Industrial control virtual device resource pool construction method, system, device and medium

Abstract

The invention belongs to the technical field of industrial control, and discloses an industrial control virtual equipment resource pool construction method, system, equipment and medium, and the method comprises the steps: optimizing parameters of a transition probability and an observation probability in a time sequence data statistical algorithm according to time sequence characteristics of an industrial control scene, and extracting resource characteristics; generating a virtual resource model, and constructing a bidirectional mapping table of physical resources and virtual resources; optimizing a clustering algorithm by using a time weight factor, dynamically aggregating the time sequence metadata in combination with a streaming computing framework, and performing anomaly detection on a resource state in combination with the aggregated time sequence metadata based on an anomaly detection technology; based on a layered decoupling flattening architecture, a resource pool is divided into a resource layer, a data layer, a link layer, a technical support layer and an interface layer. According to the method, logic-physical layered decoupling of the industrial control resource pool is realized, seamless access and dynamic optimal distribution of multi-protocol equipment are supported, and high-reliability and low-delay bottom support is provided for users.

Description

Technical Field

[0001] This invention relates to the field of industrial control technology, and in particular to a method, system, equipment and medium for constructing a virtual equipment resource pool for industrial control. Background Technology

[0002] With the rapid development of industry and intelligent manufacturing, the demand for computing resources in industrial control systems is increasing, and they are gradually transforming towards digitalization, networking, and intelligence.

[0003] In typical industrial scenarios, resource types encompass physical devices (such as PLCs and sensors), software services (such as MES and SCADA systems), and data resources (such as production logs and process parameters). These resources are typically distributed across different physical locations, provided by different vendors, and employ diverse communication protocols and data formats. Data interaction between these devices and systems requires complex protocol conversions and interface adaptations. Traditional resource management methods primarily rely on centralized architectures, integrating resources through a single database or middleware. For example, federated database technology accesses multiple heterogeneous data sources through a unified interface. However, as the number of data sources surges, the complexity and maintenance costs of the interface increase significantly, leading to system response delays. Meanwhile, while middleware integration enables cross-platform communication, its customized interface development cycle is long, compatibility is poor, and it struggles to adapt to rapidly changing business needs.

[0004] There are numerous negative examples in the current management of industrial control equipment resources: when existing resource management systems connect to new smart sensors, incompatibility in protocols leads to data acquisition delays of up to several hours, severely impacting real-time monitoring efficiency. More seriously, the "tightly coupled" design of traditional architectures limits system scalability. When a business module needs upgrading or adding new resource types, the entire system often needs to be refactored, which not only increases development costs but may also lead to decreased system stability. For example, the original system cannot flexibly expand computing resources, ultimately forcing a delay in project launch.

[0005] In recent years, with the rapid development and popularization of cloud computing technology, it has provided new ideas for industrial resource management. By pooling computing, storage, and network resources, cloud computing can achieve dynamic allocation and elastic scaling of resources. Mature industrial IoT platforms use virtualization technology to abstract dispersed device resources into unified virtual instances, supporting on-demand access. This model significantly reduces the complexity of resource management, but its initial design is more geared towards general IT scenarios and fails to fully meet the high real-time and high reliability requirements of the industrial control field. Meanwhile, virtualization technology, as the cornerstone of cloud computing, still faces many challenges in its application in industrial control. First, the heterogeneity of industrial equipment far exceeds that of traditional IT equipment. PLC control commands need to respond in milliseconds, while the storage of historical data allows for even higher latency. Existing virtualization models (such as VMware virtual machines) struggle to meet the real-time requirements of different resources. Second, industrial scenarios have extremely high requirements for data security and system stability, and traditional virtualization solutions have shortcomings in fault tolerance mechanisms and fault recovery. If a commercial virtualization platform is directly used to deploy an industrial control system, it may lead to production line downtime accidents caused by brief service interruptions due to virtual machine migration.

[0006] Besides the virtualization technology mentioned above, efficient metadata management is another core issue in resource pool construction. Metadata for industrial resources includes equipment attributes, service capabilities, and dynamic status. Traditional methods manage metadata through static database table structures, but this approach struggles to support rapid querying and intelligent scheduling in the face of dynamically changing industrial environments. Therefore, academia and industry have begun exploring metadata clustering techniques based on machine learning. The improved K-means algorithm studied here can automatically group heterogeneous resources according to functional or performance characteristics by optimizing the selection of initial cluster centers. However, existing research largely focuses on general data domains and lacks adaptation to the specific needs of industrial scenarios. The temporal, strongly correlated, and high-concurrency characteristics of industrial control resources require metadata aggregation algorithms to possess higher real-time performance and robustness.

[0007] Therefore, how to provide a method, system, equipment, and medium for constructing a virtual equipment resource pool for industrial control systems to improve the utilization, flexibility, and scalability of resources in industrial control systems, while meeting the requirements of industrial control systems for real-time performance, reliability, and security, is an urgent problem to be solved. Summary of the Invention

[0008] This invention provides a method, system, device, and medium for constructing an industrial control virtual device resource pool to solve the problems mentioned above in the prior art.

[0009] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or to describe the scope of protection of these embodiments. Its sole purpose is to present some concepts in a simple form as a prelude to the detailed description that follows.

[0010] According to a first aspect of the present invention, a method for constructing an industrial control virtual device resource pool is provided.

[0011] In one embodiment, the method for constructing the industrial control virtual device resource pool includes:

[0012] Based on the temporal characteristics of industrial control scenarios, the parameters of transition probability and observation probability in the time-series data statistical algorithm are optimized, and resource features are extracted by combining dynamic Bayesian networks and Viterbi algorithm.

[0013] The annotation results are determined based on resource characteristics. Based on the annotation results, a dynamic metadata model and ontology knowledge base are constructed. A virtual resource model is generated based on the ontology knowledge base, and a bidirectional mapping table between physical resources and virtual resources is constructed.

[0014] The clustering algorithm is optimized by using time weighting factors, and time-series metadata is dynamically aggregated by combining the streaming computing framework. Based on anomaly detection technology, the resource status is anomaly detected by combining the aggregated time-series metadata.

[0015] Based on a layered, decoupled, and flat architecture, the resource pool is divided into a resource layer, a data layer, a link layer, a technical support layer, and an interface layer.

[0016] In one embodiment, optimizing the parameters of transition probability and observation probability in the time-series data statistical algorithm based on the time-series characteristics of the industrial control scenario, and extracting resource features by combining dynamic Bayesian networks and the Viterbi algorithm, includes:

[0017] Based on the time-series characteristics of industrial control scenarios, the smoothing factor of the transition probability matrix and the parameter weights of the observation probability distribution in the Hidden Markov Model are adjusted to optimize the adaptability to non-steady-state time-series data.

[0018] Dynamic Bayesian networks are used to model the state transition process, and the Viterbi algorithm is combined to infer the hidden state of the resource operation sequence, thereby extracting resource features.

[0019] In one embodiment, the steps of determining the annotation results based on resource characteristics, constructing a dynamic metadata model and ontology knowledge base based on the annotation results, generating a virtual resource model based on the ontology knowledge base, and constructing a bidirectional mapping table between physical resources and virtual resources include:

[0020] The annotation results are determined based on resource characteristics. Based on the annotation results, combined with real-time indicators, resource names, attributes, identifiers, key query fields and mapping relationships, a six-dimensional dynamic metadata model is constructed to realize the structured storage of metadata.

[0021] An ontology knowledge base is built based on a resource description framework, defining semantic relationships between resources, and semantic associations are inferred through query language to achieve semantic unification and cross-protocol interoperability of multi-source heterogeneous resources.

[0022] Based on the semantic relationships in the ontology knowledge base, a virtual resource model is generated, consisting of attribute entities, image entities, virtual machine specification entities, and network entities.

[0023] Virtual resource instances can be quickly generated using templated configurations. Based on real-time optimized container technology, industrial control functions are encapsulated as independent microservices, and dynamic resource adjustment and load balancing are achieved through orchestration tools.

[0024] Construct a bidirectional mapping table between physical and virtual resources, record the real-time status, configuration parameters, and dependencies of resource instances, and store the bidirectional mapping table in a distributed key-value database.

[0025] In one embodiment, the container image incorporates a lightweight runtime environment to reduce resource overhead and supports rapid deployment and hot updates across hardware platforms.

[0026] In one embodiment, the step of optimizing the clustering algorithm using time weighting factors and dynamically aggregating time-series metadata using a streaming computing framework, and then performing anomaly detection on resource status based on anomaly detection technology and the aggregated time-series metadata, includes:

[0027] The partition-based unsupervised clustering algorithm is optimized by using a time weighting factor, and the distance calculation formula is optimized by combining the resource update timestamp and the time-series sensitivity coefficient.

[0028] Based on the sliding time window mechanism, the cluster centers are updated periodically, and the initial cluster centers are dynamically selected by the density peak detection algorithm in combination with the periodic characteristics of resource load.

[0029] We employ streaming metadata aggregation, deploy a streaming computing framework, and utilize time windows and water level mechanisms to process resource status data streams of device heartbeat signals and network latency indicators in real time.

[0030] Structured metadata and topological relationships between resources are stored in relational databases and graph databases respectively. Index optimization and partitioned table design are used for querying, and hot and cold data are managed in layers.

[0031] Construct a long short-term memory neural network, input resource state time series data, capture long-term dependencies through hidden layers, and output state prediction values ​​and confidence intervals;

[0032] The anomaly detection threshold is dynamically calibrated based on the prediction error distribution, and the anomaly signal is identified by comparing the state prediction value with the calibrated anomaly detection threshold.

[0033] Send alarm events, enable backup instances or migrate load to healthy nodes to achieve redundant resource switching, and record abnormal events. Combine association rules to mine and locate potential root causes.

[0034] In one embodiment, the optimized distance calculation formula is:

[0035]

[0036] In the formula, Dist(x) i ,c j (x) represents the data point. i With cluster center c j The weighted distance, w(t) i ) represents the time weighting factor, x i For the i-th data point, c j Let x be the j-th cluster center, n be the number of dimensions of the data points, and x be the number of cluster centers. ik For data point x i The k-th eigenvalue, c jk For cluster center c j The k-th eigenvalue, t i Here, α is the resource update timestamp, and α is the time-series sensitivity coefficient.

[0037] In one embodiment, the resource layer is used to integrate physical hardware resources, software services, and multimodal data streams, and generate standardized resource instances through virtualization technology; it realizes lightweight encapsulation of resources based on containerization technology, supports rapid deployment, dynamic migration, and elastic scaling of resource instances to achieve resource integration and virtualization; it supports full lifecycle management of resources, and realizes dynamic authorization and access control of interface permissions.

[0038] The data layer defines cross-platform communication protocol compatibility specifications and data format standards to ensure seamless interaction of multi-source heterogeneous data; it builds a distributed search engine to enable full-text retrieval, fuzzy matching, and multi-field joint queries of metadata; and it utilizes a hot and cold data separation strategy to achieve dynamic hierarchical data management.

[0039] The link layer is used to build transmission channels based on software-defined networks, support cross-platform resource scheduling, and ensure the bandwidth priority of critical services through traffic shaping technology; it realizes end-to-end encrypted transmission, and combines two-way certificate authentication and a perfect forward confidentiality mechanism to prevent man-in-the-middle attacks and data tampering risks.

[0040] The technical support layer is used to implement container orchestration, support automatic scaling, service mesh and resource quota management to realize toolkit virtualization; develop custom operators to realize real-time aggregation of multi-source metadata, window calculation and state management to realize metadata aggregation; collect resource pool performance indicators in real time, and trigger abnormal notifications through alarm rule engine to realize security monitoring;

[0041] The interface layer is used to define resource directory interfaces and design permission management interfaces to support resource classification retrieval, metadata subscription and version control, and to achieve fine-grained permission allocation and audit log recording; it builds a unified query gateway to support clients to obtain data fields on demand, reducing redundant data transmission; it enables efficient inter-service communication, optimizes serialization efficiency, and reduces the coupling between the business layer and the data layer.

[0042] According to a second aspect of the present invention, an industrial control virtual device resource pool construction system is provided.

[0043] In one embodiment, the industrial control virtual device resource pool construction system includes:

[0044] The resource feature extraction module is used to optimize the parameters of transition probability and observation probability in the time series data statistical algorithm according to the time series characteristics of the industrial control scenario, and to extract resource features by combining dynamic Bayesian network and Viterbi algorithm.

[0045] The virtual resource model generation module is used to determine the annotation results based on resource characteristics, construct a dynamic metadata model and ontology knowledge base based on the annotation results, generate a virtual resource model based on the ontology knowledge base, and construct a two-way mapping table between physical resources and virtual resources.

[0046] The metadata aggregation and optimization module is used to optimize the clustering algorithm using time weight factors and dynamically aggregate time-series metadata in conjunction with the streaming computing framework. Based on anomaly detection technology, it performs anomaly detection on resource status in conjunction with the aggregated time-series metadata.

[0047] The resource pool partitioning module is used to divide the resource pool into a resource layer, a data layer, a link layer, a technical support layer, and an interface layer based on a layered, decoupled, and flat architecture.

[0048] In one embodiment, when the resource feature extraction module optimizes the parameters of transition probability and observation probability in the time-series data statistical algorithm according to the time-series characteristics of the industrial control scenario, and extracts resource features by combining dynamic Bayesian networks and Viterbi algorithms, it adjusts the smoothing factor of the transition probability matrix and the parameter weights of the observation probability distribution in the hidden Markov model according to the time-series characteristics of the industrial control scenario, so as to optimize the adaptability to non-steady-state time-series data; it uses dynamic Bayesian networks to model the state transition process, and combines Viterbi algorithms to realize the hidden state inference of resource operation sequences, thereby realizing the extraction of resource features.

[0049] In one embodiment, the virtual resource model generation module determines the annotation results based on resource characteristics, constructs a dynamic metadata model and ontology knowledge base based on the annotation results, generates a virtual resource model based on the ontology knowledge base, and constructs a bidirectional mapping table between physical and virtual resources. Specifically, it determines the annotation results based on resource characteristics, constructs a six-dimensional dynamic metadata model based on the annotation results, combined with real-time indicators, resource names, attributes, identifiers, key query fields, and mapping relationships, to achieve structured storage of metadata; it constructs an ontology knowledge base based on a resource description framework, defines semantic relationships between resources, and infers semantic associations through query language to achieve semantic unification and cross-protocol interoperability of multi-source heterogeneous resources; it generates a virtual resource model composed of attribute entities, image entities, virtual machine specification entities, and network entities based on semantic associations in the ontology knowledge base; it quickly generates virtual resource instances using templated configuration, encapsulates industrial control functions into independent microservices based on real-time optimized container technology, and achieves dynamic resource adjustment and load balancing through orchestration tools; it constructs a bidirectional mapping table between physical and virtual resources, records the real-time status, configuration parameters, and dependencies of resource instances, and stores the bidirectional mapping table in a distributed key-value database.

[0050] In one embodiment, the container image incorporates a lightweight runtime environment to reduce resource overhead and supports rapid deployment and hot updates across hardware platforms.

[0051] In one embodiment, the metadata aggregation optimization module optimizes the clustering algorithm using a time weighting factor and dynamically aggregates time-series metadata using a streaming computing framework. Based on anomaly detection technology, it performs anomaly detection on resource status using the aggregated time-series metadata. Specifically, it optimizes the partition-based unsupervised clustering algorithm using a time weighting factor and optimizes the distance calculation formula by combining resource update timestamps and time-series sensitivity coefficients. Based on a sliding time window mechanism, it periodically updates cluster centers and dynamically selects initial cluster centers using a density peak detection algorithm, taking into account the periodic characteristics of resource load. It employs streaming metadata aggregation, deploys a streaming computing framework, and utilizes time windows and waterline mechanisms to process device heartbeats in real time. Resource status data streams for signal and network latency metrics; structured metadata and resource topology relationships are stored in relational databases and graph databases respectively, with index optimization and partitioned table design used for querying, and hot and cold data are managed hierarchically; a long short-term memory neural network is constructed, taking resource status time-series data as input, capturing long-term dependencies through hidden layers, and outputting status prediction values ​​and confidence intervals; anomaly judgment thresholds are dynamically calibrated based on the prediction error distribution, and anomaly signals are identified based on the comparison results between the status prediction values ​​and the calibrated anomaly judgment thresholds; alarm events are sent, backup instances are activated or the load is migrated to healthy nodes to achieve redundant resource switching, and anomaly events are recorded, and potential root causes are located by combining association rules.

[0052] In one embodiment, the optimized distance calculation formula is:

[0053]

[0054] In the formula, Dist(x) i ,c j (x) represents the data point. i With cluster center c j The weighted distance, w(t) i ) represents the time weighting factor, x i For the i-th data point, c j Let x be the j-th cluster center, n be the number of dimensions of the data points, and x be the number of cluster centers. ik For data point x i The k-th eigenvalue, c jk For cluster center c j The k-th eigenvalue, t i Here, α is the resource update timestamp, and α is the time-series sensitivity coefficient.

[0055] In one embodiment, the resource layer integrates physical hardware resources, software services, and multimodal data streams, generating standardized resource instances through virtualization technology; it achieves lightweight encapsulation of resources based on containerization technology, supporting rapid deployment, dynamic migration, and elastic scaling of resource instances to achieve resource integration and virtualization; it supports full lifecycle management of resources, enabling dynamic authorization and access control of interface permissions; the data layer defines cross-platform communication protocol compatibility specifications and data format standards to ensure seamless interaction of multi-source heterogeneous data; it builds a distributed search engine to achieve full-text retrieval, fuzzy matching, and multi-field joint queries of metadata; and it utilizes a hot and cold data separation strategy to achieve dynamic data hierarchical management; the link layer constructs transmission channels based on software-defined networks, supports cross-platform resource scheduling, and ensures bandwidth priority for critical services through traffic shaping technology. The system implements end-to-end encrypted transmission, combining two-way certificate authentication and a robust forward confidentiality mechanism to prevent man-in-the-middle attacks and data tampering risks. The technical support layer enables container orchestration, supporting automatic scaling, service mesh, and resource quota management for toolset virtualization. Custom operators are developed to achieve real-time aggregation of multi-source metadata, window calculation, and state management for metadata aggregation. Real-time collection of resource pool performance metrics and triggering of anomaly notifications via an alarm rule engine enable security monitoring. The interface layer defines resource directory interfaces and designs permission management interfaces to support resource classification retrieval, metadata subscription, and version control, enabling fine-grained permission allocation and audit log recording. A unified query gateway is built to support clients in retrieving data fields on demand, reducing redundant data transmission. Efficient inter-service communication is achieved, serialization efficiency is optimized, and the coupling between the business layer and the data layer is reduced.

[0056] According to a third aspect of the present invention, a computer device is provided.

[0057] In some embodiments, the computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described above.

[0058] According to a fourth aspect of the present invention, a computer-readable storage medium is provided.

[0059] In one embodiment, a computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the above method.

[0060] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0061] 1) This invention improves the transition and observation probability parameters of the Hidden Markov Model to meet the timing and high real-time requirements of industrial control scenarios; it adds a "real-time indicator" dimension, combines RDF semantic modeling to uniformly describe multi-source heterogeneous resources, and supports millisecond-level response quantization; it optimizes the K-means clustering algorithm by introducing a time weight factor, dynamically aggregates time-series metadata using a streaming computing framework, and embeds an anomaly detection module, significantly improving the robustness and self-healing capability of resource scheduling. These technologies synergistically achieve logical-physical layer decoupling of the industrial control resource pool, supporting seamless access and dynamic optimization allocation of multi-protocol devices, providing users with highly reliable and low-latency underlying support.

[0062] 2) Intelligent resource integration and dynamic scheduling capabilities: Through an improved Hidden Markov Model and a six-dimensional dynamic metadata model, efficient integration of industrial control equipment, services, and data is achieved, supporting unified description of multiple protocols, solving the problem of protocol fragmentation, and improving resource access efficiency by 40%. The K-means clustering algorithm is optimized based on the time weight factor, combined with a streaming computing engine to aggregate metadata in real time and dynamically perceive changes in resource status.

[0063] 3) Highly scalable layered architecture and robust security: A five-layer architecture—resource layer, data layer, link layer, technical support layer, and interface layer—is adopted to decouple logic from physical components. Containerization technology supports second-level resource scaling, tiered management of hot and cold data, and on-demand GraphQL queries, optimizing response time. End-to-end encrypted transmission, two-way authentication, and real-time Prometheus monitoring prevent data tampering and man-in-the-middle attacks.

[0064] 4) Low-cost deployment and efficient operation and maintenance: Provides RESTful API and gRPC protocol, supports seamless integration with third-party systems, and significantly shortens the access time for new devices. Closed-loop feedback mechanism combined with dynamic metadata management reduces manual intervention; templated resource configuration and lightweight microservice encapsulation reduce development complexity.

[0065] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0066] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0067] Figure 1 This is a flowchart illustrating a method for constructing a resource pool for industrial control virtual equipment according to an exemplary embodiment;

[0068] Figure 2 This is a structural block diagram illustrating an industrial control virtual device resource pool construction system according to an exemplary embodiment;

[0069] Figure 3 This is a schematic diagram illustrating the principle of a method for constructing a resource pool for industrial control virtual devices according to an exemplary embodiment;

[0070] Figure 4 This is a thumbnail diagram illustrating a heterogeneous resource virtualization and modeling technology process according to an exemplary embodiment;

[0071] Figure 5 This is a diagram illustrating a time-aware dynamic optimization and intelligent self-healing closed-loop architecture according to an exemplary embodiment;

[0072] Figure 6 This is a diagram illustrating a layered, decoupled, flattened architecture according to an exemplary embodiment;

[0073] Figure 7 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment. Detailed Implementation

[0074] The following description and accompanying drawings fully illustrate specific embodiments described herein to enable those skilled in the art to practice them. Some embodiments may include or substitute parts and features of other embodiments. The scope of the embodiments herein encompasses the entire scope of the claims and all available equivalents thereof. Throughout this document, the terms “first,” “second,” etc., are used only to distinguish one element from another without requiring or implying any actual relationship or order between the elements. Indeed, a first element can also be referred to as a second element, and vice versa. Furthermore, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a structure, apparatus, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a structure, apparatus, or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the structure, apparatus, or device that includes said element. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments; similar or identical parts between embodiments can be referred to interchangeably.

[0075] The terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer" used in this document to indicate orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings. They are used solely for the convenience of describing the document and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In the description herein, unless otherwise specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two elements; they can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0076] In this document, unless otherwise stated, the term "multiple" means two or more.

[0077] In this article, the character " / " indicates that the objects before and after it are in an "or" relationship. For example, A / B means: A or B.

[0078] In this article, the term "and / or" describes an association between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or A and B.

[0079] It should be understood that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order constraint on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the diagram may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0080] The modules in the apparatus or system of this application can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0081] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0082] Figure 1An embodiment of a method for constructing a virtual equipment resource pool for industrial control according to the present invention is shown.

[0083] In this optional embodiment, the method for constructing the industrial control virtual device resource pool includes:

[0084] Step S101: Based on the time-series characteristics of the industrial control scenario, optimize the parameters of transition probability and observation probability in the time-series data statistical algorithm, and combine dynamic Bayesian network and Viterbi algorithm to extract resource features.

[0085] Step S102: Determine the annotation results based on resource characteristics, construct a dynamic metadata model and ontology knowledge base based on the annotation results, generate a virtual resource model based on the ontology knowledge base, and construct a bidirectional mapping table between physical resources and virtual resources.

[0086] Step S103: Optimize the clustering algorithm using time weighting factors, and dynamically aggregate time-series metadata using a streaming computing framework. Based on anomaly detection technology, perform anomaly detection on resource status using the aggregated time-series metadata.

[0087] Step S104: Based on the layered, decoupled, and flattened architecture, the resource pool is divided into a resource layer, a data layer, a link layer, a technical support layer, and an interface layer.

[0088] In the above steps, step S101 extracts resource features for step S102. These resource features are further identified, classified, and semantically annotated by domain experts or knowledge base systems to obtain structured and standardized information, i.e., annotation results. Subsequently, step S102 constructs a dynamic metadata model and ontology knowledge base based on the annotation results, and generates a virtual resource model and a two-way mapping table between physical resources and virtual resources.

[0089] Step S102 provides the necessary structured data foundation and metadata model support for step S103, while step S103 further realizes dynamic real-time analysis, anomaly monitoring and reliability management of metadata. The two steps reflect the logical progression of "structured data storage and virtualization → dynamic data aggregation and analysis monitoring".

[0090] The dynamic monitoring and anomaly detection capabilities in step S103 provide the necessary technical support and functional requirements for the resource pool architecture design in step S104; while the clear architecture design in step S104 further clarifies the specific deployment location and hierarchical responsibilities of the clustering, streaming computing and anomaly monitoring technologies involved in step S103, thereby realizing the logical progression from "technical function implementation → architecture standardization implementation".

[0091] In this optional embodiment, the step of optimizing the parameters of transition probability and observation probability in the time-series data statistical algorithm based on the time-series characteristics of the industrial control scenario, and extracting resource features by combining dynamic Bayesian networks and Viterbi algorithms, includes:

[0092] Based on the time-series characteristics of industrial control scenarios, the smoothing factor of the transition probability matrix and the parameter weights of the observation probability distribution in the Hidden Markov Model are adjusted to optimize the adaptability to non-steady-state time-series data.

[0093] Dynamic Bayesian networks are used to model the state transition process, and the Viterbi algorithm is combined to infer the hidden state of the resource operation sequence, thereby extracting resource features.

[0094] In this optional embodiment, the steps of determining the annotation results based on resource characteristics, constructing a dynamic metadata model and ontology knowledge base based on the annotation results, generating a virtual resource model based on the ontology knowledge base, and constructing a bidirectional mapping table between physical resources and virtual resources include:

[0095] The annotation results are determined based on resource characteristics. Based on the annotation results, combined with real-time indicators, resource names, attributes, identifiers, key query fields and mapping relationships, a six-dimensional dynamic metadata model is constructed to realize the structured storage of metadata.

[0096] An ontology knowledge base is built based on a resource description framework, defining semantic relationships between resources, and semantic associations are inferred through query language to achieve semantic unification and cross-protocol interoperability of multi-source heterogeneous resources.

[0097] Based on the semantic relationships in the ontology knowledge base, a virtual resource model is generated, consisting of attribute entities, image entities, virtual machine specification entities, and network entities.

[0098] Virtual resource instances can be quickly generated using templated configurations. Based on real-time optimized container technology, industrial control functions are encapsulated as independent microservices, and dynamic resource adjustment and load balancing are achieved through orchestration tools.

[0099] Construct a bidirectional mapping table between physical and virtual resources, record the real-time status, configuration parameters, and dependencies of resource instances, and store the bidirectional mapping table in a distributed key-value database.

[0100] In this optional embodiment, the container image incorporates a lightweight runtime environment to reduce resource overhead and supports rapid deployment and hot updates across hardware platforms.

[0101] In this optional embodiment, the step of optimizing the clustering algorithm using time weighting factors and dynamically aggregating time-series metadata using a streaming computing framework, and then performing anomaly detection on resource status based on anomaly detection technology and the aggregated time-series metadata, includes:

[0102] The partition-based unsupervised clustering algorithm is optimized by using a time weighting factor, and the distance calculation formula is optimized by combining the resource update timestamp and the time-series sensitivity coefficient.

[0103] Based on the sliding time window mechanism, the cluster centers are updated periodically, and the initial cluster centers are dynamically selected by the density peak detection algorithm in combination with the periodic characteristics of resource load.

[0104] We employ streaming metadata aggregation, deploy a streaming computing framework, and utilize time windows and water level mechanisms to process resource status data streams of device heartbeat signals and network latency indicators in real time.

[0105] Structured metadata and topological relationships between resources are stored in relational databases and graph databases respectively. Index optimization and partitioned table design are used for querying, and hot and cold data are managed in layers.

[0106] Construct a long short-term memory neural network, input resource state time series data, capture long-term dependencies through hidden layers, and output state prediction values ​​and confidence intervals;

[0107] The anomaly detection threshold is dynamically calibrated based on the prediction error distribution, and the anomaly signal is identified by comparing the state prediction value with the calibrated anomaly detection threshold.

[0108] Send alarm events, enable backup instances or migrate load to healthy nodes to achieve redundant resource switching, and record abnormal events. Combine association rules to mine and locate potential root causes.

[0109] In this optional embodiment, the optimized distance calculation formula is:

[0110]

[0111] In the formula, Dist(x) i ,c j (x) represents the data point. i With cluster center c j The weighted distance, w(t) i ) represents the time weighting factor, x i For the i-th data point, c j Let x be the j-th cluster center, n be the number of dimensions of the data points, and x be the number of cluster centers. ik For data point x i The k-th eigenvalue, c jk For cluster center c j The k-th eigenvalue, t i Here, α is the resource update timestamp, and α is the time-series sensitivity coefficient.

[0112] In this optional embodiment, the resource layer is used to integrate physical hardware resources, software services, and multimodal data streams, and generate standardized resource instances through virtualization technology; it realizes lightweight encapsulation of resources based on containerization technology, supports rapid deployment, dynamic migration, and elastic scaling of resource instances to achieve resource integration and virtualization; it supports full lifecycle management of resources, and realizes dynamic authorization and access control of interface permissions;

[0113] The data layer defines cross-platform communication protocol compatibility specifications and data format standards to ensure seamless interaction of multi-source heterogeneous data; it builds a distributed search engine to enable full-text retrieval, fuzzy matching, and multi-field joint queries of metadata; and it utilizes a hot and cold data separation strategy to achieve dynamic hierarchical data management.

[0114] The link layer is used to build transmission channels based on software-defined networks, support cross-platform resource scheduling, and ensure the bandwidth priority of critical services through traffic shaping technology; it realizes end-to-end encrypted transmission, and combines two-way certificate authentication and a perfect forward confidentiality mechanism to prevent man-in-the-middle attacks and data tampering risks.

[0115] The technical support layer is used to implement container orchestration, support automatic scaling, service mesh and resource quota management to realize toolkit virtualization; develop custom operators to realize real-time aggregation of multi-source metadata, window calculation and state management to realize metadata aggregation; collect resource pool performance indicators in real time, and trigger abnormal notifications through alarm rule engine to realize security monitoring;

[0116] The interface layer is used to define resource directory interfaces and design permission management interfaces to support resource classification retrieval, metadata subscription and version control, and to achieve fine-grained permission allocation and audit log recording; it builds a unified query gateway to support clients to obtain data fields on demand, reducing redundant data transmission; it enables efficient inter-service communication, optimizes serialization efficiency, and reduces the coupling between the business layer and the data layer.

[0117] Figure 2 An embodiment of an industrial control virtual device resource pool construction system according to the present invention is shown.

[0118] In this optional embodiment, the industrial control virtual device resource pool construction system includes:

[0119] The resource feature extraction module 201 is used to optimize the parameters of transition probability and observation probability in the time series data statistical algorithm according to the time series characteristics of the industrial control scenario, and to extract resource features by combining dynamic Bayesian network and Viterbi algorithm.

[0120] The virtual resource model generation module 202 is used to determine the annotation results based on resource characteristics, construct a dynamic metadata model and ontology knowledge base based on the annotation results, generate a virtual resource model based on the ontology knowledge base, and construct a bidirectional mapping table between physical resources and virtual resources.

[0121] The metadata aggregation and optimization module 203 is used to optimize the clustering algorithm using time weight factors and dynamically aggregate time-series metadata in conjunction with the streaming computing framework. Based on anomaly detection technology, it performs anomaly detection on resource status in conjunction with the aggregated time-series metadata.

[0122] The resource pool partitioning module 204 is used to divide the resource pool into a resource layer, a data layer, a link layer, a technical support layer, and an interface layer based on a layered, decoupled, and flat architecture.

[0123] To facilitate understanding of the above technical solutions of the present invention, the following further describes the above technical solutions of the present invention from the perspectives of architecture and principle, as follows:

[0124] The purpose of this invention is to provide a method for constructing a resource pool for industrial control virtual equipment, such as... Figure 3 As shown, virtualization technology and resource pooling management enable unified management and dynamic allocation of computing resources in industrial control systems, improving resource utilization, flexibility, and scalability while meeting the real-time, reliability, and security requirements of industrial control systems. The resource pool construction process includes the following core sub-processes:

[0125] (1) Deploy directly on devices such as PLCs through embedded agents to collect device register data, IO signals and fault codes in real time, and support native parsing of multiple industrial protocols to avoid the protocol conversion delay of traditional gateways; (2) Use Apache Kafka to build a distributed message queue to perform timestamp alignment, out-of-order correction and missing value interpolation on the collected raw data stream to achieve time-series data stream standardization; (3) Perform resource health assessment and load balancing scheduling periodically to ensure that high-priority tasks are always allocated to low-latency instances.

[0126] This invention addresses the millisecond-level real-time requirements of industrial control systems, resolving the control command latency and jitter issues caused by the general architecture of traditional virtualization solutions. It directly embeds data into the PLC register layer via a native agent on the device side, avoiding the multi-layer protocol conversion latency associated with IoT gateways. To achieve the above objectives, this invention employs the following technical solution:

[0127] 1) Heterogeneous resource virtualization and modeling (e.g.) Figure 4 As shown):

[0128] An improved Hidden Markov Model (HMM) is adopted, combined with the temporal characteristics of industrial control scenarios. By adjusting the smoothing factor of the transition probability matrix and the parameter weights of the observation probability distribution, the model's adaptability to non-steady-state temporal data is optimized. Specifically, in this embodiment, a Dynamic Bayesian Network (DBN) is introduced to model the state transition process, and the Viterbi algorithm is used to infer the hidden state of resource operation sequences, thereby accurately extracting resource features (in this embodiment, resource features specifically refer to the temporal characteristics of equipment or components in the industrial control scenario, such as state changes, performance indicators, operation modes, and load characteristics during operation, specifically manifested as state sequences, operating trends, load modes, event modes, etc., reflecting the operating behavior mode of the equipment); a six-dimensional dynamic metadata model is constructed based on the annotation results. On the basis of the traditional five-dimensional tuple (resource name, attribute, identifier, query key field, mapping relationship), a "real-time indicator" dimension is added to quantify the dynamic performance parameters of resources. Structured storage of metadata is achieved through XML and JSON-LD formats, supporting cross-platform compatibility and dynamic expansion. An ontology knowledge base is built based on a resource description framework, defining semantic relationships between resources and using OWL to describe resource classes, attributes, and instances. Automated semantic association reasoning is achieved through the SPARQL query language, and combined with metadata specifications of industrial communication protocols, unified semantic mapping and cross-protocol interoperability of multi-source heterogeneous resources are realized.

[0129] During the operation of industrial control systems, the operating states of equipment often exhibit significant non-stationary characteristics as operating conditions change. Specifically, the transitions between equipment states are not uniform but exhibit clear phased patterns (such as equipment startup, load changes, and maintenance phases); the observed equipment state data often show time-varying distribution characteristics, with significant changes during peak load periods and relative stability during off-peak periods. Therefore, standard Hidden Markov Models (HMMs) cannot effectively adapt to the non-stationary temporal characteristics of industrial control scenarios. This invention addresses these issues by adjusting the smoothing factor of the transition probability matrix and the parameter weights of the observation probability distribution to improve the model's adaptability to non-stationary characteristics and enhance the accurate identification of resource state sequences.

[0130] Data window segmentation: Determine the appropriate data analysis window (e.g., several minutes or several hours) based on the periodic characteristics of the industrial control scenario.

[0131] Time series feature analysis: Statistical analysis of the number of state transitions and fluctuations in observed data within each time window.

[0132] Dynamically determine the smoothing factor λ: Calculate the stability of state transitions within the window (such as transition frequency and entropy value) and determine the smoothing factor: the greater the difference in stability, the greater λ should be; if the stability is better, then λ should be decreased.

[0133] Dynamic weight adjustment: The weights of each Gaussian component are calculated based on the variance changes of the observed data within the window, thereby achieving dynamic adaptive adjustment of the observation probability.

[0134] Model parameter update: Update the model parameters (transition probability matrix and observation probability parameters) after each window ends.

[0135] Through the above methods, continuous iterative updates ensure the model adapts in real-time to the non-steady-state characteristics of industrial control scenarios. Based on semantic associations in the ontology knowledge base, a virtual resource model is generated, consisting of attribute entities, image entities, virtual machine specification entities, and network entities. Resource instances are rapidly generated through templated configuration, and resource lifecycle management strategies are defined. Real-time optimized container technology replaces traditional virtual machines, encapsulating industrial control function modules as independent microservices. Elastic scaling and load balancing of resources are achieved through Kubernetes orchestration tools. Container images incorporate a lightweight runtime environment, reducing resource overhead while supporting rapid deployment and hot updates across hardware platforms. A bidirectional mapping table between physical and virtual resources is established, recording the real-time status, configuration parameters, and dependencies of resource instances. The mapping table is stored in a distributed key-value database, employing a consistent hashing algorithm for high availability and low-latency queries, and a version control mechanism ensures the consistency and traceability of the mapping relationships.

[0136] 2) Dynamic aggregation and optimization of metadata (e.g.) Figure 5 As shown):

[0137] Using an improved K-means algorithm, a weight factor w is introduced. t Furthermore, the distance calculation formula is optimized by using a time weighting factor to improve the partition-based unsupervised clustering algorithm, and by combining the resource update timestamp and time-series sensitivity coefficient to optimize the distance calculation formula; based on the difference between the update time of resource data collection and the current analysis time, a time weighting factor w(t) for the data points is defined. i This is used to adjust the weights in the distance calculation formula. The optimized distance calculation formula is as follows:

[0138]

[0139] In the formula, Dist(x) i ,c j (x) represents the data point. i With cluster center c j The weighted distance, w(t) i ) represents the time weighting factor, x i For the i-th data point, c j Let x be the j-th cluster center, n be the number of dimensions of the data points, and x be the number of cluster centers. ik For data point x i The k-th eigenvalue, cjk For cluster center c j The k-th eigenvalue, t i Let w(t) be the resource update timestamp, α be the time-sensitivity coefficient, which controls the importance of the relative time difference between data, and w(t) be the time-sensitivity coefficient. i This reflects the absolute timeliness of the data, ensuring that the clustering results more accurately reflect the actual state and timeliness characteristics of the data in the industrial control scenario.

[0140] Meanwhile, a sliding time window mechanism is adopted to periodically update cluster centers, adapting to the temporal changes in resource load fluctuations and data update frequency. Combining the periodic characteristics of resource load, an initial cluster center is dynamically selected through a density peak detection algorithm to improve classification stability. Streaming metadata aggregation is employed, deploying the Apache Flink streaming computing framework. Through time windows and watermark mechanisms, resource status data streams such as device heartbeat signals and network latency indicators are processed in real time. A checkpoint mechanism is used to achieve persistence of streaming processing status, ensuring data consistency during fault recovery. Structured metadata is stored in a relational database, with index optimization and partitioned table design improving query efficiency. Meanwhile, topological relationships between resources are stored in a graph database, supporting complex relational queries and multi-hop inference. Finally, hot and cold data are managed in layers: high-frequency data is cached in Redis, and low-frequency data is archived in HDFS, balancing performance and cost.

[0141] Specifically, the initial cluster centers are dynamically selected using the density peak detection algorithm as follows:

[0142] During the data preparation phase, determine the sliding window size and step size, and collect time-series resource data periodically;

[0143] In the density calculation stage, the local density and relative distance features of data points are calculated within each sliding window;

[0144] Cluster center selection utilizes the density peak detection method to calculate the product of local density and relative distance features, and dynamically selects cluster centers.

[0145] During the clustering execution phase, the partitioning clustering algorithm is executed using dynamically selected initial cluster centers.

[0146] During the cluster center update phase, the window slides periodically and repeats the above process to continuously and dynamically update the clustering results.

[0147] To achieve intelligent system monitoring and self-healing, an LSTM neural network is constructed. Inputting time-series resource status data, it captures long-term dependencies through hidden layers, outputting predicted status values ​​and confidence intervals. The anomaly detection threshold is dynamically calibrated based on the prediction error distribution to reduce false alarms and missed alarms. When an anomaly signal is triggered, an alarm event is sent to the monitoring center via a message queue, including the anomaly type, impact scope, and remediation suggestions. An elastic scaling interface is integrated to enable backup instances or migrate workloads to healthy nodes, ensuring service continuity and enabling redundant resource switching. Anomaly events are recorded in the log system, and association rules are used to mine and locate potential root causes. Anomaly data is fed back to the clustering model, and parameters are dynamically adjusted to optimize classification accuracy and system robustness for easier root cause analysis.

[0148] 3) Layered and decoupled flat architecture (e.g.) Figure 6 As shown):

[0149] The resource pool is divided into five parts through a layered design: resource layer, data layer, link layer, technical support layer, and interface layer.

[0150] The resource layer is responsible for integrating physical hardware resources, software services, and multimodal data streams, generating standardized resource instances through virtualization technology. It also utilizes containerization technology to achieve lightweight resource encapsulation, supporting rapid deployment, dynamic migration, and elastic scaling of resource instances to achieve resource integration and virtualization. It provides a RESTful API interface set to support full lifecycle management of resources and implements dynamic authorization and access control of interface permissions through the OAuth 2.0 protocol.

[0151] The data layer defines cross-platform communication protocol compatibility specifications and data format standards to ensure seamless interaction of multi-source heterogeneous data. It also builds a distributed search engine based on ElasticSearch, enabling full-text retrieval of metadata, fuzzy matching, and multi-field joint queries, optimizing response time to within 100ms. A hot / cold data separation strategy is employed for dynamic hierarchical data management; frequently accessed data is cached in an in-memory database, with cache hit rate optimized using the LRU algorithm. Low-frequency historical data is persistently stored in a distributed file system, supporting horizontal scaling and disaster recovery backup.

[0152] The link layer's role is to build transmission channels based on software-defined networks, support cross-platform resource scheduling, and ensure bandwidth priority for critical services through traffic shaping technology. It employs the TLS 1.3 protocol for end-to-end encrypted transmission, combined with two-way certificate authentication and a robust forward confidentiality mechanism to prevent man-in-the-middle attacks and data tampering risks. Integrated link redundancy and load balancing algorithms ensure transmission stability and low latency in high-concurrency scenarios.

[0153] The technical support layer is responsible for implementing container orchestration based on Kubernetes, supporting automatic scaling, service mesh, and resource quota management to achieve toolkit virtualization. Custom operators are developed using Apache Flink to achieve real-time aggregation of multi-source metadata, window computation, and state management for metadata aggregation. A Prometheus+Grafana monitoring stack is deployed to collect resource pool performance metrics in real time and trigger anomaly notifications through an alerting rule engine for security monitoring.

[0154] The interface layer defines resource directory interfaces and designs permission management interfaces to support resource categorization and retrieval, metadata subscription, and version control, enabling fine-grained permission allocation and audit logging. A unified query gateway is built using GraphQL, allowing clients to retrieve data fields on demand and reducing redundant data transmission. Efficient inter-service communication is achieved through the gRPC protocol, and serialization efficiency is optimized using Protocol Buffers, reducing the coupling between the business layer and the data layer.

[0155] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 7 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0156] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0157] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0158] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0159] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0160] This invention is not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this invention is limited only by the appended claims.

Claims

1. A method for constructing a resource pool for industrial control virtual equipment, characterized in that, include: Based on the temporal characteristics of industrial control scenarios, the parameters of transition probability and observation probability in the time-series data statistical algorithm are optimized, and resource features are extracted by combining dynamic Bayesian networks and Viterbi algorithm. The annotation results are determined based on resource characteristics. Based on the annotation results, a dynamic metadata model and ontology knowledge base are constructed. A virtual resource model is generated based on the ontology knowledge base, and a bidirectional mapping table between physical resources and virtual resources is constructed. The clustering algorithm is optimized by using time weighting factors, and time-series metadata is dynamically aggregated by combining the streaming computing framework. Based on anomaly detection technology, the resource status is anomaly detected by combining the aggregated time-series metadata. Based on a layered, decoupled, and flat architecture, the resource pool is divided into a resource layer, a data layer, a link layer, a technical support layer, and an interface layer.

2. The method for constructing a resource pool for industrial control virtual equipment according to claim 1, characterized in that, The optimization of the transition probability and observation probability parameters in the time-series data statistical algorithm based on the time-series characteristics of industrial control scenarios, and the extraction of resource features by combining dynamic Bayesian networks and the Viterbi algorithm, includes: Based on the time-series characteristics of industrial control scenarios, the smoothing factor of the transition probability matrix and the parameter weights of the observation probability distribution in the Hidden Markov Model are adjusted to optimize the adaptability to non-steady-state time-series data. Dynamic Bayesian networks are used to model the state transition process, and the Viterbi algorithm is combined to infer the hidden state of the resource operation sequence, thereby extracting resource features.

3. The method for constructing a resource pool for industrial control virtual equipment according to claim 1, characterized in that, The steps of determining the annotation results based on resource characteristics, constructing a dynamic metadata model and ontology knowledge base based on the annotation results, generating a virtual resource model based on the ontology knowledge base, and constructing a bidirectional mapping table between physical resources and virtual resources include: The annotation results are determined based on resource characteristics. Based on the annotation results, combined with real-time indicators, resource names, attributes, identifiers, key query fields and mapping relationships, a six-dimensional dynamic metadata model is constructed to realize the structured storage of metadata. An ontology knowledge base is built based on a resource description framework, defining semantic relationships between resources, and semantic associations are inferred through query language to achieve semantic unification and cross-protocol interoperability of multi-source heterogeneous resources. Based on the semantic relationships in the ontology knowledge base, a virtual resource model is generated, consisting of attribute entities, image entities, virtual machine specification entities, and network entities. Virtual resource instances can be quickly generated using templated configurations. Based on real-time optimized container technology, industrial control functions are encapsulated as independent microservices, and dynamic resource adjustment and load balancing are achieved through orchestration tools. Construct a bidirectional mapping table between physical and virtual resources, record the real-time status, configuration parameters, and dependencies of resource instances, and store the bidirectional mapping table in a distributed key-value database.

4. The method for constructing a resource pool for industrial control virtual equipment according to claim 3, characterized in that, Container images have a built-in lightweight runtime environment to reduce resource overhead and support rapid deployment and hot updates across hardware platforms.

5. The method for constructing a resource pool for industrial control virtual equipment according to claim 1, characterized in that, The method of optimizing the clustering algorithm using time weighting factors and dynamically aggregating time-series metadata using a streaming computing framework, and then performing anomaly detection on resource status based on anomaly detection technology and the aggregated time-series metadata, includes: The partition-based unsupervised clustering algorithm is optimized by using a time weighting factor, and the distance calculation formula is optimized by combining the resource update timestamp and the time-series sensitivity coefficient. Based on the sliding time window mechanism, the cluster centers are updated periodically, and the initial cluster centers are dynamically selected by the density peak detection algorithm in combination with the periodic characteristics of resource load. We employ streaming metadata aggregation, deploy a streaming computing framework, and utilize time windows and water level mechanisms to process resource status data streams of device heartbeat signals and network latency indicators in real time. Structured metadata and topological relationships between resources are stored in relational databases and graph databases respectively. Index optimization and partitioned table design are used for querying, and hot and cold data are managed in layers. Construct a long short-term memory neural network, input resource state time series data, capture long-term dependencies through hidden layers, and output state prediction values ​​and confidence intervals; The anomaly detection threshold is dynamically calibrated based on the prediction error distribution, and the anomaly signal is identified by comparing the state prediction value with the calibrated anomaly detection threshold. Send alarm events, enable backup instances or migrate load to healthy nodes to achieve redundant resource switching, and record abnormal events. Combine association rules to mine and locate potential root causes.

6. The method for constructing a resource pool for industrial control virtual equipment according to claim 5, characterized in that, The optimized distance calculation formula is as follows: In the formula, Dist(x) i ,c j (x) represents the data point. i With cluster center c j The weighted distance, w(t) i ) represents the time weighting factor, x i For the i-th data point, c j Let x be the j-th cluster center, n be the number of dimensions of the data points, and x be the number of cluster centers. ik For data point x i The k-th eigenvalue, c jk For cluster center c j The k-th eigenvalue, t i Here, α is the resource update timestamp, and α is the time-series sensitivity coefficient.

7. The method for constructing a resource pool for industrial control virtual equipment according to claim 1, characterized in that, The resource layer is used to integrate physical hardware resources, software services, and multimodal data streams, and generate standardized resource instances through virtualization technology; it realizes lightweight encapsulation of resources based on containerization technology, supports rapid deployment, dynamic migration, and elastic scaling of resource instances to achieve resource integration and virtualization; it supports full lifecycle management of resources, and realizes dynamic authorization and access control of interface permissions. The data layer defines cross-platform communication protocol compatibility specifications and data format standards to ensure seamless interaction of multi-source heterogeneous data; it builds a distributed search engine to enable full-text retrieval, fuzzy matching, and multi-field joint queries of metadata; and it utilizes a hot and cold data separation strategy to achieve dynamic hierarchical data management. The link layer is used to build transmission channels based on software-defined networks, support cross-platform resource scheduling, and ensure bandwidth priority for critical services through traffic shaping technology. It achieves end-to-end encrypted transmission, combining two-way certificate authentication with a perfect forward confidentiality mechanism to prevent man-in-the-middle attacks and data tampering risks; The technical support layer is used to implement container orchestration, support automatic scaling, service mesh and resource quota management, and realize toolset virtualization. Development Custom operators enable real-time aggregation of multi-source metadata, window calculation, and state management to achieve metadata aggregation; Real-time collection of resource pool performance metrics, and triggering of anomaly notifications through the alarm rule engine to achieve security monitoring; The interface layer is used to define resource directory interfaces and design permission management interfaces to support resource classification retrieval, metadata subscription and version control, and to achieve fine-grained permission allocation and audit log recording; it also builds a unified query gateway to support clients to obtain data fields on demand and reduce redundant data transmission. To achieve efficient inter-service communication, optimize serialization efficiency, and reduce the coupling between the business layer and the data layer.

8. A system for constructing a resource pool for industrial control virtual equipment, characterized in that, include: The resource feature extraction module is used to optimize the parameters of transition probability and observation probability in the time series data statistical algorithm according to the time series characteristics of the industrial control scenario, and to extract resource features by combining dynamic Bayesian network and Viterbi algorithm. The virtual resource model generation module is used to determine the annotation results based on resource characteristics, construct a dynamic metadata model and ontology knowledge base based on the annotation results, generate a virtual resource model based on the ontology knowledge base, and construct a two-way mapping table between physical resources and virtual resources. The metadata aggregation and optimization module is used to optimize the clustering algorithm using time weight factors and dynamically aggregate time-series metadata in conjunction with the streaming computing framework. Based on anomaly detection technology, it performs anomaly detection on resource status in conjunction with the aggregated time-series metadata. The resource pool partitioning module is used to divide the resource pool into a resource layer, a data layer, a link layer, a technical support layer, and an interface layer based on a layered, decoupled, and flat architecture.

9. The industrial control virtual equipment resource pool construction system according to claim 8, characterized in that, When the resource feature extraction module optimizes the parameters of transition probability and observation probability in the time series data statistical algorithm according to the time series characteristics of the industrial control scenario, and combines dynamic Bayesian network and Viterbi algorithm to extract resource features, it adjusts the smoothing factor of the transition probability matrix and the parameter weights of the observation probability distribution in the hidden Markov model according to the time series characteristics of the industrial control scenario, so as to optimize the adaptability to non-steady-state time series data. Dynamic Bayesian networks are used to model the state transition process, and the Viterbi algorithm is combined to infer the hidden state of the resource operation sequence, thereby extracting resource features.

10. The industrial control virtual equipment resource pool construction system according to claim 8, characterized in that, The virtual resource model generation module determines the annotation results based on resource characteristics, constructs a dynamic metadata model and ontology knowledge base based on the annotation results, generates a virtual resource model based on the ontology knowledge base, and constructs a bidirectional mapping table between physical and virtual resources. It then constructs a six-dimensional dynamic metadata model based on the annotation results, combined with real-time indicators, resource names, attributes, identifiers, key query fields, and mapping relationships, to achieve structured storage of metadata. Based on the resource description framework, it constructs an ontology knowledge base, defines semantic relationships between resources, and uses query language to infer semantic associations, achieving semantic unification and cross-protocol interoperability for multi-source heterogeneous resources. Based on the semantic associations in the ontology knowledge base, it generates a virtual resource model composed of attribute entities, image entities, virtual machine specification entities, and network entities. It quickly generates virtual resource instances using templated configuration, encapsulates industrial control functions into independent microservices based on real-time optimized container technology, and achieves dynamic resource adjustment and load balancing through orchestration tools. Finally, it constructs a bidirectional mapping table between physical and virtual resources, records the real-time status, configuration parameters, and dependencies of resource instances, and stores the bidirectional mapping table in a distributed key-value database.

11. The industrial control virtual equipment resource pool construction system according to claim 10, characterized in that, Container images have a built-in lightweight runtime environment to reduce resource overhead and support rapid deployment and hot updates across hardware platforms.

12. The industrial control virtual equipment resource pool construction system according to claim 8, characterized in that, The metadata aggregation and optimization module optimizes the clustering algorithm using time weighting factors and dynamically aggregates time-series metadata using a streaming computing framework. Based on anomaly detection technology, it performs anomaly detection on resource status using the aggregated time-series metadata. Specifically, it optimizes the partition-based unsupervised clustering algorithm using time weighting factors and optimizes the distance calculation formula by combining resource update timestamps and time-series sensitivity coefficients. Based on a sliding time window mechanism, it periodically updates cluster centers and dynamically selects initial cluster centers using a density peak detection algorithm, taking into account the periodic characteristics of resource load. It employs streaming metadata aggregation, deploys a streaming computing framework, and utilizes time windows and water level mechanisms to process resource status data streams of device heartbeat signals and network latency indicators in real time. Structured metadata and topological relationships between resources are stored in relational databases and graph databases, respectively. Index optimization and partitioned table design are used for querying, and hot and cold data are managed hierarchically. A long short-term memory neural network is constructed. The resource status time series data is input, and long-term dependencies are captured through hidden layers. The output status prediction value and confidence interval are then output. The system dynamically calibrates the anomaly detection threshold based on the prediction error distribution, identifies abnormal signals by comparing the predicted state value with the calibrated anomaly detection threshold, sends alarm events, enables backup instances or migrates the load to healthy nodes to achieve redundant resource switching, records abnormal events, and combines association rules to mine and locate potential root causes.

13. The industrial control virtual equipment resource pool construction system according to claim 12, characterized in that, The optimized distance calculation formula is as follows: In the formula, Dist(x) i ,c j (x) represents the data point. i With cluster center c j The weighted distance, w(t) i ) represents the time weighting factor, x i For the i-th data point, c j Let x be the j-th cluster center, n be the number of dimensions of the data points, and x be the number of cluster centers. ik For data point x i The k-th eigenvalue, c jk For cluster center c j The k-th eigenvalue, t i Here, α is the resource update timestamp, and α is the time-series sensitivity coefficient.

14. The industrial control virtual equipment resource pool construction system according to claim 8, characterized in that, The resource layer is used to integrate physical hardware resources, software services, and multimodal data streams, and generate standardized resource instances through virtualization technology; it realizes lightweight encapsulation of resources based on containerization technology, supports rapid deployment, dynamic migration, and elastic scaling of resource instances to achieve resource integration and virtualization; it supports full lifecycle management of resources, and realizes dynamic authorization and access control of interface permissions. The data layer defines cross-platform communication protocol compatibility specifications and data format standards to ensure seamless interaction of multi-source heterogeneous data; it builds a distributed search engine to enable full-text retrieval, fuzzy matching, and multi-field joint queries of metadata; and it utilizes a hot and cold data separation strategy to achieve dynamic hierarchical data management. The link layer is used to build transmission channels based on software-defined networks, support cross-platform resource scheduling, and ensure bandwidth priority for critical services through traffic shaping technology. It achieves end-to-end encrypted transmission, combining two-way certificate authentication with a perfect forward confidentiality mechanism to prevent man-in-the-middle attacks and data tampering risks; The technical support layer is used to implement container orchestration, support automatic scaling, service mesh and resource quota management to realize toolkit virtualization; develop custom operators to realize real-time aggregation of multi-source metadata, window calculation and state management to realize metadata aggregation; Real-time collection of resource pool performance metrics, and triggering of anomaly notifications through the alarm rule engine to achieve security monitoring; The interface layer is used to define resource directory interfaces and design permission management interfaces to support resource classification retrieval, metadata subscription and version control, and to achieve fine-grained permission allocation and audit log recording; it also builds a unified query gateway to support clients to obtain data fields on demand and reduce redundant data transmission. To achieve efficient inter-service communication, optimize serialization efficiency, and reduce the coupling between the business layer and the data layer.

15. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images