Environment-friendly ring main unit multi-mode operation control system and method

Abstract

The invention relates to the technical field of ring main units, in particular to an environment-friendly ring main unit multi-mode operation control system and method, and the system comprises a multi-mode operation environment perception module, a multi-physical quantity operation state collection module, an operation mode intelligent recognition and classification module, a digital twin operation simulation and prediction module, and a multi-target adaptive operation optimization module. According to the system, the technical problems of incomplete environment perception, inaccurate state evaluation, untimely early warning, asynchronous control and the like in traditional ring main unit monitoring are solved, the operation and maintenance mode change from passive response to active prediction is realized, and the operation and maintenance efficiency and the power supply reliability are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of ring main unit technology, and in particular to an environmentally friendly ring main unit multimodal operation control system and method. Background Technology

[0002] Currently, the monitoring of distribution network ring main units mainly relies on decentralized automation systems and regular manual inspections, which has significant technical shortcomings: various environmental and operational data are scattered across different systems, forming data silos and making unified analysis and in-depth mining difficult; status assessment depends on threshold alarms and human experience, lacking multimodal feature fusion and intelligent diagnosis, thus failing to achieve early fault warning and accurate status identification; maintenance strategies are mostly reactive, lacking predictive maintenance capabilities, resulting in high maintenance costs and difficulty in ensuring power supply reliability. Existing technologies lack a systematic solution integrating multi-source data fusion, intelligent status assessment, digital twin prediction, and adaptive collaborative control, which seriously restricts the transformation and upgrading of distribution network operation and maintenance management towards intelligence and precision.

[0003] Chinese patent CN120566709A discloses an event-driven intelligent ring main unit and its control method. The event-driven intelligent ring main unit includes a cabinet, which houses a circuit breaker, an isolating grounding switch, and a fusion sensor array. The array synchronously acquires mechanical and electrical quantities. The edge decision system aligns events with nanoseconds using the voltage zero-crossing point as the absolute timescale. The closed-loop control core generates protection and topology reconfiguration commands locally based on these events, driving primary equipment in milliseconds. The intelligent communication interface dynamically schedules bandwidth according to data priority to ensure real-time delivery of critical commands. This application solves the problems of fault misjudgment and local control delay caused by the lack of a unified timing reference for mechanical and electrical transient states in traditional ring main units. However, this solution still suffers from low operation and maintenance efficiency, slow fault response, and insufficient control precision due to incomplete environmental perception, inaccurate state assessment, untimely prediction and early warning, and asynchronous control execution. Summary of the Invention

[0004] To address this, the present invention provides a multimodal operation control system and method for environmental protection ring main units, which overcomes the problems of low operation and maintenance efficiency, slow fault response, and insufficient control precision caused by incomplete environmental perception, inaccurate status assessment, untimely prediction and early warning, and asynchronous control execution in the prior art.

[0005] To achieve the above objectives, on the one hand, the present invention provides a multi-modal operation control system for an environmentally friendly ring main unit, comprising: The multimodal operating environment perception module is used to collect environmental parameters and obtain environmental monitoring datasets through various environmental monitoring sensors deployed in the ring network cabinet; The multi-physical quantity operation status acquisition module is used to collect the electrical operation parameters, mechanical operation status, thermal operation status and insulation operation status of the ring main unit to obtain the operation status dataset; The intelligent identification and classification module for operation modes is used to extract features and classify operation modes based on the environmental monitoring dataset and the operation status dataset, so as to obtain the operation mode identification results. The digital twin operation simulation and prediction module is used to construct a digital twin model of the ring network cabinet, and to perform multi-physics field coupling simulation and operation trend prediction on the operation mode identification results through the digital twin model to obtain the operation status prediction results. The multi-objective adaptive operation optimization module is used to dynamically adjust the optimization target weights and generate a multi-objective optimization strategy based on the operation mode identification results and operation state prediction results. A multimodal operation control strategy generation module is used to generate an operation control strategy based on the multi-objective optimization strategy, using an expert knowledge base and a deep reinforcement learning algorithm. The actuator collaborative control module is used to issue control commands to the switch operating mechanism and the temperature and humidity control mechanism through a time-sensitive network according to the operation control strategy, so as to realize the multi-modal collaborative control of the ring main unit.

[0006] Furthermore, the multimodal operating environment sensing module collects environmental parameters through various environmental monitoring sensors deployed within the ring network cabinet, obtaining an environmental monitoring dataset including: Multiple types of environmental monitoring sensors are activated to collect environmental parameter data in parallel, resulting in raw sensor data. The raw sensor data is then verified and encapsulated to obtain a verified data packet. This verified data packet is then integrated with multi-source data to generate an environmental monitoring dataset.

[0007] Furthermore, the multi-physical quantity operation status acquisition module acquires the electrical operation parameters, mechanical operation status, thermal operation status, and insulation operation status of the ring main unit, obtaining an operation status dataset including: Start the electrical operation parameter acquisition device to collect electrical operation data and obtain the electrical operation parameter set. Start the mechanical operation status monitoring device to collect mechanical operation data and obtain the mechanical operation status set. Start the thermal operation status monitoring device to collect thermal operation data and obtain the thermal operation status set. Start the insulation operation status monitoring device to collect insulation operation data and obtain the insulation operation status set. Integrate the various operation status datasets to generate a complete operation status dataset.

[0008] Furthermore, the intelligent identification and classification module for operational modes performs feature extraction and operational mode classification based on the environmental monitoring dataset and the operational status dataset, obtaining operational mode identification results including: Multimodal feature extraction is performed on the environmental monitoring dataset and the operational status dataset to obtain a feature vector set. The feature vector set is then input into a deep neural network classifier to obtain a preliminary operational modality classification result. The confidence level of the preliminary operational modality classification result is evaluated to obtain a validated operational modality classification result. Based on a transfer learning mechanism, rare operating condition samples are adapted to obtain the final operational modality recognition result.

[0009] Furthermore, the digital twin operation simulation and prediction module constructs a digital twin model of the ring main unit, and uses the digital twin model to perform multi-physics coupling simulation and operation trend prediction on the operation mode identification results, obtaining operation status prediction results including: A high-fidelity digital twin model is constructed based on the design parameters and historical operating data of the ring main unit, resulting in a multi-physics coupling model. The operating mode identification results are then injected into the digital twin model for data-driven simulation to obtain real-time simulation results. Based on the real-time simulation results, multi-timescale operating trend prediction is performed to obtain operating status prediction data. Based on the operating status prediction data, an operating status prediction report is generated to obtain the operating status prediction results.

[0010] Furthermore, the multi-objective adaptive operation optimization module dynamically adjusts the optimization objective weights and generates a multi-objective optimization strategy based on the operation mode recognition results and operation state prediction results, including: Based on the operation mode recognition results, the current operation feature parameters are extracted to obtain the operation feature parameter set. Based on the operation feature parameter set and the operation state prediction results, the target weight coefficient is calculated to obtain the dynamic weight allocation scheme. Based on the dynamic weight allocation scheme, a multi-objective optimization function is constructed to obtain the optimization problem model. The optimization problem model is solved to generate a multi-objective optimization strategy and obtain a multi-objective optimization strategy set.

[0011] Furthermore, the multimodal operation control strategy generation module generates an operation control strategy based on the multi-objective optimization strategy, using an expert knowledge base and a deep reinforcement learning algorithm, including: Based on the multi-objective optimization strategy, expert rule matching is performed to obtain a rule-based control strategy set. Based on the deep reinforcement learning model, strategy optimization is performed to obtain a learning-optimized control strategy set. The rule-based control strategy set and the learning-optimized control strategy set are fused to obtain a fused control strategy set. The fused control strategy set is then subjected to security verification and interpretation processing to obtain the final operation control strategy.

[0012] Furthermore, the actuator collaborative control module, based on the operation control strategy, issues control commands to the switch operating mechanism and the temperature and humidity control mechanism via a time-sensitive network to achieve multimodal collaborative control of the ring main unit, including: The operation control strategy is converted into an execution instruction sequence to obtain a control instruction set. The control instruction transmission is scheduled through a time-sensitive network to obtain a timing-optimized instruction stream. Control instructions are issued to the switch operation mechanism and the temperature and humidity control mechanism, and the execution status is monitored to obtain execution status data. The control effect is evaluated and the parameters are optimized based on the execution status data to obtain the collaborative control optimization result.

[0013] Furthermore, the operation control strategy is converted into an execution instruction sequence to obtain a control instruction set. Specifically, this involves parsing the control logic and parameter settings in the operation control strategy, converting the control strategy into specific actuator operation instructions, generating an instruction execution sequence containing timing relationships, and outputting a standardized format control instruction set. By scheduling control command transmission through time-sensitive networking, a timing-optimized command stream is obtained. Specifically, clock synchronization of each node is achieved based on the IEEE 1588 precise time protocol, a time-aware shaper is used to arrange command transmission time slots, high-priority transmission channels are allocated for critical control commands, and a timing-optimized command stream with deterministic delay guarantee is output. The system sends control commands to the switch operating mechanism and the temperature and humidity control mechanism and monitors their execution status to obtain execution status data. Specifically, it sends opening and closing commands to the switch operating mechanism and heater and fan control commands to the temperature and humidity control mechanism through the TSN network. It also collects the action feedback signals of the switch operating mechanism and the temperature and humidity control mechanism in real time and outputs execution status data that includes the command execution status and effect. Based on the execution status data, the control effect is evaluated and the parameters are optimized to obtain the collaborative control optimization results. Specifically, the difference between the actual control effect and the expected target is analyzed, the control parameters are dynamically adjusted according to the execution status, the timing of the coordination between multiple actuators is optimized, and the collaborative control optimization results containing parameter optimization suggestions are output.

[0014] On the other hand, the present invention also provides a method for a multimodal operation control system for an environmental protection ring main unit, comprising: Step S1: Collect environmental parameters using various environmental monitoring sensors deployed in the ring network cabinet to obtain an environmental monitoring dataset; Step S2: Collect the electrical operating parameters, mechanical operating status, thermal operating status and insulation operating status of the ring main unit to obtain the operating status dataset; Step S3: Based on the environmental monitoring dataset and the operation status dataset, perform feature extraction and operation mode classification to obtain the operation mode identification result; Step S4: Construct a digital twin model of the ring main unit, and use the digital twin model to perform multi-physics coupling simulation and operation trend prediction on the operation mode identification results to obtain the operation status prediction results; Step S5: Based on the operation mode recognition results and operation state prediction results, dynamically adjust the optimization target weights and generate a multi-objective optimization strategy; Step S6: Based on the multi-objective optimization strategy, generate an operation control strategy through an expert knowledge base and a deep reinforcement learning algorithm; Step S7 is used to send control commands to the switch operating mechanism and the temperature and humidity control mechanism through the time-sensitive network according to the operation control strategy, so as to realize the multi-modal collaborative control of the ring main unit.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: the system achieves comprehensive environmental parameter acquisition through a multi-modal operating environment perception module, solving the problems of single environmental factor consideration and incomplete data acquisition in traditional monitoring, and providing a comprehensive environmental data foundation for intelligent control. Furthermore, the system achieves multi-dimensional status monitoring of electrical, mechanical, thermal, and insulation dimensions through a multi-physical quantity operating status acquisition module, overcoming the technical bottleneck of traditional monitoring parameters being singular and unable to fully reflect the true operating status of equipment. Additionally, the system achieves accurate identification of operating conditions through a modal intelligent identification and classification module that performs deep feature extraction and intelligent classification based on multi-source data. Finally, the system utilizes digital twin operation... The simulation and prediction module constructs a high-fidelity virtual model and performs multiphysics simulation, enabling accurate prediction of equipment status and early warning of faults. This transforms the maintenance strategy from "periodic inspection" to "predictive maintenance." The system also dynamically adjusts the weights of optimization objectives through a multi-objective adaptive operation optimization module, achieving balanced optimization of multiple objectives such as safety, economy, and environmental protection. Furthermore, the system generates control strategies that are both reliable and adaptable through a multi-modal operation control strategy generation module that combines expert knowledge and deep learning, significantly improving the intelligence level of decision-making. Finally, the system ensures the coordinated operation of multiple actuators by issuing precise commands based on time-sensitive networks through an actuator collaborative control module. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the multi-modal operation control system for the environmental protection ring main unit in this embodiment; Figure 2 This is a flowchart illustrating the method of the multi-modal operation control system for the environmental protection ring main unit in this embodiment. Detailed Implementation

[0017] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0018] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0019] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0020] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0021] Please see Figure 1 The diagram shown is a structural schematic of the multi-modal operation control system for the environmental protection ring main unit in this embodiment. The system includes: The multimodal operating environment perception module is used to collect environmental parameters and obtain environmental monitoring datasets through various environmental monitoring sensors deployed in the ring network cabinet; The multi-physical quantity operation status acquisition module is used to collect the electrical operation parameters, mechanical operation status, thermal operation status and insulation operation status of the ring main unit to obtain the operation status dataset. The multi-physical quantity operation status acquisition module is connected to the multi-modal operation environment sensing module. The intelligent identification and classification module for operation modes is used to extract features and classify operation modes based on the environmental monitoring dataset and the operation status dataset to obtain operation mode identification results. The intelligent identification and classification module for operation modes is connected to the multi-physical quantity operation status acquisition module. The digital twin operation simulation and prediction module is used to construct a digital twin model of the ring network cabinet, and to perform multi-physics field coupling simulation and operation trend prediction on the operation mode identification results through the digital twin model to obtain the operation status prediction results. The digital twin operation simulation and prediction module is connected to the operation mode intelligent identification and classification module. A multi-objective adaptive operation optimization module is used to dynamically adjust the optimization target weights and generate a multi-objective optimization strategy based on the operation mode identification results and operation state prediction results. The multi-objective adaptive operation optimization module is connected to the digital twin operation simulation and prediction module. A multimodal operation control strategy generation module is used to generate an operation control strategy based on the multi-objective optimization strategy, through an expert knowledge base and a deep reinforcement learning algorithm. The multimodal operation control strategy generation module is connected to the multi-objective adaptive operation optimization module. The actuator collaborative control module is used to issue control commands to the switch operating mechanism and the temperature and humidity control mechanism through a time-sensitive network according to the operation control strategy, so as to realize the multi-modal collaborative control of the ring main unit. The actuator collaborative control module is connected to the multi-modal operation control strategy generation module.

[0022] Specifically, this system constructs a closed-loop intelligent monitoring system for ring main units, achieving intelligent management from environmental perception to collaborative control. The system utilizes a multi-modal operating environment perception module to collect comprehensive environmental parameters, solving the problems of single-factor consideration and incomplete data collection in traditional monitoring, thus providing a comprehensive environmental data foundation for intelligent control. Furthermore, the system employs a multi-physical quantity operating status acquisition module to monitor electrical, mechanical, thermal, and insulation conditions, overcoming the technical bottleneck of traditional monitoring parameters being singular and unable to fully reflect the true operating status of equipment. Finally, the system utilizes an intelligent operating mode recognition and classification module to perform deep feature extraction and intelligent classification based on multi-source data, achieving accurate identification of operating conditions. The system also constructs a high-fidelity virtual model and performs multiphysics simulation through a digital twin operation simulation and prediction module, enabling accurate prediction of equipment status and early warning of faults, thus transforming the maintenance strategy from "periodic maintenance" to "predictive maintenance." Furthermore, the system dynamically adjusts the weights of optimization objectives through a multi-objective adaptive operation optimization module, achieving balanced optimization of multiple objectives such as safety, economy, and environmental protection. Additionally, the system generates a control strategy that combines reliability and adaptability through a multi-modal operation control strategy generation module that integrates expert knowledge and deep learning, significantly improving the intelligence level of decision-making. Finally, the system ensures the coordinated operation of multiple actuators by issuing precise commands based on a time-sensitive network through an actuator collaborative control module.

[0023] Specifically, the multimodal operating environment sensing module collects environmental parameters through various environmental monitoring sensors deployed in the ring network cabinet, and obtains an environmental monitoring dataset including: Multiple types of environmental monitoring sensors are activated to collect environmental parameter data in parallel, resulting in raw sensor data. The raw sensor data is then verified and encapsulated to obtain a verified data packet. This verified data packet is then integrated with multi-source data to generate an environmental monitoring dataset.

[0024] Specifically, the various environmental monitoring sensors include microclimate environmental monitoring sensors, electrical environmental monitoring sensors, mechanical environmental monitoring sensors, and chemical environmental monitoring sensors. The system activates these sensors to collect environmental parameter data in parallel, resulting in the initial sensor count. Specifically: the microclimate environmental monitoring sensor is activated to collect data on temperature gradient, humidity distribution, condensation risk, and air pressure changes; the electrical environmental monitoring sensor is activated to collect data on electromagnetic interference intensity, partial discharge signals, and ozone concentration; the mechanical environmental monitoring sensor is activated to collect data on mechanical vibration spectrum, structural displacement, and structural stress; and the chemical environmental monitoring sensor is activated to collect data on characteristic gas concentrations, SF6 decomposition products, and dust concentrations. The raw sensor data is verified and encapsulated to obtain a verified data packet. Specifically, the raw data collected by various sensors is subjected to CRC verification to remove abnormal data. A unified timestamp is added to all sensor data using the NTP protocol. The verified data is encapsulated in a standardized manner according to the sensor type. A sensor ID and device location identifier are added to each data packet. The verified data packets are integrated from multiple sources to generate an environmental monitoring dataset. Specifically, all verified data packets are aggregated through a multi-protocol interface, and the multi-source data are spatiotemporally aligned according to a preset time window. The four types of environmental data—microclimate, electrical, mechanical, and chemical—are integrated into a structured dataset, and an environmental monitoring dataset containing complete environmental parameters is output.

[0025] Specifically, the multi-physical quantity operation status acquisition module collects the electrical operation parameters, mechanical operation status, thermal operation status, and insulation operation status of the ring main unit, and obtains an operation status dataset including: Start the electrical operation parameter acquisition device to collect electrical operation data and obtain the electrical operation parameter set. Start the mechanical operation status monitoring device to collect mechanical operation data and obtain the mechanical operation status set. Start the thermal operation status monitoring device to collect thermal operation data and obtain the thermal operation status set. Start the insulation operation status monitoring device to collect insulation operation data and obtain the insulation operation status set. Integrate the various operation status datasets to generate a complete operation status dataset.

[0026] Specifically, the electrical operating parameter acquisition device is activated to collect electrical operating data and obtain an electrical operating parameter set. Specifically, the electronic current transformer is activated to collect current waveform data of each circuit, the electronic voltage transformer is activated to collect system voltage waveform data, the power quality analysis device is activated to calculate harmonic distortion rate, power factor and voltage sag parameters, and output an electrical operating parameter set containing current, voltage and power quality parameters. Start the mechanical operation status monitoring device to collect mechanical operation data and obtain a mechanical operation status set. Specifically, start the high-resolution encoder to collect the opening and closing stroke-time characteristic data, start the strain gauge torque sensor to collect the torque curve data of the operating mechanism, start the contact resistance measuring device to collect the contact resistance data of the contact, and output a mechanical operation status set containing stroke characteristics, torque curve and contact resistance. Start the thermal operation status monitoring device, collect thermal operation data, and obtain a thermal operation status set. Specifically, start the distributed fiber optic temperature measurement system to collect conductor temperature distribution data, start the infrared thermal imager to collect infrared image data of key connection points, start the thermal circuit model calculation engine to calculate conductor temperature rise distribution data, and output a thermal operation status set that includes temperature distribution, thermal image, and temperature rise calculation. The insulation operation status monitoring device is activated to collect insulation operation data and obtain an insulation operation status set. Specifically, the UHF partial discharge monitoring system is activated to collect partial discharge signal data, the digital lock-in amplifier is activated to collect dielectric loss factor measurement data, the DC superposition monitoring device is activated to collect insulation resistance change data, and the insulation operation status set containing partial discharge, dielectric loss factor, and insulation resistance is output. Integrate the various operating status datasets to generate a complete operating status dataset. Specifically, align the electrical operating parameter set, mechanical operating status set, thermal operating status set, and insulation operating status set in time, fuse the data according to a unified data structure, and output a complete operating status dataset containing four categories of operating status: electrical, mechanical, thermal, and insulation.

[0027] Specifically, the intelligent identification and classification module for operational modes performs feature extraction and operational mode classification based on the environmental monitoring dataset and the operational status dataset, and obtains operational mode identification results including: Multimodal feature extraction is performed on the environmental monitoring dataset and the operational status dataset to obtain a feature vector set. The feature vector set is then input into a deep neural network classifier to obtain a preliminary operational modality classification result. The confidence level of the preliminary operational modality classification result is evaluated to obtain a validated operational modality classification result. Based on a transfer learning mechanism, rare operating condition samples are adapted to obtain the final operational modality recognition result.

[0028] Specifically, multimodal feature extraction is performed on the environmental monitoring dataset and the operational status dataset to obtain a feature vector set. Specifically, time-domain feature extraction is performed on electrical operating parameters to obtain statistical features such as mean, variance, peak value, and kurtosis; frequency-domain feature extraction is performed on mechanical vibration data to obtain spectral features based on FFT transformation; nonlinear feature extraction is performed on temperature time-series data to calculate Lyapunov exponent, correlation dimension, and sample entropy, and output a feature vector set containing 128-dimensional feature vectors. The feature vector set is input into a deep neural network classifier to obtain preliminary classification results of the running modalities. Specifically, the 128-dimensional feature vector is input into an improved ResNet-50 architecture classifier, and feature learning is performed through a 50-layer residual network and attention mechanism module. Preliminary classification probabilities of 8 running modalities are obtained in the output layer, and preliminary classification results of the running modalities containing the probability distribution of each modality are output. The confidence level of the preliminary operational modality classification results is evaluated to obtain the verified operational modality classification results. Specifically, the classification confidence level is calculated based on the probability distribution output by Softmax. The classification results are directly output for high-confidence samples, and a manual review mechanism is triggered for low-confidence samples to output the verified operational modality classification results. The rare operating condition samples are adapted based on the transfer learning mechanism to obtain the final operating mode recognition result. Specifically, the prototype network is used to learn small samples for uncertain samples, and the domain adaptation and parameter fine-tuning are performed based on the pre-trained model to output the final operating mode recognition result containing 8 operating mode classifications.

[0029] Specifically, the digital twin operation simulation and prediction module constructs a digital twin model of the ring main unit, and uses the digital twin model to perform multi-physics coupling simulation and operation trend prediction on the operation mode identification results, obtaining operation status prediction results including: A high-fidelity digital twin model is constructed based on the design parameters and historical operating data of the ring main unit, resulting in a multi-physics coupling model. The operating mode identification results are then injected into the digital twin model for data-driven simulation to obtain real-time simulation results. Based on the real-time simulation results, multi-timescale operating trend prediction is performed to obtain operating status prediction data. Based on the operating status prediction data, an operating status prediction report is generated to obtain the operating status prediction results.

[0030] Specifically, a high-fidelity digital twin model is constructed based on the design parameters and historical operating data of the ring main unit, resulting in a multi-physics coupling model. Specifically, a geometric model is constructed based on a 3D CAD model with an accuracy of 0.1mm; a physical model is established using the Modelica multi-domain unified modeling language; an electromagnetic field model is constructed based on the finite element method; a thermal field model is constructed based on computational fluid dynamics; a mechanical model is constructed based on multibody dynamics; and a thermal aging model of the insulation material is established using the Arrhenius equation. The output is a multi-physics coupling model that includes the geometric model, physical model, electromagnetic field model, thermal field model, mechanical model, and material thermal aging model. The operational mode recognition results are injected into the digital twin model for data-driven simulation to obtain real-time simulation results. Specifically, the data synchronization between the physical entity and the digital twin is achieved through the OPC UA protocol, the model parameters are updated in real time using the recursive least squares method, the state variables that cannot be directly measured are estimated using the extended Kalman filter, and the real-time simulation results synchronized with the physical entity are output. Based on the real-time simulation results, multi-timescale operation trend prediction is performed to obtain operation status prediction data. Specifically, LSTM network is used to predict the state evolution in the next 30 minutes, combined with the degradation model to predict the remaining lifespan of the equipment, Monte Carlo simulation is used to evaluate the prediction confidence interval, and operation status prediction data including short-term state prediction and long-term lifespan prediction is output. Based on the operational status prediction data, an operational status prediction report is generated to obtain the operational status prediction results. Specifically, when the prediction results of multiple time scales are integrated, the prediction uncertainty is quantified and the confidence level is marked, a structured operational status prediction report is generated, and the operational status prediction results containing the prediction results and confidence intervals are output.

[0031] Specifically, the multi-objective adaptive operation optimization module dynamically adjusts the optimization objective weights and generates a multi-objective optimization strategy based on the operation mode recognition results and operation state prediction results, including: Based on the operation mode recognition results, the current operation feature parameters are extracted to obtain the operation feature parameter set. Based on the operation feature parameter set and the operation state prediction results, the target weight coefficient is calculated to obtain the dynamic weight allocation scheme. Based on the dynamic weight allocation scheme, a multi-objective optimization function is constructed to obtain the optimization problem model. The optimization problem model is solved to generate a multi-objective optimization strategy and obtain a multi-objective optimization strategy set.

[0032] Specifically, based on the operational modality recognition results, current operational feature parameters are extracted to obtain an operational feature parameter set. Specifically, this involves: parsing the modality type and confidence parameters in the operational modality recognition results, extracting key performance indicator threshold parameters related to the current modality, analyzing modality transition trend characteristics and duration parameters, and outputting an operational feature parameter set containing modality type, confidence level, performance indicators, and trend characteristics. Based on the set of operational characteristic parameters and the operational status prediction results, the target weight coefficients are calculated to obtain a dynamic weight allocation scheme. Specifically, the safety target weight coefficients are adjusted according to the risk level parameters in the operational status prediction results, the economic target weight coefficients are adjusted according to the life prediction result parameters, and the environmental protection target weight coefficients are adjusted in combination with the environmental characteristic parameters. The relative weight coefficients of each target are calculated using the analytic hierarchy process, and a dynamic weight allocation scheme containing the weights of multiple targets such as safety, economy, and environmental protection is output. Based on the dynamic weight allocation scheme, a multi-objective optimization function is constructed to obtain the optimization problem model. Specifically, the objective function is constructed according to the weight coefficients in the dynamic weight allocation scheme, a set of constraint functions is established based on the running constraints, the decision variables and the range of the feasible solution space are determined, and a complete mathematical model of the optimization problem is output. Solving the optimization problem model generates a multi-objective optimization strategy set. Specifically, a multi-objective evolutionary algorithm is used to solve the optimization problem model, the optimal compromise solution is selected from the Pareto solution set, the feasibility and effectiveness of the optimization strategy are verified, and a multi-objective optimization strategy set containing control parameter settings is output.

[0033] Specifically, the multimodal operation control strategy generation module generates an operation control strategy based on the multi-objective optimization strategy, using an expert knowledge base and a deep reinforcement learning algorithm, including: Based on the multi-objective optimization strategy, expert rule matching is performed to obtain a rule-based control strategy set. Based on the deep reinforcement learning model, strategy optimization is performed to obtain a learning-optimized control strategy set. The rule-based control strategy set and the learning-optimized control strategy set are fused to obtain a fused control strategy set. The fused control strategy set is then subjected to security verification and interpretation processing to obtain the final operation control strategy.

[0034] Specifically, based on the multi-objective optimization strategy, expert rule matching is performed to obtain a set of basic rule control strategies. Specifically, the optimization objectives in the multi-objective optimization strategy are converted into expert knowledge base query conditions, rule matching is performed in an expert knowledge base containing 200+ control rules, a preliminary control strategy is generated based on a generative rule reasoning engine, and a set of basic rule control strategies that conforms to the current optimization objective is output. Policy optimization is performed based on a deep reinforcement learning model to obtain a set of optimized control policies. Specifically, a state space of 50+ dimensions including operating modes, environmental parameters, and device states is constructed. Policy search is performed in the action space of 16 combinations of control actions. Policy optimization is performed using a deep deterministic policy gradient algorithm, and the optimized control policy set is output. The rule-based control strategy set and the learning-optimized control strategy set are fused to obtain the fused control strategy set: the confidence of the two strategies is calculated based on the DS evidence theory, the high confidence strategy is given a higher weight, the conservative and safe strategy is adopted when the strategies conflict, and the optimized control strategy set after decision fusion is output. The fusion control strategy set is subjected to security verification and interpretation to obtain the final operation control strategy. Specifically, formal security verification is performed using temporal logic, the effectiveness of the strategy is tested in a digital twin environment, the strategy decision traceability and visualization explanation are generated, and the final operation control strategy that has been verified by security is output.

[0035] Specifically, the actuator collaborative control module, based on the operation control strategy, issues control commands to the switch operating mechanism and the temperature and humidity control mechanism through a time-sensitive network to achieve multimodal collaborative control of the ring main unit, including: The operation control strategy is converted into an execution instruction sequence to obtain a control instruction set. The control instruction transmission is scheduled through a time-sensitive network to obtain a timing-optimized instruction stream. Control instructions are issued to the switch operation mechanism and the temperature and humidity control mechanism, and the execution status is monitored to obtain execution status data. The control effect is evaluated and the parameters are optimized based on the execution status data to obtain the collaborative control optimization result.

[0036] Specifically, the operation control strategy is converted into an execution instruction sequence to obtain a control instruction set. Specifically, the control logic and parameter settings in the operation control strategy are parsed, the control strategy is converted into specific actuator operation instructions, an instruction execution sequence containing timing relationships is generated, and a standardized format control instruction set is output. By scheduling control command transmission through time-sensitive networking, a timing-optimized command stream is obtained. Specifically, clock synchronization of each node is achieved based on the IEEE 1588 precise time protocol, a time-aware shaper is used to arrange command transmission time slots, high-priority transmission channels are allocated for critical control commands, and a timing-optimized command stream with deterministic delay guarantee is output. The system sends control commands to the switch operating mechanism and the temperature and humidity control mechanism and monitors their execution status to obtain execution status data. Specifically, it sends opening and closing commands to the switch operating mechanism and heater and fan control commands to the temperature and humidity control mechanism through the TSN network. It also collects the action feedback signals of the switch operating mechanism and the temperature and humidity control mechanism in real time and outputs execution status data that includes the command execution status and effect. Based on the execution status data, the control effect is evaluated and the parameters are optimized to obtain the collaborative control optimization results. Specifically, the difference between the actual control effect and the expected target is analyzed, the control parameters are dynamically adjusted according to the execution status, the timing of the coordination between multiple actuators is optimized, and the collaborative control optimization results containing parameter optimization suggestions are output.

[0037] Please see Figure 2 As shown, it is a flowchart illustrating the method of the multi-modal operation control system for the environmental protection ring main unit in this embodiment. The method includes: Step S1: Collect environmental parameters using various environmental monitoring sensors deployed in the ring network cabinet to obtain an environmental monitoring dataset; Step S2: Collect the electrical operating parameters, mechanical operating status, thermal operating status and insulation operating status of the ring main unit to obtain the operating status dataset; Step S3: Based on the environmental monitoring dataset and the operation status dataset, perform feature extraction and operation mode classification to obtain the operation mode identification result; Step S4: Construct a digital twin model of the ring main unit, and use the digital twin model to perform multi-physics coupling simulation and operation trend prediction on the operation mode identification results to obtain the operation status prediction results; Step S5: Based on the operation mode recognition results and operation state prediction results, dynamically adjust the optimization target weights and generate a multi-objective optimization strategy; Step S6: Based on the multi-objective optimization strategy, generate an operation control strategy through an expert knowledge base and a deep reinforcement learning algorithm; Step S7 is used to send control commands to the switch operating mechanism and the temperature and humidity control mechanism through the time-sensitive network according to the operation control strategy, so as to realize the multi-modal collaborative control of the ring main unit.

[0038] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A multi-modal operation control system for an environmentally friendly ring main unit, characterized in that, include: The multimodal operating environment perception module is used to collect environmental parameters and obtain environmental monitoring datasets through various environmental monitoring sensors deployed in the ring network cabinet; The multi-physical quantity operation status acquisition module is used to collect the electrical operation parameters, mechanical operation status, thermal operation status and insulation operation status of the ring main unit to obtain the operation status dataset; The intelligent identification and classification module for operation modes is used to extract features and classify operation modes based on the environmental monitoring dataset and the operation status dataset, so as to obtain the operation mode identification results. The digital twin operation simulation and prediction module is used to construct a digital twin model of the ring network cabinet, and to perform multi-physics field coupling simulation and operation trend prediction on the operation mode identification results through the digital twin model to obtain the operation status prediction results. The multi-objective adaptive operation optimization module is used to dynamically adjust the optimization target weights and generate a multi-objective optimization strategy based on the operation mode identification results and operation state prediction results. A multimodal operation control strategy generation module is used to generate an operation control strategy based on the multi-objective optimization strategy, using an expert knowledge base and a deep reinforcement learning algorithm. The actuator collaborative control module is used to issue control commands to the switch operating mechanism and the temperature and humidity control mechanism through a time-sensitive network according to the operation control strategy, so as to realize the multi-modal collaborative control of the ring main unit.

2. The multi-modal operation control system for environmental protection ring main units according to claim 1, characterized in that, The multimodal operating environment sensing module collects environmental parameters through various environmental monitoring sensors deployed within the ring network cabinet, resulting in an environmental monitoring dataset including: Multiple types of environmental monitoring sensors are activated to collect environmental parameter data in parallel, resulting in raw sensor data. The raw sensor data is then verified and encapsulated to obtain a verified data packet. This verified data packet is then integrated with multi-source data to generate an environmental monitoring dataset.

3. The multi-modal operation control system for environmental protection ring main units according to claim 1, characterized in that, The multi-physical quantity operation status acquisition module collects the electrical operation parameters, mechanical operation status, thermal operation status, and insulation operation status of the ring main unit, resulting in an operation status dataset including: Start the electrical operation parameter acquisition device to collect electrical operation data and obtain the electrical operation parameter set. Start the mechanical operation status monitoring device to collect mechanical operation data and obtain the mechanical operation status set. Start the thermal operation status monitoring device to collect thermal operation data and obtain the thermal operation status set. Start the insulation operation status monitoring device to collect insulation operation data and obtain the insulation operation status set. Integrate the various operation status datasets to generate a complete operation status dataset.

4. The multi-modal operation control system for environmental protection ring main units according to claim 1, characterized in that, The intelligent identification and classification module for operational modes performs feature extraction and operational mode classification based on the environmental monitoring dataset and the operational status dataset, and obtains operational mode identification results including: Multimodal feature extraction is performed on the environmental monitoring dataset and the operational status dataset to obtain a feature vector set. The feature vector set is then input into a deep neural network classifier to obtain a preliminary operational modality classification result. The confidence level of the preliminary operational modality classification result is evaluated to obtain a validated operational modality classification result. Based on a transfer learning mechanism, rare operating condition samples are adapted to obtain the final operational modality recognition result.

5. The multi-modal operation control system for environmental protection ring main units according to claim 1, characterized in that, The digital twin operation simulation and prediction module constructs a digital twin model of the ring main unit, and uses the digital twin model to perform multi-physics coupling simulation and operation trend prediction on the operation mode identification results, obtaining operation status prediction results including: A high-fidelity digital twin model is constructed based on the design parameters and historical operating data of the ring main unit, resulting in a multi-physics coupling model. The operating mode identification results are then injected into the digital twin model for data-driven simulation to obtain real-time simulation results. Based on the real-time simulation results, multi-timescale operating trend prediction is performed to obtain operating status prediction data. Based on the operating status prediction data, an operating status prediction report is generated to obtain the operating status prediction results.

6. The multi-modal operation control system for environmental protection ring main units according to claim 1, characterized in that, The multi-objective adaptive operation optimization module dynamically adjusts the optimization objective weights and generates a multi-objective optimization strategy based on the operation mode recognition results and operation state prediction results, including: Based on the operation mode recognition results, the current operation feature parameters are extracted to obtain the operation feature parameter set. Based on the operation feature parameter set and the operation state prediction results, the target weight coefficient is calculated to obtain the dynamic weight allocation scheme. Based on the dynamic weight allocation scheme, a multi-objective optimization function is constructed to obtain the optimization problem model. The optimization problem model is solved to generate a multi-objective optimization strategy and obtain a multi-objective optimization strategy set.

7. The multi-modal operation control system for environmental protection ring main units according to claim 1, characterized in that, The multimodal operation control strategy generation module generates operation control strategies based on the multi-objective optimization strategy, using an expert knowledge base and deep reinforcement learning algorithms, including: Based on the multi-objective optimization strategy, expert rule matching is performed to obtain a rule-based control strategy set. Based on the deep reinforcement learning model, strategy optimization is performed to obtain a learning-optimized control strategy set. The rule-based control strategy set and the learning-optimized control strategy set are fused to obtain a fused control strategy set. The fused control strategy set is then subjected to security verification and interpretation processing to obtain the final operation control strategy.

8. The multi-modal operation control system for environmental protection ring main units according to claim 1, characterized in that, The actuator collaborative control module, based on the operation control strategy, issues control commands to the switch operating mechanism and the temperature and humidity control mechanism via a time-sensitive network to achieve multimodal collaborative control of the ring main unit, including: The operation control strategy is converted into an execution instruction sequence to obtain a control instruction set. The control instruction transmission is scheduled through a time-sensitive network to obtain a timing-optimized instruction stream. Control instructions are issued to the switch operation mechanism and the temperature and humidity control mechanism, and the execution status is monitored to obtain execution status data. The control effect is evaluated and the parameters are optimized based on the execution status data to obtain the collaborative control optimization result.

9. The multi-modal operation control system for environmental protection ring main units according to claim 8, characterized in that, The operation control strategy is converted into an execution instruction sequence to obtain a control instruction set. Specifically, the control logic and parameter settings in the operation control strategy are parsed, the control strategy is converted into specific actuator operation instructions, an instruction execution sequence containing timing relationships is generated, and a standardized format control instruction set is output. By scheduling control command transmission through time-sensitive networking, a timing-optimized command stream is obtained. Specifically, clock synchronization of each node is achieved based on the IEEE 1588 Precision Time Protocol, a time-aware shaper is used to arrange command transmission time slots, high-priority transmission channels are allocated for critical control commands, and a timing-optimized command stream with deterministic delay guarantee is output. The system sends control commands to the switch operating mechanism and the temperature and humidity control mechanism and monitors their execution status to obtain execution status data. Specifically, it sends opening and closing commands to the switch operating mechanism and heater and fan control commands to the temperature and humidity control mechanism through the TSN network. It also collects the action feedback signals of the switch operating mechanism and the temperature and humidity control mechanism in real time and outputs execution status data that includes the command execution status and effect. Based on the execution status data, the control effect is evaluated and the parameters are optimized to obtain the collaborative control optimization results. Specifically, the difference between the actual control effect and the expected target is analyzed, the control parameters are dynamically adjusted according to the execution status, the timing of the coordination between multiple actuators is optimized, and the collaborative control optimization results containing parameter optimization suggestions are output.

10. A method applied to the multimodal operation control system of an environmental protection ring main unit as described in any one of claims 1-9, characterized in that, include: Step S1: Collect environmental parameters using various environmental monitoring sensors deployed in the ring network cabinet to obtain an environmental monitoring dataset; Step S2: Collect the electrical operating parameters, mechanical operating status, thermal operating status and insulation operating status of the ring main unit to obtain the operating status dataset; Step S3: Based on the environmental monitoring dataset and the operation status dataset, perform feature extraction and operation mode classification to obtain the operation mode identification result; Step S4: Construct a digital twin model of the ring main unit, and use the digital twin model to perform multi-physics coupling simulation and operation trend prediction on the operation mode identification results to obtain the operation status prediction results; Step S5: Based on the operation mode recognition results and operation state prediction results, dynamically adjust the optimization target weights and generate a multi-objective optimization strategy; Step S6: Based on the multi-objective optimization strategy, generate an operation control strategy through an expert knowledge base and a deep reinforcement learning algorithm; Step S7 is used to send control commands to the switch operating mechanism and the temperature and humidity control mechanism through the time-sensitive network according to the operation control strategy, so as to realize the multi-modal collaborative control of the ring main unit.

Images