Remote equipment programming control and state monitoring method

Abstract

The invention relates to the technical field of equipment remote control, and discloses a remote equipment programming control and state monitoring method. Constructing a virtual mapping model, and integrating multi-source sensor data streams to synchronize equipment states in real time; a distributed communication link is established, low-delay two-way transmission is realized through a 5G network and an edge computing node, multi-protocol adaptation is supported, and routing is optimized; implementing multi-modal state monitoring, acquiring data by using a heterogeneous sensor, and generating a comprehensive health index through a spatial-temporal feature fusion algorithm; designing a self-adaptive programming control strategy, dynamically generating a control code according to an equipment state and an external instruction, and verifying the control code; and executing exception collaborative response, and carrying out collaborative positioning and exception repair on the edge node and the cloud based on a federated learning training model. The method solves the problem of remote equipment management, improves equipment operation reliability and control accuracy, realizes efficient fault processing, and has wide application value in the fields of industry, medical treatment and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of remote device control, in particular to a remote device programming control and status monitoring method. Background Art

[0002] In today's era of rapid digitalization and intelligence development, remote devices are increasingly used in many fields, including industry, energy, medical care, home, etc. However, existing remote device control and monitoring technologies face many challenges, which seriously restrict the efficient operation and management of remote devices.

[0003] From the perspective of equipment status monitoring, traditional methods rely on a single type of sensor, making it difficult to fully obtain equipment operating information. Only monitoring the temperature of the equipment, ignoring other key parameters such as vibration and current, makes it impossible to detect potential equipment failures in a timely manner. Moreover, the data collected by different sensors lack effective fusion processing, and it is impossible to form a comprehensive assessment of the health of the equipment. In industrial production, the data formats and frequencies of various sensors are different, making it difficult to integrate and analyze, resulting in an inability to accurately judge whether the equipment is operating normally, increasing the risk of sudden equipment failures and affecting production continuity.

[0004] In terms of data transmission, there are problems of high latency and poor stability in data transmission between remote devices and control terminals. In the industrial Internet of Things scenario, a large number of devices generate a huge amount of data, and the existing network infrastructure is difficult to meet the real-time transmission needs. Insufficient network coverage and weak signals in remote areas lead to frequent packet loss and transmission interruptions. Although 5G networks are gradually becoming popular, in complex industrial environments, multiple devices and multiple protocols coexist, and network adaptation is difficult. It is impossible to give full play to the advantages of 5G, which affects the timeliness and accuracy of remote real-time control and status monitoring of equipment.

[0005] In terms of programming control strategies, traditional control methods lack flexibility and adaptability. Control codes are often pre-set and difficult to dynamically adjust according to the real-time status of the equipment and changes in the external environment. In smart factories, production tasks and equipment conditions are constantly changing. Fixed control codes cannot meet the needs of efficient production, which may lead to inefficient equipment operation, energy waste, and even equipment damage.

[0006] There are also deficiencies in the abnormality handling link. When an abnormality occurs in a device, existing technologies cannot quickly and accurately locate the root cause of the fault, and it is difficult to achieve effective collaborative processing between edge nodes and the cloud. In a large-scale power system, when a node device fails, due to the lack of a global unified abnormality detection model and collaborative mechanism, troubleshooting takes a long time, and repair instructions are not issued in time, causing large-scale power outages and causing huge losses to social production and life. Summary of the invention

[0007] The purpose of the present invention is to provide a remote device programming control and status monitoring method to solve the problems raised in the above background technology.

[0008] To achieve the above object, the present invention provides the following technical solution: a remote device programming control and status monitoring method, the method comprising: Constructing a remote device virtual mapping model, including generating a virtual mapping model based on physical device attributes, control logic and communication protocol; the virtual mapping model integrates multi-source sensor data streams and synchronizes device status in real time; Establishing a distributed communication link, including realizing low-latency bidirectional data transmission between the device and the control terminal through the 5G network and edge computing nodes; the distributed communication link supports multi-protocol adaptation and uses a dynamic routing optimization algorithm to adjust the data transmission path; Implement multimodal condition monitoring, including collecting equipment vibration, temperature, current and position data through heterogeneous sensors, and generating comprehensive equipment health indicators based on spatiotemporal feature fusion algorithms; Design adaptive programming control strategies, including dynamically generating control codes based on the real-time status of the equipment and external instructions, and verifying the feasibility of the control logic through virtual mapping models; Execute coordinated response to anomalies, including training a global anomaly detection model based on a federated learning framework, and locating anomalies and issuing repair instructions through collaboration between edge nodes and the cloud.

[0009] Preferably, the multi-source sensor data stream includes: collecting internal status data of the device through the industrial Internet of Things protocol, acquiring external environment data of the device through a visual sensor, and capturing the equipment operation noise spectrum through an acoustic sensor.

[0010] Preferably, the objective function of the dynamic routing optimization algorithm is: in, For the The transmission delay of the path, is the path energy consumption coefficient, and are the weight factors of delay and energy consumption, is the total number of optional paths.

[0011] Preferably, the spatiotemporal feature fusion algorithm includes: extracting frequency domain features from vibration data using wavelet packet decomposition, applying sliding window mean filtering to temperature data, and associating the spatiotemporal dependencies of multiple sensors through a graph convolutional network.

[0012] Preferably, the dynamic generation of the control code includes: inputting device state data into a pre-trained Transformer model, outputting a control code sequence that complies with the device instruction set, and ensuring the legality of the code through syntax tree parsing.

[0013] Preferably, the model update rule of the federated learning framework is: in, For the Wheel global model parameters, For the The local model parameters of the edge nodes, is the amount of local data, is the total amount of data, is the total number of edge nodes participating in federated learning.

[0014] Preferably, the verification process of the virtual mapping model includes: checking the completeness of the control logic based on a formal verification method, and evaluating the robustness boundary of the control strategy through Monte Carlo simulation.

[0015] Preferably, the calculation formula of the comprehensive health index of the equipment is: in, For time Comprehensive equipment health indicators, For the Real-time data after normalization of class sensors, For the The minimum safety threshold of class sensor data, For the The maximum safety threshold of class sensor data, is the weight coefficient, is the total number of sensor categories.

[0016] Preferably, the abnormal coordinated response further includes: solving the root cause combinatorial optimization problem of equipment failure based on a quantum annealing algorithm, and its objective function is: in, Indicates Whether a potential root cause is activated, Indicates Whether a potential root cause is activated, The cost is incurred for the root cause alone. Root cause and The strength of association, is the weight coefficient of the associated term, and are the total number of root causes and the number of associated logarithms, respectively.

[0017] Compared with the prior art, the present invention has the following beneficial effects: At the equipment status monitoring level, the constructed remote equipment virtual mapping model integrates multi-source sensor data streams and can synchronize equipment status in real time. Through multi-source collection such as industrial Internet of Things protocols, visual sensors, acoustic sensors, etc., comprehensive internal and external information of the equipment can be obtained. In industrial equipment monitoring, not only can the internal temperature, pressure and other parameters of the equipment be grasped in real time, but also the external environment of the equipment can be observed for abnormalities, and the changes in the operating noise spectrum can be captured, providing rich data for equipment failure prediction. The spatiotemporal feature fusion algorithm further improves the accuracy of state assessment. Wavelet packet decomposition, sliding window mean filtering and graph convolutional networks work together to process multi-source data, generate comprehensive equipment health indicators, discover potential failure risks in advance, reduce sudden equipment failures, and improve equipment operation reliability.

[0018] In terms of data transmission, the established distributed communication links use 5G networks and edge computing nodes to achieve low-latency two-way data transmission. The high-speed, low-latency characteristics of the 5G network, combined with the local data processing capabilities of edge computing nodes, greatly improve the efficiency of data transmission. In the telemedicine surgery scenario, the doctor's instructions for operating the control terminal can be quickly and accurately transmitted to the telemedicine equipment, and the real-time status data of the equipment can also be fed back to the doctor in time to ensure the smooth progress of the operation. The dynamic routing optimization algorithm optimizes the path according to transmission delay and energy consumption, and the multi-protocol adaptation function adapts to complex network environments, ensuring the stability of data transmission and reducing data packet loss and transmission interruptions.

[0019] In terms of programming control strategy, the adaptive programming control strategy dynamically generates control code based on the real-time status of the equipment and external instructions. The Transformer model combines the equipment status data to generate code that conforms to the instruction set, and the syntax tree parsing ensures the legality of the code. In a smart factory, when the production task changes, the system can quickly adjust the control code, optimize the equipment operating parameters, improve production efficiency, reduce energy consumption, and achieve precise control and intelligent management of equipment.

[0020] The exception handling capability is significantly enhanced. Based on the federated learning framework, the global anomaly detection model is trained, and the edge node and the cloud collaborate to realize the location of the anomaly and the issuance of repair instructions. In a large energy network, when a device is abnormal, the model trained by the local data of the edge node collaborates with the global model in the cloud to quickly locate the root cause of the fault, solve the combinatorial optimization problem of the root cause of the fault through the quantum annealing algorithm, and accurately determine the cause of the fault. Then, the repair instructions are quickly issued to reduce equipment downtime, reduce economic losses, and ensure the stable operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1This is a working principle diagram of the remote device programming control and status monitoring method of the present invention; Figure 2 Flowchart for multi-source sensor data stream acquisition; Figure 3 Flowchart for dynamic generation and verification of control codes; Figure 4 Flowchart updated for the Federated Learning Framework model. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] See also Figure 1-4 The present invention provides a remote device programming control and status monitoring method to achieve efficient management and precise control of remote devices. The specific steps are as follows: Build a virtual mapping model for remote devices: Generate a virtual mapping model based on the properties of physical devices, such as the device model, specifications, working principle, etc., as well as its control logic, including the device's operating procedures, instruction rules, etc., combined with communication protocols, such as common Modbus, MQTT and other protocols. This virtual mapping model can integrate multi-source sensor data streams, such as device internal status data, external environment data, operating noise spectrum, etc., to synchronize device status in real time and provide an accurate data basis for subsequent monitoring and control.

[0024] Establish a distributed communication link: With the high speed and low latency characteristics of the 5G network and the local processing capabilities of the edge computing nodes, low-latency two-way data transmission between the device and the control terminal is achieved. This distributed communication link supports the adaptation of multiple communication protocols. For example, in industrial scenarios, different devices may use different protocols, and it can all be compatible. At the same time, a dynamic routing optimization algorithm is used to adjust the data transmission path to improve the efficiency and stability of data transmission.

[0025] Implement multimodal status monitoring: Use heterogeneous sensors, such as vibration sensors, temperature sensors, current sensors, position sensors, etc., to collect vibration, temperature, current and position data of the equipment. Then, based on the spatiotemporal feature fusion algorithm, these multi-source data are processed to generate comprehensive health indicators of the equipment, so as to comprehensively and accurately reflect the operating status of the equipment.

[0026] Design adaptive programming control strategy: dynamically generate control code based on the real-time status of the equipment and the external instructions received. The generated control code will be verified through the virtual mapping model to ensure the feasibility of the control logic and avoid equipment failure or abnormal operation due to erroneous instructions.

[0027] Execute abnormal collaborative response: Based on the federated learning framework, the global anomaly detection model is trained, and the edge nodes and the cloud work together to quickly locate the anomaly and issue repair instructions in a timely manner. In this way, the efficiency of equipment fault handling is improved, equipment downtime is reduced, and stable operation of the equipment is guaranteed.

[0028] The implementation of the present invention is further described below in conjunction with Examples 1 to 6.

[0029] Embodiment 1: This embodiment elaborates on the specific collection method of multi-source sensor data streams, and collects data from different dimensions through multiple sensors to enrich the device status information. When building a remote device virtual mapping model, the collection of multi-source sensor data streams is crucial. The internal status data of the device is collected through industrial Internet of Things protocols, such as the OPC UA protocol. The OPC UA protocol has the advantages of platform independence and high security, and can communicate with various types of industrial equipment. Taking the motor equipment on an industrial production line as an example, the internal status data of the motor, such as speed, torque, and power, can be obtained in real time using the OPC UA protocol. These data directly reflect the operating performance of the motor and are an important basis for evaluating the health of the motor.

[0030] Obtain the external environment data of the equipment through visual sensors. Visual sensors can use high-definition cameras and be installed in appropriate locations around the equipment. For example, for robotic arm equipment on an automated production line, the camera can capture the working scene of the robotic arm in real time and obtain information such as whether there are obstacles around the robotic arm and whether the material placement is correct. These external environment data play a key role in determining whether the equipment can work properly. If there are obstacles around the robotic arm, it may cause the robotic arm to collide and be damaged.

[0031] The acoustic sensor can be used to capture the noise spectrum of the equipment. The acoustic sensor can be a high-sensitivity microphone installed close to the equipment. When the equipment is running, the microphone collects the sound signal emitted by the equipment and converts it into an electrical signal. By performing spectrum analysis on these electrical signals, the noise spectrum of the equipment operation can be obtained. For example, for a fan equipment, its noise spectrum has specific characteristics during normal operation. When the fan fails, such as blade wear or bearing damage, the noise spectrum will change significantly. By monitoring the changes in the noise spectrum, potential equipment failure hazards can be discovered in a timely manner.

[0032] Embodiment 2: This embodiment introduces a dynamic routing optimization algorithm in detail. Through this algorithm, the data transmission path can be optimized according to factors such as transmission delay and energy consumption, the data transmission efficiency can be improved, the network resource consumption can be reduced, and the stability of data transmission between the device and the control terminal can be guaranteed.

[0033] In the process of establishing distributed communication links, the dynamic routing optimization algorithm plays a key role. The objective function of the algorithm is: .in, For the The transmission delay of the path reflects the transmission delay of data from the device to the control terminal through the The time required for a path. In actual network environments, transmission delay is affected by many factors such as network congestion and link quality. For example, when the data traffic on a path is too large, network congestion will occur, resulting in increased transmission delay.

[0034] is the path energy consumption coefficient, which measures the The energy consumption during the transmission of each path. In some battery-powered devices, energy consumption is an important consideration. For example, for some wireless sensor devices in remote areas, reducing data transmission energy consumption can extend the service life of the device and reduce maintenance costs.

[0035] and are the weight factors of latency and energy consumption, respectively. Their values ​​are adjusted according to the actual application scenario and needs. If the real-time requirements are extremely high, such as the control of remote medical surgical equipment, Set a larger value to prioritize minimizing transmission delay; in some IoT devices that are sensitive to energy consumption, it may increase The weight of .

[0036] is the total number of optional paths. In an actual network, there may be multiple data transmission paths between the device and the control terminal. The algorithm will evaluate and screen these optional paths based on the above objective function and select the optimal transmission path. In specific implementation, the algorithm will monitor the transmission delay and energy consumption coefficient of each path in real time, and continuously adjust the path selection to adapt to the dynamic changes of the network environment.

[0037] Embodiment 3: This embodiment describes in detail a spatiotemporal feature fusion algorithm, which can more accurately reflect the comprehensive health status of the equipment by extracting and associating features of different types of sensor data, and provide strong support for equipment fault prediction and diagnosis.

[0038] When implementing multimodal condition monitoring, the spatiotemporal feature fusion algorithm is used to process the collected equipment data. For vibration data, wavelet packet decomposition is used to extract frequency domain features. Wavelet packet decomposition is a time-frequency analysis method that can decompose vibration signals into different frequency bands, thereby analyzing the frequency components of the signal in more detail. For example, for the vibration signal of a rotating machine, the distribution of vibration energy in different frequency ranges can be obtained through wavelet packet decomposition. During normal operation, the vibration energy of each frequency band is within a certain range. When the equipment fails, such as unbalanced or misaligned, the vibration energy of certain frequency bands will increase significantly. By monitoring the energy changes in these characteristic frequency bands, it is possible to determine whether the equipment has a fault and the type of fault.

[0039] For temperature data, sliding window mean filtering is used. Sliding window mean filtering is a simple and effective data smoothing method. It calculates the average value of the data in the window instead of the data at the center of the window by sliding a window of fixed length on the time series. For example, for the temperature monitoring data of the device, the temperature sensor may be interfered by noise, resulting in data fluctuations. After using sliding window mean filtering, these noises can be removed to obtain smoother data that better reflects the actual temperature change trend of the device.

[0040] The spatiotemporal dependencies of multiple sensors are associated through graph convolutional networks. Graph convolutional networks are a type of neural network specifically designed to process graph-structured data. In the equipment status monitoring scenario, different sensors are regarded as nodes in the graph, and the associations between sensors are regarded as edges to construct a graph structure. Graph convolutional networks can automatically learn the spatiotemporal dependencies between these nodes. For example, in a complex industrial system, temperature sensors, vibration sensors, pressure sensors and other sensors are distributed in different locations, and there is a certain spatiotemporal correlation between them. Graph convolutional networks can mine the intrinsic connections between these sensor data by learning from them, thereby generating more accurate comprehensive equipment health indicators.

[0041] Embodiment 4: This embodiment introduces in detail the dynamic generation process of the control code to ensure that the generated control code complies with the device instruction set and is legal and valid, thereby achieving precise control of the device according to the real-time status of the device and external instructions.

[0042] When designing adaptive programming control strategies, the dynamic generation of control codes is the core link. Input the equipment status data into the pre-trained Transformer model. The Transformer model has powerful language understanding and generation capabilities. After being trained with a large amount of equipment status data and corresponding control codes, it can learn the mapping relationship between equipment status and control codes. For example, input the temperature, pressure, flow and other status data of a chemical production equipment into the pre-trained Transformer model, and the model can output the corresponding control code based on these data, such as adjusting the valve opening, adjusting the pump speed and other instructions.

[0043] The model outputs a control code sequence that conforms to the device instruction set. Different devices have different instruction sets. During the training process, the Transformer model will learn the instruction specifications of a specific device to ensure that the output control code can be correctly recognized and executed by the device. For example, for a CNC lathe, its instruction set specifies the code format for various machining operations, and the control code output by the Transformer model must conform to these format requirements.

[0044] The code legitimacy is ensured through syntax tree parsing. A syntax tree is a tree data structure that represents the syntax structure of the code. After the generated control code sequence is converted into a syntax tree, the syntax structure of the code can be checked. For example, check whether there are syntax errors in the code, such as missing brackets, misspelling of keywords, etc. If a syntax error is found, the code will be corrected or regenerated to ensure the legitimacy of the control code and ensure that the device can correctly execute the control instructions.

[0045] Embodiment 5: This embodiment elaborates on the model update rules of the federated learning framework and the verification process of the virtual mapping model. The federated learning framework can effectively utilize the data of edge nodes to train the global anomaly detection model. At the same time, the verification of the virtual mapping model ensures the reliability and stability of the control logic.

[0046] When performing anomaly coordinated response, the model update rule of the federated learning framework is: .in, For the The global model parameters are continuously updated during multiple rounds of training. In each round of training, the edge nodes upload the model parameters obtained through local training to the cloud, which aggregates the data volume and model parameters of each edge node to obtain new global model parameters.

[0047] For the The local model parameters of each edge node. The edge node uses the device data collected locally to train the model and obtain the local model parameters. Since the data of different edge nodes have certain differences, the federated learning framework can comprehensively utilize these scattered data to improve the generalization ability of the global model.

[0048] is the amount of local data, which reflects the The size of the data owned by each edge node. During the model update process, edge nodes with larger data volumes have a relatively greater impact on the global model parameters because their data contain more information.

[0049] is the total data volume, which is the sum of the data volumes of all edge nodes. By calculating the proportion of the data volume of each edge node to the total data volume, the weight of the local model parameters of each edge node in the global model parameter update is determined.

[0050] is the total number of edge nodes participating in federated learning. These edge nodes are distributed in different geographical locations, collect data from devices in different regions, and jointly participate in the training of the global model.

[0051] The verification process of the virtual mapping model includes: checking the completeness of the control logic based on the formal verification method. The formal verification method is a verification technology based on mathematical logic. It establishes a mathematical model of the control logic to reason and verify the model. For example, the control logic of the device is modeled using Petri nets, which can clearly describe the state transition and event triggering relationship of the system. By analyzing the Petri net model, it can be verified whether the control logic meets all the expected conditions, such as whether there is a deadlock, whether it can cover all operation scenarios, etc.

[0052] Evaluate the robustness bounds of the control strategy through Monte Carlo simulation. Monte Carlo simulation is a method of estimating results through random sampling. When evaluating the robustness of the control strategy, various device states and external interference factors are randomly generated within a certain range, and then these data are input into the virtual mapping model to simulate the execution process of the control strategy. After repeated simulations, the execution results of the control strategy are statistically analyzed to evaluate the stability and reliability of the control strategy under different circumstances, thereby determining the robustness bounds of the control strategy.

[0053] Embodiment 6: This embodiment describes in detail the calculation method of the comprehensive health index of the equipment and the combination optimization problem of solving the root causes of equipment failures based on the quantum annealing algorithm in the abnormal coordinated response, providing a quantitative basis for equipment status evaluation and fault diagnosis, and improving the accuracy and efficiency of fault handling.

[0054] When implementing multimodal status monitoring, the calculation formula for the comprehensive health index of the equipment is: This formula is used to comprehensively evaluate the equipment’s The health status of the equipment is the key quantitative basis for judging whether the equipment is operating normally.

[0055] Representative time It is a comprehensive health indicator of the equipment, which integrates the data of multiple sensors into a single value to intuitively reflect the overall health of the equipment. For example, when evaluating the operating status of a wind turbine, It can comprehensively reflect the operating status of multiple key components such as blades, gearboxes, generators, etc. The closer it is to 1, the better the equipment is running. If it is close to 0, it means that the device may be at risk of serious failure.

[0056] It is Normalized real-time data from different sensors. Since the data collected by different sensors vary greatly in dimension and value range, normalization is an important step to ensure that the data can be comprehensively calculated. Taking temperature sensors and vibration sensors as examples, the temperature is usually in degrees Celsius, while the vibration data may be acceleration values, which are obviously different. Through the normalization formula , all kinds of sensor data can be uniformly mapped to the [0, 1] interval, making different types of data comparable.

[0057] and They are The minimum and maximum safety thresholds of the sensor data are determined based on the design specifications, historical operating data, and industry standards of the equipment. For example, for a certain type of transformer, the normal operating range of the oil temperature is 30℃-80℃. Set to 30℃, Once the oil temperature exceeds this range, it means that the equipment may be abnormal. The health indicator score of the device will be reduced accordingly.

[0058] is the weight coefficient, which reflects the The degree of influence of sensor data of different types on the overall health status of the equipment. Different types of sensors have different importance in reflecting the health status of the equipment. When evaluating rotating machinery, vibration sensor data is more critical to determine equipment failure, so its weight coefficient is There are many methods to determine the weight coefficient, such as hierarchical analysis method, principal component analysis method, etc. These methods can be used to scientifically and reasonably allocate weights. More accurately reflects the actual health status of the equipment.

[0059] The total number of sensor categories includes various types of sensor data, which can comprehensively evaluate the operating status of equipment from multiple dimensions. For example, when monitoring large industrial boilers, the integrated temperature sensor, pressure sensor, flow sensor, gas composition sensor and other sensor data can more accurately grasp the boiler's combustion efficiency, heat exchange conditions and potential safety hazards, and avoid misjudgment or omission of important information by a single sensor.

[0060] When performing abnormal coordinated response, the root cause combinatorial optimization problem of equipment failure is solved based on the quantum annealing algorithm, and the objective function is: The algorithm and objective function are used to quickly and accurately determine the root cause of equipment failure, providing strong support for the formulation of targeted maintenance strategies.

[0061] and Respectively represent and Whether a potential root cause is activated. When, representing A potential root cause is identified as one of the factors that lead to equipment failure; when , it means that the root cause is not activated. For example, when analyzing a CNC machine tool failure, It may represent different potential root causes such as tool wear, motor failure, control system abnormality, etc., and the algorithm is used to calculate whether these factors actually caused the failure.

[0062] is the cost of the root cause occurring alone, which measures the The magnitude of the loss caused by a potential root cause. Failures caused by different root causes have different impacts on equipment operation, production efficiency, and maintenance costs. For example, a motor failure may cause equipment to shut down for a long time, affecting production progress. The value is higher; while the minor failures of some small parts have little impact on production, The value is relatively low. , when determining the root cause of the fault, priority can be given to factors that may cause greater losses.

[0063] The root cause and The strength of the correlation reflects the degree of mutual influence between two potential root causes. In actual equipment operation, multiple root causes of failure are often interrelated. For example, in a car engine failure, poor fuel supply and abnormal spark plug ignition may affect each other and jointly lead to reduced engine performance. The larger the value of is, the more significant the impact of the two root causes on the fault is when they occur simultaneously.

[0064] is the weight coefficient of the correlation item, which is used to adjust the importance of the root cause correlation item in the objective function. When the correlation between the causes of equipment failure is strong, it can be appropriately increased. The value of makes the algorithm pay more attention to the relationship between the root causes; on the contrary, if the correlation is weak, it can be reduced By setting , can make the algorithm more consistent with the actual situation of equipment failure and improve the accuracy of fault root cause analysis.

[0065] and are the total number of root causes and the number of associated logarithms, respectively. Determines the number of potential root causes of failure that need to be considered, This reflects the complexity of the interrelationships between these root causes. and The value of may be large, which requires the use of the powerful computing power of the quantum annealing algorithm to quickly find the optimal solution among many possible root cause combinations and determine the most likely root cause combination of the fault, thereby improving the efficiency of fault detection and repair, reducing equipment downtime, and reducing production losses.

[0066] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0067] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A remote device programming control and status monitoring method, characterized in that: include: Constructing a remote device virtual mapping model, including generating a virtual mapping model based on physical device attributes, control logic, and communication protocol; The virtual mapping model integrates multi-source sensor data streams and synchronizes device status in real time; Establishing a distributed communication link, including realizing low-latency bidirectional data transmission between the device and the control terminal through the 5G network and edge computing nodes; the distributed communication link supports multi-protocol adaptation and uses a dynamic routing optimization algorithm to adjust the data transmission path; Implement multimodal condition monitoring, including collecting equipment vibration, temperature, current and position data through heterogeneous sensors, and generating comprehensive equipment health indicators based on spatiotemporal feature fusion algorithms; Design adaptive programming control strategies, including dynamically generating control codes based on the real-time status of the equipment and external instructions, and verifying the feasibility of the control logic through virtual mapping models; Execute coordinated response to anomalies, including training a global anomaly detection model based on a federated learning framework, and locating anomalies and issuing repair instructions through collaboration between edge nodes and the cloud.

2. The remote device programming control and status monitoring method according to claim 1, characterized in that: The multi-source sensor data stream includes: collecting internal device status data through the industrial Internet of Things protocol, obtaining external device environment data through visual sensors, and capturing device operation noise spectrum through acoustic sensors.

3. The remote device programming control and status monitoring method according to claim 1, characterized in that: The objective function of the dynamic routing optimization algorithm is: in, For the The transmission delay of the path, is the path energy consumption coefficient, and are the weight factors of delay and energy consumption, is the total number of optional paths.

4. The remote device programming control and status monitoring method according to claim 1, characterized in that: The spatiotemporal feature fusion algorithm includes: extracting frequency domain features from vibration data using wavelet packet decomposition, applying sliding window mean filtering to temperature data, and associating the spatiotemporal dependencies of multiple sensors through a graph convolutional network.

5. The remote device programming control and status monitoring method according to claim 1, characterized in that: The dynamic generation of the control code includes: inputting device status data into a pre-trained Transformer model, outputting a control code sequence that complies with the device instruction set, and ensuring the legality of the code through syntax tree parsing.

6. The remote device programming control and status monitoring method according to claim 1, characterized in that: The model update rule of the federated learning framework is: in, For the Wheel global model parameters, For the The local model parameters of the edge nodes, is the amount of local data, is the total data volume, is the total number of edge nodes participating in federated learning.

7. The remote device programming control and status monitoring method according to claim 1, characterized in that: The verification process of the virtual mapping model includes: checking the completeness of the control logic based on a formal verification method, and evaluating the robustness boundary of the control strategy through Monte Carlo simulation.

8. The remote device programming control and status monitoring method according to claim 1, characterized in that: The calculation formula of the comprehensive health index of the equipment is: in, For time Comprehensive equipment health indicators, For the Real-time data after normalization of class sensors, For the The minimum safety threshold of class sensor data, For the The maximum safety threshold of class sensor data, is the weight coefficient, is the total number of sensor categories.

9. The remote device programming control and status monitoring method according to claim 1, characterized in that: The abnormal coordinated response further includes: solving the root cause combination optimization problem of equipment failure based on the quantum annealing algorithm, and its objective function is: in, Indicates Whether a potential root cause is activated, Indicates Whether a potential root cause is activated, The cost is incurred for the root cause alone. Root Cause and The strength of association, is the weight coefficient of the associated term, and are the total number of root causes and the number of associated logarithms, respectively.

Images