A device information acquisition device and method for device safety inspection

Abstract

The application discloses a device information collection device and method for device safety inspection, relates to the technical field of data processing, and comprises the following steps: collecting device operation data, environment data and artificial preliminary inspection data, and generating preliminary collection data; based on the preliminary collection data and set device normal working state data, analyzing and processing whether the device working state is normal; the device information collection device and method for device safety inspection, when the device is abnormal but the specific abnormal type cannot be analyzed, instructing the inspection personnel to perform deep detection on the unknown abnormal device to determine the abnormal type or the device normal (detection failure), and updating to the analysis reference sample, thereby expanding the sample library, reducing the probability of detection of unknown abnormalities and device normality judgment failure, and without wasting too much manpower, so that the automatic data collection and analysis and the artificial inspection reach a good balance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a device information acquisition device and method for device safety inspection. BACKGROUND

[0002] With the continuous development of modern industry, the number and complexity of equipment are increasing, and the traditional manual inspection method is inefficient. Modern device safety inspection systems integrate multiple cutting-edge technologies to build a comprehensive and multi-level device health protection system. These systems fundamentally change the traditional manual inspection mode and make device safety management enter a new stage of digitization and intelligence. For example, the prior art with publication number CN118278920A discloses an intelligent inspection system suitable for industrial equipment operation and maintenance, which includes an inspection management platform, an information acquisition unit, a potential interference risk unit, a maintenance residual risk unit, a fault self-checking unit, a detailed supervision unit and a management response unit. The prior art analyzes potential interference and maintenance interference to understand the fault risk situation during the operation of the industrial equipment, thereby ensuring the safety and stability of the operation of the industrial equipment. Under the premise of normal potential interference and maintenance interference, the working state data is analyzed for fault risk safety supervision, thereby reducing the risk of unexpected downtime, improving the continuity and stability of production, and periodically predicting and controlling analysis through information feedback and fusion to adjust the supervision intensity of the next industrial equipment management, thereby ensuring the rationality of the supervision of the industrial equipment and the stability of the operation.

[0003] The prior art only relies on sensor automatic data acquisition for analysis, but the accuracy of the analysis is affected by the accuracy of the sensor, the complexity of the equipment working condition, and the existence of interference factors, resulting in low accuracy of the analysis, resulting in many false positives and unexpected equipment downtime. The efficiency of manual inspection is also low, so how to balance automatic data acquisition and analysis and manual inspection is a problem to be solved. SUMMARY

[0004] The purpose of the present application is to provide a device information acquisition device and method for device safety inspection to solve the above problems in the prior art.

[0005] In order to achieve the above-mentioned purpose, the present application provides the following technical scheme: a device information acquisition method for device safety inspection, comprising the following steps:

[0006] S1, collecting equipment operation data, environmental data and manual preliminary inspection data to generate preliminary acquisition data;

[0007] S2, based on the preliminary acquisition data and the set normal working state data of the equipment, analyzing and processing whether the working state of the equipment is normal;

[0008] S3, if the device working state is not normal, based on the preliminary collected data and the set device abnormal type characteristic data, device abnormal type analysis processing is performed, device abnormal type analysis data is generated, the device abnormal type analysis data includes multiple device determined abnormal types, multiple device presumed abnormal types and device unknown abnormality;

[0009] S4, based on the abnormal device running data, the position of the corresponding device of the abnormal device running data is determined, and device abnormal position analysis data is generated;

[0010] S5, if the device abnormal type analysis data is device unknown abnormality, based on the preliminary collected data corresponding to the device unknown abnormality and the device normal working state data, similarity analysis of the device running data and the device normal working state data is performed, and the priority level of the device detection is determined based on the size of the similarity, and device detection priority level analysis data is generated;

[0011] S6, based on the device detection priority level analysis data corresponding to the device unknown abnormality and the device abnormal position analysis data, the inspection personnel is instructed to perform deep detection on the device abnormal position within the specified time, and deep inspection data and deep inspection result data are generated; wherein, the deep detection includes using professional detection tools to detect the inside and outside of the device or the abnormal position, the data obtained in the detection process is the corresponding deep inspection data, such as ultrasonic flaw detection data, image data, high-definition audio data, etc., and the result of analyzing the deep inspection data is the deep inspection result data, such as cracks in the device shell, loose bolts, excessive bearing wear, etc.

[0012] Further, the S1 includes the following steps:

[0013] S1.1, based on the data collection device, the device running data and the environment data can be collected by various sensors, detection instruments (such as ultrasonic flaw detector) and other devices;

[0014] Among them, the device running data can be: device temperature, pressure, vibration, current, load, device abnormal sound spectrum, device image, cooling liquid pH value, transformer oil chromatographic data, flow, speed, valve opening, etc.; The environment data can be the temperature, humidity, dust concentration, combustible gas concentration, etc. of the space where the device is located.

[0015] S1.2, record and save the artificial preliminary inspection data; wherein, the artificial preliminary inspection data can include the device sensory description (such as normal, surface crack, xx connection gap, loose bolt, xx color abnormality is xx color, etc.) obtained when the device is artificially inspected, and the artificial detection evidence certificate (such as corresponding photo, video, audio, etc.).

[0016] S1.3, collecting, combining, and generating preliminary collection data by running data, environment data, and artificial preliminary inspection data of the sensor collection device.

[0017] Further, the S2 includes the following steps:

[0018] S2.1, collecting historical device normal working state data by related device historical working log, obtaining device normal working state data;

[0019] S2.2, standardizing the preliminary collection data and the device normal working state data, generating first preliminary collection standard data and device normal working state standard data;

[0020] S2.3, calculating the first cosine similarity of the first preliminary collection standard data and the device normal working state standard data based on the cosine similarity algorithm, and judging whether the first cosine similarity is greater than the set first cosine similarity threshold;

[0021] S2.4, if yes, the device working state is normal, and if no, the device working state is not normal.

[0022] Further, the S3 includes the following steps:

[0023] S3.1, collecting historical preliminary collection data corresponding to the device working state not normal, and performing corresponding abnormal type marking, collecting preliminary collection data corresponding to each abnormal type marking respectively, generating device abnormal type-preliminary collection data set, the abnormal type marking includes multiple device determined abnormal types, multiple device speculative abnormal types, and device unknown abnormality, wherein if the current preliminary collection data matches the historical preliminary collection data corresponding to the device determined abnormal type marking, the abnormal type of the device is more likely (such as confidence greater than or equal to 95%) to correspond to the corresponding device determined abnormal type marking; if the current preliminary collection data matches the historical preliminary collection data corresponding to the device speculative abnormal type marking, the abnormal type of the device has a certain possibility (such as confidence greater than or equal to 90%) to correspond to the corresponding device speculative abnormal type marking; if the current preliminary collection data does not match the historical preliminary collection data corresponding to any device determined abnormal type and device speculative abnormal type marking, the abnormal type of the device is unknown, that is, matches the device unknown abnormality marking; wherein the device abnormal type can be, for example: high-frequency impact and envelope spectrum characteristic frequency exist in the position of the motor bearing, indicating bearing loss; 1 times frequency amplitude is prominent and phase is stable in the position of the rotor system, indicating rotor imbalance, etc.

[0024] S3.2, based on the DBSCAN clustering algorithm, the historical preliminary collection data is clustered and analyzed to generate a plurality of preliminary collection data feature clusters, and the number of device abnormal types corresponding to the historical preliminary collection data in each preliminary collection data feature cluster is counted;

[0025] S3.3, for each preliminary collection data feature cluster, the minimum real probability of the device abnormal type is calculated based on the exact binomial method when the confidence is greater than the set first confidence threshold and the lower limit of the confidence interval of the preliminary collection data feature cluster for the device abnormal type is greater than the set probability threshold; A feature cluster first threshold is obtained. Further, when judging, the effect amount of the corresponding device abnormal type can also be calculated to assist in judging the relationship strength between the device abnormal type and the corresponding preliminary collection data feature cluster.

[0026] S3.4, determine whether there is a device abnormal type in the preliminary collection data feature cluster whose true proportion is greater than or equal to the feature cluster first threshold, if yes, the device abnormal type is divided into a device determined abnormal type, and the device abnormal type is used to mark the preliminary collection data feature cluster, and the corresponding device abnormal type feature data is generated;

[0027] S3.5, if not, the minimum real probability of the device abnormal type is calculated based on the exact binomial method when the confidence is greater than the set second confidence threshold and the lower limit of the confidence interval of the preliminary collection data feature cluster for the device abnormal type is greater than the set probability threshold; A feature cluster second threshold is obtained. Wherein, the second confidence threshold is less than the first confidence threshold, for example, the first confidence threshold is 95%, and the second confidence threshold is 90%.

[0028] S3.6, determine whether there is a device abnormal type in the preliminary collection data feature cluster whose true proportion is greater than or equal to the feature cluster second threshold and less than the feature cluster first threshold, if yes, the device abnormal type is divided into a device suspected abnormal type, and the device abnormal type is used to mark the preliminary collection data feature cluster, and the corresponding device abnormal type feature data is generated;

[0029] S3.7, if not, the device abnormal type is divided into a device unknown abnormal type, and the device abnormal type is used to mark the preliminary collection data feature cluster, and the corresponding device abnormal type feature data is generated;

[0030] S3.8, standardize the device abnormal type feature data and the preliminary collection data to generate device abnormal type feature standard data and first preliminary collection standard data, and match the first preliminary collection standard data with the corresponding device abnormal type feature standard data based on the DBSCAN clustering algorithm to generate device abnormal type analysis data. The parameters of the DBSCAN clustering algorithm in this step are the same as those in step S3.2. The first preliminary collection standard data can be classified into the corresponding preliminary collection data feature cluster through this step, that is, the first preliminary collection standard data is matched with the device abnormal type feature standard data corresponding to the corresponding preliminary collection data feature cluster, and the matched device abnormal type feature standard data corresponding device abnormal type is output to generate device abnormal type analysis data.

[0031] Further, the S4 includes the following steps:

[0032] S4.1, collect the ID and location data of the data collection device to generate data collection device ID-location mapping data;

[0033] S4.2, obtain the data collection device ID corresponding to the abnormal device operation data, search the location data corresponding to the data collection device ID corresponding to the abnormal device operation data based on the data collection device ID-location mapping data, and generate device abnormal position analysis data. Further, the specific position of the abnormal position on the abnormal device can be further determined according to acoustic imaging technology, vibration transmission path analysis technology, infrared thermal imaging positioning technology, etc.

[0034] Further, if the device abnormal type analysis data is a device determined abnormal type or a device presumed abnormal type, the worker can independently determine the timing of detecting, reviewing and maintaining the abnormal device according to the obtained device abnormal type analysis data and device abnormal position analysis data.

[0035] Further, the S5 includes the following steps:

[0036] S5.1, if the device abnormal type analysis data is a device unknown abnormal type, standardize the preliminary collection data and device normal working state data corresponding to the device unknown abnormal type to generate second preliminary collection standard data and device normal working state standard data;

[0037] S5.2, calculate the second cosine similarity of the second preliminary collection standard data and the device normal working state standard data based on the cosine similarity algorithm;

[0038] S5.3, searching out the priority level corresponding to the second cosine similarity based on the second cosine similarity and the priority level-second cosine similarity interval mapping data, generating device detection priority level analysis data, wherein search methods such as brute force search, hash table search, regular expression search, etc. can be used; wherein, when setting the priority level-second cosine similarity interval mapping data, the higher the second cosine similarity, the lower the priority level.

[0039] Further, the method further comprises:

[0040] S7.1, collecting and combining the preliminary collection data and the deep inspection result data to generate device abnormal type feature update data;

[0041] S7.2, setting an update subset in the device abnormal type feature data, and when the device abnormal type feature update data reaches the set update quantity threshold, updating the device abnormal type feature update data to the update subset of the device abnormal type feature data, and returning to S3, that is, there is a subset in the device abnormal type feature data with a sample quantity of the set update quantity threshold, which is used to update the device abnormal type feature update data.

[0042] Further, the S3.7 comprises the following steps:

[0043] If not, the latest set number of corresponding deep inspection result data corresponding to the preliminary collection data feature cluster is obtained, and the number of corresponding device states in the corresponding deep inspection result data is counted;

[0044] The confidence degree of the preliminary collection data feature cluster corresponding to the device state normal is calculated, and feature cluster corresponding device normal confidence data is generated;

[0045] Determine whether the feature cluster corresponding device normal confidence data is less than the second confidence threshold, if yes, the device abnormal type is divided into device unknown abnormal, and the preliminary collection data feature cluster is marked with the device abnormal type to generate corresponding device abnormal type feature data;

[0046] If not, based on the preliminary collection data feature cluster corresponding device normal confidence data, the first confidence threshold and the second confidence threshold, the relationship strength of the preliminary collection data feature cluster corresponding device normal is determined, specifically: determine whether the device normal confidence data is less than the first confidence threshold, if yes, mark the preliminary collection data feature cluster as "device bias normal", if not, mark the preliminary collection data feature cluster as "device normal".

[0047] A device information collection device for device safety inspection, comprising a data collection device, a storage, and a processor;

[0048] The data collection device is used for collecting device operation data, environment data, and artificial preliminary inspection data and storing them into the storage;

[0049] The storage is used for storing the computer program;

[0050] The processor is used for executing the computer program to realize a device information collection method for device safety inspection.

[0051] 1. Compared with the prior art, the device information collection device and collection method for device safety inspection provided by the application analyze whether the device is abnormal and the abnormal type of the device by collecting device operation data, environment data, and artificial preliminary inspection data, and give the possible abnormal type when the abnormal type is uncertain, so as to facilitate the staff to determine the time of detecting and maintaining the device according to the abnormal type and the possible abnormal type of the device.

[0052] 2. Compared with the prior art, the device information collection device and collection method for device safety inspection provided by the application instruct the inspection personnel to perform deep detection on the device with unknown abnormality to determine the abnormal type or the normality of the device (detection error) and update the reference sample of the analysis when the device is abnormal but the specific abnormal type cannot be analyzed, so as to expand the sample library, reduce the probability of detecting unknown abnormality and the judgment error of whether the device is normal, and not waste too much manpower, so that the automatic data collection and analysis and the artificial inspection reach a good balance. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art according to these drawings.

[0054] Figure 1 The overall method step diagram provided for the embodiments of the present application;

[0055] Figure 2 The S3 specific step diagram provided for the embodiments of the present application;

[0056] Figure 3 The system structure block diagram provided for the embodiments of the present application. DETAILED DESCRIPTION

[0057] In order to make those skilled in the art better understand the technical solutions of the present application, the present application will be further described in detail with reference to the drawings.

[0058] In the description of the present application, the terms "first", "second", "third", etc. are used only for descriptive purpose and are not to be construed as indicating or implying relative importance or an indicated order of limiting the number of features indicated. Therefore, features defined with "first", "second", "third" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "plurality" is two or more, unless otherwise explicitly specified. In addition, the terms "mounting", "connecting", "connection" should be interpreted broadly, for example, it can be fixed connection, or detachable connection, or integral connection; it can be mechanical connection, or electrical connection; it can be direct connection, or indirect connection through intermediate medium, or internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0059] In the following, example embodiments will be described more fully with reference to the accompanying drawings, in which example embodiments can be embodied in different forms and should not be construed as being limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0060] In the case of no conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other.

[0061] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0062] The terms used herein are only used to describe specific embodiments, and are not intended to limit the present disclosure. As used herein, the singular forms "a" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms "comprise" and / or "consist of" are used in the specification, the specified features, integers, steps, operations, elements, and / or components are present, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0063] The embodiments described herein can be described with reference to plan views and / or cross-sectional views by virtue of the idealized schematic illustrations of the present disclosure. Thus, the example illustrations can be modified according to manufacturing techniques and / or tolerances. Therefore, the embodiments are not limited to the examples illustrated in the drawings, but include modifications based on manufacturing processes. Therefore, the regions illustrated in the drawings have a schematic property, and the shape of the regions shown in the drawings illustrates the specific shape of the regions of the elements, but is not intended to be limiting.

[0064] Please refer to Figures 1-2 A device information acquisition method for device safety inspection, comprising the following steps:

[0065] S1, collect equipment running data, environment data, and artificial preliminary inspection data to generate preliminary collection data, including the following steps:

[0066] S1.1, based on the data collection equipment, various sensors, detection instruments (such as ultrasonic flaw detector) and other equipment can collect equipment running data and environment data;

[0067] The equipment running data can be: equipment temperature, pressure, vibration, current, load, equipment abnormal sound spectrum, equipment image, cooling liquid pH value, transformer oil chromatographic data, flow, speed, valve opening, etc. The environment data can be the temperature, humidity, dust concentration, and flammable gas concentration of the space where the equipment is located.

[0068] S1.2, record and save the artificial preliminary inspection data; wherein the artificial preliminary inspection data can include the sensory description of the equipment obtained during artificial inspection of the equipment (such as normal, surface cracks, xx connection gap, loose bolts, xx color abnormal xx color, etc.), artificial detection evidence (such as corresponding photos, videos, and audio recordings).

[0069] S1.3, collect and combine the sensor collected equipment running data, environment data, and artificial preliminary inspection data to generate preliminary collection data.

[0070] S2, based on the preliminary collection data and the set equipment normal working state data, analyze and process whether the equipment working state is normal, including the following steps:

[0071] S2.1, collect the working state data of the equipment when it is working normally through the related equipment historical working log to obtain the equipment normal working state data;

[0072] S2.2, standardize the preliminary collection data and the equipment normal working state data to generate the first preliminary collection standard data and the equipment normal working state standard data;

[0073] S2.3, based on the cosine similarity algorithm, calculate the first cosine similarity of the first preliminary collection standard data and the equipment normal working state standard data, and determine whether the first cosine similarity is greater than the set first cosine similarity threshold;

[0074] S2.4, if yes, the equipment working state is normal, if not, the equipment working state is not normal.

[0075] S3, if the equipment working state is not normal, based on the preliminary collection data and the set equipment abnormal type characteristic data, analyze and process the equipment abnormal type to generate equipment abnormal type analysis data, the equipment abnormal type analysis data includes multiple equipment determined abnormal types, multiple equipment inferred abnormal types, and equipment unknown abnormalities, including the following steps:

[0076] S3.1, collect historical preliminary acquisition data corresponding to the abnormal working state of the acquisition device, and mark the corresponding abnormal type, collect preliminary acquisition data corresponding to each abnormal type mark respectively, generate a device abnormal type-preliminary acquisition data set, and the abnormal type mark includes a plurality of device determined abnormal types, a plurality of device suspected abnormal types, and a device unknown abnormality. If the current preliminary acquisition data matches the historical preliminary acquisition data corresponding to the device determined abnormal type mark, the abnormal type of the device is more likely (e.g., confidence greater than or equal to 95%) to correspond to the corresponding device determined abnormal type mark. If the current preliminary acquisition data matches the historical preliminary acquisition data corresponding to the device suspected abnormal type mark, the abnormal type of the device has a certain possibility (e.g., confidence greater than or equal to 95%) to correspond to the corresponding device suspected abnormal type mark. If the current preliminary acquisition data does not match the historical preliminary acquisition data corresponding to any device determined abnormal type and device suspected abnormal type mark, the device abnormal type is unknown, that is, it matches the device unknown abnormality mark. For example, the device abnormal type can be: the motor bearing position exists high-frequency impact and envelope spectrum characteristic frequency, indicating bearing loss; the rotor system position exists 1 times frequency amplitude prominence and stable phase, indicating rotor imbalance, etc.

[0077] S3.2, based on the DBSCAN clustering algorithm, the historical preliminary acquisition data is analyzed by clustering to generate a plurality of preliminary acquisition data feature clusters, and the number of device abnormal types corresponding to the historical preliminary acquisition data in each preliminary acquisition data feature cluster is counted.

[0078] S3.3, for each preliminary acquisition data feature cluster, the minimum real probability of the device abnormal type is calculated based on the exact binomial method when the confidence is greater than a set first confidence threshold and the lower limit of the confidence interval of the preliminary acquisition data feature cluster is greater than a set probability threshold, to obtain a feature cluster first threshold. In one embodiment, the number of first preliminary acquisition standard data in the preliminary acquisition data feature cluster is m, the target device abnormal type corresponds to k marks, the first confidence threshold is 95%, the probability threshold is 90%, and the lower limit of the confidence interval is L. It is required that L is greater than or equal to 90%. Based on the exact binomial method, the 1-95% quantile of the Beta distribution is constructed. It is calculated that when k is 95, L is about 0.891, which is less than 90%, that is, unqualified, and when k is 96, L is about 0.916, which is greater than 90%, that is, qualified. Therefore, the feature cluster first threshold of this embodiment is 96. That is, when there are 100 first preliminary acquisition standard data in the preliminary acquisition data feature cluster, and at least 96 first preliminary acquisition standard data correspond to the target device abnormal type, it is believed that the probability of the preliminary acquisition data feature cluster corresponding to the target device abnormal type is greater than 90% with 95% confidence.

[0079] Further, in the judgment, the effect amount corresponding to the equipment abnormal type can also be calculated to assist in judging the relationship strength between the equipment abnormal type and the corresponding preliminary collection data feature cluster.

[0080] S3.4, judging whether there is an equipment abnormal type with a real proportion greater than or equal to a feature cluster first threshold value in the preliminary collection data feature cluster, if yes, dividing the equipment abnormal type into an equipment determined abnormal type, and using the equipment abnormal type to mark the preliminary collection data feature cluster to generate corresponding equipment abnormal type feature data;

[0081] S3.5, if no, calculating a confidence degree greater than a set second confidence threshold value based on the exact binomial method, and the minimum real probability of the equipment abnormal type when the lower limit of the confidence interval of the preliminary collection data feature cluster for the equipment abnormal type is greater than a set probability threshold value, to obtain a feature cluster second threshold value; wherein the second confidence threshold value is less than the first confidence threshold value, for example, the first confidence threshold value is 95%, and the second confidence threshold value is 90%.

[0082] S3.6, judging whether there is an equipment abnormal type with a real proportion greater than or equal to the feature cluster second threshold value and less than the feature cluster first threshold value in the preliminary collection data feature cluster, if yes, dividing the equipment abnormal type into an equipment presumed abnormal type, and using the equipment abnormal type to mark the preliminary collection data feature cluster to generate corresponding equipment abnormal type feature data;

[0083] S3.7, if no, dividing the equipment abnormal type into an equipment unknown abnormal type, and using the equipment abnormal type to mark the preliminary collection data feature cluster to generate corresponding equipment abnormal type feature data;

[0084] S3.8, performing standardization processing on the equipment abnormal type feature data and the preliminary collection data to generate equipment abnormal type feature standard data and first preliminary collection standard data, and matching the first preliminary collection standard data with corresponding equipment abnormal type feature standard data based on a DBSCAN clustering algorithm to generate equipment abnormal type analysis data. The parameters of the DBSCAN clustering algorithm in this step are the same as the parameters of the DBSCAN clustering algorithm in step S3.2, the first preliminary collection standard data can be classified into the corresponding preliminary collection data feature cluster in this step, that is, the first preliminary collection standard data is matched with the equipment abnormal type feature standard data corresponding to the corresponding preliminary collection data feature cluster, and the equipment abnormal type corresponding to the matched equipment abnormal type feature standard data is output to generate equipment abnormal type analysis data.

[0085] S4, based on the abnormal equipment operation data, determining the position of the equipment corresponding to the abnormal equipment operation data to generate equipment abnormal position analysis data, including the following steps:

[0086] S4.1, collect ID and location data of the data collection device, and generate data collection device ID-location mapping data;

[0087] S4.2, obtain the data collection device ID corresponding to the abnormal device operation data, search the location data corresponding to the data collection device ID corresponding to the abnormal device operation data based on the data collection device ID-location mapping data, and generate device abnormal position analysis data. Further, the specific position of the abnormal position in the abnormal device can be further determined according to acoustic imaging technology, vibration transmission path analysis technology, infrared thermal imaging positioning technology, etc.

[0088] Further, if the device abnormal type analysis data is a device determined abnormal type or a device inferred abnormal type, the worker can independently determine the timing of detecting and reviewing and maintaining the abnormal device according to the obtained device abnormal type analysis data and device abnormal position analysis data.

[0089] S5, if the device abnormal type analysis data is a device unknown abnormality, based on the preliminary collection data corresponding to the device unknown abnormality and the device normal working state data, the similarity analysis of the device operation data and the device normal working state data is performed, and the priority level of the device detection is determined based on the size of the similarity, and device detection priority level analysis data is generated, including the following steps:

[0090] S5.1, if the device abnormal type analysis data is a device unknown abnormality, the preliminary collection data corresponding to the device unknown abnormality and the device normal working state data are standardized to generate second preliminary collection standard data and device normal working state standard data;

[0091] S5.2, based on the cosine similarity algorithm, the second cosine similarity of the second preliminary collection standard data and the device normal working state standard data is calculated;

[0092] S5.3, based on the second cosine similarity and the set priority level-second cosine similarity interval mapping data, the priority level corresponding to the second cosine similarity is searched out, and the device detection priority level analysis data is generated, wherein search methods such as brute force search, hash table search, regular expression search, etc. can be used; wherein when setting the priority level-second cosine similarity interval mapping data, the higher the second cosine similarity, the lower the priority level.

[0093] S6, based on the device unknown abnormality corresponding device detection priority level analysis data and device abnormal position analysis data, instruct the inspection personnel in the specified time to the device abnormal position depth detection, generate depth inspection data and depth inspection result data; For example, there are priority level 1, priority level 2 two priority levels, priority level 1 priority is greater than priority level 2, priority level 1 represents the inspection personnel immediately to the device abnormal position depth detection, priority level 2 represents the inspection personnel in the next inspection to the device abnormal position depth detection.

[0094] Among them, the depth detection includes using professional detection tool to detect the equipment inside and outside or abnormal position, the data obtained in the detection process is the corresponding depth inspection data, such as ultrasonic flaw detection data, image data, high definition recording data, etc., the result of analyzing the depth inspection data is the depth inspection result data, such as the existence of cracks in the equipment shell, bolt loosening, bearing wear excessive, etc.

[0095] S7, the device abnormal type feature data is updated, including the following steps:

[0096] S7.1, the preliminary collection data and the depth inspection result data are collected and combined, and the device abnormal type feature update data is generated;

[0097] S7.2, an update subset is set in the device abnormal type feature data, and when the device abnormal type feature update data reaches the set update quantity threshold, the device abnormal type feature update data is updated to the update subset of the device abnormal type feature data, and returns to S3, that is, there is a subset in the device abnormal type feature data, the number of samples of which is the set update quantity threshold, which is used to update the device abnormal type feature update data.

[0098] One embodiment of S3.7 can also be the following steps:

[0099] If not, the latest set number of corresponding depth inspection result data corresponding to the preliminary collection data feature cluster is obtained from S6 step, and the number of corresponding device state normal in the corresponding depth inspection result data is counted;

[0100] The confidence degree of the preliminary collection data feature cluster corresponding to the device state normal is calculated, and the feature cluster corresponding device normal confidence degree data is generated;

[0101] Determine whether the feature cluster corresponding device normal confidence degree data is less than the second confidence threshold, if yes, the device abnormal type is divided into device unknown abnormality, and the device abnormal type is marked in the preliminary collection data feature cluster, and the corresponding device abnormal type feature data is generated;

[0102] If no, based on the preliminary acquisition data feature cluster corresponding to the equipment normal confidence data, the first confidence threshold and the second confidence threshold, the relationship strength of the preliminary acquisition data feature cluster corresponding to the equipment normal is determined, specifically: whether the equipment normal confidence data is less than the first confidence threshold is judged, if yes, the preliminary acquisition data feature cluster is marked as "equipment bias normal", if no, the preliminary acquisition data feature cluster is marked as "equipment normal".

[0103] Please refer to Figure 3 The application further provides a device information acquisition device for device safety inspection, comprising a data acquisition device, a storage, and a processor.

[0104] The data acquisition device is used for acquiring device operation data, environmental data, and artificial preliminary inspection data and storing them into the storage.

[0105] The storage is used for storing a computer program.

[0106] The processor is used for executing the computer program to realize the device information acquisition method for device safety inspection provided by the application.

[0107] The above only describes some exemplary embodiments of the application by way of illustration, and it is needless to say that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the application. Therefore, the above drawings and descriptions are illustrative in nature and should not be understood as limiting the scope of protection of the claims of the application.

Claims

1. A method for collecting equipment information for equipment safety inspection, characterized in that: Includes the following steps: S1. Collect equipment operation data, environmental data, and preliminary manual inspection data to generate preliminary collected data; S2. Based on the preliminary collected data and the set normal operating status data of the equipment, perform analysis and processing to determine whether the equipment is operating normally. S3. If the equipment is not working properly, based on the preliminary collected data and the set equipment anomaly type feature data, perform equipment anomaly type analysis processing to generate equipment anomaly type analysis data. The equipment anomaly type analysis data includes multiple confirmed equipment anomaly types, multiple inferred equipment anomaly types, and unknown equipment anomalies. S4. Based on abnormal equipment operation data, determine the location of the equipment corresponding to the abnormal equipment operation data, and generate abnormal equipment location analysis data; S5. If the equipment anomaly type analysis data is an unknown equipment anomaly, then based on the preliminary collected data and normal equipment status data corresponding to the unknown equipment anomaly, perform a similarity analysis between the equipment operation data and the normal equipment status data, and determine the priority level of equipment detection based on the magnitude of the similarity, and generate equipment detection priority level analysis data. S6. Based on the equipment detection priority analysis data and equipment anomaly location analysis data corresponding to the unknown equipment anomaly, instruct the inspection personnel to conduct in-depth inspection of the equipment anomaly location within a specified time, and generate in-depth inspection data and in-depth inspection result data. S3 includes the following steps: S3.1 Collect historical preliminary data when the equipment is not in normal working condition, and mark the corresponding abnormality type. Collect the preliminary data corresponding to each abnormality type mark separately to generate a set of equipment abnormality type-preliminary data. The abnormality type mark includes multiple determined abnormality types of equipment, multiple inferred abnormality types of equipment, and unknown abnormality of equipment. S3.

2. Based on the DBSCAN clustering algorithm, perform cluster analysis on the historical preliminary data collection to generate multiple preliminary data collection feature clusters, and count the number of device anomaly types corresponding to the historical preliminary data collection in each preliminary data collection feature cluster. S3.3 For each preliminary data feature cluster, calculate the minimum true probability of the device anomaly type when the confidence level of the preliminary data feature cluster is greater than the set first confidence threshold, based on the exact binomial method. S3.4 Determine whether there are any device anomaly types in the preliminary data feature clusters where the actual proportion is greater than or equal to the first threshold of the feature cluster. If so, classify the device anomaly type into a device-determined anomaly type, and use the device anomaly type to mark the preliminary data feature cluster, generating corresponding device anomaly type feature data. S3.5 If not, then based on the exact binomial method, calculate the confidence level when it is greater than the set second confidence level threshold, and initially collect the data feature clusters. The lower limit of the confidence interval for the device abnormality type is greater than the set probability threshold, and obtain the minimum true probability of the device abnormality type. S3.6 Determine whether there are any device anomaly types in the preliminary data feature clusters whose actual proportion is greater than or equal to the second threshold of the feature cluster and less than the first threshold of the feature cluster. If so, classify the device anomaly type as a device speculation anomaly type, and use the device anomaly type to mark the preliminary data feature cluster, and generate corresponding device anomaly type feature data. S3.7 If not, the device anomaly type is classified as an unknown device anomaly, and the device anomaly type is used to mark the feature cluster of the initially collected data to generate corresponding device anomaly type feature data. S3.

8. Standardize the equipment anomaly type feature data and the preliminary collected data to generate equipment anomaly type feature standard data and first preliminary collected standard data. Based on the DBSCAN clustering algorithm, match the first preliminary collected standard data with the corresponding equipment anomaly type feature standard data to generate equipment anomaly type analysis data.

2. The equipment information collection method for equipment safety inspection according to claim 1, characterized in that: S1 includes the following steps: S1.1 Collect equipment operation data and environmental data; S1.2 Record and save the preliminary manual inspection data; S1.3 Collect and combine the operating data of the acquisition equipment, environmental data, and preliminary manual inspection data to generate preliminary acquisition data.

3. The equipment information collection method for equipment safety inspection according to claim 1, characterized in that: S2 includes the following steps: S2.1 Collect historical data on the working status of the equipment during normal operation to obtain normal operating status data of the equipment; S2.

2. Standardize the preliminary collected data and the normal operating status data of the equipment to generate the first preliminary collected standard data and the normal operating status standard data of the equipment; S2.

3. Based on the cosine similarity algorithm, calculate the first cosine similarity between the first preliminary collected standard data and the standard data of normal working status of the equipment, and determine whether the first cosine similarity is greater than the set first cosine similarity threshold. S2.4 If yes, the equipment is working normally; otherwise, the equipment is not working normally.

4. The equipment information collection method for equipment safety inspection according to claim 1, characterized in that: S4 includes the following steps: S4.1 Collect the ID and location data of the data acquisition device, and generate data acquisition device ID-location mapping data; S4.2 Obtain the data acquisition device ID corresponding to the abnormal equipment operation data. Based on the data acquisition device ID-location mapping data, search for the location data corresponding to the data acquisition device ID corresponding to the abnormal equipment operation data, and generate abnormal equipment location analysis data.

5. The equipment information collection method for equipment safety inspection according to claim 1, characterized in that: S5 includes the following steps: S5.1 If the equipment anomaly type analysis data is an unknown equipment anomaly, then the preliminary collected data and normal operating status data corresponding to the unknown equipment anomaly are standardized to generate second preliminary collected standard data and normal operating status standard data. S5.

2. Based on the cosine similarity algorithm, calculate the second cosine similarity between the second preliminary collected standard data and the standard data of normal working status of the equipment; S5.

3. Based on the second cosine similarity and the priority level-second cosine similarity interval mapping data, search for the priority level corresponding to the second cosine similarity and generate device detection priority level analysis data.

6. The equipment information collection method for equipment safety inspection according to claim 1, characterized in that: The method further includes: S7.1 Collect and combine the preliminary data and in-depth inspection results to generate equipment anomaly type feature update data; S7.

2. Set an update subset in the device anomaly type feature data. When the device anomaly type feature update data reaches the set update quantity threshold, update the device anomaly type feature update data to the update subset of the device anomaly type feature data, and return to S3.

7. The equipment information collection method for equipment safety inspection according to claim 1, characterized in that: S3.7 Includes the following steps: If not, then obtain the latest set number of corresponding depth inspection result data corresponding to the feature cluster of the preliminary data collection, and count the number of devices with normal status in the corresponding depth inspection result data; Calculate the confidence level of the device when the device status is normal according to the feature clusters of the preliminary data collection, and generate the confidence level data of the device status when the feature clusters are normal. Determine whether the normal confidence data of the device corresponding to the feature cluster is less than the second confidence threshold. If so, classify the device anomaly type as an unknown device anomaly, and use the device anomaly type to mark the feature cluster of the initially collected data to generate the corresponding device anomaly type feature data. If not, the relationship strength between the device normality corresponding to the feature clusters of the preliminary data is determined based on the device normality confidence data corresponding to the feature clusters of the preliminary data, the first confidence threshold, and the second confidence threshold.

8. An equipment information collection device for equipment safety inspection, used to execute the equipment information collection method for equipment safety inspection as described in any one of claims 1-7, characterized in that: Includes data acquisition devices, storage devices, and processors; The data acquisition device is used to collect equipment operation data, environmental data, and preliminary manual inspection data and store them in the storage device; The storage device is used to store computer programs; The processor is used to execute a computer program to implement a method for collecting equipment information for equipment safety inspection as described in any one of claims 1-7.

Images