Industrial operation data visual display method and system based on artificial intelligence platform

Abstract

The invention discloses an industrial operation data visualization display method and system based on an artificial intelligence platform, and relates to the technical field of industrial equipment operation data visualization. Multi-dimensional operation data of industrial equipment are collected and associated identifiers are marked, and a part physical association relation table and a global unique identifier are established; binding the data and the global unique identifier, converting the causal analysis conclusion of the artificial intelligence platform into a global unique identifier association chain, generating a numerical label and a thermodynamic diagram based on the binding relationship and the global unique identifier association chain, and performing highlight labeling and global unique identifier association chain display when the fault data reaches the standard. According to the method, the relevance and intuition of industrial equipment operation data visualization are improved, a user is helped to quickly grasp the equipment operation state and the abnormal cause, and the actual requirements of industrial operation and maintenance are met.

Description

Technical Field

[0001] This invention relates to the field of industrial equipment operation data visualization technology, specifically to a method and system for visualizing industrial operation data based on an artificial intelligence platform. Background Technology

[0002] Existing methods for visualizing industrial equipment operation data often involve relatively simple data collection types, making it difficult to comprehensively capture various aspects of the equipment's operational status. Furthermore, the collected data is not associated with the coupling relationships of different components of the industrial equipment, making it difficult to establish accurate and structurally consistent binding relationships between the data and the geometric objects of the industrial model. At the same time, visualization is often limited to the presentation of isolated operational data, failing to intuitively reflect the scope of the impact of equipment failures and the cause-effect relationship of abnormal data. This makes it difficult for users to quickly grasp the overall operational status of the equipment and the causes of anomalies, affecting operational efficiency and decision-making accuracy. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a method for visualizing industrial operation data based on an artificial intelligence platform, comprising the following steps: S1. Collect multi-dimensional operating data of industrial equipment, including numerical data, image data and acoustic data; label the multi-dimensional operating data of the same component or related components with association tags according to the coupling relationship of each component of the industrial equipment. S2. Establish a physical relationship table for each component of the industrial equipment, and assign globally unique identifiers to geometric objects at the macro, meso, and micro levels in the industrial model. Each globally unique identifier corresponds one-to-one with a component of the industrial equipment. S3. By combining the physical relationship table and the association identifiers of multi-dimensional operational data, establish the binding relationship between multi-dimensional operational data and globally unique identifiers; obtain the causal analysis conclusions output by the artificial intelligence platform, and transform the causal analysis conclusions into a globally unique identifier association chain; where the causal analysis conclusions are the cause-effect correspondence of abnormal data obtained by the artificial intelligence platform based on multi-dimensional operational data; S4. Based on the binding relationship and the association chain of globally unique identifiers, generate numerical labels containing current values, predicted values ​​and status indicators at the corresponding geometric positions of the industrial model; map multi-dimensional operating data into color gradients and generate a continuous heat map on the surface of the industrial model; when the fault-related data contained in the multi-dimensional operating data reaches the preset conditions, highlight the corresponding geometric area of ​​the industrial model and mark the corresponding globally unique identifier association chain.

[0004] Furthermore, after step S1 and before step S2, the following steps are also included: Receive data analysis feedback from the AI ​​platform based on multi-dimensional operational data output. The data analysis feedback includes data missing information, data noise information, and key data requirement information. Based on data analysis feedback, we adjusted the data collection strategy, supplemented the corresponding types of multi-dimensional operational data collection for missing data prompts, optimized the collection parameters of corresponding collection points for data noise prompts, and increased the frequency of multi-dimensional operational data collection for corresponding components for key data requirements prompts. The multi-dimensional operational data acquired after adjusting the acquisition strategy is preprocessed to extract time-series features and core features of image and acoustic data. The time-series features include data periodic fluctuation features and trend change features, while the core features include the appearance status features of components in the image and the abnormal frequency band features in the acoustic wave.

[0005] Furthermore, step S3 establishes the binding relationship between multi-dimensional operational data and globally unique identifiers, specifically including the following steps: By associating time-series features and core features with corresponding multi-dimensional operational data, combined data is formed. Based on the physical association table and association identifier, the combined data is bound to the corresponding globally unique identifier, so that the industrial model geometric object corresponding to each globally unique identifier is synchronously associated with the multi-dimensional operation data and time series features and core features of the industrial equipment components. The combined data is synchronized to the artificial intelligence platform to assist the platform in optimizing the causal analysis conclusions, and the globally unique identifier association chain is updated based on the optimized causal analysis conclusions.

[0006] Furthermore, the preprocessing of the collected multi-dimensional operational data also includes the following steps: Data filtering conditions are established based on the physical operation rules of industrial equipment. The physical operation rules include the correlation logic of the operating parameters of each component of the industrial equipment and the target operating range. The collected multi-dimensional operational data is filtered according to the data filtering criteria. Interference data that does not conform to the physical operation rules is filtered out, and valid data related to the operating status of industrial equipment and valid abnormal data that reflect potential anomalies are retained. The filtered valid data and valid abnormal data are associated with the extracted time-series features and core features to form a preprocessed data feature set, which is then bound to a globally unique identifier.

[0007] Furthermore, in step S3, the process of establishing the binding relationship between multi-dimensional operational data and globally unique identifiers also includes the following steps: Based on the physical association table and the association identifier of multi-dimensional operational data, the effective data, effective abnormal data, time series features and core features in the data feature set are associated with the corresponding globally unique identifier one by one, so that each globally unique identifier stores the complete data feature information of a single component or related components of industrial equipment. The bound globally unique identifier and the corresponding data feature set are synchronized to the artificial intelligence platform to assist the artificial intelligence platform in analyzing the operating status of industrial equipment and generating causal analysis conclusions. When the industrial model is visualized, the geometric object corresponding to the globally unique identifier is synchronously associated with and calls the time-series features and core features in the data feature set. When the user views the display information corresponding to the geometric object, the operating trend reflected by the time-series features and the component status details reflected by the core features are presented synchronously.

[0008] Furthermore, in step S4, when the fault-related data contained in the multi-dimensional operational data reaches a preset condition, the method further includes the following steps: Based on the physical association table, query the associated components corresponding to the target globally unique identifier bound to the fault-related data that meets the preset conditions; Extend the binding scope of the target globally unique identifier to the globally unique identifiers corresponding to the associated components, and establish an extended association chain of globally unique identifiers corresponding to the fault propagation path; In the industrial model, the target globally unique identifier is connected to the globally unique identifier of the associated component by a preset identification line, and the multi-dimensional operation data and status indicators corresponding to the extended association chain of the globally unique identifier are displayed simultaneously.

[0009] Furthermore, the method further includes the following steps: Based on the globally unique identifiers associated with the macro, meso, and micro levels of the industrial model, the globally unique identifiers are extended and the associated chains are synchronously mapped to the industrial model at each level. In the macro-level industrial model, the production line or equipment cluster corresponding to the globally unique identifier extended association chain is identified as a whole; in the meso-level industrial model, the equipment and related components corresponding to the globally unique identifier extended association chain are highlighted; in the micro-level industrial model, the geometric position and multi-dimensional operation data of each component in the globally unique identifier extended association chain are displayed. When a user clicks on the corresponding icon or label in the extended association chain of the industrial model at any level, they will be redirected to the industrial model at other related levels.

[0010] Furthermore, the method further includes the following steps: Based on the association chain of globally unique identifiers, the globally unique identifiers corresponding to the causal data in the causal analysis conclusion are labeled with causal identifiers, and the globally unique identifiers corresponding to the result data are labeled with result identifiers. Based on the strength of the correlation between cause and effect in the causal analysis conclusion, set the display style of the preset label line connecting cause and effect labels; On the industrial model, the labeled cause markers, result markers, and corresponding preset marker lines are displayed simultaneously. When the user clicks on the preset marker line, the basis for the formation of the causal analysis conclusion is presented.

[0011] Furthermore, after establishing the binding relationship between multi-dimensional operational data and globally unique identifiers, the following steps are also included: Based on the real-time operating conditions of industrial equipment, the mapping priority of the globally unique identifiers corresponding to each component is determined. When the equipment is under heavy load or in a critical process stage, the mapping priority of the globally unique identifiers corresponding to the core components is higher than that of other components. When the fault-related data of multiple components all reach the preset conditions, the mapping areas corresponding to each fault are distinguished by different styles of status identifiers on the industrial model according to the mapping priority and the association chain of the globally unique identifiers corresponding to each fault. Based on changes in the operating conditions or fault status updates of industrial equipment, the mapping priority and display style of the corresponding globally unique identifiers are adjusted in real time.

[0012] Secondly, the present invention also provides an industrial operation data visualization system based on an artificial intelligence platform, comprising: The data acquisition module is used to collect multi-dimensional operational data of industrial equipment, including numerical data, image data, and acoustic data; and to label the multi-dimensional operational data of the same or related components according to the coupling relationship between the various components of the industrial equipment. The allocation module is used to establish a physical relationship table for each component of industrial equipment. It assigns globally unique identifiers to geometric objects at the macro, meso, and micro levels in the industrial model, with each globally unique identifier corresponding one-to-one with a component of the industrial equipment. The module combines the physical relationship table and the association identifiers of multi-dimensional operational data to establish a binding relationship between multi-dimensional operational data and globally unique identifiers; it obtains the causal analysis conclusions output by the artificial intelligence platform and transforms the causal analysis conclusions into a globally unique identifier association chain; where the causal analysis conclusions are the cause-effect correspondence of abnormal data derived by the artificial intelligence platform based on multi-dimensional operational data. The generation module is used to generate numerical labels containing current values, predicted values, and status indicators at the corresponding geometric locations of the industrial model based on binding relationships and globally unique identifier association chains; map multi-dimensional operational data into color gradients to generate continuous heat maps on the surface of the industrial model; when the fault-related data contained in the multi-dimensional operational data reaches preset conditions, it highlights and marks the corresponding geometric areas of the industrial model and marks the corresponding globally unique identifier association chains.

[0013] The method provided by this invention has the following beneficial effects: This method collects multi-dimensional operational data from industrial equipment, comprising numerical, image, and acoustic data, and labels them with correlation identifiers based on component coupling relationships. Compared to single-type data collection, this method can more comprehensively capture the equipment's operating status and clearly define the component correlation attributes between data. By establishing a physical correlation table, globally unique identifiers corresponding to components are assigned to geometric objects at each level of the industrial model, achieving accurate binding of multi-dimensional operational data with the model's geometric location and adapting to multi-level operation and maintenance viewing needs at the macro, meso, and micro levels. Data binding relationships are established by combining the physical correlation table and correlation identifiers, and the causal analysis conclusions of the artificial intelligence platform are transformed into globally unique identifier correlation chains, enabling the cause-effect correspondence of abnormal data to be associated with the model's location. By generating numerical labels, heatmaps, and fault highlighting, and simultaneously displaying the globally unique identifier correlation chains, industrial equipment operational data is presented in an intuitive form. When fault-related data meets the criteria, the corresponding areas and correlations are clearly identified, helping users quickly grasp the equipment's operating status, locate abnormal parts, and understand the causes of abnormalities. This effectively achieves the visualization of industrial equipment operational data, providing support for equipment operation and maintenance. Attached Figure Description

[0014] Figure 1 A schematic diagram of the process for visualizing industrial operation data based on an artificial intelligence platform, provided in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the hierarchical association mapping and fault propagation between multi-dimensional operating data of industrial equipment and industrial models, provided in an embodiment of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0016] refer to Figure 1 This invention provides a method for visualizing industrial operation data based on an artificial intelligence platform, comprising the following steps: S1. Collect multi-dimensional operating data of industrial equipment, including numerical data, image data and acoustic data; label the multi-dimensional operating data of the same component or related components with association tags according to the coupling relationship of each component of the industrial equipment. S2. Establish a physical relationship table for each component of the industrial equipment, and assign globally unique identifiers to geometric objects at the macro, meso, and micro levels in the industrial model. Each globally unique identifier corresponds one-to-one with a component of the industrial equipment. S3. By combining the physical relationship table and the association identifiers of multi-dimensional operational data, establish the binding relationship between multi-dimensional operational data and globally unique identifiers; obtain the causal analysis conclusions output by the artificial intelligence platform, and transform the causal analysis conclusions into a globally unique identifier association chain; where the causal analysis conclusions are the cause-effect correspondence of abnormal data obtained by the artificial intelligence platform based on multi-dimensional operational data; S4. Based on the binding relationship and the association chain of globally unique identifiers, generate numerical labels containing current values, predicted values ​​and status indicators at the corresponding geometric positions of the industrial model; map multi-dimensional operating data into color gradients and generate a continuous heat map on the surface of the industrial model; when the fault-related data contained in the multi-dimensional operating data reaches the preset conditions, highlight the corresponding geometric area of ​​the industrial model and mark the corresponding globally unique identifier association chain.

[0017] Multidimensional operational data is a comprehensive collection of various data reflecting the operating status of industrial equipment, including numerical data, image data, and acoustic data. Numerical data consists of quantifiable data collected by various sensors during equipment operation, such as temperature, pressure, and rotational speed. Image data consists of images of the equipment components taken by cameras, such as images of bearing surfaces and valve sealing surfaces. Acoustic data consists of sound signals generated during equipment operation, such as motor running sounds and fluid sounds in pipelines.

[0018] When collecting multi-dimensional operational data, it is necessary to analyze the coupling relationships between various components based on the actual structure of the industrial equipment. These coupling relationships include mechanical transmission coupling, pipeline connection coupling, and electrical circuit coupling. For example, there is mechanical transmission coupling between the motor and the drive shaft, and pipeline connection coupling between the storage tank and the connecting pipes. Based on these coupling relationships, the same association identifier should be used to label different types of multi-dimensional operational data for the same component. Similarly, a unified association identifier should be used to label the multi-dimensional operational data of related components with coupling relationships. This association identifier can be a combination of letters and numbers and must be unique to ensure accurate mapping between data and equipment components in the future.

[0019] After data collection, a physical relationship table is established for each component of the industrial equipment. This table records the name, model, installation location, connection method, and transmission path of each component, such as the connection between a gearbox and a motor, or the installation location of a valve and a pipeline. The industrial model is a digital reconstruction of equipment, production lines, and factory areas in an industrial production scenario. It is divided into three levels based on the scope and level of detail: macro, meso, and micro. The macro level corresponds to the overall layout of the entire factory or multiple production lines; the meso level corresponds to a single production line or equipment cluster; and the micro level corresponds to a single piece of equipment or its core components. A globally unique identifier is assigned to each geometric object in each level. This globally unique identifier is a unique identifier used to distinguish different geometric objects. Each globally unique identifier corresponds to only one component of the industrial equipment, allowing for accurate association with the corresponding equipment component and related data.

[0020] Subsequently, combining the established physical relationship table and the association identifiers of multi-dimensional operational data, the collected multi-dimensional operational data is bound to its corresponding globally unique identifier. This allows each geometric object in the industrial model to be associated with the multi-dimensional operational data of its corresponding equipment component through a globally unique identifier, achieving a one-to-one correspondence between data and geometric objects. Simultaneously, the causal analysis conclusions output by the artificial intelligence platform are obtained. These conclusions are derived from the analysis and calculation of the collected multi-dimensional operational data, identifying the cause-effect relationships of abnormal data. Specifically, they clarify which data anomalies caused a particular abnormal result; for example, an abnormal result of excessive equipment vibration may be caused by an anomaly in bearing wear. This causal analysis conclusion is then transformed into a globally unique identifier association chain, where the globally unique identifiers corresponding to the cause data and the globally unique identifiers corresponding to the result data are associated in causal order, forming a chain-like identifier combination.

[0021] In the visualization phase, based on the binding relationship between multi-dimensional operational data and globally unique identifiers, and the transformed globally unique identifier association chain, numerical labels are generated at the positions of corresponding geometric objects in the industrial model. These numerical labels can include current values, predicted values, and status indicators. Current values ​​are the actual measured data from real-time equipment operation; predicted values ​​are the estimated results of subsequent operational data derived from historical data and operational trends; and status indicators are used to indicate the current operating status of the equipment, such as normal or abnormal. Simultaneously, multi-dimensional operational data are mapped to different color gradients according to their numerical magnitude. The depth or hue of the color gradient corresponds to the high or low level or trend of the data, generating a continuous heatmap on the surface of the industrial model. This color distribution visually presents the spatial distribution of the data. When fault-related data contained in the multi-dimensional operational data reaches a preset condition—that is, operational data directly related to equipment failure, with the preset condition being a fault judgment benchmark set based on equipment safety operation standards—the corresponding geometric area in the industrial model is highlighted. This highlighting can use colors or lines that clearly distinguish it from the surrounding area, along with the corresponding globally unique identifier association chain, allowing relevant personnel to quickly locate the fault and understand its causal relationship.

[0022] Through the above methods, this approach realizes a complete process from data collection to visualization of industrial equipment operation data. Data collection closely matches the coupling characteristics of equipment components, the mapping between data and industrial models accurately corresponds to the physical structure of the equipment, and causal relationships are presented through a globally unique identifier association chain. This makes the visualization results more in line with actual industrial needs, helping relevant personnel to quickly and accurately grasp the operating status and abnormal correlations of industrial equipment, and providing effective data support for equipment maintenance and fault diagnosis.

[0023] In some implementations, the following steps are included after step S1 and before step S2: Receive data analysis feedback from the AI ​​platform based on multi-dimensional operational data output. The data analysis feedback includes data missing information, data noise information, and key data requirement information. Based on data analysis feedback, we adjusted the data collection strategy, supplemented the corresponding types of multi-dimensional operational data collection for missing data prompts, optimized the collection parameters of corresponding collection points for data noise prompts, and increased the frequency of multi-dimensional operational data collection for corresponding components for key data requirements prompts. The multi-dimensional operational data acquired after adjusting the acquisition strategy is preprocessed to extract time-series features and core features of image and acoustic data. The time-series features include data periodic fluctuation features and trend change features, while the core features include the appearance status features of components in the image and the abnormal frequency band features in the acoustic wave.

[0024] Data analysis feedback refers to the optimization suggestions output by the AI ​​platform based on the initially received multi-dimensional operational data and its internal analysis logic. These suggestions primarily include data missing information, data noise warnings, and critical data requirement alerts. Data missing information alerts indicate the absence of a specific type or item of data that significantly impacts the analysis results; for example, missing acoustic data when analyzing bearing failures, or missing image data for a specific area when analyzing pipeline leaks. Data noise warnings indicate the presence of interference in some of the collected data, which may stem from sensor errors or external environmental influences, preventing the data from accurately reflecting the equipment's operating status. Critical data requirement alerts identify key equipment components or data types that require focused data collection, such as more detailed numerical data to support fatigue wear analysis for core transmission components.

[0025] After receiving data analysis feedback from the AI ​​platform, the data acquisition strategy needs to be adjusted accordingly. For data missing prompts, supplement the data with corresponding multi-dimensional operational data acquisition. If image data is missing, add a camera to the corresponding component; if sound wave data is missing, adjust the sensor acquisition type. For data noise prompts, optimize the acquisition parameters at the corresponding acquisition points. For example, adjust the sensor acquisition frequency band to avoid external interference frequencies, or adjust the sensor installation position to reduce errors caused by equipment vibration. For prompts indicating critical data requirements, increase the frequency of multi-dimensional operational data acquisition for the corresponding components to ensure more comprehensive data coverage and richer details for key analysis objects.

[0026] After adjusting the data acquisition strategy, the newly acquired multi-dimensional operational data undergoes preprocessing. The preprocessing focuses on extracting effective features to provide more accurate input data for the AI ​​platform analysis. Time-series features are extracted based on the time-series variation patterns of numerical data, including data periodic fluctuation features and trend change features. Data periodic fluctuation features reflect the fluctuations of data within a fixed time interval, such as the fluctuation pattern of motor speed within its operating cycle. Trend change features reflect the overall trend of data change over a period of time, such as the rising or falling trend of equipment temperature with operating time. Core features are key information extracted from image and acoustic data. Core features of image data include the appearance and condition of components, such as whether there are wear marks on the sealing surface or rust spots on the component surface—features that can be identified through images. Core features of acoustic data include abnormal frequency band features, i.e., sound signals in the equipment's operating sound waves that exceed the normal frequency range, such as the abnormal resonance frequency band caused by gear wear.

[0027] Through the above process, the data collection strategy can be optimized in response to the analysis needs of the artificial intelligence platform, effectively making up for data gaps, reducing data noise, and strengthening the collection of key data. The time-series features and core features extracted by preprocessing further uncover the effective information of multi-dimensional operational data, making the data input into the artificial intelligence platform more targeted and practical. This not only provides high-quality data for the artificial intelligence platform to optimize causal analysis conclusions, but also lays the foundation for the accurate binding of subsequent multi-dimensional operational data with globally unique identifiers.

[0028] In some implementations, in step S3, the generation rules and structure of the association identifier are as follows: the association identifier is not merely a simple combination of letters and numbers; its structure includes information about the location and coupling relationship of the component within the device topology. In this invention, the association identifier has a hierarchical coding system: device code - subsystem code - component serial number - coupling group number. The coupling group number is specifically used to identify a group of components with a physical coupling relationship. Thus, identifiers with the same coupling group number but different component serial numbers can clearly point to a group of associated components.

[0029] The physical relationship table cannot only record basic information about components; it also contains a core dictionary of mapping relationships. The two key fields in this core dictionary are as follows: component_global_id: Stores the globally unique identifier assigned to this component.

[0030] `association_group_id`: Stores the association identifier to which this component belongs, specifically the coupling group number. By querying this core dictionary, the system can: Find the association group to which it belongs based on a globally unique identifier: association_group_id.

[0031] Based on the association_group_id of an association identifier, find the list of globally unique identifiers corresponding to all components within that group.

[0032] Secondly, a clear mapping decision logic should be established in the binding process.

[0033] Scenario 1: The data is labeled with an association identifier pointing to a single component. The binding can be completed directly by querying the physical association table to find the unique `component_global_id` corresponding to that identifier.

[0034] Scenario 2: The data is labeled with association identifiers pointing to a group of related components. In this case, the physical association table should be queried to obtain a list of all component_global_ids under that association_group_id. Subsequently, the binding operation needs to select the most accurate one or several globally unique identifiers from the list for binding, based on the specific data type and the sensor installation location.

[0035] Based on the above logic, step S3 establishes the binding relationship between multi-dimensional operational data and globally unique identifiers, specifically including the following steps: By associating time-series features and core features with corresponding multi-dimensional operational data, combined data is formed. Based on the physical association table and association identifier, the combined data is bound to the corresponding globally unique identifier, so that the industrial model geometric object corresponding to each globally unique identifier is synchronously associated with the multi-dimensional operation data and time series features and core features of the industrial equipment components. The combined data is synchronized to the artificial intelligence platform to assist the platform in optimizing the causal analysis conclusions, and the globally unique identifier association chain is updated based on the optimized causal analysis conclusions.

[0036] After preprocessing the multi-dimensional operational data and extracting the temporal and core features, establishing the binding relationship between the multi-dimensional operational data and the globally unique identifier requires combining these feature information to form a more complete data association logic, so that the geometric objects in the industrial model can be more accurately associated with the comprehensive operational information of the equipment components.

[0037] First, the extracted time-series features and core features need to be correlated with the corresponding multi-dimensional operational data to form combined data. Time-series features originate from the numerical data in the multi-dimensional operational data, reflecting the periodic fluctuations and trend changes of the data. Core features come from image and acoustic data, reflecting the appearance status of components and abnormal frequency band information. These features have a natural affiliation with the original multi-dimensional operational data; the time-series feature of a motor's rotational speed corresponds to the motor's numerical rotational speed data; the core features of a bearing's appearance status also match the bearing's image data. During the correlation process, based on the equipment component affiliation information at the time of data acquisition, the multi-dimensional operational data, time-series features, and core features of the same or related components need to be integrated to ensure the consistency of information within the combined data and avoid confusion between data and features from different components. The formation of combined data supplements the operational patterns and state details behind individual data points, transforming them from isolated values ​​or signals.

[0038] Subsequently, based on the physical association table of each component of the industrial equipment and the association identifiers of the multi-dimensional operational data, the combined data is bound to the corresponding globally unique identifier. The physical association table clarifies basic information such as the installation location and connection method of each component, while the association identifier marks the component affiliation for the multi-dimensional operational data. The combination of the two can accurately locate the equipment component corresponding to each set of combined data. For example, the association identifier determines that a certain set of combined data belongs to a hydraulic pump, and the physical association table confirms the corresponding geometric object of the hydraulic pump in the industrial model. Then, the combined data is bound to the globally unique identifier of that geometric object. After binding, the industrial model geometric object corresponding to each globally unique identifier is not only associated with the multi-dimensional operational data of the industrial equipment component, but also with temporal and core features. That is, through the geometric object, it is possible to directly associate with the component's real-time operational data, historical change trends, appearance status, and sound signal characteristics, achieving deep binding between data and geometric objects.

[0039] After the binding operation is completed, the combined data is synchronized to the artificial intelligence platform. Compared to the standalone raw multi-dimensional operational data, the combined data, including time-series features and core features, provides the artificial intelligence platform with richer analytical basis. The trend changes of time-series features can help determine the development direction of the equipment's operating status, and the abnormal information of core features can point to potential problems in components. Based on this information, the artificial intelligence platform can more accurately identify the causes and impact range of abnormal data, thereby optimizing the output causal analysis conclusions. Based on the optimized causal analysis conclusions, the globally unique identifier association chain is updated: the original globally unique identifier association chain may have been established only based on the raw multi-dimensional operational data, resulting in ambiguous causal attribution. The optimized conclusions clarify a more accurate correspondence between abnormal causes and results. Transforming these relationships into new globally unique identifier association chains makes the subsequent visualized fault causal logic more reliable and avoids misleading equipment maintenance personnel.

[0040] The entire process integrates time-series features, core features, and original multi-dimensional operational data, accurately binds them to globally unique identifiers, and then helps the artificial intelligence platform optimize its analysis conclusions. This not only makes the association information of globally unique identifiers more comprehensive, but also improves the accuracy of causal analysis conclusions and the corresponding association chains of globally unique identifiers. This provides high-quality data support for the subsequent visualization of industrial models, enabling the displayed content to more realistically reflect the operating status and abnormal associations of equipment components.

[0041] In some implementations, the preprocessing of the collected multi-dimensional operational data also includes the following steps: Data filtering conditions are established based on the physical operation rules of industrial equipment. The physical operation rules include the correlation logic of the operating parameters of each component of the industrial equipment and the target operating range. The collected multi-dimensional operational data is filtered according to the data filtering criteria. Interference data that does not conform to the physical operation rules is filtered out, and valid data related to the operating status of industrial equipment and valid abnormal data that reflect potential anomalies are retained. The filtered valid data and valid abnormal data are associated with the extracted time-series features and core features to form a preprocessed data feature set, which is then bound to a globally unique identifier.

[0042] When preprocessing the collected multi-dimensional operational data, in order to further improve the relevance and reliability of the data, we consider combining the operating rules of the industrial equipment itself to complete the data screening, and integrate the screened data with the extracted features to provide a high-quality data foundation for subsequent data binding.

[0043] The physical operating rules of industrial equipment are the inherent operating laws formed through long-term practice and design specifications. Their core includes the correlation logic of the operating parameters of each component and the target operating range. The correlation logic of operating parameters is reflected in the mutual constraints and matching relationships between different components or different parameters of the same component. For example, there is a clear linkage between the steam pressure of a boiler and the fuel supply of a burner; when the pressure increases, the fuel supply needs to be adjusted accordingly to maintain balance. The target operating range is the reasonable fluctuation range of each parameter under normal operating conditions. For example, the target operating range of the temperature of a fan bearing is determined by factors such as the bearing material and lubrication conditions; exceeding this range may cause wear or damage. Establishing data screening conditions based on these physical operating rules essentially transforms the inherent laws of the equipment into quantifiable and executable data judgment standards, providing a clear basis for screening operations.

[0044] When filtering multi-dimensional operational data based on established data filtering criteria, the core is to distinguish between interfering data and valid data. Interfering data refers to data that deviates from physical operating rules and cannot truly reflect the equipment's operating status. This type of data may originate from numerical jumps caused by temporary sensor malfunctions, signal distortion caused by external electromagnetic interference, or environmental data accidentally mixed in during the acquisition process, such as the instantaneous low temperature value generated by rain splashing on a temperature sensor. It is irrelevant to the actual operating status of the equipment and needs to be filtered out. Valid data, on the other hand, conforms to physical operating rules and reflects the normal operating status of the equipment, such as motor speed within the target operating range and images that clearly reflect the normal appearance of components. Valid abnormal data, although not completely exceeding the target operating range, shows a trend of deviation from the norm and may be an early signal of equipment failure. For example, the gearbox vibration frequency may be within the allowable range, but it has increased significantly compared to the historical average. This type of data needs to be retained to provide early warning of potential failures.

[0045] After the filtering process is complete, the obtained valid and abnormal data need to be correlated and integrated with the previously extracted time-series features and core features. Valid and abnormal data are the basic information reflecting the equipment status, time-series features reveal the changing patterns of these data, and core features supplement the intuitive status information of components from the dimensions of images and sound waves. The correlation among the three must be based on the equipment component attribution information at the time of data acquisition, so that various types of data and features of the same component form a complete association. For example, the effective pressure value of a certain pump and the effective abnormal pressure value close to the upper limit of the target range need to correspond one-to-one with the periodic fluctuation characteristics of the pump's pressure data, the appearance status characteristics of the pump body shell image, and the frequency band characteristics of the operating sound waves, ultimately forming a complete and comprehensive data feature set. This data feature set will serve as the core data carrier for subsequent binding with globally unique identifiers. Its data quality determines the accuracy of subsequent data binding and the reliability of the analysis conclusions of the artificial intelligence platform. Through filtering and integration operations, interference from invalid data to subsequent processes is avoided, while retaining complete information on the equipment's operating status, providing reliable data support for the entire visualization process.

[0046] In some implementations, the method further includes the following steps in step S3, during the process of establishing the binding relationship between multi-dimensional operational data and globally unique identifiers: Based on the physical association table and the association identifier of multi-dimensional operational data, the effective data, effective abnormal data, time series features and core features in the data feature set are associated with the corresponding globally unique identifier one by one, so that each globally unique identifier stores the complete data feature information of a single component or related components of industrial equipment. The method also includes: The bound globally unique identifier and the corresponding data feature set are synchronized to the artificial intelligence platform to assist the artificial intelligence platform in analyzing the operating status of industrial equipment and generating causal analysis conclusions. When the industrial model is visualized, the geometric object corresponding to the globally unique identifier is synchronously associated with and calls the time-series features and core features in the data feature set. When the user views the display information corresponding to the geometric object, the operating trend reflected by the time-series features and the component status details reflected by the core features are presented synchronously.

[0047] After filtering and integrating multi-dimensional operational data to form a data feature set, when establishing the binding relationship between multi-dimensional operational data and globally unique identifiers, the data feature set should be used as a carrier to achieve accurate association based on the physical association information of equipment components.

[0048] The physical association table records basic information such as the installation location, connection method, and transmission path of each component of industrial equipment. The association identifier clearly marks the component affiliation of various data types in the data feature set. Together, they form the basis for associating data with globally unique identifiers (GUIDs). In specific association operations, valid data, valid abnormal data, time-series features, and core features in the data feature set can be treated as a whole and matched one by one with the corresponding GUID. For example, for a certain wind turbine, its valid wind speed data within the target operating range, valid abnormal temperature data slightly above the historical average, time-series features reflecting periodic wind speed changes, and the core feature of the complete appearance of the turbine blades must all point to the corresponding GUID for that wind turbine in the industrial model. Through this association method, each GUID no longer corresponds to only a single type of data, but completely stores comprehensive data feature information of a single component of the industrial equipment or related components with coupling relationships, making the binding of data with GUIDs both accurate and comprehensive.

[0049] After binding the data feature set with the globally unique identifier, the bound globally unique identifier and the corresponding data feature set can be synchronized to the artificial intelligence platform. Valid data in the data feature set provides the AI ​​platform with baseline information for the normal operation of the equipment, while valid abnormal data contains early signals of potential faults. Time-series features and core features supplement the data details from the perspective of change patterns and intuitive states. This information significantly improves the comprehensiveness of the AI ​​platform's analysis. Based on this integrated information, the AI ​​platform can more clearly identify the correlations between the operating data of different components, more accurately determine the causes of abnormal data, and thus generate more reliable causal analysis conclusions, avoiding analytical biases caused by incomplete data.

[0050] During the visualization of industrial models, the geometric objects corresponding to globally unique identifiers are simultaneously associated with temporal and core features in the data feature set. When relevant personnel view the display information corresponding to a specific geometric object, they can extract and present the associated temporal and core features. Temporal features, such as trend curves, can show the periodic fluctuations and long-term trends of equipment component operating parameters, helping personnel predict changes in operating status. Core features can present the appearance of components in the image, such as the presence of cracks, rust, or other traces, and whether abnormal frequency bands are included in the sound waves, allowing personnel to grasp the details of the component's physical state. In this way, visualization is no longer limited to the presentation of surface data, but delves into the operating patterns and status information behind the data, helping relevant personnel to more comprehensively understand the operating status of industrial equipment, accurately locate the causes of anomalies, and provide a more sufficient basis for equipment maintenance decisions.

[0051] The entire process binds the complete set of data features with globally unique identifiers, which not only strengthens the connection between data and device components, but also provides high-quality input for the analysis of the artificial intelligence platform, while enhancing the information value of the visualization.

[0052] In some implementations, when the fault-related data included in the multi-dimensional operational data reaches a preset condition in step S4, the method further includes the following steps: Based on the physical association table, query the associated components corresponding to the target globally unique identifier bound to the fault-related data that meets the preset conditions; Extend the binding scope of the target globally unique identifier to the globally unique identifiers corresponding to the associated components, and establish an extended association chain of globally unique identifiers corresponding to the fault propagation path; In the industrial model, the target globally unique identifier is connected to the globally unique identifier of the associated component by a preset identification line, and the multi-dimensional operation data and status indicators corresponding to the extended association chain of the globally unique identifier are displayed simultaneously.

[0053] When the fault-related data contained in the multi-dimensional operational data reaches the preset conditions, in order to enable relevant personnel to clearly understand the scope of the fault's impact and its transmission path, it is necessary to further improve the mapping logic in the visualization of the industrial model to achieve a comprehensive presentation of fault-related information.

[0054] Fault-related data are multi-dimensional operational data that directly reflect the fault status of industrial equipment, such as vibration values ​​exceeding the target operating range, image data showing component damage, and sound wave data containing abnormal frequency bands. Preset conditions are fault judgment benchmarks set based on the safety operation standards, design parameters, and maintenance experience of industrial equipment. When the fault-related data meets these benchmarks, it is determined that the equipment has a fault risk or has already failed. At this point, based on the established physical association table, the associated components corresponding to the target globally unique identifier bound to the fault-related data are queried. The physical association table records the physical association information between various components of the industrial equipment, such as mechanical connections, transmission coordination, and pipeline connections. The target globally unique identifier is a unique identifier bound to the fault-related data, which can accurately locate the corresponding faulty component. Associated components are those that have a direct or indirect physical connection with the faulty component and may be affected by the fault, such as a gearbox connected to a faulty motor via a drive shaft, or a pipeline supplying fluid to a faulty pump.

[0055] After identifying the associated components, the binding scope of the target globally unique identifier is expanded to include the globally unique identifiers corresponding to these associated components. Originally, the target globally unique identifier only bound to the multi-dimensional operational data of the faulty component; after expansion, its binding scope covers the globally unique identifiers of the faulty component and all associated components, thus establishing an extended association chain of globally unique identifiers corresponding to the fault propagation path. This extended association chain starts with the target globally unique identifier and sequentially links the globally unique identifiers of associated components according to the physical association order, forming a clear chain structure that reflects the path that a fault may propagate from the faulty component to other associated components.

[0056] In the visualization of the industrial model, preset marker lines connect the target globally unique identifier with the globally unique identifiers corresponding to associated components. These marker lines can use specific colors, line types, or thicknesses to clearly distinguish them from other lines in the industrial model, enabling rapid identification of fault propagation paths. Additionally, multi-dimensional operational data and status indicators corresponding to the extended association chain of the globally unique identifier are displayed simultaneously: for faulty components, the specific content of their fault-related data is displayed, and the status indicator indicates the fault status; for associated components, their current multi-dimensional operational data is displayed, and the status indicator indicates normal, warning, or affected status. For example, if the vibration value of an associated component has not reached the fault standard but is close to the critical value, the status indicator indicates a warning.

[0057] Through this operation, the industrial model no longer simply displays the location and data of a single faulty component in isolation. Instead, it extends the association chain through a globally unique identifier, integrating and presenting the relationship between the faulty component and related components, operational data, and status indicators. This allows relevant personnel to intuitively understand the transmission path of the fault and the range of components that may be affected, avoiding focusing only on a single fault point while ignoring potential chain failure risks. It provides a more comprehensive reference for fault diagnosis and maintenance plan development, improving the pertinence and effectiveness of equipment operation and maintenance.

[0058] In some implementations, the method further includes the following steps: Based on the globally unique identifiers associated with the macro, meso, and micro levels of the industrial model, the globally unique identifiers are extended and the associated chains are synchronously mapped to the industrial model at each level. In the macro-level industrial model, the production line or equipment cluster corresponding to the globally unique identifier extended association chain is identified as a whole; in the meso-level industrial model, the equipment and related components corresponding to the globally unique identifier extended association chain are highlighted; in the micro-level industrial model, the geometric position and multi-dimensional operation data of each component in the globally unique identifier extended association chain are displayed. When a user clicks on the identifier or label corresponding to the globally unique identifier extended association chain in any level of the industrial model, they will be redirected to the industrial model of other related levels.

[0059] After the globally unique identifier (GUID) extended association chain is established, to ensure that personnel with different needs can clearly understand the fault propagation situation, it is necessary to synchronously map this GUID extended association chain to the macro, meso, and micro levels of the industrial model based on the relationships between the various levels of the industrial model. The macro, meso, and micro levels of the industrial model are divided according to the display scope and detail precision. The three levels form a whole by associating with the globally unique identifier. The globally unique identifier of a production line in the macro level will be associated with the globally unique identifiers of all the equipment contained in that production line (i.e., meso level identifiers), and the globally unique identifier of each equipment will be associated with the globally unique identifiers of its internal core components (i.e., micro level identifiers). This hierarchical association ensures that the globally unique identifier extended association chain can be accurately synchronized between different levels.

[0060] In macro-level industrial models, the focus is on the overall layout and large-scale connections. Therefore, production lines or equipment clusters corresponding to the extended association chains of globally unique identifiers are identified holistically. This overall identification does not require showing the details of individual components, but rather uses eye-catching color blocks or outlines to distinguish production line clusters containing fault propagation paths from the entire factory model. This allows production scheduling personnel to quickly locate the overall scope of the fault's impact and determine whether production plans need to be adjusted. In meso-level industrial models, the display precision is further improved. At this point, the focus is on individual equipment and related components corresponding to the extended association chains of globally unique identifiers. These are highlighted with bold borders or special markings. For example, uniform style markings are added to the geometric objects corresponding to the faulty motor and its connected gearbox and conveyor belt. This allows equipment managers to intuitively identify the specific equipment group involved in the fault and coordinate maintenance resources accordingly. In the micro-level industrial model, the most detailed information needs to be presented. It will show the geometric position of each component in the extended association chain of the globally unique identifier, such as the specific installation position of the bearing in the motor and the accurate coordinates of the pipe interface. At the same time, it will simultaneously display the multi-dimensional operating data of each component, such as the vibration frequency of the bearing and the pressure value of the pipe, to provide the basis for on-site maintenance personnel to carry out maintenance.

[0061] To ensure seamless information flow across different levels, when a user clicks on an identifier or label corresponding to a globally unique identifier in the extended association chain of an industrial model at any level, the system will jump to other related industrial models based on the globally unique identifiers associated with those levels. For example, after a production scheduler clicks on a production line identified as a whole at the macro level, they can directly jump to the meso level model of that production line to see which specific piece of equipment is malfunctioning; after an equipment supervisor clicks on a highlighted motor in the meso level model, they can further jump to the micro level model of the motor to view the detailed status of core components in the fault propagation path. This cross-level jumping allows users to trace the fault propagation from the whole to the part or from the part to the whole according to their needs, avoiding judgment errors caused by information dispersion.

[0062] Through this cross-level mapping and linkage method, personnel in different positions can obtain the necessary fault association information in the corresponding industrial model level, which not only ensures the efficiency of overall scheduling, but also ensures the accuracy of on-site maintenance. This allows the value of the globally unique identifier extended association chain to be fully utilized in various management and operation links, thereby improving the overall efficiency of fault handling.

[0063] In some implementations, the method further includes the following steps: Based on the association chain of globally unique identifiers, cause identifiers are marked for the globally unique identifiers corresponding to the cause data in the causal analysis conclusion, and result identifiers are marked for the globally unique identifiers corresponding to the result data. According to the correlation strength between cause and result in the causal analysis conclusion, the display style of the preset identifier lines connecting cause and result identifiers is set. On the industrial model, the marked cause identifiers, result identifiers and corresponding preset identifier lines are displayed synchronously. When the user clicks on the preset identifier lines, the basis for the formation of the causal analysis conclusion is presented.

[0064] The globally unique identifier association chain is a digital transformation of causal analysis conclusions. This chain contains various identifiers corresponding to the causes and effects of abnormal data, enabling a visual representation of causal relationships and allowing relevant personnel to more clearly grasp the logic behind the anomalies. Causal analysis conclusions are derived by the AI ​​platform based on multi-dimensional operational data, clarifying which specific data anomalies caused a particular abnormal result. For example, a decrease in equipment output power may stem from multiple causes such as abnormal motor speed or bearing wear. The globally unique identifier association chain connects the globally unique identifiers corresponding to these causal and effect data according to their correlation relationships.

[0065] In practice, the first step is to label the globally unique identifiers (GUIDs) corresponding to the causal data and the result data in the causal analysis conclusions with different identifiers. The causal identifier points to the GUID associated with the root cause data leading to the abnormal result; for example, the GUID corresponding to abnormal motor speed data would be labeled as the causal identifier. The result identifier corresponds to the GUID associated with the abnormal result data; for example, the GUID corresponding to decreased equipment output power data would be labeled as the result identifier. These identifiers can be distinguished using different graphics or colors; for example, causal identifiers use circular marks, and result identifiers use square marks, to ensure quick identification in the industrial model and avoid confusion in the correspondence between cause and effect.

[0066] Subsequently, based on the strength of the correlation between cause and effect in the causal analysis conclusions, the display style of the preset marker lines connecting the two is set. Correlation strength reflects the degree of influence of a particular cause on the abnormal result. Some cause data are direct and critical factors leading to the abnormal result, resulting in a high correlation strength; others are indirect influencing factors, with a relatively low correlation strength. The display style of the preset marker lines will differ for different correlation strengths. For example, high correlation strength uses a thicker solid line in a striking red color to highlight the core causal relationship; medium correlation strength uses a regular thickness solid line in orange; and low correlation strength uses a dashed line in yellow. This differentiated style allows personnel to intuitively judge the weight of each cause's influence on the result.

[0067] After completing the labeling and line style settings, these elements are synchronously displayed on the industrial model. At the corresponding geometric locations on the model, labeled cause and effect labels, along with pre-defined label lines connecting them, will appear, forming an intuitive causal network. When a user clicks on a pre-defined label line on the industrial model, the system will present the basis for that causal analysis conclusion. This basis includes the multi-dimensional data types used in the analysis process, the changing characteristics of key data, and the analysis logic of the artificial intelligence platform. For example, clicking on the line connecting abnormal motor speed and power reduction will display the time-series characteristics of motor speed, power numerical data, and the correlation analysis process between the two.

[0068] This visualization method transforms abstract causal analysis conclusions into intuitively perceptible graphical elements on an industrial model, avoiding the problem of ambiguous causal relationships in traditional data reports. By viewing the correspondence between icons and lines, relevant personnel can quickly pinpoint the core cause of anomalies; clicking on lines reveals the basis for their formation, verifying the rationality of the analysis conclusions. This provides logical support for determining the cause of anomalies and developing targeted solutions, effectively improving the efficiency of troubleshooting and equipment maintenance.

[0069] In some implementations, after establishing the binding relationship between multi-dimensional operational data and globally unique identifiers, the method further includes the following steps: Based on the real-time operating conditions of industrial equipment, the mapping priority of the globally unique identifiers corresponding to each component is determined. When the equipment is under heavy load or in a critical process stage, the mapping priority of the globally unique identifiers corresponding to the core components is higher than that of other components. When the fault-related data of multiple components all reach the preset conditions, the mapping areas corresponding to each fault are distinguished by different styles of status identifiers on the industrial model according to the mapping priority and the association chain of the globally unique identifiers corresponding to each fault. Based on changes in the operating conditions or fault status updates of industrial equipment, the mapping priority and display style of the corresponding globally unique identifiers are adjusted in real time.

[0070] After establishing a binding relationship between multi-dimensional operational data and globally unique identifiers, the operating conditions of industrial equipment are constantly changing. Under different operating conditions, the degree of impact of each component on production varies. Therefore, the mapping strategy of data on the industrial model can be adjusted in combination with real-time conditions to ensure that the displayed content matches actual needs.

[0071] Real-time operating conditions of industrial equipment include core information such as load status and process stage. For example, the high-temperature smelting stage of metallurgical equipment and the precision machining process of machine tools are critical operating conditions. At this time, the core components of the equipment bear the main production tasks, and their operating status determines the production quality and safety. Core components usually refer to components that play a decisive role in the overall operation of the equipment, such as the crankshaft of an engine and the main drive gear of a production line. When the equipment is under heavy load or in a critical process stage, the globally unique identifier mapping priority of these core components should be higher than that of non-core components such as cooling fans and auxiliary supports. This is because even minor abnormalities in core components can trigger cascading failures. It is necessary to prioritize their operating data and status in the industrial model so that relevant personnel can capture critical information as soon as possible. The process of determining priorities should be based on real-time parameters of equipment operation, such as load rate and process progress, so that the priority division matches the actual operating conditions.

[0072] When the fault-related data of multiple components all meet the preset conditions, displaying all fault areas in a single style on the industrial model can easily lead to information confusion and affect fault handling efficiency. In this case, differentiated status identifiers can be designed for the mapping areas corresponding to different faults, based on the established mapping priorities and the globally unique identifier association chains for each fault. For faults of core components with high mapping priority, the status identifier can use high-contrast colors and dynamic effects, such as a flashing dark red border; for faults of non-core components with lower priority, the identifier uses relatively soft colors and static effects, such as a stable orange fill. Simultaneously, the status identifiers of mapping areas corresponding to the globally unique identifier association chains of the same fault propagation path will maintain a consistent style. For example, for a fault in a certain transmission system, the mapping areas from the motor to the gearbox to the conveyor belt will all use identifiers with diagonal textures, distinguishing them from other fault areas while also reflecting the fault propagation associations and avoiding interference between the displayed information of different faults.

[0073] The operating conditions of industrial equipment are not static. Process switching, load adjustment, or fault repair will all cause equipment status updates. At this time, it is necessary to adjust the mapping priority and display style of the corresponding globally unique identifier in real time. If the equipment switches from a normal process to a critical process, the mapping priority of the main controller, which was originally of medium priority, will be increased, and the display style will change from the conventional numerical label to a highlighted card with a trend curve. If the core component recovers from a fault, its status indicator will change from red indicating a fault to green indicating normal operation, and the mapping priority will also drop from the highest level to a level that matches the current operating condition. If the equipment load decreases, the mapping priority of auxiliary components will be further reduced, and their display information can be appropriately simplified to reduce redundant content in the industrial model.

[0074] This dynamically adjusted mapping strategy ensures that the visualization of the industrial model is always synchronized with the actual operating status of the equipment. It guarantees the priority presentation of core information and avoids information chaos in multiple fault scenarios. This helps relevant personnel quickly focus on key points in complex operating data, improves their control over the operating status of the equipment, and enhances the targeted nature of fault handling.

[0075] In certain scenarios, during the shutdown and maintenance phase of industrial equipment, artificial intelligence platforms can be used to accurately map operational data with maintenance requirements, thereby improving the targeting of maintenance. Multi-dimensional operational data can supplement maintenance scenario-specific data at this stage, such as torque data during component disassembly and image data from flaw detection. This data, along with routine operational data, forms the basis for maintenance analysis.

[0076] The AI ​​platform can access the equipment's historical maintenance data and similar equipment failure cases, compare them with current multi-dimensional operating data, and extract discrepancies. For example, it can analyze whether the abnormal frequency bands in the current acoustic data of a bearing match the characteristics of bearing wear in historical maintenance. These discrepancies are then bound to globally unique identifiers for the corresponding components, marking maintenance priorities, with discrepancies in core components marked as high priority.

[0077] In the industrial model, the geometric objects of high-priority components are highlighted with a flashing blue border. Clicking on them reveals detailed differences: for example, the current bearing image is presented side-by-side with historical images of intact bearings, and wear areas are marked; the deviation of torque data from standard values ​​is also displayed simultaneously. This approach allows maintenance personnel to quickly locate components requiring key maintenance without having to check each component individually, clarifying the direction of maintenance, avoiding blind disassembly, and adapting to the practical needs of industrial equipment maintenance.

[0078] When multiple industrial devices operate collaboratively, local anomalies in the data of a single device can easily be overlooked. In such cases, establishing cross-device data association mapping can be considered. When collecting data, in addition to the multi-dimensional data of a single device, supplementary linkage parameters between devices should be included, such as the speed matching values ​​of the upstream motor and downstream conveyor belt in a production line, and the pressure coordination data of multiple pump groups in a hydraulic system, etc., and collaborative association identifiers should be added.

[0079] Artificial intelligence platforms can analyze the temporal and core characteristics of multiple devices based on the collaborative operation logic of the devices, extract collaborative anomaly characteristics, such as the correlation between upstream motor speed fluctuations and downstream conveyor belt tension anomalies, and identify potential problems where the data of a single device has not reached the fault threshold but the linkage of multiple devices has become unbalanced.

[0080] By binding collaborative anomaly characteristics to globally unique identifiers of associated devices, a cross-device collaborative association chain is formed. In the industrial model, dual lines can connect the geometric objects of associated devices, synchronously displaying matching curves of the linkage parameters, with abnormal intervals marked in red. Clicking on the curve allows viewing the causes of collaborative imbalances, such as tension fluctuations caused by motor speed regulation lag.

[0081] This mapping method, which adapts to multi-device collaborative scenarios, can accurately capture the unique linkage risks in industrial production and avoid the limitations of judgment from a single device perspective.

[0082] Figure 2 This is a schematic diagram illustrating the hierarchical association mapping and fault propagation between multi-dimensional operating data of industrial equipment and industrial models, provided in an embodiment of the present invention. Figure 2The data association mapping link presented in the document restores the complete logic of equipment components-data processing-model display in the overall solution, providing a reference for understanding the entire process of data acquisition and visualization. Motors, gearboxes, pipes, and bearings in the industrial equipment component layer illustrate the core structure of the industrial equipment. The physical association labels between components correspond to the coupling relationships such as mechanical transmission and pipeline connections in the solution. This is a prerequisite for labeling data associations; only by clarifying the inherent connections between components can the data of the same or related components be accurately classified.

[0083] The ID1 to ID4 identifiers in the data-identifier binding layer are the embodiment of globally unique identifiers in the solution. Each identifier corresponds one-to-one with the component on the left, reflecting the binding relationship between multi-dimensional operational data, temporal characteristics, core features, etc., after processing, and the device component. This binding relationship allows the ID1 corresponding to the motor to not only be associated with the real-time speed value, but also to include the temporal characteristics of its speed fluctuations and the core characteristics of its appearance, providing an integrated data foundation for the AI ​​platform analysis. The causal analysis conclusion of the AI ​​platform within the dashed box is the globally unique identifier association chain generated by the AI ​​platform based on this integrated data, and the fault transmission arrows show the correspondence between the cause and effect of the anomaly.

[0084] The industrial model's hierarchical display layer is divided into macro, meso, and micro levels, with bidirectional cross-level navigation labels, corresponding to the visualization design in the solution that meets the needs of different positions. Macro-level production line / equipment cluster identifiers allow dispatchers to grasp the overall scope; meso-level equipment and related component labels assist supervisors in overall maintenance coordination; and micro-level component details provide maintenance personnel with accurate information. The linkage between levels allows data to be accessed as needed.

[0085] This invention also provides an industrial operation data visualization system based on an artificial intelligence platform, including: The data acquisition module is used to collect multi-dimensional operational data of industrial equipment, including numerical data, image data, and acoustic data; and to label the multi-dimensional operational data of the same or related components according to the coupling relationship between the various components of the industrial equipment. The allocation module is used to establish a physical relationship table for each component of industrial equipment. It assigns globally unique identifiers to geometric objects at the macro, meso, and micro levels in the industrial model, with each globally unique identifier corresponding one-to-one with a component of the industrial equipment. The module combines the physical relationship table and the association identifiers of multi-dimensional operational data to establish a binding relationship between multi-dimensional operational data and globally unique identifiers; it obtains the causal analysis conclusions output by the artificial intelligence platform and transforms the causal analysis conclusions into a globally unique identifier association chain; where the causal analysis conclusions are the cause-effect correspondence of abnormal data derived by the artificial intelligence platform based on multi-dimensional operational data. The generation module is used to generate numerical labels containing current values, predicted values, and status indicators at the corresponding geometric locations of the industrial model based on binding relationships and globally unique identifier association chains; map multi-dimensional operational data into color gradients to generate continuous heat maps on the surface of the industrial model; when the fault-related data contained in the multi-dimensional operational data reaches preset conditions, it highlights and marks the corresponding geometric areas of the industrial model and marks the corresponding globally unique identifier association chains.

[0086] The above description is merely a preferred embodiment of the present invention and the technical principles employed. The present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention.

Claims

1. A method for visualizing industrial operation data based on an artificial intelligence platform, characterized in that: Includes the following steps: S1. Collect multi-dimensional operating data of industrial equipment, including numerical data, image data and acoustic data; label the multi-dimensional operating data of the same component or related components with association tags according to the coupling relationship of each component of the industrial equipment. S2. Establish a physical relationship table for each component of the industrial equipment, and assign globally unique identifiers to geometric objects at the macro, meso, and micro levels in the industrial model. Each globally unique identifier corresponds one-to-one with a component of the industrial equipment. S3. By combining the physical relationship table and the association identifiers of multi-dimensional operational data, establish the binding relationship between multi-dimensional operational data and globally unique identifiers; obtain the causal analysis conclusions output by the artificial intelligence platform, and transform the causal analysis conclusions into a globally unique identifier association chain; where the causal analysis conclusions are the cause-effect correspondence of abnormal data obtained by the artificial intelligence platform based on multi-dimensional operational data; S4. Based on the binding relationship and the association chain of globally unique identifiers, generate numerical labels containing the current value, predicted value and status identifier at the corresponding geometric position of the industrial model. Multi-dimensional operational data is mapped to color gradients to generate a continuous heatmap on the surface of the industrial model. When the fault-related data contained in the multi-dimensional operational data reaches the preset conditions, the corresponding geometric area of ​​the industrial model is highlighted and marked with the corresponding globally unique identifier association chain.

2. The method according to claim 1, characterized in that, The steps following step S1 and before step S2 are also included: Receive data analysis feedback from the AI ​​platform based on multi-dimensional operational data output. The data analysis feedback includes data missing information, data noise information, and key data requirement information. Based on data analysis feedback, we adjusted the data collection strategy, supplemented the corresponding types of multi-dimensional operational data collection for missing data prompts, optimized the collection parameters of corresponding collection points for data noise prompts, and increased the frequency of multi-dimensional operational data collection for corresponding components for key data requirements prompts. The multi-dimensional operational data acquired after adjusting the acquisition strategy is preprocessed to extract time-series features and core features of image and acoustic data. The time-series features include data periodic fluctuation features and trend change features, while the core features include the appearance status features of components in the image and the abnormal frequency band features in the acoustic wave.

3. The method according to claim 2, characterized in that, Step S3 establishes the binding relationship between multi-dimensional runtime data and globally unique identifiers, specifically including the following steps: By associating time-series features and core features with corresponding multi-dimensional operational data, combined data is formed. Based on the physical association table and association identifier, the combined data is bound to the corresponding globally unique identifier, so that the industrial model geometric object corresponding to each globally unique identifier is synchronously associated with the multi-dimensional operation data and time series features and core features of the industrial equipment components. The combined data is synchronized to the artificial intelligence platform to assist the platform in optimizing the causal analysis conclusions, and the globally unique identifier association chain is updated based on the optimized causal analysis conclusions.

4. The method according to claim 2, characterized in that, The preprocessing of the collected multi-dimensional operational data also includes the following steps: Data filtering conditions are established based on the physical operation rules of industrial equipment. The physical operation rules include the correlation logic of the operating parameters of each component of the industrial equipment and the target operating range. The collected multi-dimensional operational data is filtered according to the data filtering criteria. Interference data that does not conform to the physical operation rules is filtered out, and valid data related to the operating status of industrial equipment and valid abnormal data that reflect potential anomalies are retained. The filtered valid data and valid abnormal data are associated with the extracted time-series features and core features to form a preprocessed data feature set, which is then bound to a globally unique identifier.

5. The method according to claim 4, characterized in that, Step S3, in establishing the binding relationship between multi-dimensional runtime data and globally unique identifiers, also includes the following steps: Based on the physical association table and the association identifier of multi-dimensional operational data, the effective data, effective abnormal data, time series features and core features in the data feature set are associated with the corresponding globally unique identifier one by one, so that each globally unique identifier stores the complete data feature information of a single component or related components of industrial equipment. The bound globally unique identifier and the corresponding data feature set are synchronized to the artificial intelligence platform to assist the artificial intelligence platform in analyzing the operating status of industrial equipment and generating causal analysis conclusions. When the industrial model is visualized, the geometric object corresponding to the globally unique identifier is synchronously associated with and calls the time-series features and core features in the data feature set. When the user views the display information corresponding to the geometric object, the operating trend reflected by the time-series features and the component status details reflected by the core features are presented synchronously.

6. The method according to claim 1, characterized in that, In step S4, when the fault-related data contained in the multi-dimensional operational data reaches a preset condition, the method further includes the following steps: Based on the physical association table, query the associated components corresponding to the target globally unique identifier bound to the fault-related data that meets the preset conditions; Extend the binding scope of the target globally unique identifier to the globally unique identifiers corresponding to the associated components, and establish an extended association chain of globally unique identifiers corresponding to the fault propagation path; In the industrial model, the target globally unique identifier is connected to the globally unique identifier of the associated component by a preset identification line, and the multi-dimensional operation data and status indicators corresponding to the extended association chain of the globally unique identifier are displayed simultaneously.

7. The method according to claim 6, characterized in that, The method further includes the following steps: Based on the globally unique identifiers associated with the macro, meso, and micro levels of the industrial model, the globally unique identifiers are extended and the associated chains are synchronously mapped to the industrial model at each level. In the macro-level industrial model, the production line or equipment cluster corresponding to the globally unique identifier extended association chain is identified as a whole; in the meso-level industrial model, the equipment and related components corresponding to the globally unique identifier extended association chain are highlighted; in the micro-level industrial model, the geometric position and multi-dimensional operation data of each component in the globally unique identifier extended association chain are displayed. When a user clicks on the corresponding icon or label in the extended association chain of the industrial model at any level, they will be redirected to the industrial model at other related levels.

8. The method according to claim 1, characterized in that, The method further includes the following steps: Based on the association chain of globally unique identifiers, the globally unique identifiers corresponding to the causal data in the causal analysis conclusion are labeled with causal identifiers, and the globally unique identifiers corresponding to the result data are labeled with result identifiers. Based on the strength of the correlation between cause and effect in the causal analysis conclusion, set the display style of the preset label line connecting cause and effect labels; On the industrial model, the labeled cause markers, result markers, and corresponding preset marker lines are displayed simultaneously. When the user clicks on the preset marker line, the basis for the formation of the causal analysis conclusion is presented.

9. The method according to claim 1, characterized in that, After establishing the binding relationship between multi-dimensional operational data and globally unique identifiers, the following steps are also included: Based on the real-time operating conditions of industrial equipment, the mapping priority of the globally unique identifiers corresponding to each component is determined. When the equipment is under heavy load or in a critical process stage, the mapping priority of the globally unique identifiers corresponding to the core components is higher than that of other components. When the fault-related data of multiple components all reach the preset conditions, the mapping areas corresponding to each fault are distinguished by different styles of status identifiers on the industrial model according to the mapping priority and the association chain of the globally unique identifiers corresponding to each fault. Based on changes in the operating conditions or fault status updates of industrial equipment, the mapping priority and display style of the corresponding globally unique identifiers are adjusted in real time.

10. An industrial operation data visualization system based on an artificial intelligence platform, characterized in that: For implementing the method as described in any one of claims 1-9, comprising: The data acquisition module is used to collect multi-dimensional operational data of industrial equipment, including numerical data, image data, and acoustic data; and to label the multi-dimensional operational data of the same or related components according to the coupling relationship between the various components of the industrial equipment. The allocation module is used to establish a physical relationship table for each component of industrial equipment. It assigns globally unique identifiers to geometric objects at the macro, meso, and micro levels in the industrial model, with each globally unique identifier corresponding one-to-one with a component of the industrial equipment. The module combines the physical relationship table and the association identifiers of multi-dimensional operational data to establish a binding relationship between multi-dimensional operational data and globally unique identifiers; it obtains the causal analysis conclusions output by the artificial intelligence platform and transforms the causal analysis conclusions into a globally unique identifier association chain; where the causal analysis conclusions are the cause-effect correspondence of abnormal data derived by the artificial intelligence platform based on multi-dimensional operational data. The generation module is used to generate numerical labels containing current values, predicted values, and status indicators at the corresponding geometric locations of the industrial model based on binding relationships and globally unique identifier association chains; map multi-dimensional operational data into color gradients to generate continuous heat maps on the surface of the industrial model; when the fault-related data contained in the multi-dimensional operational data reaches preset conditions, it highlights and marks the corresponding geometric areas of the industrial model and marks the corresponding globally unique identifier association chains.

Images