A method for real-time monitoring of operating state data of field devices in a production industry

Abstract

This invention relates to the field of data monitoring technology and discloses a method for real-time monitoring of the operating status data of on-site equipment in the production industry. The method includes: acquiring production-level master data; constructing a regional list table and a collection point list table; establishing a dynamic mapping mechanism to filter the effective collection point set; completing adaptive frequency conversion acquisition and data quality verification; generating standard data messages and sending them to a message queue; caching the effective information from the regional list table and the collection point list table to edge-side memory; differentiating anomaly types to complete differentiated data repair and full standardization processing; and encapsulating the processed data messages; performing dual-dimensional anomaly identification combining fixed absolute safety thresholds and adaptive dynamic thresholds based on operating conditions; implementing hierarchical control and root cause convergence for identified alarms; completing the closed-loop binding of the entire process from alarm triggering to handling and archiving; providing feedback to optimize the core configuration and processing rules of the first three steps; and updating the collection point list table to build an equipment fault knowledge base.

Description

Technical Field

[0001] This invention relates to the field of data monitoring technology, specifically to a method for real-time monitoring of the operating status data of on-site equipment in the production industry. Background Technology

[0002] In the field of equipment operation status monitoring in the production industry, existing technologies are based on the traditional manual inspection mode and have evolved into information-based mobile inspection and monitoring solutions. Both types of solutions have systemic defects in adaptability, real-time performance, accuracy, and closed-loop operation, and cannot meet the refined management and control needs of multi-factory and multi-level production scenarios.

[0003] In existing technologies, data acquisition configurations are fixed and rigid, master data is not synchronized with production organization adjustments, and regions and acquisition points are statically bound. Manual configuration modifications are required during production line changes and equipment maintenance, easily leading to invalid data redundancy and monitoring blind spots. Furthermore, the use of fixed sampling frequencies fails to balance resource consumption and fault feature capture, and the lack of sensor status verification mechanisms makes it easy to confuse sensor malfunctions with equipment operational abnormalities, making it impossible to proactively avoid false alarms. Data preprocessing capabilities are weak, lacking cross-device time-series alignment mechanisms; misalignment of data from multiple acquisition points can easily lead to the failure of correlation analysis, particularly for missing values. Outliers are directly discarded, which easily leads to the loss of early fault characteristics of equipment; the full data cloud processing mode introduces additional transmission and processing latency, which cannot meet the low-latency monitoring requirements of industrial sites; anomaly identification relies on fixed thresholds, which cannot be dynamically adapted to fluctuations in operating conditions and equipment aging, and the false alarm and missed alarm rates continue to rise with the operating time of equipment; it can identify faults that exceed parameter limits, but cannot capture early deterioration signals of equipment; at the same time, it lacks an alarm convergence mechanism, which can easily trigger alarm storms; the monitoring and operation and maintenance processes are completely disconnected, and no closed loop for handling is formed; the system cannot achieve self-optimization and iteration, and the long-term operating effect continues to decline.

[0004] Therefore, there is a need to provide a method for real-time monitoring of the operating status data of on-site equipment in the production industry. Summary of the Invention

[0005] The purpose of this invention is to provide a method for real-time monitoring of the operating status data of on-site equipment in the production industry. To solve the above-mentioned problems in the prior art, this invention achieves this through the following technical solution:

[0006] In a first aspect, the present invention provides a method for real-time monitoring of the operating status data of on-site equipment in the production industry, which specifically includes the following steps:

[0007] Step 1: Synchronize production-level master data with unique identifiers from the enterprise master data platform, construct regional list table and collection point list table, establish a dynamic mapping mechanism to filter the effective collection point set, complete adaptive frequency conversion collection and data quality verification by grouping by device ID, and generate standard data messages to send to the message queue;

[0008] Step 2: Cache the valid information of the area list table and the collection point list table into the edge memory, continuously subscribe to the standard data packets in the message queue, parse them and group them by device ID and collection timestamp, perform single-device and cross-device two-level spatiotemporal alignment, distinguish the anomaly type to complete differentiated data repair and full standardization processing, and encapsulate them into processed data packets;

[0009] Step 3: Based on the processed data message, perform dual-dimensional anomaly identification by combining fixed absolute safety threshold and working condition adaptive dynamic threshold, simultaneously complete the trend anomaly judgment of the continuous collection cycle, perform hierarchical control and root cause convergence on the identified alarms, and complete the type-oriented hierarchical data push.

[0010] Step 4: Automatically generate standardized operation and maintenance work orders, complete the closed-loop binding of the entire process from alarm triggering to handling and archiving, and use the archived handling records to feed back and optimize the core configuration and processing rules of the first three steps, synchronously update the collection point list to build an equipment fault knowledge base, and regularly generate equipment operation status analysis reports.

[0011] Secondly, the real-time monitoring system for the operating status data of on-site equipment in the production industry provided by the embodiments of the present invention specifically includes the following modules:

[0012] Verification module: Synchronizes production-level master data with unique identifiers from the enterprise master data platform, constructs regional list table and collection point list table, establishes a dynamic mapping mechanism to filter the effective collection point set, completes adaptive frequency conversion collection and data quality verification by grouping by device ID, and generates standard data messages to be sent to the message queue;

[0013] Repair module: Caches the valid information of the area list table and the collection point list table into the edge memory, continuously subscribes to the standard data packets in the message queue, parses them and stores them in groups according to device ID and collection timestamp, performs two-level spatiotemporal alignment for single device and cross device, distinguishes the anomaly type to complete differentiated data repair and full standardization processing, and encapsulates them into processed data packets;

[0014] Push module: Based on the processed data message, it performs dual-dimensional anomaly identification by combining fixed absolute safety threshold and working condition adaptive dynamic threshold, simultaneously completes trend anomaly judgment for continuous collection cycle, performs hierarchical control and root cause convergence for identified alarms, and completes type-oriented hierarchical data push.

[0015] Generation module: Automatically generates standardized operation and maintenance work orders, completes the closed-loop binding of the entire process from alarm triggering to handling and archiving, and optimizes the core configuration and processing rules of the first three steps based on the archived handling records. It also synchronously updates the collection point list to build an equipment fault knowledge base and regularly generates equipment operation status analysis reports.

[0016] The beneficial effects of this invention are:

[0017] 1. Construct a dynamic mapping and adaptive frequency conversion acquisition mechanism with dual lists linked to master data to address the core pain points of existing technologies, such as static binding of acquisition configurations, manual modification of configurations required for production line changes / equipment maintenance, inability of fixed sampling frequencies to balance resource consumption and fault feature capture, and confusion between sensor faults and equipment anomalies. This is achieved by adding regional lists with new operating status and permission matching fields, and collection point lists with new semantic standardization and mechanism association fields. Effective collection points are dynamically filtered, and the basic sampling frequency is differentiated according to equipment fault mechanisms. The sampling frequency is adaptively adjusted based on operating condition fluctuations, and false alarms are avoided through mechanism cross-validation and sensor status monitoring. A two-level spatiotemporal alignment and differentiated data preprocessing workflow is designed to address the pain points of existing technologies, such as cross-device temporal misalignment leading to failed correlation analysis, direct discarding of abnormal / missing data resulting in loss of early fault features, and excessive latency in full cloud processing. Through two-level timestamp alignment within and across devices, false sensor anomalies and genuine equipment anomalies are accurately distinguished, and differentiated interpolation repair and original feature preservation are performed. Non-core data aggregation processing is completed at the edge, balancing data integrity and real-time processing.

[0018] 2. A dual-dimensional anomaly identification and hierarchical convergence targeted push mode is proposed to address the pain points of existing technologies, such as fixed thresholds drifting with equipment aging, only being able to identify out-of-limit faults, alarm storms overwhelming core faults, and front-end lag caused by full push. Through dual-dimensional identification of absolute safety red lines and adaptive dynamic thresholds, continuous trend anomaly warnings are completed for 5 consecutive cycles. Root cause alarm convergence filtering and derived alarms are executed, and targeted hierarchical push is performed based on user permissions, adapting to weak network and network outage scenarios on mobile devices. A full-process self-optimization system driven by a closed-loop operation and maintenance system is built to address the pain points of existing technology's disconnect between monitoring and operation and maintenance, long-term system performance degradation, and the inability to accumulate and reuse operation and maintenance experience. An operation and maintenance work order is automatically generated by alarms to form a closed-loop handling system. Based on archived handling records, the entire process configuration is optimized, an equipment fault knowledge base is built, and the system achieves self-iteration throughout its entire lifecycle. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of the steps of a method for real-time monitoring of the operating status data of on-site equipment in the production industry, provided in Embodiment 1 of the present invention.

[0021] Figure 2This is a schematic diagram of the structure of a real-time monitoring system for the operating status data of on-site equipment in the production industry, provided in Embodiment 2 of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0023] Example 1: As Figure 1 As shown in the figure, an embodiment of the present invention provides a method for real-time monitoring of the operating status data of field equipment in the production industry, which specifically includes the following steps:

[0024] Step 1: Synchronize production-level master data with unique identifiers from the enterprise master data platform, construct regional list table and collection point list table, establish a dynamic mapping mechanism to filter the effective collection point set, complete adaptive frequency conversion collection and data quality verification by grouping by device ID, and generate standard data messages to send to the message queue;

[0025] In a specific embodiment, the production-level master data is fully and updated in real time from the enterprise master data platform. The production-level master data includes: uniquely identified branch plant ID, workshop ID, production line ID, equipment ID, personnel permission information, production plan time period, and equipment maintenance / replacement work order information.

[0026] Based on synchronized production-level master data, two sets of configuration lists are constructed, and all configuration information is synchronized and updated in real time with the master data:

[0027] The first configuration list is a regional list table, which combines the branch factory ID, workshop ID, production line ID and regional name, and adds three core fields: regional operating status, regional effective production time period and job permission matching. The regional operating status automatically synchronizes production plan and work order information, and is divided into four categories: normal production, equipment maintenance, production line changeover and downtime waiting for materials. The job permission matching field stores the set of job personnel IDs that can be accessed in the corresponding region, so as to achieve precise binding between region and personnel permissions.

[0028] The second set of configuration lists is a list of data collection points. It combines the area ID, data collection point address, data collection point description, and alarm upper and lower limits, and adds four core fields: data collection point semantic standardization, data collection point effective status, equipment fault mechanism association, and working condition adaptation label. Among them, the data collection point semantic standardization is defined in accordance with the industry's unified specifications, with the order format of equipment type, parameter type, monitoring location, and data unit.

[0029] The valid status of the data collection point is automatically updated according to the regional operating conditions, and is divided into four categories: normal activation, maintenance shielding, type change shutdown, and fault disable. The equipment fault mechanism association field marks the core fault mode of the equipment corresponding to the data collection point, distinguishing between key data collection points and auxiliary data collection points. The operating condition adaptation label field marks the sensitive operating condition types of the data collection point, including: start-up and shutdown, load increase, load decrease, roller change, and type change.

[0030] Based on two sets of configuration lists, a real-time dynamic mapping and matching mechanism is established to automatically filter the set of valid collection points corresponding to each region and each logged-in user:

[0031] Collection points that simultaneously meet the three triggering conditions are included in the effective collection and push scope; otherwise, they are automatically blocked. The triggering conditions include: the region to which the collection point belongs is within the effective time period of the production plan, the currently logged-in user has access permissions to the region, and the collection point is in an effective enabled state under the current working conditions in the region.

[0032] For example, after a production line initiates a changeover work order, the system automatically updates the corresponding area's operating conditions to the production line changeover, automatically blocks non-changeover-compatible data collection points, and retains the core data collection points related to the changeover; after equipment submits a maintenance work order, it automatically blocks non-maintenance-related data collection points, completing the filtering of invalid data and avoiding invalid data from occupying bandwidth and storage resources.

[0033] After the data acquisition program starts, it reads the set of valid acquisition points, groups them by equipment ID, establishes a stable acquisition channel with the on-site SMC system, obtains real-time operating data of the corresponding equipment and acquisition points, and synchronously acquires real-time equipment operating conditions, production load, and process setpoints. It also performs dynamic adaptive adjustment of the acquisition frequency and data quality verification.

[0034] Based on the device fault mechanism association field of the collection points, a basic sampling frequency is set: for key collection points associated with the core fault mode of the device, the basic sampling frequency is set to once per second; for non-core auxiliary collection points, the basic sampling frequency is set to once every 5 seconds, so as to avoid resource waste caused by all collection points collecting at the same frequency.

[0035] Based on the real-time fluctuations in equipment operating conditions and the degree of data anomalies, the real-time sampling frequency is dynamically adjusted: when the equipment is in a state of drastic fluctuation or when the collected data shows obvious anomalies, the sampling frequency is automatically increased to ensure complete capture of the weak characteristics of early equipment failures; when the equipment is in stable production and the load fluctuation range is within [-5%, 5%], the sampling frequency is automatically decreased to reduce resource consumption under normal operating conditions.

[0036] During each data acquisition process, data quality cross-validation and sensor status monitoring are performed simultaneously: for data acquisition points that are physically related within the same device, such as rolling mill current, torque and furnace displacement parameters; the rationality of the data is verified through the physical mechanism of the device. If the data of a single data acquisition point deviates from the range of the mechanism correlation, it is marked as a state to be verified.

[0037] The system synchronously collects the power supply voltage and signal strength data of the sensors to determine the sensor's own operating status. If the sensor loses power, the signal is interrupted, or the power supply is abnormal, the corresponding collection point is marked as fault disabled and a sensor fault alarm is pushed, thus completely solving the problem of false alarms caused by confusion between sensor faults and abnormal equipment operation.

[0038] The data acquisition program combines each device ID, the IDs of all valid data collection points under the corresponding device, the data collection value, the data collection timestamp, the data quality label, and the operating condition label into a standard data message, and sends it into the message queue to complete a complete data acquisition cycle.

[0039] Step 2: Cache the valid information of the area list table and the collection point list table into the edge memory, continuously subscribe to the standard data packets in the message queue, parse them and group them by device ID and collection timestamp, perform single-device and cross-device two-level spatiotemporal alignment, distinguish the anomaly type to complete differentiated data repair and full standardization processing, and encapsulate them into processed data packets;

[0040] In a specific embodiment, the processing is completed at the edge by combining the standard data packets collected in step one and sent to the message queue, the area list table cached in memory, and the collection point list table with all valid configuration information, thus avoiding processing delays caused by uploading all data to the cloud.

[0041] When the background service starts, it caches the valid information of the area list table and the collection point list table in the edge side memory. At the same time, it continuously subscribes to the message queue, receives standard data packets sent by the collection program, parses the packets, groups them by device ID and collection timestamp, and stores them in the local cache on the edge side.

[0042] To address the timing discrepancy issue that is easily overlooked in conventional solutions, a two-level spatiotemporal alignment process is implemented to ensure the timing consistency of data across the entire production line and all equipment:

[0043] The first level is timestamp alignment of collection points within a single device: using the start timestamp of each collection cycle as the reference, all collection point data within the same cycle and under the same device are uniformly marked as the reference timestamp of that time window, thus solving the problem of timing asynchrony caused by the collection time deviation of different collection points;

[0044] The second level is the alignment of linkage points across equipment and regions: For the equipment collection points related to upstream and downstream processes across the entire production line, the unified system time of the enterprise master data platform is used as the benchmark to perform unified alignment of timestamps across equipment, ensuring the consistency of linkage data between upstream and downstream processes and solving the problem of correlation analysis errors caused by time sequence misalignment;

[0045] For the aligned time-series data, differentiated processing is performed based on the type of data anomaly to ensure data integrity and avoid covering key fault characteristics:

[0046] For missing data at a single acquisition point within a time window, interpolation repair based on device mechanism association is adopted: real-time valid values ​​of reference acquisition points that have a clear physical mechanism association with the corresponding missing acquisition point are selected, and the repair value of the missing data is calculated by combining the fixed mechanism association ratio determined by the device's factory design parameters, thereby completing the data completion and avoiding time series breaks caused by missing data.

[0047] For abnormal data marked as pending verification, the abnormality type is first distinguished through mechanism correlation verification: if it is determined to be a false abnormality caused by transient sensor interference, the interpolation repair method is used to complete the repair.

[0048] If the anomaly is determined to be a genuine anomaly caused by a change in the equipment's operating status, the original anomaly value is retained, marked as anomaly feature data, and no repair processing is performed to ensure that the subtle features of early equipment failures are not covered up.

[0049] For the valid data after repair, perform full standardization processing: unify the data units, formats and naming rules of all collection points according to the industry standards corresponding to the semantic standardization fields of the collection points;

[0050] For non-core data collection points under normal operating conditions, data aggregation is performed at the edge, and only the aggregated feature values ​​are sent to the subsequent processing stage, reducing the computing power consumption at the edge and the pressure on subsequent data processing, and avoiding the latency problem caused by full data processing in conventional solutions.

[0051] The processed device ID, collection point ID, collection value, repair mark, anomaly mark, reference timestamp and corresponding alarm configuration information are grouped and encapsulated into processed data packets by device ID;

[0052] Step 3: Based on the processed data message, perform dual-dimensional anomaly identification by combining fixed absolute safety threshold and working condition adaptive dynamic threshold, simultaneously complete the trend anomaly judgment of the continuous collection cycle, perform hierarchical control and root cause convergence on the identified alarms, and complete the type-oriented hierarchical data push.

[0053] In a specific embodiment, a dual-dimensional anomaly recognition mode is adopted to achieve accurate alarm for out-of-limit faults and early warning for early anomalies, adapting to scenarios of frequent fluctuations in operating conditions and full life cycle operation of equipment in the production industry.

[0054] Specifically, the first dimension is emergency fault identification with a fixed safety red line: based on the equipment's factory design parameters, industry safety standards and historical fault data, a fixed absolute safety threshold range is set for each data collection point, including an absolute upper limit and an absolute lower limit. The absolute safety threshold is the safety red line for equipment operation. If the real-time data exceeds the absolute safety threshold range, the highest level of emergency alarm is triggered to ensure that the core safety risks of the equipment are identified.

[0055] The second dimension is dynamic threshold anomaly identification that adapts to operating conditions: within the absolute safety threshold range, a dynamically adjusted alarm threshold range is set for each data acquisition point, and the alarm threshold is updated in real time according to the on-site operating conditions.

[0056] The specific adjustment rules are as follows: Based on the theoretical parameter range set by the current production process, the alarm threshold range is adjusted in combination with the real-time production load of the equipment. The higher the actual production load, the more the alarm threshold range should be adjusted accordingly.

[0057] The alarm threshold offset is corrected based on the equipment's full life cycle runtime. The longer the equipment runs, the larger the corresponding threshold range is corrected. This adapts to the normal parameter offset caused by equipment aging and solves the problem of soaring false alarm and missed alarm rates after long-term operation of fixed alarm thresholds.

[0058] Synchronous execution of early warning identification of abnormal trends: For time series data of 5 consecutive collection cycles, it is determined whether the data change trend is consistent with the normal operation of the equipment. If the data shows a continuous unidirectional upward / downward trend, and does not exceed the dynamic alarm threshold range, but the trend slope exceeds the maximum range of historical normal operation, an early abnormal warning is directly triggered, solving the core defect of being unable to capture early equipment degradation.

[0059] For identified abnormal alarms, implement tiered control and root cause convergence to resolve the alarm storm problem that is prone to occur:

[0060] Based on the degree of data anomaly and the level of fault risk, alarms are divided into four levels: data exceeding the absolute safety red line is classified as an emergency alarm; data exceeding the dynamic threshold by more than 30% is classified as an important alarm; data exceeding the dynamic threshold by less than 30% is classified as a general alarm; and data with abnormal trends but not exceeding the threshold is classified as a level four early warning. Different alarm levels correspond to different push and handling rules.

[0061] When multiple alarms occur within the same device, the alarm with the highest alarm level is designated as the root alarm, and the other lower-level alarms triggered at the same time are designated as derivative alarms; when multiple devices in the same area experience cascading alarms, the highest-level alarm that is triggered first in sequence is designated as the root alarm, and the other alarms triggered subsequently are designated as derivative alarms.

[0062] Derivative alarms are only recorded in the background and are not actively pushed to the front end. They are displayed when the user actively views them, thus avoiding the problem of alarm storms overwhelming the core fault when an anomaly occurs.

[0063] Establish a long-lived WebSocket connection with the mobile frontend, abandoning the conventional approach of pushing all data across the entire connection, and adopting a targeted push strategy based on permissions and data types, while also adapting to weak network environments on the mobile device in the field:

[0064] When a long connection is established, the identity verification and permission matching of the logged-in user are completed first. The valid region and valid collection point set corresponding to the user are obtained. The mapping relationship between user ID-Websocket connection-valid collection point set is established. Only valid data within the user's permission range is pushed to the user, reducing bandwidth consumption and the risk of front-end page lag.

[0065] For different types of data, a differentiated push strategy is adopted: For root cause alarm data, a real-time priority push mode is adopted. Once triggered, it is immediately pushed to all mobile front-ends with permissions, and the alarm level, abnormal location and abnormal description are marked simultaneously to ensure that core fault information is delivered.

[0066] For normal equipment operation data, an aggregated push mode is adopted, which aggregates parameters of the same device and the same type into a single message for push, rather than pushing a single message from a single collection point, thereby further reducing the amount of data pushed.

[0067] For scenarios involving weak or disconnected mobile devices in industrial settings, adaptive keep-alive and resume transmission after network outages are implemented: when the network is in good condition, a heartbeat interval of 30 seconds is set to maintain a stable long connection; when the network is in poor condition, the heartbeat interval is automatically shortened to 10 seconds to ensure uninterrupted connection.

[0068] When a network disconnection occurs, all device operation data and alarm data during the disconnection period are cached locally on the edge side. After the network is restored, the latest real-time data and the highest level alarm data are pushed first, and then the historical data during the disconnection period is retransmitted to avoid network congestion and front-end lag caused by pushing all data when the network is restored.

[0069] After receiving the push data, the front-end mobile terminal displays the device's real-time operating data on the corresponding area page according to the division of the area list table. The display is differentiated according to the alarm level and dynamic threshold range: data above the dynamic alarm upper limit is displayed in red, data below the dynamic alarm lower limit is displayed in blue, data within the normal range is displayed in the default style, and early warning data is displayed in yellow.

[0070] Authorized users can click on the corresponding data collection point to view the threshold parameters and data trend curves of the point, and can modify the absolute safety threshold and process theoretical parameters of the point. The modified information will be updated synchronously to the data collection point list.

[0071] Step 4: Automatically generate standardized operation and maintenance work orders, complete the closed-loop binding of the entire process from alarm triggering to handling and archiving, and use the archived handling records to feed back and optimize the core configuration and processing rules of the first three steps, synchronously update the collection point list to build an equipment fault knowledge base, and regularly generate equipment operation status analysis reports;

[0072] In a specific embodiment, a closed-loop mechanism is established to handle and archive alarms. When the system triggers a level 2 or higher alarm, a standardized maintenance work order is automatically generated. The work order fully includes the device ID, anomaly collection point information, alarm level, anomaly data, historical trends, and related fault mechanism information. Based on the field matching according to the regional job authority, the work order is assigned to the corresponding equipment maintenance personnel. After receiving the work order, the maintenance personnel perform on-site handling and submit a complete handling record after the handling is completed. The complete handling record includes: the root cause of the fault, the handling method, the handling result, and the equipment recovery status. The system binds the handling record with the original alarm event and archives it, completing a complete alarm handling closed loop.

[0073] Based on archived operation and maintenance records, the system automatically feeds back and optimizes all core configurations and processing rules in the first three steps, achieving self-optimization and iteration throughout the entire system process and solving the problem of performance degradation after long-term system operation.

[0074] For data collection points that trigger false alarms multiple times, the system automatically analyzes the causes of the false alarms and optimizes and updates the working condition adaptation tags and mechanism association fields of the data collection points; for data collection points that experience multiple failures but do not provide early warnings, the system automatically increases their basic sampling frequency to ensure that key fault characteristics are fully captured; for locations where sensor failures are frequent, the system automatically pushes sensor verification and replacement reminders.

[0075] For confirmed real fault events, extract equipment operation data, operating condition data, and threshold parameters before the fault occurs, optimize the dynamic threshold adaptation rules, and improve the accuracy of anomaly identification; for confirmed early equipment degradation events, optimize the judgment rules for trend anomaly identification and improve the accuracy of early warning; for false alarm events, adjust the threshold adaptation rules accordingly to reduce the false alarm rate.

[0076] Based on the handling habits and job requirements of maintenance personnel, the push strategy is automatically optimized. For users with different positions and permissions, the granularity of push data and the priority of alarm push are adjusted to ensure that the information received by users matches their own responsibilities and avoid interference from irrelevant information.

[0077] The archived fault events, handling records, and root cause analysis information are synchronously updated to the equipment fault mechanism association field of the collection point list table to build an equipment fault knowledge base. When similar abnormal alarms occur in the future, the system will automatically push the corresponding fault root cause analysis and handling suggestions to assist maintenance personnel in quickly completing fault handling.

[0078] Based on full data of equipment operation, alarm records, handling records, and production plan data, regularly generate equipment operation status analysis reports, including equipment health assessment, root cause analysis of high-frequency failures, maintenance plan suggestions, and production process optimization directions, and submit them to equipment management and production decision-making personnel.

[0079] Example 2: As Figure 2 As shown in the figure, the real-time monitoring system for the operating status data of field equipment in the production industry provided by this embodiment of the invention specifically includes the following modules:

[0080] Verification module: Synchronizes production-level master data with unique identifiers from the enterprise master data platform, constructs regional list table and collection point list table, establishes a dynamic mapping mechanism to filter the effective collection point set, completes adaptive frequency conversion collection and data quality verification by grouping by device ID, and generates standard data messages to be sent to the message queue;

[0081] Repair module: Caches the valid information of the area list table and the collection point list table into the edge memory, continuously subscribes to the standard data packets in the message queue, parses them and stores them in groups according to device ID and collection timestamp, performs two-level spatiotemporal alignment for single device and cross device, distinguishes the anomaly type to complete differentiated data repair and full standardization processing, and encapsulates them into processed data packets;

[0082] Push module: Based on the processed data message, it performs dual-dimensional anomaly identification by combining fixed absolute safety threshold and working condition adaptive dynamic threshold, simultaneously completes trend anomaly judgment for continuous collection cycle, performs hierarchical control and root cause convergence for identified alarms, and completes type-oriented hierarchical data push.

[0083] Generation module: Automatically generates standardized operation and maintenance work orders, completes the closed-loop binding of the entire process from alarm triggering to handling and archiving, and optimizes the core configuration and processing rules of the first three steps based on the archived handling records. It also synchronously updates the collection point list to build an equipment fault knowledge base and regularly generates equipment operation status analysis reports.

[0084] The above provides a detailed description of one embodiment of the present invention, but the content described is only a preferred embodiment of the present invention and should not be considered as limiting the scope of the present invention. The above formulas are all dimensionless numerical calculations, and the formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world situation. The preset parameters in the formulas are set by those skilled in the art based on actual conditions and historical experience, and can be adjusted according to actual conditions. The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. All equivalent changes and improvements made in accordance with the scope of the present invention should still fall within the patent coverage of the present invention.

Claims

1. A method for real-time monitoring of on-site equipment operating status data in a production industry, characterized in that, Includes the following steps: Synchronize production-level master data with unique identifiers from the enterprise master data platform, construct regional list tables and collection point list tables, establish a dynamic mapping mechanism to filter the effective collection point set, complete adaptive frequency conversion collection and data quality verification by grouping by device ID, and generate standard data messages to send to the message queue. The method for establishing the dynamic mapping mechanism is as follows: Synchronize the production-level master data with unique identifiers from the enterprise master data platform in real time and fully update it. Combine the master data to build regional list table and collection point list table. The regional list table adds fields for regional working condition status, regional production effective time period and job permission matching. The collection point list table adds fields for collection point semantic standardization, collection point effective status, equipment failure mechanism association and working condition adaptation label. Establish a real-time dynamic mapping and matching mechanism. Collection points that meet three conditions simultaneously are included in the effective collection and push scope: the area where the collection point belongs is within the effective period of the production plan, the currently logged-in user has regional access permissions, and the collection point is in an effective enabled state under the current regional working conditions. Otherwise, they are automatically blocked. The effective information of the regional list table and the collection point list table is cached in the edge memory. The standard data packets in the message queue are continuously subscribed to, parsed and grouped by device ID and collection timestamp. The spatiotemporal alignment is performed at two levels: single device and cross device. Differentiated data repair and full standardization processing are completed by distinguishing the anomaly type and encapsulating it into processed data packets. Based on the processed data packets, a two-dimensional anomaly identification is performed, combining a fixed absolute safety threshold and an adaptive dynamic threshold for operating conditions. The trend anomaly judgment of the continuous collection cycle is completed simultaneously. The identified alarms are subject to hierarchical control and root cause convergence, and the data is pushed in a categorized, targeted, and hierarchical manner. It automatically generates standardized operation and maintenance work orders, completes the closed-loop binding of alarm triggering to handling and archiving, and optimizes the core configuration and processing rules of the first three steps based on the archived handling records. It also synchronously updates the collection point list to build an equipment fault knowledge base and regularly generates equipment operation status analysis reports.

2. The method for real-time monitoring of on-site equipment operating status data in the production industry according to claim 1, characterized in that, The adaptive frequency conversion acquisition method is as follows: The data acquisition program reads the set of valid data acquisition points, groups them by equipment ID, establishes a data acquisition channel with the on-site SMC system, and obtains the real-time operating data, real-time operating status of the equipment, production load and process setpoints of the corresponding equipment and data acquisition points. By combining the equipment failure mechanism association fields of the collection points, a basic sampling frequency of once per second is set for key collection points, and a basic sampling frequency of once every 5 seconds is set for non-core auxiliary collection points. The real-time sampling frequency is dynamically adjusted based on the real-time operating condition fluctuations of the equipment and the degree of data anomalies. Data quality cross-validation and sensor status monitoring are completed synchronously during each collection process, and finally, a standard data message is generated and sent to the message queue.

3. The method for real-time monitoring of on-site equipment operating status data in the production industry according to claim 1, characterized in that, The method for two-level spatiotemporal alignment is as follows: When the background service starts, it caches the valid information of the area list table and the collection point list table in the edge side memory, continuously subscribes to the message queue to receive standard data packets, and after parsing the packets, it groups and stores them in the local cache on the edge side according to the device ID and collection timestamp. Using the start timestamp of each collection cycle as a reference, all data points collected from the same device within the same cycle are uniformly marked as the reference timestamp of that time window, thus completing the timestamp alignment of collection points within a single device. Based on the unified system time of the enterprise master data platform, the timestamps of the equipment collection points associated with upstream and downstream processes across the entire production line are uniformly aligned.

4. The method for real-time monitoring of on-site equipment operating status data in the production industry according to claim 1, characterized in that, The method for repairing the differential data is as follows: For the aligned time-series data, for missing data at a single acquisition point within the time window, the real-time valid value of a reference acquisition point that has a clear physical mechanism association with the missing acquisition point is selected. Combined with the fixed mechanism association ratio determined by the equipment's factory design parameters, the repair value of the missing data is calculated to complete the data filling. For abnormal data marked as pending verification, the abnormality type is distinguished by mechanism correlation verification. False abnormalities caused by instantaneous sensor interference are repaired by interpolation. For real abnormalities caused by changes in equipment operating status, the original abnormal values ​​are retained and marked as abnormal feature data. Full standardization processing is performed on the restored valid data, and data aggregation processing is completed on the edge side for non-core collection point data under normal operating conditions.

5. The method for real-time monitoring of on-site equipment operating status data in the production industry according to claim 1, characterized in that, The method for two-dimensional anomaly identification is as follows: Based on the equipment's factory design parameters, industry safety standards, and historical fault data, a fixed absolute safety threshold range containing absolute upper and lower limits is set for each data collection point. When real-time data exceeds the absolute safety threshold range, the highest level emergency alarm is triggered. Within the absolute safety threshold range, a dynamically adjustable alarm threshold range is set for each data collection point. Based on the theoretical parameter range set for the current production process, the alarm threshold range is adjusted in conjunction with the real-time production load of the equipment, and the alarm threshold offset is corrected according to the equipment's full life cycle runtime.

6. The method for real-time monitoring of operating condition data of field devices in a process industry according to claim 1, wherein, The method for judging trend anomalies is as follows: For time-series data collected over five consecutive collection cycles, the system determines whether the data trend conforms to the normal operating pattern of the equipment. For collection points where the data shows a continuous upward or downward trend, does not exceed the dynamic alarm threshold range, or has a trend slope exceeding the maximum range of historical normal operation, an early anomaly warning is directly triggered.

7. The method for real-time monitoring of operating condition data of field devices in a process industry according to claim 1, wherein, The method for hierarchical control and root cause convergence of alarms is as follows: Based on the degree of data anomaly and the level of fault risk, alarms are divided into four levels: emergency alarm, important alarm, general alarm, and early warning. When multiple alarms occur in the same device, the alarm with the highest alarm level is the root alarm, and the other lower-level alarms triggered at the same time are marked as derivative alarms. When multiple devices in the same area trigger cascading alarms, the highest-level alarm that is triggered first in the sequence is designated as the root alarm, and the remaining alarms that are triggered subsequently are designated as derivative alarms. Derivative alarms are only recorded in the background and are not actively pushed to the front end.

8. The method for real-time monitoring of on-site equipment operating status data in the production industry according to claim 1, characterized in that, The method for targeted and hierarchical data push is as follows: Establish a long WebSocket connection with the mobile frontend. When the connection is established, complete the identity verification and permission matching of the logged-in user, obtain the valid region and valid collection point set of the corresponding user, and establish a mapping relationship between user ID, WebSocket connection and valid collection point set. Real-time priority push mode is adopted for root cause alarm data, and aggregated push mode is adopted for normal device operation data; an adaptive heartbeat packet interval is set to maintain long connection stability. When the network is disconnected, the device operation data and alarm data during the disconnection period are cached locally on the edge side, and the data is pushed according to priority after the network is restored.

9. A method for real-time monitoring of on-site equipment operating status data in a production industry according to claim 1, characterized in that, The method for full-process closed-loop binding is as follows: When the system triggers a level 2 or higher alarm, it automatically generates a standardized operation and maintenance work order containing the device ID, abnormal data collection point information, alarm level, abnormal data, historical trends, and related fault mechanism information. The work order is then assigned to the corresponding equipment operation and maintenance personnel based on the field matching of the region's job permissions. After completing on-site handling, maintenance personnel submit a handling record that includes the root cause of the fault, handling method, handling result, and equipment recovery status. The system will bind and archive the handling record with the original alarm event.

Images