation index threshold, and record it as each type of power equipment belonging to each type corresponding to the fault; Compare the fault risk indexes of each type of power equipment belonging to each type within the set period with the preset fault risk index threshold respectively, screen out each type of power equipment belonging to each type in which the fault risk index within the set period is greater than the fault risk index threshold, and record it as each type of power equipment belonging to each type corresponding to the potential fault; Furthermore, record the power equipment belonging to each type after excluding the power equipment belonging to each type corresponding to the potential fault and the fault as the power equipment belonging to each type corresponding to normal operation.

8. The multi-modal fault prediction system for power equipment based on a knowledge base according to claim 7, characterized in that: The specific acquisition method for each fault of each type of power equipment belonging to each type corresponding to the fault is as follows: Screen the probability evaluation indexes of each fault type of each type of power equipment belonging to each type corresponding to the fault from the probability evaluation indexes of each fault type of each type of power equipment belonging to each type within the set period, compare them with the probability evaluation index threshold respectively, screen out the fault types of each type of power equipment belonging to each type corresponding to the fault in which the probability evaluation index is greater than the probability evaluation index threshold, and record it as each fault of each type of power equipment belonging to each type corresponding to the fault.

9. The multimodal fault prediction system for power equipment based on a knowledge base according to claim 5, characterized in that: The specific analysis method for the matching degree between the power equipment fault clustering and the manual verification result is as follows: Manually inspect each type of power equipment belonging to normal operation, potential faults, and faults, and obtain the consistent quantities of power equipment belonging to each type of normal operation, potential faults, and faults, which are respectively recorded as Manually inspect each fault of each power equipment belonging to each type corresponding to faults, and obtain the consistent quantity of faults of each power equipment belonging to each type corresponding to faults Among them , is the number of each power equipment belonging to the corresponding fault type, is the quantity of power equipment belonging to the corresponding fault type; according to and perform fusion analysis on the matching degree between power equipment fault clustering and manual verification results 10. The multi-modal fault prediction system for power equipment based on a knowledge base according to claim 1, characterized in that: The system uses a database during the execution process to store the critical parameter values when no fault occurs corresponding to the fault parameters of each fault type of each type of power equipment belonging to each type, and store the fault risk index corresponding to each maintenance times, each maintenance frequency, each maintenance content, and each maintenance cost.

Images