Method and device for optimizing application of opc ua protocol in industrial automation field

Abstract

The application belongs to the technical field of industrial automation, and discloses an optimization application method and equipment of an OPC UA protocol in the field of industrial automation, which comprises the following steps: obtaining multi-dimensional context information in an industrial field, analyzing future communication requirements and behavior trends of a target OPC UA node by using a preset prediction model, obtaining predictive behavior characteristics, dynamically generating and adjusting an OPC UA communication optimization strategy according to the predictive behavior characteristics and current context information, and applying the optimization strategy to communication sessions between an OPC UA server and a client, wherein the optimization strategy comprises a data sampling period, a publishing interval, a message queue size, a data encoding mode and a security policy level. Through intelligent sensing and prediction analysis, the application dynamically optimizes OPC UA communication parameters, can significantly improve data transmission efficiency, reduce system resource consumption, enhance the real-time performance and reliability of communication, and is particularly suitable for complex and dynamic industrial automation environments.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of industrial automation, and relates to an optimized application method and device of an OPC UA protocol in the field of industrial automation. BACKGROUND

[0002] OPC UA is a secure, reliable, platform-independent information exchange standard based on a service-oriented architecture for the field of industrial automation. It is widely used for data integration between devices, devices to control systems, and enterprise management systems, realizing interoperability between devices and systems of different suppliers. OPC UA defines a rich information model, various communication services, and built-in security mechanisms, such as data access, historical data access, alarm and condition, etc.

[0003] In traditional OPC UA applications, communication parameters such as data sampling interval, publishing interval, message queue size, data encoding method, and security policy of servers and clients, such as encryption algorithm, signature algorithm, and security mode, are usually statically set according to empirical values or expected working conditions during system deployment or configuration. Once set, these parameters are rarely or never actively adjusted during operation.

[0004] However, modern industrial automation systems are increasingly complex and exhibit the following characteristics:

[0005] 1) Large scale and heterogeneity: a large number of connected devices of various types generate a large amount of diverse data.

[0006] 2) Dynamic environment: production tasks are frequently switched, device operating states are dynamically changed, network load fluctuates greatly, especially in wireless communication scenarios.

[0007] 3) High real-time and reliability requirements: many industrial control and monitoring applications have strict requirements for data real-time and communication reliability.

[0008] Under the above background, the statically configured OPC UA communication strategy exposes some problems:

[0009] Waste or lack of resources: when the system load is low or the data changes slowly, too high a sampling and publishing frequency will result in unnecessary network bandwidth occupation and CPU resource consumption; conversely, when critical data changes rapidly or the network is congested, fixed communication parameters may not guarantee the real-time and integrity of data.

[0010] Poor adaptability: unable to dynamically adjust communication priority and resource allocation according to real-time working condition changes, such as device fault warning and urgent production instruction issuance, which may cause delay in transmission of critical information.

[0011] Optimization configuration difficulty: In complex systems, manually optimizing the communication parameters of all OPC UA nodes is a daunting task and it is difficult to achieve global optimization.

[0012] There are currently some studies trying to optimize the performance of OPC UA, such as improving data compression algorithms, optimizing information models, etc., but most of them still focus on static optimization of a certain aspect, lacking a systematic method and device that can comprehensively perceive the dynamic changes of the industrial site and intelligently, predictively, and adaptively adjust the overall communication strategy of OPC UA.

[0013] Therefore, there is an urgent need for a new method and device that can overcome the defects of the prior art and achieve dynamic optimization of OPC UA communication strategy, in order to improve the efficiency, real-time performance and intelligent level of information interaction in industrial automation systems. SUMMARY

[0014] The present application aims to solve the technical problems of static communication strategy configuration in existing OPC UA applications, which is difficult to adapt to dynamic changes in industrial environments, resulting in low resource utilization and difficulty in ensuring real-time performance of critical data. Specifically, the present application aims to provide a method and device that can dynamically optimize OPC UA communication parameters based on real-time multi-dimensional context information and predictive behavior analysis.

[0015] To achieve the above-mentioned purpose, the present application provides an optimization application method of OPC UA protocol in the field of industrial automation, which comprises the following steps:

[0016] Step S1: Real-time acquisition of multi-dimensional context information of the industrial site. These information is the basis for subsequent analysis and decision-making, at least including:

[0017] Device operating state data: For example, CPU load, memory usage, key sensor readings, device fault codes or warning signals, etc.

[0018] Network environment parameters: For example, network delay, available bandwidth, data packet loss rate, signal strength of OPC UA communication link, etc., where signal strength corresponds to wireless conditions.

[0019] Current business priority information: For example, priority levels of different data streams or OPC UA services defined according to current production tasks and operation modes, etc., where operation modes are, for example, normal production, emergency shutdown, and device maintenance.

[0020] OPC UA historical interaction data: For example, data publishing frequency, data size, subscription changes, connection establishment / disconnection events of specific OPC UA nodes in the past period of time, etc.

[0021] Step S2: Based on the acquired multi-dimensional context information, a pre-set prediction model is used to perform predictive analysis on the future communication needs and behavior trends of the target OPC UA node, obtaining predictive behavior characteristics.

[0022] The pre-set prediction model can be obtained through machine learning or deep learning methods, where machine learning can be a time series model ARIMA, LSTM long short-term memory network, trained using historical OPC UA interaction data and corresponding multi-dimensional context information. This model can learn the pattern of data changes, periodicity of network state, and relevance of business needs.

[0023] The predictive behavior characteristics can include: data generation rate prediction value, data subscription request frequency prediction value, network bandwidth demand prediction value, potential connection instability prediction, etc. of the target OPC UA node in one or more future time windows.

[0024] Step S3: According to the predictive behavior characteristics and the current real-time multi-dimensional context information, dynamically generate or adjust the OPC UA communication optimization strategy. This strategy is targeted and aims to balance performance, resources and reliability. The communication optimization strategy at least includes adjustment of one or more of the following:

[0025] Data sampling period: the frequency at which the server collects data.

[0026] Publishing interval: the frequency at which the server sends notifications to the client.

[0027] Message queue size: the queue length used by the server or client to cache notification messages.

[0028] Data encoding method: for example, prefer UA-Binary when bandwidth is limited, or consider XML or JSON when good readability or integration with other systems is needed, where system integration requires bandwidth allowance and XML or JSON needs to be supported.

[0029] Security policy level: for example, on an internal trusted network and resource-limited edge device, the security level can be appropriately reduced for non-sensitive data to reduce computational overhead, where reducing the security level can be only signing without encryption or using a lighter encryption algorithm; for high-sensitive data or in untrusted networks, high security level must be used.

[0030] Step S4: Apply the generated OPC UA communication optimization strategy to the communication session between the target OPC UA server and the client. This may involve updating the parameters of the server or client through OPC UA service calls or specific configuration interfaces.

[0031] Further, in order to form a closed loop optimization, the method can further comprise a step S5 of monitoring the actual communication effect after the application of the OPC UA communication optimization strategy, such as the actual data throughput, delay, resource consumption change, and feeding back the monitoring result to the prediction model of step S2 for online learning or parameter fine-tuning of the model, and / or to the communication optimization strategy generation and adjustment process of step S3 to realize the iterative optimization of the model and the adaptive adjustment of the strategy.

[0032] The application also provides an optimization application device for OPC UA protocol in the field of industrial automation, which is used to execute the above method and comprises:

[0033] a context information acquisition module for executing step S1.

[0034] a predictive analysis module connected with the context information acquisition module, which is internally provided with a preset prediction model and is used to execute step S2.

[0035] a strategy generation and adjustment module connected with the predictive analysis module and the context information acquisition module, which is used to execute step S3.

[0036] a strategy application module connected with the strategy generation and adjustment module, which is used to execute step S4.

[0037] Optionally, the device can further comprise an effect monitoring and feedback module for executing step S5 and connected with the predictive analysis module and / or the strategy generation and adjustment module to form a closed loop feedback.

[0038] Compared with the prior art, the application has the following beneficial effects:

[0039] 1. The application can avoid excessive communication in the slow data change or non-critical period by dynamically adjusting the sampling and publishing frequency, data encoding, etc., reduce unnecessary network bandwidth and computing resource consumption, and guarantee sufficient communication resources in the data demand peak period, effectively improving the communication efficiency and resource utilization.

[0040] 2. The application can prioritize the timely and reliable transmission of critical data according to the business priority and the predicted network condition, for example, by shortening the publishing interval of critical data, or reserving bandwidth or using more robust transmission strategies for critical data when there is a network congestion warning, to enhance its real-time performance and reliability.

[0041] 3. The application makes OPC UA communication no longer static configuration, but can actively perceive environmental changes and predict future trends, intelligently adjust its own behavior, better adapt to complex and variable industrial field environment, and improve the adaptability and intelligent level of the system.

[0042] 4. The present invention automates the optimization process, reduces the need for manual intervention, and lowers the difficulty of system configuration and maintenance, which has significant advantages especially for large-scale OPC UA deployment.

[0043] 5. By predicting communication behavior and network status, the present invention can take preventive measures before potential problems occur, thereby improving the overall availability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 This is a flow chart of an optimized application method of the OPC UA protocol in the field of industrial automation according to the present invention;

[0046] Figure 2 This is a structural block diagram of an optimized application device for the OPC UA protocol in the field of industrial automation according to the present invention;

[0047] Figure 3 Schematic diagram of the training and application of the prediction model in the present invention;

[0048] Figure 4 The figure is a schematic diagram of the decision logic for dynamic adjustment of OPC UA communication parameters in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0050] Example 1: OPC UA Communication Optimization Method and System Based on LSTM Network

[0051] See also Figure 1 and Figure 2 As shown, this embodiment describes an OPC UA communication optimization method and system based on long short-term memory (LSTM) network for predictive analysis. The specific method steps are as follows:

[0052] Step S1: Real-time acquisition of multi-dimensional context information, corresponding to Figure 1 Step S1 inFigure 2 Context information acquisition module 701 in the execution. Context information acquisition module 701 continuously collects data from various sources of industrial sites:

[0053] Device running state data: read from PLC, DCS, SCADA system or directly from intelligent devices through OPC UA or other protocols, such as the joint motor current and temperature of a certain production line robot arm, the temperature is high-frequency sampling requirement, the vibration frequency of a certain large compressor, the vibration frequency is required for high frequency when the device is abnormal, and the CPU / memory occupancy of each device.

[0054] Network environment parameters: monitored by network management tools or built-in probes in OPC UA communication endpoints, which can also be OPC UA servers / clients, to obtain parameters such as OPC UA session round-trip time (RTT), packet loss rate, wireless AP signal strength, and channel occupancy.

[0055] Current business priority information: obtain current work order information, product type, and production rhythm requirements from MES system or production scheduling system, and map them to different OPC UA data source priorities, such as P1-highest, P3-ordinary, and quality traceability-related key parameter points (KPCs) marked as P1 in the production process.

[0056] OPC UA historical interaction data: record the historical sampling interval, publishing interval, subscription quantity, published data volume, message queue overflow events, and corresponding context information at the time of each OPC UA session, and store them in a time series database.

[0057] Step S2: Predictive behavior analysis based on LSTM model, corresponding to step S2 in Figure 1 , performed by predictive analysis module 702 in Figure 2 . Predictive analysis module 702 has a pre-trained LSTM model built-in.

[0058] Please refer to Figure 3 , model training:

[0059] Offline phase, use the accumulated OPC UA historical interaction data in step S1 as the target sequence, here the historical interaction data can be data publishing volume or subscription request number,

[0060] and the contemporaneous device state, network parameter, and business priority as the feature sequence.

[0061] For example, train a model M_{\text{load}}=\text{LSTM}(\text{Historical_Load,Context_Features})to predict the data release volume (denoted as \text{Predicted_Load}) in the next T time steps (for example, the next 5 minutes or 10 minutes), covering multiple predicted values ​​from time step t+1 to t+T (for example, multiple predicted values ​​from t+1 to t+T: \text{Predicted_Load}(t+1), \text{Predicted_Load}(t+2) to \text{Predicted_Load}(t+T)).

[0062] Another model M_{\text{net}}=\text{LSTM}(\text{Historical_RTT,Historical_Loss,Context_Features})predicts future network delay and packet loss rate.

[0063] The training process can use optimizers such as Adam, with the goal of minimizing the mean squared error (MSE) or other loss functions between the predicted and actual values.

[0064] Mathematical formula example, simplified representation of LSTM unit, which is actually more complicated:

[0065] Input Gate:

[0066] Forget Gate:

[0067] Output Gate:

[0068] Candidate cell states:

[0069] Cell status update:

[0070] Hidden state, output:

[0071] In the update calculation formula of the above LSTM unit, the specific meanings of each symbol are as follows:

[0072] : Represents the feature vector input to the LSTM unit at the current time step t, which may contain multi-dimensional context information such as the current device status, network parameters, and service priority.

[0073] : Represents the hidden state output of the LSTM unit at the previous time step t-1.

[0074] : denotes the cell state of the LSTM unit at the previous time step t-1.

[0075] : denote the weight matrices for input gate, forget gate, output gate and cell state candidate, respectively.

[0076] : denote the bias vectors for input gate, forget gate, output gate and cell state candidate, respectively.

[0077] : denotes the Sigmoid activation function, whose output value range is (0, 1), used to control the degree of opening and closing of the threshold.

[0078] : denotes the hyperbolic tangent activation function, whose output value range is (-1, 1), used to generate the candidate cell state and the final hidden state.

[0079] : denotes the output of the input gate at time step t, which determines how much information in the current input can be written into the cell state.

[0080] : denotes the output of the forget gate at time step t, which determines how much information in the previous cell state needs to be forgotten.

[0081] : denotes the output of the output gate at time step t, which determines how much information in the current cell state can be output as the current hidden state .

[0082] : denotes the candidate cell state calculated at time step t.

[0083] : denotes the updated cell state at time step t, which combines the filtering of the old state by the forget gate and the admission of new information by the input gate.

[0084] : denotes the final hidden state output of the LSTM unit at time step t, which is based on the current cell state and obtained through the regulation of the output gate.

[0085] ​​​​​​Through these gating mechanisms, LSTM models can effectively learn and memorize long-term dependencies in time-series data, enabling accurate predictions of future communication demand and behavioral trends. For example, model M_load=LSTM(Historical_Load, Context_Features) is used to predict future data publishing volume based on historical load data (Historical_Load) and context features (Context_Features); model M_net=LSTM(Historical_RTT, Historical_Loss, Context_Features) is used to predict future network latency and packet loss based on historical network round-trip time (Historical_RTT), historical packet loss rate (Historical_Loss), and context features.

[0086] Online prediction:

[0087] At runtime, the latest multi-dimensional context information obtained in step S1 is input into the trained LSTM model to obtain the predicted behavior characteristics of the target OPC UA node, where the target OPC UA node can be a certain data source on a specific OPC UA server or the subscription behavior of a certain client.

[0088] For example, it is predicted that within the next 5 minutes, the data publishing volume of node A will increase by 30%, and the network connection quality of node B may decrease by 10%.

[0089] Step S3: dynamically generating and adjusting the OPC UA communication optimization strategy, corresponding to Figure 1 Step S3 in Figure 2 is performed by the strategy generation and adjustment module 703 in

[0090] The strategy generation and adjustment module 703 generates a communication optimization strategy based on the predictive behavior characteristics obtained in step S2 and the current context information in step S1 through a set of predefined rule engines or a small decision optimization model, where the small decision optimization model can be a reinforcement learning agent or a decision maker based on a utility function.

[0091] Please refer to Figure 4 for an example of decision logic:

[0092] In step S3, the process of dynamically generating and adjusting the OPC UA communication optimization policy is performed, and the decision logic can be based on a series of input parameters. These input parameters include, for example: the predicted data rate of node A (denoted as Predicted_DataRate_NodeA), the current traffic priority of node A (denoted as Current_Priority_NodeA), the predicted network quality (denoted as Predicted_NetworkQuality), and the current device load of server X (denoted as Current_DeviceLoad_ServerX).

[0093] Based on these input parameters, the system can make decisions according to pre-set rules. The following are some examples of decision rules:

[0094] Rule 1: When the system predicts that the data generation rate of node A will exceed a pre-set high threshold (Predicted_DataRate_NodeA>Threshold_High), and the current traffic priority of node A is the highest level (e.g. P1), and the predicted network quality is good (Predicted_NetworkQuality==Good), and the current load of server X is lower than its maximum capacity (Current_DeviceLoad_ServerX<Load_Max), the system will adjust and apply a high-performance communication strategy for node A. This strategy may specifically include: setting the data sampling interval (SamplingInterval) to "fast" mode, setting the data publishing interval (PublishingInterval) to "fast" mode, increasing the message queue size (QueueSize set to Large), using binary data encoding (Encoding=Binary), and enabling high-level security policy (Security=High).

[0095] Rule 2: If the system predicts that the data generation rate of node A is lower than a pre-set low threshold (Predicted_DataRate_NodeA<Threshold_Low), and the current traffic priority of node A is a normal level (e.g. P3), the system will adjust and apply a power-saving or low-overhead communication strategy for node A. This strategy may specifically include: setting the data sampling interval to "slow" mode, setting the data publishing interval to "slow" mode, and reducing the message queue size (QueueSize=Small).

[0096] Rule Three: When the predicted network quality is poor (Predicted_NetworkQuality == Poor), the system will take differentiated measures for different types of nodes. For all critical nodes (Critical_Nodes), the system will increase the redundancy of communication or use more robust data encoding (Increase_Redundancy_Or_RobustEncoding) to ensure the reliability of data transmission. For all non-critical nodes (NonCritical_Nodes), the system can choose to increase their data publishing interval (Increase_PublishingInterval) or, if allowed, reduce their security level (Decrease_SecurityLevel_If_Allowed) to reduce network congestion.

[0097] The parameters mentioned in the above strategies, such as "fast" or "slow", will be quantified into specific values in actual application. For example, the "fast" (Fast) sampling interval can be set to 100 milliseconds (ms), while the "slow" (Slow) sampling interval can be set to 1000 milliseconds (ms). Similarly, the "high level" (High) security policy can correspond to a combination of security protocols such as AES256 encryption, SHA256 signature, and RsaPss key agreement, while the "low level" (Low) security policy may only require data signing (Sign) or use more lightweight encryption measures.

[0098] Regarding the data encoding method, the system usually defaults to the UA-Binary format to improve efficiency. However, in specific situations, such as when the network bandwidth is very abundant and detailed debugging analysis is required, if the OPC UA server supports and the corresponding XML configuration is made, the system can temporarily switch the data encoding method to XML format.

[0099] For more complex decision-making scenarios, the system can introduce an optimization objective function U to guide the selection and adjustment of communication strategies, aiming to maximize the function value. The optimization objective function U can be expressed as:

[0100] In this function:

[0101] represents the real-time performance indicator, which can be quantified by the inverse of the normalized average data transmission delay, for example.

[0102] represents the communication reliability indicator, which can be measured by the success delivery rate of data packets within a statistical period.

[0103] representative of resource consumption metrics, such as the normalized value of CPU occupancy of a key device (e.g. OPC UA server) or network bandwidth usage of a specific communication link.

[0104] representative of security risk assessment value, such as a pre-set risk score corresponding to the selected security policy level, or a comprehensive result of assessing the occurrence probability and impact degree of potential security threats.

[0105] are the weight factors corresponding to each metric, which can be dynamically evaluated and adjusted according to the current overall policy, prediction results, and current context information (such as business priority, device status, etc.). By maximizing the U value, the system aims to seek the best balance between real-time performance, reliability, resource consumption, and security risk.

[0106] Step S4: Apply OPC UA communication optimization policy, corresponding to step S4 in Figure 1 , performed by the policy application module 704 in Figure 2 . The policy application module 704 is responsible for applying the specific parameter settings generated in step S300 to the actual OPC UA communication.

[0107] For OPC UA server-side parameter adjustment, such as sampling interval, it can be done through internal API or management interface of the server.

[0108] For session-related parameters, it can be done through standard OPC UA service calls, such as the client sending a ModifySubscription request to the server, which contains new revisedPublishingInterval, revisedLifetimeCount, etc., where the session-related parameters are specifically the publishing interval and queue size. ModifySubscription is a standard service name in OPC UA, revisedPublishingInterval is a parameter in the ModifySubscription service, and revisedLifetimeCount is also a parameter in the ModifySubscription service.

[0109] Adjustment of security policy may involve renegotiation of security channel.

[0110] Adjustment of data encoding method may require both server and client to support the encoding and negotiation.

[0111] ​​​Step S5: Effect monitoring and feedback, corresponding to the effect monitoring and feedback module 705 in Figure 2

[0112] Monitoring: Continuously collect actual performance indicators after policy application, such as actual data throughput, end-to-end delay, server CPU / memory occupancy changes, message queue overflow rate, etc.

[0113] Feedback:

[0114] Compare actual performance indicators with predicted values, calculate errors, and use them for online adjustment or retraining of the LSTM model in step S2, for example, adjust learning rate or trigger batch retraining after accumulating to a certain amount.

[0115] Feedback the current system state and policy execution effect to the policy generation module in step S3, so that it can make self-correction and optimization of the policy, for example, adjust the threshold of the rule or the reward function of the reinforcement learning agent.

[0116] Please refer to Figure 2 The optimization application device 70 of the present application can be a separate hardware entity, such as an industrial gateway, an edge computing device, or a software module embedded in an existing OPC UA server, client or management platform.

[0117] Context information acquisition module 701: includes various sensor interfaces, network monitoring interfaces, and communication interfaces with other industrial systems (MES, SCADA).

[0118] Predictive analysis module 702: usually contains one or more processors, specifically CPU / GPU / TPU, for running LSTM or other machine learning models, and memory for storing models.

[0119] Policy generation and adjustment module 703: contains the execution unit of the rule engine or decision optimization algorithm.

[0120] Policy application module 704: contains the interface of OPC UA client / server stack, used to execute OPC UA service calls or configure parameters.

[0121] Effect monitoring and feedback module 705: contains performance data collection and analysis units. These modules interact with each other through internal buses or networks. The device also includes necessary memory, power module and communication interface, where the memory is used to store programs, data, models, and the communication interface can be Ethernet, Wi-Fi, 5G, etc.

[0122] Embodiment two: specific scene optimization - OPC UA optimization for large-scale distributed robot collaboration​

[0123] In this scenario, a large number of robots need to synchronize states and coordinate tasks through OPC UA.

[0124] Contextual information: current task of the robot, motion state, battery level, surrounding environment, relative position to other robots, where the current task can be carrying or welding specifically, the motion state can be position or speed specifically, and the surrounding environment can be whether there is an obstacle.

[0125] Prediction model: predict potential collision risk between robots, synchronization demand at task handover points, and data burstiness, such as task completion signals.

[0126] Optimization strategy: when the robot approaches a task handover point or predicts high coordination demand, dynamically shorten its OPC UA publishing interval with the coordination server, which can also be other robots, to increase the refresh rate of position and state data.

[0127] For non-urgent state information, such as slow battery level decline, a longer publishing interval is adopted.

[0128] When the network is congested, prioritize the transmission of collision avoidance and critical task synchronization data, and temporarily reduce the publishing frequency of non-critical robot state data or use more aggressive data compression.

[0129] When the robot is idle or performing independent tasks, reduce its OPC UA communication frequency to save energy and network resources.

[0130] The core of the present application is its "perception-prediction-decision-execution-feedback" closed-loop intelligent optimization mechanism, which enables OPC UA communication to dynamically adapt to complex industrial environments, rather than relying solely on static configuration. In this way, the utilization efficiency of system resources can be maximized while ensuring the performance of key business.

[0131] Although the specific embodiments of the present application are described above, those skilled in the art should understand that these specific embodiments are only illustrative, and those skilled in the art can make various omissions, substitutions and changes to the details of the above method and system without departing from the principles and essence of the present application. For example, combining the above method steps to perform substantially the same function to achieve substantially the same result in a substantially identical manner is within the scope of the present application. Therefore, the scope of the present application is only limited by the appended claims.

Claims

1. A method for optimizing the application of OPC UA protocols in the field of industrial automation, characterized in that, The method comprises the following steps: Step S1: Real-time acquisition of multi-dimensional context information of an industrial site, the multi-dimensional context information comprising at least device running state data, network environment parameters, current service priority information and OPC UA historical interaction data; Step S2: Based on the multi-dimensional context information, prediction and analysis of future communication demand and behavior trend of a target OPC UA node by using a preset prediction model to obtain predictive behavior characteristics of the target OPC UA node; Step S3: Dynamic generation or adjustment of an OPC UA communication optimization strategy according to the predictive behavior characteristics of the target OPC UA node and the current multi-dimensional context information, the communication optimization strategy comprising adjustment of at least one of an OPC UA data sampling period, a publishing interval, a message queue size, a data encoding mode and a security policy level; Wherein, the dynamic generation or adjustment of the OPC UA communication optimization strategy specifically comprises: if it is predicted that the data generation rate of the target OPC UA node will significantly increase and the current service priority is high, the data sampling period and the publishing interval are correspondingly shortened, and a data encoding mode with a higher compression rate can be selectively used; if it is predicted that the network bandwidth will become crowded or the network delay will increase, the publishing interval is appropriately increased, or the security policy level of non-critical data is reduced to reduce the overhead; Step S4: Application of the OPC UA communication optimization strategy to a communication session between the target OPC UA server and the client; Step S5: Monitoring of actual communication effect after application of the OPC UA communication optimization strategy, and feedback of the monitoring result to the prediction model of step S2 and / or the communication optimization strategy generation and adjustment process of step S3 to realize iterative optimization of the model and adaptive adjustment of the strategy.

2. The method of claim 1, wherein the method further comprises: The device running state data comprises device load rate, key parameter value and fault warning information; the network environment parameters comprise network delay, bandwidth utilization rate and data packet loss rate; the service priority information comprises data importance level associated with the current industrial application scenario.

3. The method of claim 1, wherein the method further comprises: The preset prediction model is a machine learning model or a deep learning model based on time series analysis, which is trained by learning the OPC UA historical interaction data and the corresponding multi-dimensional context information.

4. The method of claim 1, wherein the method further comprises: The predictive behavior characteristics comprise data generation rate prediction value, data subscription request frequency prediction value or network bandwidth demand prediction value of the target OPC UA node in one or more future time windows.

5. A device for optimizing the application of OPC UA protocols in the field of industrial automation according to any of claims 1 to 4, characterized in that, The method comprises: A context information acquisition module is configured to acquire multi-dimensional context information of an industrial site in real time, the multi-dimensional context information comprising at least device running state data, network environment parameters, current service priority information and OPC UA historical interaction data; A predictive analysis module is connected with the context information acquisition module and is configured to perform prediction and analysis of future communication demand and behavior trend of a target OPC UA node by using a preset prediction model based on the multi-dimensional context information to obtain predictive behavior characteristics of the target OPC UA node; The strategy generation and adjustment module is connected with the predictive analysis module and the context information acquisition module, and is configured to dynamically generate or adjust an OPC UA communication optimization strategy according to the predictive behavior characteristics of the target OPC UA node and the current multi-dimensional context information, the communication optimization strategy including adjustment of at least one of an OPC UA data sampling period, a publishing interval, a message queue size, a data encoding mode, and a security policy level; and the strategy application module is connected with the strategy generation and adjustment module, and is configured to apply the OPC UA communication optimization strategy to a communication session between a target OPC UA server and a client. The effect monitoring and feedback module is connected with the strategy application module, and is configured to monitor an actual communication effect after the OPC UA communication optimization strategy is applied, and feed back the monitoring result to the predictive analysis module and / or the strategy generation and adjustment module, so as to realize iterative optimization of the preset prediction model and adaptive adjustment of the communication optimization strategy.

6. The device for optimized application of OPC UA protocol in the field of industrial automation according to claim 5, characterized in that, The predictive analysis module is internally provided with a machine learning model or a deep learning model based on time series analysis, which is trained by learning OPC UA historical interaction data and corresponding multi-dimensional context information.

7. The device for optimized application of OPC UA protocol in the field of industrial automation according to claim 5, characterized in that, The strategy generation and adjustment module is further configured to, when receiving predictive behavior characteristics indicating that a data generation rate of a target OPC UA node will be significantly increased and a current service priority is high, generate an instruction to shorten a data sampling period and a publishing interval, and selectively instruct to use a data encoding mode with a higher compression rate; and when receiving predictive behavior characteristics indicating that a network bandwidth will become crowded or a network delay will be increased, generate an instruction to appropriately increase a publishing interval, or instruct to reduce a security policy level of non-critical data.

Images