A positioning method of abnormal industrial equipment in a multi-source heterogeneous scene

Abstract

The present disclosure relates to a positioning method of an abnormal industrial equipment in a multi-source heterogeneous scene, and relates to the technical field of big data processing. The method comprises the following steps: collecting data sources of industrial equipment in a multi-source heterogeneous scene, the industrial equipment being a plurality of devices with different device types, and the industrial equipment with different device types being accessed to an edge computing terminal through different protocol types; pre-processing the data sources to obtain data items carrying unified semantic identifiers, and correcting clock bias and time alignment of the data items to obtain multi-source data sequences; performing real-time state discrimination on the multi-source data sequences to generate alarm events, and establishing a correlation relationship between the industrial equipment and the device monitoring points associated with the industrial equipment based on the unified semantic identifiers; and positioning the devices in an abnormal operating state in the industrial equipment according to the correlation relationship, the alarm events and the multi-source data sequences to obtain an abnormal device positioning result. The present disclosure improves the accuracy of the abnormal device positioning result.

Description

Technical Field

[0001] This disclosure relates to the field of big data processing technology, and more specifically, to a method for locating abnormal industrial equipment in a multi-source heterogeneous scenario. Background Technology

[0002] In the field of industrial equipment monitoring, common solutions may include, but are not limited to: process monitoring systems based on SCADA (Supervisory Control And Data Acquisition) systems or DCS (Distributed Control System); status acquisition and threshold alarm systems based on PLC (Programmable Logic Controller) or sensors; status monitoring and analysis systems for vibration, temperature, current, acoustics, etc., for predictive maintenance; and operation and maintenance management systems that achieve multi-plant data aggregation, dashboard display, alarm push, and operation and maintenance management through industrial gateways + cloud platforms or edge computing. At the same time, in some scenarios, the corresponding systems also introduce multimodal data such as video and thermal imaging to enhance the ability to identify safety, quality, and operational anomalies.

[0003] However, the above solutions have the following drawbacks: on the one hand, in distributed and / or heterogeneous scenarios, it is often difficult to simultaneously meet the requirements of high real-time performance and high reliability; on the other hand, the high heterogeneity of devices and protocols leads to high costs for device access adaptation and expansion maintenance, as well as problems with inconsistent data semantics, which further makes it difficult to align and merge multi-source data in time, and easily results in false alarms / missed alarms and difficulty in anomaly location.

[0004] It should be noted that the information in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this disclosure is to provide a method for locating abnormal industrial equipment in multi-source heterogeneous scenarios, thereby overcoming, to at least a certain extent, the problems of difficulty in simultaneously satisfying high real-time performance and high reliability, as well as the easy occurrence of false alarms / missed alarms and difficulty in abnormal location caused by the limitations and defects of related technologies.

[0006] According to one aspect of this disclosure, a method for locating abnormal industrial equipment in a multi-source heterogeneous scenario is provided, configured in an edge computing terminal. The edge computing terminal includes multiple protocol adapters corresponding to transmission control protocols with different protocol types. The method includes: collecting data sources from industrial equipment in a multi-source heterogeneous scenario, where the industrial equipment consists of multiple devices with different device types, and the different device types of industrial equipment access the edge computing terminal through different protocol types; preprocessing the data sources to obtain data items carrying unified semantic identifiers, and performing clock deviation correction and time alignment on the data items to obtain a multi-source data sequence; performing real-time status discrimination on the multi-source data sequence to generate alarm events, and establishing an association relationship between the industrial equipment and the associated device monitoring points based on the unified semantic identifiers; and locating the devices in the industrial equipment that are in an abnormal operating state according to the association relationship, the alarm events, and the multi-source data sequence to obtain an abnormal equipment location result.

[0007] In one exemplary embodiment of this disclosure, collecting data sources from industrial equipment in a multi-source heterogeneous scenario includes: determining, based on the equipment type of the industrial equipment, the transmission control protocol required to establish a communication connection between the edge computing terminal and the industrial equipment, and determining, based on the protocol type of the transmission control protocol, the protocol adapter required to establish the communication connection; establishing a communication connection between the edge computing terminal and the industrial equipment based on the transmission control protocol through the protocol adapter; and receiving data sources sent by the industrial equipment based on the communication connection; wherein the data sources include a communication register definition table corresponding to the industrial equipment, a device identifier, and current operating status information.

[0008] In one exemplary embodiment of this disclosure, preprocessing the data source to obtain data items carrying unified semantic identifiers includes: adapting and parsing the current operating status information according to a communication register definition table to obtain a standard data source; the standard data source includes multiple elements such as the equipment identifier of the industrial equipment, the register address corresponding to the equipment monitoring point associated with the industrial equipment, the meaning of the physical quantity, the data type, the conversion rule, and the original data value; obtaining a measurement point dictionary table corresponding to the industrial equipment, and semantically mapping the standard data source based on the measurement point dictionary table to obtain data items carrying unified semantic identifiers; the data items include multiple elements such as the equipment identifier, the measurement point identifier of the equipment monitoring point, the dimensional identifier corresponding to the equipment monitoring point, the data acquisition completion timestamp of the data source, and the data writing timestamp.

[0009] In one exemplary embodiment of this disclosure, semantic mapping of a standard data source based on the measurement point dictionary table to obtain a data item carrying a unified semantic identifier includes: converting the original data value based on the conversion rule to obtain a target data value, and determining the measurement point identifier of the device monitoring point based on the measurement point dictionary table and the register address; determining the data type corresponding to the device monitoring point and the dimension identifier corresponding to the data type based on the measurement point dictionary table, and determining the allowed value range of the device monitoring point; verifying the target data value based on the allowed value range, and generating the data item according to the device identifier, measurement point identifier, and dimension identifier in response to the target data value passing the verification.

[0010] In one exemplary embodiment of this disclosure, clock skew correction and time alignment are performed on the data items to obtain a multi-source data sequence, including: determining a reference clock based on a preset time server, and determining a current clock skew parameter based on the reference clock and the local clock in the edge computing terminal; correcting the data acquisition timestamp and / or data writing timestamp based on the current clock skew parameter to obtain a time correction result, and aligning the data items to event time based on the time correction result to obtain a multi-source data sequence.

[0011] In an exemplary embodiment of this disclosure, the process of aligning data items to event time to obtain a multi-source data sequence based on time correction results includes: using a first preset time period as the unit for dividing event time alignment windows, and using a preset time point within the first preset time period as the starting time point for dividing the event time alignment windows; dividing the windows according to the starting time point and a preset window length to obtain multiple event time alignment windows, and matching window numbers to the event time alignment windows according to the starting time point; aligning the data items to each of the event time alignment windows according to the window numbers and time correction results, and generating a multi-source data sequence based on the data items contained in each of the event time alignment windows at intervals of a second preset time period.

[0012] In an exemplary embodiment of this disclosure, generating a multi-source data sequence based on data items in each event time alignment window at intervals of a second preset time period includes: determining a target number of data items that each event time alignment window can match based on the window number and time correction result; determining a first current number of existing data items in each event time alignment window at intervals of the second preset time period; at this moment, the existing data items in each event time alignment window are in a normal state; in response to any event time alignment window having a first current number less than the target number, controlling the event time alignment window to enter a late waiting stage; determining a second current number of existing data items other than data in a normal state in each event time alignment window at intervals of a third preset time period; at this moment, the existing data items other than data in a normal state in each event time alignment window are in a late state; in response to any event time alignment window having a sum of a first current number and a second current number less than the target number, determining data items in a missing state, and generating the multi-source data sequence based on data items in a normal state, data items in a late state, and data items in a missing state.

[0013] In an exemplary embodiment of this disclosure, real-time status discrimination of a multi-source data sequence to generate an alarm event includes: reading the first window measurement value of a device monitoring point included in three or more consecutive window numbers according to the window number of the event time alignment window in the multi-source data sequence; determining a threshold judgment condition corresponding to the device monitoring point, and determining whether the first window measurement value in each event time alignment window meets the threshold judgment condition; in response to the first window measurement value of any window number not meeting the threshold judgment condition, marking the event time alignment window that does not meet the threshold judgment condition as an over-limit window; in response to the number of over-limit windows being greater than a preset threshold, generating the alarm event according to the device monitoring point, the device identifier of the industrial equipment associated with the device monitoring point, the measurement point identifier, the dimensional identifier, the alarm type, the threshold judgment condition, the window number and the corresponding first window measurement value, and the alarm generation time.

[0014] In one exemplary embodiment of this disclosure, the method further includes: reading the device monitoring points included in the window number of the event time alignment window in the multi-source data sequence, and the second window measurement value within the event time alignment window corresponding to the window number; generating a measurement value sequence based on the second window measurement value, and generating a feature summary of the event time alignment window corresponding to the window number based on the measurement value sequence; and / or uploading the feature summary of the event time alignment window and the multi-source data sequence to a cloud server, so that the cloud server performs normalized aggregation on the multi-source data sequence and feature summary based on the unified semantic identifier, and performs cross-scenario fusion analysis on the normalized aggregated data to obtain a global monitoring result.

[0015] In one exemplary embodiment of this disclosure, generating a feature summary of an event time-aligned window corresponding to the window number based on the measurement value sequence includes: determining the mean, standard deviation, peak value, and rate of change of the event time-aligned window corresponding to the window number based on the measurement value sequence; determining a mean-reduced sequence based on the measurement value sequence and the mean, and performing a discrete Fourier transform on the mean-reduced sequence to obtain the frequency domain energy of the event time-aligned window corresponding to the window number; and generating a feature summary of the event time-aligned window corresponding to the window number based on the mean, standard deviation, peak value, rate of change, and frequency domain energy.

[0016] This disclosure provides a method for locating abnormal industrial equipment in a multi-source heterogeneous scenario. Firstly, it collects data from industrial equipment in a multi-source heterogeneous scenario, preprocesses the data to obtain data items carrying unified semantic identifiers, and performs clock deviation correction and time alignment on the data items to obtain a multi-source data sequence. Then, it performs real-time status discrimination on the multi-source data sequence to generate alarm events, and establishes a correlation between industrial equipment and associated equipment monitoring points based on the unified semantic identifiers. Finally, based on the correlation, alarm events, and multi-source data sequence, it locates the equipment in an abnormal operating state within the industrial equipment, obtaining the abnormal equipment location result. This method achieves real-time abnormal location based on the data source of industrial equipment in a multi-source heterogeneous scenario, solving the problem of existing methods. In distributed and / or heterogeneous scenarios, it is often difficult to simultaneously meet the requirements of high real-time performance and high reliability, thus improving the accuracy of abnormal device location results. On the other hand, it also solves the problems of high access adaptation and expansion maintenance costs and inconsistent data semantics caused by the high heterogeneity of devices and protocols in existing technologies. Furthermore, it also solves the problems of difficulty in time alignment and fusion of multi-source data in existing technologies, which easily leads to false alarms / missed alarms and difficulty in anomaly location. Moreover, the method described in this disclosure realizes the consistent expression of multi-source industrial data in distributed heterogeneous scenarios on a unified time base and semantic dimension, improving the real-time performance and accuracy of industrial equipment operating status perception, and avoiding monitoring distortion caused by protocol differences, clock deviations, and data out-of-order issues.

[0017] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0019] Figure 1 The flowchart illustrates a method for locating abnormal industrial equipment in a multi-source heterogeneous scenario according to an exemplary embodiment of the present disclosure.

[0020] Figure 2 The diagram illustrates an example of a positioning system for abnormal industrial equipment in a multi-source heterogeneous scenario according to an exemplary embodiment of the present disclosure.

[0021] Figure 3The diagram illustrates a scenario example of a protocol adapter included in an edge computing terminal according to an exemplary embodiment of the present disclosure.

[0022] Figure 4 An example diagram of an industrial device according to an exemplary embodiment of the present disclosure is shown schematically.

[0023] Figure 5 This diagram illustrates a scenario example showing the correspondence between an industrial device and a protocol type according to an example embodiment of this disclosure.

[0024] Figure 6 The illustration shows a scenario example of an event time alignment window obtained according to an example embodiment of the present disclosure.

[0025] Figure 7 The illustration schematically depicts a scenario example of window encoding associated with an event time-aligned window according to an exemplary embodiment of the present disclosure.

[0026] Figure 8 The illustration shows a scenario example diagram illustrating an event timing alignment result obtained according to an example embodiment of the present disclosure.

[0027] Figure 9 The diagram illustrates an example of a positioning device for abnormal industrial equipment in a multi-source heterogeneous scenario according to an exemplary embodiment of the present disclosure.

[0028] Figure 10 The diagram schematically illustrates an example electronic device for implementing a method for locating abnormal industrial equipment in a multi-source heterogeneous scenario, according to an example embodiment of the present disclosure. Detailed Implementation

[0029] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this disclosure more comprehensive and complete, and to fully convey the concept of the example embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more of the specific details omitted, or other methods, components, apparatus, steps, etc., can be employed. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0030] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0031] This exemplary embodiment first provides a method for locating abnormal industrial equipment in a multi-source heterogeneous scenario. This method can run on an edge computing terminal. Of course, those skilled in the art can also run the method disclosed herein on other platforms as needed, and this exemplary embodiment does not impose any special limitations on this. Specifically, refer to... Figure 1 As shown, the method for locating abnormal industrial equipment in this multi-source heterogeneous scenario may include the following steps: Step S110. Collect data sources from industrial equipment in a multi-source heterogeneous scenario. The industrial equipment consists of multiple devices with different equipment types, and the industrial equipment of different equipment types accesses the edge computing terminal through different protocol types. Step S120. Preprocess the data source to obtain data items carrying unified semantic identifiers, and perform clock skew correction and time alignment on the data items to obtain a multi-source data sequence; Step S130. Perform real-time status discrimination on the multi-source data sequence to generate alarm events, and establish the association relationship between industrial equipment and the equipment monitoring points associated with the industrial equipment based on the unified semantic identifier; Step S140. Based on the association, the alarm event, and the multi-source data sequence, locate the equipment in the industrial equipment that is in an abnormal operating state, and obtain the abnormal equipment location result.

[0032] In the above-described method for locating abnormal industrial equipment in a multi-source heterogeneous scenario, on the one hand, data sources of industrial equipment in a multi-source heterogeneous scenario are collected, and the data sources are preprocessed to obtain data items carrying unified semantic identifiers. Clock deviation correction and time alignment are then performed on the data items to obtain a multi-source data sequence. Real-time status discrimination is then performed on the multi-source data sequence to generate alarm events, and the association between industrial equipment and its associated monitoring points is established based on the unified semantic identifiers. Finally, based on the association relationships, alarm events, and multi-source data sequences, the equipment in an abnormal operating state within the industrial equipment is located, yielding the abnormal equipment location result. This method achieves real-time abnormal location based on the data source of industrial equipment in a multi-source heterogeneous scenario, solving the problems of existing methods. In distributed and / or heterogeneous scenarios, it is often difficult to simultaneously meet the requirements of high real-time performance and high reliability, thus improving the accuracy of abnormal device location results. On the other hand, it also solves the problems of high access adaptation and expansion maintenance costs and inconsistent data semantics caused by the high heterogeneity of devices and protocols in existing technologies. Furthermore, it also solves the problems of difficulty in time alignment and fusion of multi-source data in existing technologies, which easily leads to false alarms / missed alarms and difficulty in anomaly location. Moreover, the method described in this disclosure realizes the consistent expression of multi-source industrial data in distributed heterogeneous scenarios on a unified time base and semantic dimension, improving the real-time performance and accuracy of industrial equipment operating status perception, and avoiding monitoring distortion caused by protocol differences, clock deviations, and data out-of-order issues.

[0033] The following will provide a detailed explanation and description of the method for locating abnormal industrial equipment in a multi-source heterogeneous scenario as described in the exemplary embodiments of this disclosure, with reference to the accompanying drawings.

[0034] First, the positioning system for abnormal industrial equipment in a multi-source heterogeneous scenario involved in the exemplary embodiments of this disclosure will be explained and described. Specifically, refer to... Figure 2 As shown, the localization system for abnormal industrial equipment in a multi-source heterogeneous scenario may include industrial equipment 210, an edge computing terminal 220, and a cloud server 230. The edge computing terminal can communicate with both the industrial equipment and the cloud server. In practical applications, the industrial equipment can feed back data sources to the edge computing terminal, which can be used to implement the localization method for abnormal industrial equipment in a multi-source heterogeneous scenario described in this example embodiment. The cloud server can be used to normalize and aggregate multi-source data sequences and feature summaries, and perform cross-scenario fusion analysis on the normalized and aggregated data to obtain global monitoring results.

[0035] In one example embodiment, the edge computing terminal described above includes multiple protocol adapters corresponding to transmission control protocols of different protocol types; for example, referring to... Figure 3 As shown, protocol adapter 1 corresponds to protocol type 1, protocol adapter 2 corresponds to protocol type 2, protocol adapter 3 corresponds to protocol type 3, and so on. In practical applications, the protocol types described here can be determined according to the access protocol type required between industrial equipment and edge computing terminals. For example, Modbus protocol type, industrial Ethernet protocol type, high-performance industrial Ethernet protocol type, general industrial Ethernet protocol type, high-level protocol type based on CAN (Controller Area Network) bus, and traditional fieldbus protocol type, etc. This example does not make any special limitations on this.

[0036] In one example embodiment, reference is made to... Figure 4 As shown, the industrial equipment involved in this exemplary embodiment includes multiple devices with different equipment types; furthermore, one protocol type can correspond to an industrial device of that protocol type, and the industrial devices of the same protocol type can include one or more, depending on the actual situation. This example does not impose any special restrictions in this regard. A scenario example diagram illustrating the correspondence between industrial equipment and protocol types can be found in [reference needed]. Figure 5 As shown.

[0037] In one example embodiment, after receiving multi-source data sequences and feature summaries, the cloud server can normalize and aggregate the multi-source data sequences and feature summaries based on a unified semantic identifier. On this basis, combined with the event time alignment result, cross-scenario fusion analysis is performed on the normalized and aggregated data to output global monitoring results. The feature summary described here is a set of features extracted from the multi-source data sequences within the event time alignment window corresponding to the alignment window number. The feature set may include, but is not limited to, mean, variance, peak value, rate of change, and frequency domain energy, etc. In practical applications, when communication conditions are below a preset bandwidth threshold, the feature summary can be written to the storage medium and reported to the cloud server first when reporting data. It should also be noted that the specific process required during normalization and aggregation is: unit conversion and dimensional consistency verification. Under this premise, the cloud server can form a cross-scenario alignment set based on the unified semantic identifier, matching data items from different industrial field scenarios with the same measurement point identifier within the alignment window of the event time alignment result. Finally, during the cross-scenario fusion analysis, a cross-scenario consistency score can be calculated based on the cross-scenario alignment set. At the same time, the cross-scenario consistency score can be determined based on the degree of matching between the data deviation and change trend of the equipment monitoring points corresponding to each scenario within the same alignment window. The global monitoring results include: equipment status indicators output according to industrial field scenarios and cross-scenario consistency scores.

[0038] The following will combine Figures 2-5 right Figure 1 The method for locating abnormal industrial equipment in a multi-source heterogeneous scenario is explained and illustrated in detail. Specifically: In step S110, data sources of industrial equipment in a multi-source heterogeneous scenario are collected. The industrial equipment consists of multiple devices with different equipment types, and the industrial equipment of different equipment types accesses the edge computing terminal through different protocol types.

[0039] In this embodiment of the disclosure, the specific data source acquisition process can be implemented as follows: Based on the device type of the industrial equipment, determine the transmission control protocol required to establish a communication connection between the edge computing terminal and the industrial equipment, and determine the protocol adapter required to establish the communication connection based on the protocol type of the transmission control protocol; establish a communication connection between the edge computing terminal and the industrial equipment based on the transmission control protocol using the protocol adapter; receive the data source sent by the industrial equipment based on the communication connection; the data source includes a communication register definition table corresponding to the industrial equipment, a device identifier, and current operating status information. Specifically, when determining the device type, it can be based on the information registered by the industrial equipment in the edge computing terminal; for example, the corresponding protocol adapter can be loaded based on the device type identifier. Furthermore, taking a circulating water pump controller as an example of industrial equipment, the circulating water pump controller establishes a communication connection with the edge computing terminal through the Modbus communication protocol based on the transmission control protocol. At the same time, the circulating water pump controller provides the "WP-3000 Communication Register Definition Table" with the equipment. Meanwhile, the circulating water pump controller provides operating data (i.e., the current operating status information in the data source) to the outside world through the Modbus communication protocol based on the transmission control protocol. All process quantities are stored in the holding register area, corresponding to Modbus function code 03. The physical meaning, data type and conversion coefficient of each register are explained in detail in the communication register definition table mentioned above.

[0040] In step S120, the data source is preprocessed to obtain data items carrying unified semantic identifiers, and the data items are clock offset correction and time alignment are performed to obtain a multi-source data sequence.

[0041] In this embodiment of the disclosure, firstly, the data source is preprocessed to obtain data items carrying unified semantic identifiers. Specifically, this can be achieved as follows: the current operating status information is adapted and parsed according to the communication register definition table to obtain a standard data source; the standard data source includes multiple elements such as the equipment identifier of the industrial equipment, the register address corresponding to the equipment monitoring point associated with the industrial equipment, the meaning of the physical quantity, the data type, the conversion rule, and the original data value; a measurement point dictionary table corresponding to the industrial equipment is obtained, and semantic mapping is performed on the standard data source based on the measurement point dictionary table to obtain data items carrying unified semantic identifiers; the data items include multiple elements such as the equipment identifier, the measurement point identifier of the equipment monitoring point, the dimensional identifier corresponding to the equipment monitoring point, the data acquisition completion timestamp of the data source, and the data writing timestamp.

[0042] In one example embodiment, semantic mapping of a standard data source based on the measurement point dictionary table to obtain data items carrying unified semantic identifiers can be achieved as follows: The original data value is converted based on the conversion rules to obtain a target data value; the measurement point identifier of the device monitoring point is determined based on the measurement point dictionary table and the register address; the data type corresponding to the device monitoring point and the dimension identifier corresponding to the data type are determined based on the measurement point dictionary table, and the allowed value range of the device monitoring point is determined; the target data value is verified based on the allowed value range, and in response to the target data value passing the verification, the data item is generated according to the device identifier, measurement point identifier, and dimension identifier.

[0043] The following will further explain and illustrate the specific generation process of the data items. Specifically, the specific determination process of the data items described here can be implemented based on the heterogeneous access standardization module in the edge computing terminal. Specifically, after receiving the data source, the heterogeneous access standardization module can perform protocol adaptation and data parsing on the industrial equipment data sources with at least two different communication protocols or interfaces, and perform semantic mapping on the collected data sources according to the preset measurement point dictionary table to generate data items carrying unified semantic identifiers. The unified semantic identifiers described here can include equipment identifiers, measurement point identifiers, and dimensional identifiers, etc. The dimensional identifiers described here can be used to characterize the unit of the data value of the monitoring point of the equipment, such as pressure corresponding to MPa, current corresponding to A, speed corresponding to r / min, etc.

[0044] Furthermore, the heterogeneous access standardization module can be implemented based on the corresponding protocol adapter during the specific data source parsing process; each protocol adapter is used to complete the message parsing, field extraction and data type conversion of a certain type of communication protocol or interface, and output data items containing device identifier, device monitoring point measurement point identifier, dimensional identifier and timestamp fields.

[0045] It should also be noted that the measurement point dictionary table configures the measurement point identifier, data type, unit identifier, allowed value range and sampling period for each device monitoring measurement point. Based on this, the heterogeneous access standardization module verifies the field names and units of the collected data source according to the measurement point dictionary table, and maps the original point number on the device side to the measurement point identifier of the device monitoring point through the identifier mapping table, thereby generating a unified semantic identifier.

[0046] Secondly, clock skew correction and time alignment are performed on the data items to obtain a multi-source data sequence. Specifically, this can be achieved as follows: a reference clock is determined based on a preset time server, and the current clock skew parameter is determined based on the reference clock and the local clock in the edge computing terminal; the data acquisition timestamp and / or data writing timestamp are corrected based on the current clock skew parameter to obtain a time correction result, and the event time alignment of the data items is performed based on the time correction result to obtain a multi-source data sequence.

[0047] In one example embodiment, the multi-source data sequence is obtained by aligning the data items according to the time correction result. This can be achieved as follows: A first preset time period is used as the unit for dividing the event time alignment window, and a preset time point within the first preset time period is used as the starting time point for dividing the event time alignment window; the window is divided according to the starting time point and a preset window length to obtain multiple event time alignment windows, and the starting time point is used as the matching window number for each event time alignment window; the data items are aligned to each event time alignment window according to the window number and the time correction result, and after a second preset time period, a multi-source data sequence is generated based on the data items contained in each event time alignment window.

[0048] In one example embodiment, generating a multi-source data sequence based on the data items in each event time alignment window at a second preset time interval can be achieved as follows: Based on the window number and time correction result, determine the target number of data items that each event time alignment window can match; at a second preset time interval, determine the first current number of existing data items in each event time alignment window; at this moment, the existing data items in each event time alignment window are in a normal state; in response to any event time alignment window having a first current number less than the target number, control the event time alignment window to enter a late waiting stage; at a third preset time interval, determine the second current number of existing data items in each event time alignment window excluding data in a normal state; at this moment, the existing data items in each event time alignment window excluding data in a normal state are in a late state; in response to any event time alignment window having a sum of the first current number and the second current number less than the target number, determine the data items in a missing state, and generate a multi-source data sequence based on the data items in a normal state, the data items in a late state, and the data items in a missing state.

[0049] The following section will further explain and illustrate the specific generation process of multi-source data sequences. Specifically, the determination process of multi-source data sequences can rely on the time alignment module in the edge computing terminal. During the time alignment process, the time alignment module can achieve the following: obtain the source timestamp (i.e., the data acquisition completion timestamp) and acquisition timestamp (i.e., the data writing timestamp) of the data items; then correct the clock deviation of the data items based on the reference clock (the determined clock offset parameter); and perform event time alignment on the data items from different data sources according to a preset event time alignment window to obtain the event time alignment result. Finally, output the aligned multi-source data sequence based on the event time alignment result. The event time alignment result includes clock offset correction information and the data matching relationship within the event time alignment window.

[0050] Furthermore, during the alignment process, the time alignment module establishes a synchronization session with a reference clock (in a preset time server) to obtain the clock offset (i.e., the current clock offset parameter), and corrects the source timestamp and acquisition timestamp based on the current clock offset parameter. Based on this, the event time alignment result further includes: an alignment window number and a corrected timestamp corresponding to the alignment window number, used to identify the event time alignment window to which each data item belongs. It should also be noted that the event time alignment window is a fixed-duration window. Based on this, during event time alignment, the time alignment module can establish data matching relationships within the corresponding event time alignment window according to the window number, marking data items that fail to enter the corresponding event time alignment window within the preset lateness tolerance time as late data items, and marking missing data items within the event time alignment window as missing data items, and writing the late data items and missing data items into the event time alignment result.

[0051] In step S130, real-time status discrimination is performed on the multi-source data sequence to generate alarm events, and the association between industrial equipment and the associated equipment monitoring points is established based on unified semantic identifiers.

[0052] In this embodiment of the disclosure, firstly, real-time status discrimination is performed on the multi-source data sequence to generate an alarm event. Specifically, this can be achieved as follows: based on the window number of the event time alignment window in the multi-source data sequence, the device monitoring points included in three or more consecutive window numbers are read, and the first window measurement value within the event time alignment window corresponding to the window number is read; a threshold judgment condition corresponding to the device monitoring point is determined, and it is determined whether the first window measurement value within each event time alignment window meets the threshold judgment condition; in response to the first window measurement value of any window number not meeting the threshold judgment condition, the event time alignment window that does not meet the threshold judgment condition is marked as an over-limit window; in response to the number of over-limit windows being greater than a preset threshold, the alarm event is generated based on the device monitoring point, the device identifier of the industrial equipment associated with the device monitoring point, the measurement point identifier, the dimension identifier, the alarm type, the threshold judgment condition, the window number and the corresponding first window measurement value, and the alarm generation time.

[0053] The following will further explain and illustrate the specific generation process of alarm events. Specifically, in practical applications, within each event time alignment window, for each measurement point identifier, the window measurement value (i.e., the first window measurement value) corresponding to that measurement point identifier within the event time alignment window is obtained. The first window measurement value is compared with the threshold judgment condition composed of the upper limit threshold or the lower limit threshold corresponding to the measurement point identifier. Based on this, if the first window measurement value exceeds the upper limit threshold or is lower than the lower limit threshold, the event time alignment window is determined to be an out-of-limit window. When there are no fewer than 3 consecutive out-of-limit windows corresponding to the same measurement point identifier, an alarm event is generated and output. Here, the window measurement value recorded is the measurement value of the last data item within the event time alignment window.

[0054] Secondly, a relationship is established between industrial equipment and its associated monitoring points based on a unified semantic identifier. Specifically, in practical applications, the construction of this relationship can rely on the anomaly localization module in the edge computing terminal. For example, the anomaly localization module can determine the monitoring points of the industrial equipment based on the unified semantic identifier and identify the monitoring point identifier. On this basis, a relationship is established between the monitoring point identifier and the equipment identifier of the industrial equipment. Furthermore, if it is necessary to obtain the data of the equipment monitoring points of the industrial equipment, it can be obtained based on this relationship.

[0055] In step S140, based on the association, the alarm event, and the multi-source data sequence, the equipment in the industrial equipment that is in an abnormal operating state is located to obtain the abnormal equipment location result.

[0056] Specifically, the determination process for locating abnormal equipment can rely on the abnormal location module in the edge computing terminal. For example, the abnormal location module can output abnormal location results based on the correlation between industrial equipment and equipment monitoring points, alarm events, and event time alignment results (i.e., multi-source data sequences). The abnormal location results include the candidate abnormal source equipment number and the corresponding alarm event. It should be noted that the determination process for locating abnormal equipment can rely on a corresponding abnormal equipment location model. This model can be based on Transformer, BERT, or a large language model; this example does not impose any special restrictions.

[0057] In one example embodiment, the method further includes: reading the device monitoring points included in the window number of the event time alignment window in the multi-source data sequence, and the second window measurement value within the event time alignment window corresponding to the window number; generating a measurement value sequence based on the second window measurement value, and generating a feature summary of the event time alignment window corresponding to the window number based on the measurement value sequence; and / or uploading the feature summary of the event time alignment window and the multi-source data sequence to a cloud server, so that the cloud server performs normalization aggregation on the multi-source data sequence and feature summary based on the unified semantic identifier, and performs cross-scene fusion analysis on the normalized and aggregated data to obtain a global monitoring result. Specifically, the reporting process described here can rely on the real-time sensing and reporting module in the edge computing terminal. In practical applications, the real-time sensing and reporting module can perform real-time status discrimination on multi-source data sequences to generate alarm events or feature summaries, then write the alarm events or feature summaries to the storage medium, and send them to the cloud server through a reporting channel with acknowledgment and retransmission mechanisms. During this process, when the communication conditions do not meet the preset reporting requirements, the local output of alarm events is maintained and the alarm events or feature summaries are continued to be cached.

[0058] In one example embodiment, generating a feature summary of the event time-aligned window corresponding to the window number based on the measurement value sequence can be achieved as follows: determining the mean, standard deviation, peak value, and rate of change of the event time-aligned window corresponding to the window number based on the measurement value sequence; determining a mean-reduced sequence based on the measurement value sequence and the mean, and performing a discrete Fourier transform on the mean-reduced sequence to obtain the frequency domain energy of the event time-aligned window corresponding to the window number; and generating a feature summary of the event time-aligned window corresponding to the window number based on the mean, standard deviation, peak value, rate of change, and frequency domain energy.

[0059] In one example embodiment, the global monitoring results described above may include equipment status indicators and cross-scenario consistency scores output according to industrial site scenarios. The equipment status indicators characterize the operating status of corresponding industrial equipment in each industrial site scenario, while the cross-scenario consistency score characterizes the degree of data deviation and the consistency of change trends of equipment monitoring points with the same or related equipment in different industrial site scenarios within the same event time alignment window. In practical applications, based on the global monitoring results, the system can further distinguish between local equipment anomalies, single industrial site scenario anomalies, and cross-scenario common operating condition fluctuations, based on single equipment alarm judgment, thereby improving the accuracy and reliability of anomaly identification results.

[0060] In one example embodiment, equipment status indicators and cross-scenario consistency scores can also serve as input for the anomaly localization module to generate anomaly localization results. For instance, when an alarm event is generated by a monitoring point corresponding to a certain industrial equipment, and the cross-scenario consistency score of the industrial site scene where the equipment is located is lower than a preset score threshold, the anomaly localization module can combine the aforementioned correlations, alarm events, and event time alignment results (i.e., multi-source data sequences) to identify the industrial equipment as a candidate anomaly source device. Thus, the global monitoring results are used for unified monitoring and display on the server side, and also to assist in anomaly localization and maintenance decisions, improving the real-time performance, accuracy, and interpretability of industrial equipment monitoring in distributed heterogeneous scenarios.

[0061] In one possible example embodiment, the abnormal equipment in the industrial equipment is located based on the correlation, alarm events, and multi-source data sequences to obtain the abnormal equipment location result. This can also be achieved by receiving the global monitoring results fed back from the cloud server and generating an input feature sequence based on the global monitoring results, correlation, alarm events, and multi-source data sequences. On this basis, the input feature sequence is input into a preset abnormal equipment location model to obtain the abnormal equipment location result.

[0062] The following will further explain and illustrate the method for locating abnormal industrial equipment in a multi-source heterogeneous scenario as described in the exemplary embodiments of this disclosure, with reference to specific examples.

[0063] Example 1: This example verifies whether the heterogeneous access standardization module, under actual industrial equipment operating conditions, can perform standardized parsing, numerical conversion, and semantic mapping of raw collected data according to the device communication register definition file, and generate data items with unified semantic identifiers. The specific implementation method is as follows: S101, Implementation Environment and Equipment Basis. Specifically, the heterogeneous access standardization module is deployed on the edge computing terminal in the production workshop. The edge computing terminal is an industrial-grade computing device running an industrial-grade Linux operating system. Based on this, it is assumed that one of the field access devices (i.e., industrial equipment) is a circulating water pump controller, model WP-3000. The circulating water pump controller establishes a communication connection with the edge computing terminal through the Modbus communication protocol based on the Transmission Control Protocol. Simultaneously, the circulating water pump controller provides a "WP-3000 Communication Register Definition Table," which clearly stipulates that the controller provides operational data externally through the Modbus communication protocol based on the Transmission Control Protocol. All process quantities are stored in the holding register area, corresponding to Modbus function code 03. The physical meaning, data type, and conversion coefficient of each register are given in the register definition table. Based on this, the heterogeneous access standardization module parses the data according to the aforementioned register definition table.

[0064] S102, Register Configuration and Data Source Description. Specifically, according to the "WP-3000 Communication Register Definition Table," the register configurations related to the operating status in the circulating water pump controller can be shown in Table 1 below: Table 1: Register Configuration Data Table.

[0065]

[0066] Based on Table 1, during the device online phase, the heterogeneous access standardization module parses the above register definitions into a device parsing configuration file and uses it as the basis for data parsing of the protocol adapter.

[0067] S103, raw data from an actual data acquisition process. Specifically, assuming that at 14:26:08.412 on xx / xx / 20xx, the Modbus communication protocol adapter based on the transmission control protocol initiated a register read request to device WP-01, and returned the raw data (i.e., source data) shown in Table 2 below; where device WP-01 is the field instance number of the WP-3000 circulating water pump controller: Table 2: Register read raw data table.

[0068]

[0069] The original register values ​​mentioned above are all within the valid value range specified in the register definition table, and are all data frame contents directly returned by the device without any processing.

[0070] S104, Data Conversion Basis and Calculation Process. Specifically, the heterogeneous access standardization module performs physical quantity conversion on the original values ​​according to the conversion rules in the register definition table. For example, the original value of the circulating water outlet pressure is 294, the conversion factor is 0.01 MPa, and the actual pressure value is 294. 0.01 = 2.94 MPa; For the motor operating current, the original value is 167, the conversion factor is 0.1A, and the actual current value is 167. 0.1 = 16.7A; For the water pump speed, the original value is 1465. According to the register definition table, the register directly represents the speed, and the actual speed value is 1465 r / min.

[0071] It should be noted that the above calculation process is automatically completed by the protocol adapter, and the conversion coefficients are derived from the register definition file provided by the equipment manufacturer.

[0072] S105, Measurement Point Dictionary Table Configuration Basis. Specifically, the measurement point dictionary table is pre-configured in the heterogeneous access standardization module. The dictionary table is formulated based on the equipment design parameters and on-site operating specifications, and its contents are shown in Table 3 below: Table 3: Standardized Formulation Table for Measurement Point Dictionary Table.

[0073]

[0074] Specifically, the heterogeneous access standardization module maps register addresses to corresponding measurement point identifiers based on the aforementioned measurement point dictionary table, and performs range verification on the converted values. The verification showed that 2.94 MPa, 16.7 A, and 1465 r / min were all within their respective valid ranges.

[0075] S106 includes the formation result of data items with unified semantic identifiers. Specifically, after parsing, conversion, and verification, the heterogeneous access standardization module generates data items with unified semantic identifiers. Among them, the data item corresponding to the circulating water outlet pressure is: equipment identifier is circulating water pump-01, measuring point identifier is outlet pressure, dimension identifier is MPa, measured value is 2.94, and timestamp is 20xx-xx-xx-14:26:08.412. The measured value of this data item originates from the original value 294 of register 40010 and is obtained according to the conversion rules in the device register definition table; simultaneously, the timestamp is the local system time obtained by the edge computing terminal when completing the register read.

[0076] In summary, the implementation results show that the heterogeneous access standardization module can reliably acquire raw register data according to the register definition file provided by the equipment manufacturer, and convert the raw values ​​into corresponding engineering quantity data based on preset conversion rules. Simultaneously, the module can combine the measurement point dictionary table to complete measurement point identification mapping, dimensional consistency verification, and value range determination. The generated data items have clear data sources, calculation bases, and a unified data structure, and are repeatable under the same equipment and configuration conditions. This indicates that under the implementation conditions, the heterogeneous access standardization module can meet the application requirements for standardized access to heterogeneous equipment data in industrial sites.

[0077] Example 2: This example demonstrates a real-world operation of the time alignment module under industrial conditions where multiple devices have inconsistent local clocks and communication experiences millisecond-level latency and jitter. It showcases the process of performing event time correction and alignment on multi-source data items. The specific implementation is as follows: S201, Setting up the implementation environment. Specifically, this embodiment is a real-world example of the time alignment module operating continuously in an industrial site. In this scenario, the time alignment module can be deployed on an edge computing industrial control computer (i.e., an edge computing terminal) to perform event time alignment processing on data items from two industrial devices to obtain multi-source data sequences. Assume the two industrial devices are device A and device B, where device A is a vibration monitoring device and device B is a temperature monitoring device. Device A and device B are connected via communication links of different protocol types, and their respective local clocks are not uniformly calibrated.

[0078] S202, determine the reference clock and clock deviation data (i.e., the current clock deviation parameter). Specifically, first, determine the source of the reference clock; the time alignment module uses the built-in industrial control computer system clock as the reference clock (i.e., the local clock); simultaneously, the system clock synchronizes with the factory time server via a network time synchronization protocol, with a synchronization period set to 10 seconds, and the time error of the synchronization result during the network time synchronization process is within the millisecond range. Second, determine the basis for obtaining the current clock deviation parameter; for example, in a time synchronization process completed at 12:00:00, the time alignment module records the difference between the system clock and the return time from the factory time server as +18ms; during this process, the difference can be automatically calculated by the synchronization program in the time alignment module and written to the operation log as the clock deviation parameter used in the time period from 12:00:00 to 12:00:10. It should be noted here that +18ms is one of the stable deviation values ​​repeatedly recorded in multiple synchronization processes, belonging to the common deviation range recorded in time synchronization processes under industrial field network conditions. Furthermore, in this embodiment, the reference clock does not experience time rollback or abrupt changes during the synchronization period, and the time alignment module performs time correction processing based on the acquired clock deviation parameters under this condition.

[0079] S203, Acquisition of the source timestamp (i.e., the data acquisition completion timestamp) and the acquisition timestamp (i.e., the data write timestamp). Specifically, the timestamps are generated as follows: both Device A and Device B carry source timestamp and acquisition timestamp fields in their data packets. The source timestamp is the time recorded by the device's local clock when the physical quantity sampling is completed; the acquisition timestamp is the time recorded when the device writes the sampled data into the communication buffer. Furthermore, the firmware time resolution of both Device A and Device B is 1ms.

[0080] Actual data collected: Around 12:00:05, the time alignment module received the following two data items, with the relevant timestamps directly derived from the device's reported message content: Device A data item: source timestamp is 12:00:05.236, acquisition timestamp is 12:00:05.241. Device B data item: source timestamp is 12:00:05.251, acquisition timestamp is 12:00:05.256; the approximately 5ms difference between the two timestamps conforms to the processing delay range recorded in the corresponding device's sampling and data encapsulation process.

[0081] S204, Timestamp Correction Processing. Specifically, the correction is based on the following: After receiving the above data item, the time alignment module calls the clock deviation parameter recorded within the current synchronization cycle + 18ms to correct the source timestamp in the data item. Correction results: The corrected event time for device A is 12:00:05.218, and the corrected event time for device B is 12:00:05.233. In this embodiment, since the source timestamp is complete and does not exceed a reasonable time range, the time alignment module preferentially uses the corrected source timestamp as the event time.

[0082] S205, Event Time Alignment Window Division. Specifically, the window parameter settings are based on the following: the event time alignment window uses a fixed duration window, with a window length of 100ms. The window length is set based on the field operating conditions: the minimum sampling period for both device A and device B is not less than 50ms, the end-to-end jitter of the field network under normal load conditions is less than 20ms, and the end-to-end jitter data comes from field operating statistics.

[0083] Under the above conditions, to ensure that data items corresponding to the same physical event can stably fall into the same window, the event time alignment window length is set to 100ms. 100ms is a set of actual operating parameters used in this embodiment to illustrate the setting method of the event time alignment window and does not constitute a limitation on the event time alignment window length. Next, window number calculation: the alignment start reference time T0 is set to the system startup time 12:00:00.000. Based on the corrected event time calculation described above, the time offset corresponding to device A is 5218ms, and the window number (i.e., window code) is 52; the time offset corresponding to device B is 5233ms, and the window number is also 52. Based on this, the time alignment module assigns the above two data items to the same event time alignment window with the number 52. The window number is obtained by dividing the time difference between the event time and the start reference time by the window length and rounding down. The obtained event time alignment window can be referenced... Figure 6 As shown, the window coding associated with the event time alignment window can be referenced. Figure 7As shown, the obtained event timing alignment results (e.g., window encoding 52) can be referenced. Figure 8 As shown.

[0084] S206, Lateness Tolerance Time and Window Closure. Specifically, the lateness tolerance time is set as follows: 30ms, exceeding the 20ms upper limit for network jitter on-site, while also allowing for a safety margin. Window Processing: After the event time alignment window numbered 52 ends, the time alignment module enters the lateness waiting phase. If no new data item belonging to the window number is received within the 30ms waiting time, the time alignment module closes the window after the lateness tolerance time ends. Although no late data items appear during the operation of this embodiment, the judgment logic corresponding to the lateness tolerance time remains effective in the time alignment module and participates in the event time alignment window closure determination process.

[0085] S207, Generation of Event Time Alignment Results. Specifically, firstly, result confirmation is performed: when the window is closed, the time alignment module confirms that window number 52 contains valid data items from device A and device B, no missing data items are detected, and no late data items are marked. Secondly, result output: the time alignment module generates the corresponding event time alignment result, records window number 52, the corrected event time of each data item, and the clock offset parameters used in this synchronization cycle, and outputs the aligned data in the order of window numbers, thereby obtaining the corresponding multi-source data sequence.

[0086] Through the above implementation steps, based on the operational results of the embodiment, the time alignment module synchronously obtains clock deviation parameters through a reference clock and corrects the timestamps in the data items. Based on this, it divides the data items according to a fixed-length event time alignment window, enabling data items from different devices to be grouped into the same alignment window according to the corrected event time. Within a set lateness tolerance time, the time alignment module completes the matching and confirmation of data items within the window and outputs information including the window number, corrected timestamp, and related identifiers when the window is closed. During the above operation, the time alignment module corrects the timestamps of multi-source data items based on the clock deviation parameters obtained from the reference clock, and completes the division and confirmation of data items according to a fixed-length event time alignment window, generating corresponding event time alignment results even when device clocks are inconsistent and communication is uncertain.

[0087] Example 3: This example illustrates the process by which the real-time sensing and reporting module performs real-time status determination and generates alarm events based on the measurement point data within the event time alignment window. The specific implementation is as follows: S301, Data Source and Window Configuration. Specifically, the real-time sensing and reporting module is deployed on the edge computing terminal in the industrial field to receive multi-source data sequences output by the time alignment module. Assume one of the devices connected to the industrial field is a circulating water pump controller, identified as circulating water pump-01. Measuring points include outlet pressure, operating current, and rotational speed, with corresponding dimensions of MPa, A, and r / min, respectively. Based on this, a fixed-length window of 100ms is used for the event time alignment. The window number is obtained by dividing the time difference between the corrected event time and the alignment start point by the window length and rounding down. The real-time sensing and reporting module reads the window measurement value of each measuring point within the corresponding event time alignment window according to the window number. The window measurement value is the measurement value of the last data item within the window.

[0088] S302, Threshold Comparison and Over-Limit Window Determination. Specifically, for equipment monitoring points identified as outlet pressure, the preset upper threshold is 3.20 MPa, and the preset lower threshold is 0.50 MPa. The outlet pressure data obtained by the real-time sensing and reporting module within three consecutive event time-aligned windows are as follows: window number 1052 corresponds to 3.31 MPa, window number 1053 corresponds to 3.36 MPa, and window number 1054 corresponds to 3.42 MPa. The outlet pressures of the above three windows all exceed 3.20 MPa; therefore, windows numbered 1052, 1053, and 1054 are all marked as over-limit windows.

[0089] S303, Alarm Event Generation. Specifically, assuming that three consecutive out-of-limit windows corresponding to the outlet pressure of the same measuring point are identified, satisfying the alarm condition of at least three consecutive out-of-limit windows, the real-time sensing and reporting module generates a pressure upper limit out-of-limit alarm event accordingly. The alarm event includes the equipment identifier circulating water pump-01, the outlet pressure of the measuring point identifier, the unit identifier (MPa), the alarm type pressure upper limit out-of-limit, the trigger threshold (3.20MPa), the consecutive out-of-limit window numbers 1052 to 1054, the corresponding window measured values ​​(3.31MPa, 3.36MPa, and 3.42MPa), and the alarm generation time. The alarm generation time is the local system time recorded by the edge computing terminal when window number 1054 is closed and the continuous out-of-limit judgment is completed.

[0090] S304, Alarm Event Storage and Reporting. Specifically, after an alarm event is generated, the real-time sensing and reporting module writes it to the local storage medium of the edge computing terminal and sends it to the cloud server through a reporting channel with acknowledgment and retransmission mechanisms. After the cloud server returns acknowledgment information, the alarm event is marked as reported. Meanwhile, if the communication bandwidth, packet loss rate, or link availability does not meet the preset reporting requirements, the alarm event remains output locally and is written to a local cache queue. Once communication conditions are restored, the real-time sensing and reporting module re-executes the reporting based on the alarm event identifier in the cache queue.

[0091] Through the above processing, the real-time sensing and reporting module can complete the reading of measurement point data, threshold comparison, over-limit window determination, and continuous over-limit identification within the event time alignment window, and generate an alarm event when the preset alarm conditions are met. The alarm event has a clear device source, measurement point source, trigger threshold, window number, and measurement value basis. Under communication abnormal conditions, the alarm event can still be output and cached locally, thereby ensuring the real-time performance and traceability of alarms in the industrial field.

[0092] Example 4: This example illustrates the process by which the real-time sensing and reporting module extracts feature summaries within an event time alignment window and prioritizes reporting these feature summaries when communication conditions are limited. The specific implementation is as follows: S401, Data Source and Window Data. Specifically, the real-time sensing and reporting module is deployed on the edge computing terminal in the industrial site, receiving multi-source data sequences output by the time alignment module. Assume one of the industrial site devices is identified as circulating water pump-02, the measurement point is identified as operating current, and the unit of measurement is A. The event time alignment window uses a fixed duration window with a length of 100ms. Based on this, within window number 2060, the operating current measurement point corresponds to five aligned measurement values, namely 16.4A, 16.7A, 17.1A, 16.9A, and 17.3A. The real-time sensing and reporting module extracts features from the measurement value sequence within the window, generating a feature summary corresponding to window number 2060.

[0093] S402, In the process of determining the feature summary, statistical feature extraction is required first. Specifically, the extracted statistical features may include, but are not limited to: mean, which is obtained by summing all measurements within the window and dividing by the number of measurements, resulting in 16.88A; variance, which is obtained by summing the squares of the differences between each measurement within the window and the mean, divided by the number of measurements, resulting in 0.0976A²; peak value, which is the maximum measurement within the window, resulting in 17.3A; and rate of change, which is calculated by the difference between the last measurement and the first measurement within the window, resulting in 0.9A / window.

[0094] S403. In the process of determining the feature summary, frequency domain energy calculation is also required. Specifically, the frequency domain energy is calculated using the mean-reduced Discrete Fourier Transform (DFT) result. For example, the mean of 16.88A can be subtracted from the five measured values ​​respectively to obtain the mean-reduced sequence: -0.48A, -0.18A, 0.22A, 0.02A, and 0.42A. Based on this, a DFT is performed on the mean-reduced sequence, and the frequency domain energy is calculated according to the normalized energy of the non-DC frequency components. For example, under the normalization calculation rule, the sum of the energy of the non-DC frequency components corresponds to the sum of squares of the mean-reduced sequence, and the calculated result is 0.488A².

[0095] S404 prioritizes reporting under limited communication conditions during the reporting process. Specifically, the real-time sensing and reporting module has a preset bandwidth threshold of 1 Mbps. When the currently available uplink bandwidth is detected to be 0.45 Mbps, it is determined that the current communication condition is below the preset bandwidth threshold. To reduce the amount of reported data and transmission latency, the real-time sensing and reporting module prioritizes writing the feature summary corresponding to window number 2060 to the local storage medium and sends the feature summary to the cloud server. The complete measurement value sequence within the window continues to be stored in the local cache and will be uploaded again after the available uplink bandwidth recovers to no less than 1 Mbps. Furthermore, after receiving the feature summary, the cloud server judges the trend of equipment operating status changes based on the mean, peak, rate of change, and frequency domain energy of the operating current. On this basis, if the mean and peak values ​​of the operating current within multiple consecutive event time-aligned windows increase and the rate of change remains positive, the cloud server marks the corresponding industrial equipment as having an increased load. If the status occurs simultaneously with abnormal changes in pressure or speed measurement points, a monitoring prompt indicating abnormal motor load or increased pump operating resistance is output.

[0096] The above process demonstrates that feature summaries can replace complete measurement sequences for priority reporting when communication resources are limited. This allows the cloud server to obtain statistical, trend, and frequency domain characteristics reflecting the equipment's operating status even without immediately receiving all the original sampled data. This processing method reduces the amount of reported data, mitigates the impact of fluctuating communication conditions on monitoring real-time performance, and retains the crucial data required for subsequent status assessments.

[0097] Example 5: This example illustrates how the cloud-side fusion and collaboration module performs cross-scenario fusion analysis based on unified semantic identifiers and event time alignment results, and generates global monitoring results including device status indicators and cross-scenario consistency scores. The specific implementation is as follows: S501, Cross-Scenario Data Reception and Dimension Normalization. Specifically, the cloud server-side fusion and collaboration module receives multi-source data sequences and feature summaries from three (or more) industrial field scenarios. Assume the device identifier in scenario A is circulating water pump-A01, in scenario B it's circulating water pump-B01, and in scenario C it's circulating water pump-C01. All three devices have a measurement point identifier indicating outlet pressure, but their access protocols, sampling periods, and deployment environments differ.

[0098] The data items received by the cloud server all include device identifier, measurement point identifier, unit identifier, alignment window number, and measured value. For the outlet pressure measurement point, the unit identifier uploaded by scenarios A and C is MPa, while that uploaded by scenario B is kPa. When the fusion and collaboration module on the cloud server performs normalization aggregation, it first converts the pressure value uploaded by scenario B from kPa to MPa and performs a unit consistency check on the data from the three scenarios. After the conversion and check are completed, the outlet pressure data of the three scenarios all adopt MPa as the unified unit. Specifically, the original pressure values ​​uploaded by scenario B in windows 3049 and 3050 are 2930 kPa and 2950 kPa, respectively, which, after unit conversion, are 2.93 MPa and 2.95 MPa, respectively.

[0099] S502, cross-scenario alignment sets are formed. Specifically, within alignment window number 3049, the normalized measured values ​​of the outlet pressure for scenarios A, B, and C are 2.90 MPa, 2.93 MPa, and 3.33 MPa, respectively. Within alignment window number 3050, the normalized measured values ​​of the outlet pressure for scenarios A, B, and C are 2.91 MPa, 2.95 MPa, and 3.48 MPa, respectively.

[0100] The cloud server's fusion and collaboration module will form a cross-scenario alignment set by aligning data items with the same measuring point identifier outlet pressure within window number 3050. The cross-scenario alignment set includes 2.91MPa for circulating water pump-A01, 2.95MPa for circulating water pump-B01, and 3.48MPa for circulating water pump-C01.

[0101] S503, Cross-Scenario Consistency Score Calculation. Specifically, the cross-scenario consistency score is determined by combining data deviation and the degree of matching of change trends. For any scenario, first calculate the deviation between the pressure value of the scenario within the current alignment window and the median pressure value of each scenario; then calculate the deviation between the pressure change of the scenario relative to the previous alignment window and the median pressure change of each scenario. For example, within alignment window number 3050, the median pressure value of the three scenarios is 2.95 MPa, and the pressure changes of the three scenarios relative to window number 3049 are 0.01 MPa, 0.02 MPa, and 0.15 MPa, respectively, with a median change of 0.02 MPa.

[0102] In a preset scoring rule, the normalized upper limit for pressure deviation is set at 0.50 MPa, the normalized upper limit for change deviation is set at 0.20 MPa, the weight for data deviation is 0.7, and the weight for trend deviation is 0.3. The cross-scenario consistency score is determined by subtracting the weighted deviation value from 1, with the score range limited to 0 to 1. Based on this, according to the preset scoring rule, when the pressure deviation or change deviation exceeds the corresponding normalized upper limit, the corresponding normalized deviation value is counted as 1. Scenario A has a pressure deviation of 0.04 MPa and a change deviation of 0.01 MPa, corresponding to a cross-scenario consistency score of approximately 0.93. Scenario B has a pressure deviation of 0 MPa and a change deviation of 0 MPa, corresponding to a cross-scenario consistency score of 1.00. Scenario C has a pressure deviation of 0.53 MPa, exceeding the normalized upper limit for pressure deviation, and a change deviation of 0.13 MPa, corresponding to a cross-scenario consistency score of approximately 0.11.

[0103] S504, Global Monitoring Result Generation and Anomaly Location. Specifically, firstly, the fusion and collaboration module generates equipment status indicators based on equipment status judgment rules. The upper limit threshold for pressure is set to 3.20 MPa, and the cross-scenario consistency score threshold is set to 0.60. When the outlet pressure corresponding to the equipment does not exceed 3.20 MPa and the cross-scenario consistency score is not lower than 0.60, the equipment status indicator is normal. When the outlet pressure corresponding to the equipment exceeds 3.20 MPa, or the cross-scenario consistency score is lower than 0.60, the equipment status indicator is of abnormal concern.

[0104] Based on the above rules, the status indicators of the circulating water pump-A01 in scenario A are normal, and the cross-scenario consistency score is 0.93. The status indicators of the circulating water pump-B01 in scenario B are normal, and the cross-scenario consistency score is 1.00. The status indicators of the circulating water pump-C01 in scenario C are abnormal and require attention because the outlet pressure is 3.48MPa and the cross-scenario consistency score is 0.11.

[0105] Secondly, the integration and collaboration module combines the device status indicators and cross-scenario consistency scores corresponding to each scenario into a global monitoring result.

[0106] Global monitoring results show that the outlet pressure values ​​in scenarios A and B are similar and their trends are consistent, while the outlet pressure in scenario C is significantly higher than in other scenarios and its trend deviates, indicating a risk of abnormal pressure rise in circulating water pump-C01. The global monitoring results are written to the monitoring logs on the cloud server and provided to the anomaly localization module as input. The anomaly localization module, combining the correlation between industrial equipment and equipment monitoring points, equipment status indicators, cross-scenario consistency scores, and event time alignment results (i.e., multi-source data sequences), identifies circulating water pump-C01 as a candidate anomaly source device.

[0107] The above process demonstrates that the cloud server-side fusion and collaboration module can form a cross-scenario alignment set based on unified semantic identifiers and event time alignment results. It then generates global monitoring results through unit normalization, dimensional consistency verification, data deviation comparison, and trend comparison. These global monitoring results not only characterize the operating status of equipment in various industrial scenarios but also distinguish between local equipment anomalies, single-scenario anomalies, and cross-scenario common operating condition fluctuations. Furthermore, they provide data support for the anomaly localization module to identify candidate anomaly source equipment.

[0108] Thus, the method for locating abnormal industrial equipment in multi-source heterogeneous scenarios described in the exemplary embodiments of this disclosure has been fully implemented. Based on the foregoing description, the method for locating abnormal industrial equipment in multi-source heterogeneous scenarios described in the exemplary embodiments of this disclosure has at least the following advantages: On the one hand, by unifying data semantics and dimensional constraints, and performing consistency verification and anomaly identification during the access and preprocessing stages, ambiguity and distortion of multi-source heterogeneous data during the fusion process are reduced, improving data integrity, reliability, and the credibility of analysis results; on the other hand, by aligning and windowing multi-source data based on event time, and completing state determination and feature extraction at the edge, stable and timely state monitoring and anomaly identification can still be achieved under complex network and multi-sampling frequency conditions, balancing real-time performance and system load; furthermore, through the decoupling design of device access and protocol adaptation, as well as the unified data model and configurable mapping mechanism, the system can easily support the expansion of device types, communication protocols, and deployment scale, reducing system evolution and maintenance costs.

[0109] The following are embodiments of the apparatus disclosed herein, which can be used to execute embodiments of the method disclosed herein. For details not disclosed in the apparatus embodiments of this disclosure, please refer to the embodiments of the method disclosed herein.

[0110] Specifically, this disclosure also provides a positioning device for abnormal industrial equipment in a multi-source heterogeneous scenario, configured on an edge computing terminal, which includes multiple protocol adapters corresponding to transmission control protocols with different protocol types. (See reference) Figure 9 As shown, the positioning device for abnormal industrial equipment in this multi-source heterogeneous scenario may include a multi-source heterogeneous data source acquisition module 910, a multi-source data sequence determination module 920, an alarm event generation module 930, and an abnormal equipment positioning result determination module 940.

[0111] The multi-source heterogeneous data source acquisition module 910 can be used to acquire data sources from industrial equipment in multi-source heterogeneous scenarios. The industrial equipment consists of multiple devices with different equipment types, and these different equipment types access the edge computing terminal through different protocol types. The multi-source data sequence determination module 920 can be used to preprocess the data sources to obtain data items carrying unified semantic identifiers, and to perform clock deviation correction and time alignment on the data items to obtain a multi-source data sequence. The alarm event generation module 930 can be used to perform real-time status discrimination on the multi-source data sequence to generate alarm events, and to establish the association relationship between the industrial equipment and the associated equipment monitoring points based on the unified semantic identifiers. The abnormal equipment location result determination module 940 can be used to locate the equipment in the industrial equipment that is in an abnormal operating state according to the association relationship, the alarm events, and the multi-source data sequence to obtain the abnormal equipment location result.

[0112] In one exemplary embodiment of this disclosure, collecting data sources from industrial equipment in a multi-source heterogeneous scenario includes: determining, based on the equipment type of the industrial equipment, the transmission control protocol required to establish a communication connection between the edge computing terminal and the industrial equipment, and determining, based on the protocol type of the transmission control protocol, the protocol adapter required to establish the communication connection; establishing a communication connection between the edge computing terminal and the industrial equipment based on the transmission control protocol through the protocol adapter; and receiving data sources sent by the industrial equipment based on the communication connection; wherein the data sources include a communication register definition table corresponding to the industrial equipment, a device identifier, and current operating status information.

[0113] In one exemplary embodiment of this disclosure, preprocessing the data source to obtain data items carrying unified semantic identifiers includes: adapting and parsing the current operating status information according to a communication register definition table to obtain a standard data source; the standard data source includes multiple elements such as the equipment identifier of the industrial equipment, the register address corresponding to the equipment monitoring point associated with the industrial equipment, the meaning of the physical quantity, the data type, the conversion rule, and the original data value; obtaining a measurement point dictionary table corresponding to the industrial equipment, and semantically mapping the standard data source based on the measurement point dictionary table to obtain data items carrying unified semantic identifiers; the data items include multiple elements such as the equipment identifier, the measurement point identifier of the equipment monitoring point, the dimensional identifier corresponding to the equipment monitoring point, the data acquisition completion timestamp of the data source, and the data writing timestamp.

[0114] In one exemplary embodiment of this disclosure, semantic mapping of a standard data source based on the measurement point dictionary table to obtain a data item carrying a unified semantic identifier includes: converting the original data value based on the conversion rule to obtain a target data value, and determining the measurement point identifier of the device monitoring point based on the measurement point dictionary table and the register address; determining the data type corresponding to the device monitoring point and the dimension identifier corresponding to the data type based on the measurement point dictionary table, and determining the allowed value range of the device monitoring point; verifying the target data value based on the allowed value range, and generating the data item according to the device identifier, measurement point identifier, and dimension identifier in response to the target data value passing the verification.

[0115] In one exemplary embodiment of this disclosure, clock skew correction and time alignment are performed on the data items to obtain a multi-source data sequence, including: determining a reference clock based on a preset time server, and determining a current clock skew parameter based on the reference clock and the local clock in the edge computing terminal; correcting the data acquisition timestamp and / or data writing timestamp based on the current clock skew parameter to obtain a time correction result, and aligning the data items to event time based on the time correction result to obtain a multi-source data sequence.

[0116] In an exemplary embodiment of this disclosure, the process of aligning data items to event time to obtain a multi-source data sequence based on time correction results includes: using a first preset time period as the unit for dividing event time alignment windows, and using a preset time point within the first preset time period as the starting time point for dividing the event time alignment windows; dividing the windows according to the starting time point and a preset window length to obtain multiple event time alignment windows, and matching window numbers to the event time alignment windows according to the starting time point; aligning the data items to each of the event time alignment windows according to the window numbers and time correction results, and generating a multi-source data sequence based on the data items contained in each of the event time alignment windows at intervals of a second preset time period.

[0117] In an exemplary embodiment of this disclosure, generating a multi-source data sequence based on data items in each event time alignment window at intervals of a second preset time period includes: determining a target number of data items that each event time alignment window can match based on the window number and time correction result; determining a first current number of existing data items in each event time alignment window at intervals of the second preset time period; at this moment, the existing data items in each event time alignment window are in a normal state; in response to any event time alignment window having a first current number less than the target number, controlling the event time alignment window to enter a late waiting stage; determining a second current number of existing data items other than data in a normal state in each event time alignment window at intervals of a third preset time period; at this moment, the existing data items other than data in a normal state in each event time alignment window are in a late state; in response to any event time alignment window having a sum of a first current number and a second current number less than the target number, determining data items in a missing state, and generating the multi-source data sequence based on data items in a normal state, data items in a late state, and data items in a missing state.

[0118] In an exemplary embodiment of this disclosure, real-time status discrimination of a multi-source data sequence to generate an alarm event includes: reading the first window measurement value of a device monitoring point included in three or more consecutive window numbers according to the window number of the event time alignment window in the multi-source data sequence; determining a threshold judgment condition corresponding to the device monitoring point, and determining whether the first window measurement value in each event time alignment window meets the threshold judgment condition; in response to the first window measurement value of any window number not meeting the threshold judgment condition, marking the event time alignment window that does not meet the threshold judgment condition as an over-limit window; in response to the number of over-limit windows being greater than a preset threshold, generating the alarm event according to the device monitoring point, the device identifier of the industrial equipment associated with the device monitoring point, the measurement point identifier, the dimensional identifier, the alarm type, the threshold judgment condition, the window number and the corresponding first window measurement value, and the alarm generation time.

[0119] In one exemplary embodiment of this disclosure, the apparatus further includes: a second window measurement value reading module, configured to read the second window measurement value of the device monitoring point included in the window number within the event time alignment window corresponding to the window number, based on the window number of the event time alignment window in the multi-source data sequence; a feature summary generation module, configured to generate a measurement value sequence based on the second window measurement value, and generate a feature summary of the event time alignment window corresponding to the window number based on the measurement value sequence; and a reporting module, configured to upload the feature summary of the event time alignment window and the multi-source data sequence to a cloud server, so that the cloud server performs normalization aggregation on the multi-source data sequence and feature summary based on the unified semantic identifier, and performs cross-scenario fusion analysis on the normalized aggregated data to obtain a global monitoring result.

[0120] In one exemplary embodiment of this disclosure, generating a feature summary of an event time-aligned window corresponding to the window number based on the measurement value sequence includes: determining the mean, standard deviation, peak value, and rate of change of the event time-aligned window corresponding to the window number based on the measurement value sequence; determining a mean-reduced sequence based on the measurement value sequence and the mean, and performing a discrete Fourier transform on the mean-reduced sequence to obtain the frequency domain energy of the event time-aligned window corresponding to the window number; and generating a feature summary of the event time-aligned window corresponding to the window number based on the mean, standard deviation, peak value, rate of change, and frequency domain energy.

[0121] The specific details of each module in the positioning device for abnormal industrial equipment in the above-mentioned multi-source heterogeneous scenario have been described in detail in the corresponding positioning method for abnormal industrial equipment in multi-source heterogeneous scenario, so they will not be repeated here.

[0122] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0123] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0124] In an exemplary embodiment of this disclosure, an electronic device capable of implementing the above-described method is also provided.

[0125] Those skilled in the art will understand that various aspects of this disclosure can be implemented as a system, method, or program product. Therefore, various aspects of this disclosure can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software aspects, collectively referred to herein as a "circuit," "module," or "system."

[0126] The following reference Figure 10 To describe an electronic device 1000 according to such an embodiment of the present disclosure. Figure 10 The electronic device 1000 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0127] like Figure 10 As shown, the electronic device 1000 is manifested in the form of a general-purpose computing device. The components of the electronic device 1000 may include, but are not limited to: at least one processing unit 1010, at least one storage unit 1020, a bus 1030 connecting different system components (including storage unit 1020 and processing unit 1010), and a display unit 1040.

[0128] The storage unit stores program code that can be executed by the processing unit 1010, causing the processing unit 1010 to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure. For example, the processing unit 1010 can perform actions such as... Figure 1 Step S110: Collect data sources from industrial equipment in a multi-source heterogeneous scenario. The industrial equipment consists of multiple devices with different equipment types, and these devices access the edge computing terminal through different protocol types. Step S120: Preprocess the data sources to obtain data items carrying unified semantic identifiers, and perform clock deviation correction and time alignment on the data items to obtain a multi-source data sequence. Step S130: Perform real-time status discrimination on the multi-source data sequence to generate alarm events, and establish the association relationship between industrial equipment and associated equipment monitoring points based on the unified semantic identifiers. Step S140: Based on the association relationship, the alarm events, and the multi-source data sequence, locate the devices in the industrial equipment that are in an abnormal operating state to obtain the abnormal device location result.

[0129] Storage unit 1020 may include readable media in the form of volatile storage units, such as random access memory (RAM) 10201 and / or cache memory 10202, and may further include read-only memory (ROM) 10203.

[0130] Storage unit 1020 may also include a program / utility 10204 having a set (at least one) program module 10205, such program module 10205 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0131] Bus 1030 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the multiple bus structures.

[0132] Electronic device 1000 can also communicate with one or more external devices 1100 (e.g., keyboard, pointing device, Bluetooth device, etc.), one or more devices that enable a user to interact with electronic device 1000, and / or any device that enables electronic device 1000 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 1050. Furthermore, electronic device 1000 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 1060. As shown, network adapter 1060 communicates with other modules of electronic device 1000 via bus 1030. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 1000, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0133] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0134] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible implementations, various aspects of this disclosure may also be implemented as a program product including program code that, when the program product is run on a terminal device, causes the terminal device to perform the steps of the various exemplary embodiments of this disclosure described in the "Exemplary Methods" section above.

[0135] The program product for implementing the above-described method according to embodiments of the present disclosure may employ a portable compact disc read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto. In this document, the readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0136] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0137] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0138] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0139] Program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0140] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of this disclosure and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0141] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention described herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not invented by this disclosure. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

Claims

1. A method for locating abnormal industrial equipment in a multi-source heterogeneous scenario, characterized in that, Configured in an edge computing terminal, the edge computing terminal includes multiple protocol adapters corresponding to transmission control protocols with different protocol types, the method includes: Data sources are collected from industrial equipment in multi-source heterogeneous scenarios. The industrial equipment consists of multiple devices with different equipment types, and these different equipment types are connected to the edge computing terminal through different protocol types. The data source is preprocessed to obtain data items carrying unified semantic identifiers, and the data items are then clock skew correction and time alignment to obtain a multi-source data sequence. Real-time status discrimination is performed on multi-source data sequences to generate alarm events, and the association between industrial equipment and the associated equipment monitoring points is established based on the unified semantic identifier. Based on the aforementioned correlation, alarm events, and multi-source data sequences, the equipment in the industrial equipment that is in an abnormal operating state is located, and the abnormal equipment location result is obtained.

2. The method according to claim 1, characterized in that, Data sources for collecting industrial equipment in multi-source heterogeneous scenarios include: Based on the equipment type of the industrial equipment, determine the transmission control protocol required to establish a communication connection between the edge computing terminal and the industrial equipment, and determine the protocol adapter required to establish the communication connection based on the protocol type of the transmission control protocol; The communication connection between the edge computing terminal and the industrial equipment is established through the protocol adapter based on the transmission control protocol; The data source sent by the industrial equipment is received based on the communication connection; the data source includes the communication register definition table corresponding to the industrial equipment, the equipment identifier, and the current operating status information.

3. The method according to claim 2, characterized in that, Preprocessing the data source yields data items carrying unified semantic identifiers, including: The current operating status information is adapted and parsed according to the communication register definition table to obtain a standard data source; the standard data source includes the equipment identifier of the industrial equipment, the register address corresponding to the equipment monitoring point associated with the industrial equipment, the meaning of the physical quantity, the data type, the conversion rule, and various original data values; Obtain a measurement point dictionary table corresponding to the industrial equipment, and perform semantic mapping on the standard data source based on the measurement point dictionary table to obtain data items carrying a unified semantic identifier; the data item includes multiple elements such as the equipment identifier, the measurement point identifier of the equipment monitoring point, the dimensional identifier corresponding to the equipment monitoring point, the data acquisition completion timestamp of the data source, and the data writing timestamp.

4. The method according to claim 3, characterized in that, Based on the aforementioned measurement point dictionary table, semantic mapping is performed on the standard data source to obtain data items carrying unified semantic identifiers, including: The original data value is converted based on the conversion rule to obtain the target data value, and the measurement point identifier of the device monitoring point is determined based on the measurement point dictionary table and the register address. Based on the measurement point dictionary, determine the data type corresponding to the equipment monitoring point and the dimension identifier corresponding to the data type, and determine the allowable value range of the equipment monitoring point; The target data value is verified based on the allowed value range, and in response to the target data value verification passing, the data item is generated according to the device identifier, measurement point identifier, and unit identifier.

5. The method according to claim 3, characterized in that, Clock skew correction and time alignment are performed on the data items to obtain a multi-source data sequence, including: A reference clock is determined based on a preset time server, and the current clock deviation parameter is determined based on the reference clock and the local clock in the edge computing terminal. Based on the current clock deviation parameter, the data acquisition timestamp and / or data writing timestamp are corrected to obtain a time correction result. Then, the data items are aligned to the event time according to the time correction result to obtain a multi-source data sequence.

6. The method according to claim 5, characterized in that, Based on the time correction results, the data items are aligned to event time to obtain a multi-source data sequence, including: The first preset time period is used as the unit for dividing the event time alignment window, and the preset time point in the first preset time period is used as the starting time point for dividing the event time alignment window. The window is divided according to the starting division time point and the preset window length to obtain multiple event time aligned windows, and the event time aligned windows are matched with window numbers according to the starting division time point. The data items are aligned to each event time alignment window according to the window number and time correction result, and a multi-source data sequence is generated according to the data items in each event time alignment window at a second preset time interval.

7. The method according to claim 6, characterized in that, At a second preset time interval, a multi-source data sequence is generated based on the data items present in each event time alignment window, including: Based on the window number and the time correction result, determine the target number of data items that each event time alignment window can match; At a second preset time interval, determine the first current number of existing data items in each event time alignment window; at this moment, the state of existing data items in each event time alignment window is normal. In response to the first current number in any event time alignment window being less than the target number, the event time alignment window is controlled to enter a late waiting phase; At a third preset time interval, determine the second current number of existing data items (excluding data in the normal state) in each event time alignment window; at this moment, the existing data items (excluding data in the normal state) in each event time alignment window are in a late state. In response to the fact that the sum of the first current number and the second current number in any of the event time alignment windows is less than the target number, a data item in a missing state is determined, and the multi-source data sequence is generated based on the data items in a normal state, the data items in a late state, and the data items in a missing state.

8. The method according to claim 1, characterized in that, Real-time status assessment of multi-source data sequences to generate alarm events, including: Based on the window number of the event time alignment window in the multi-source data sequence, read the device monitoring point included in three or more consecutive window numbers, and the first window measurement value within the event time alignment window corresponding to the window number; Determine the threshold judgment condition corresponding to the detection point of the device, and determine whether the first window measurement value within each event time alignment window meets the threshold judgment condition; In response to the first window measurement value of any window number not meeting the threshold determination condition, the event time alignment window that does not meet the threshold determination condition is marked as an out-of-limit window; In response to the number of windows exceeding the limit being greater than a preset threshold, the alarm event is generated based on the equipment monitoring point, the equipment identifier of the industrial equipment associated with the equipment monitoring point, the measurement point identifier, the dimension identifier, the alarm type, the threshold judgment condition, the window number and the corresponding first window measurement value, and the alarm generation time.

9. The method according to claim 1, characterized in that, The method further includes: Based on the window number of the event time alignment window in the multi-source data sequence, read the device monitoring point included in the window number, and the second window measurement value within the event time alignment window corresponding to the window number; Generate a sequence of measurement values ​​based on the second window's measurement values, and generate a feature summary of the event time-aligned window corresponding to the window number based on the sequence of measurement values; and / or The feature summary of the event time alignment window and the multi-source data sequence are uploaded to the cloud server so that the cloud server can normalize and aggregate the multi-source data sequence and feature summary based on the unified semantic identifier, and perform cross-scenario fusion analysis on the normalized and aggregated data to obtain global monitoring results.

10. The method according to claim 9, characterized in that, Based on the measured value sequence, a feature summary of the event time-aligned window corresponding to the window number is generated, including: The mean, standard deviation, peak value, and rate of change of the event time alignment window corresponding to the window number are determined based on the measured value sequence. Based on the measured value sequence and the mean, a mean-reduced sequence is determined, and a discrete Fourier transform is performed on the mean-reduced sequence to obtain the frequency domain energy of the event time-aligned window corresponding to the window number. Based on the mean, standard deviation, peak value, rate of change, and frequency domain energy, a feature summary of the event time-aligned window corresponding to the window number is generated.

Images