A high-speed rail four-electric information edge cooperative control method and system

Abstract

The application discloses a kind of high-speed rail four electric power information edge collaborative control method and system, belong to high-speed rail four electric power technical field, comprising: in high-speed rail four electric power system inside deployment edge control node;The original data of four electric power equipment are acquired and preprocessed to obtain the optimal state estimate;Calculate theoretical electrical parameter, generate physical residual vector and generate fault identification;Calculate optimal energy flow allocation reference value, regional net load;Define integrated control utility function, generate final local actuator output instruction vector;Define the local regulation feasible region of node i, according to regional net load, local regulation feasible region Calculate collaborative demand vector, dynamic communication topology;Generate consistent collaborative regulation instruction;Generate whole network synchronous execution instruction;Upload cloud, and integrate upload content.The prior art, there are many defects in high-speed rail four electric power management.Compared with prior art, the application is effectively optimized, thereby effectively improving the management performance.

Description

Technical Field

[0001] This invention relates to the field of high-speed rail electrical engineering technology, and in particular to a method and system for edge collaborative control of electrical information in high-speed rail electrical systems. Background Technology

[0002] my country boasts the world's longest high-speed railway network, and its four key electrical systems (traction power supply, electrical power, communication, and signaling) are the core infrastructure ensuring safe and efficient train operation. With the advancement of "dual-carbon" goals, increasing demands for intelligent and energy-saving railways, and the continuous densification of the high-speed rail network and rapid expansion of equipment scale, the four electrical systems exhibit significant characteristics such as numerous points of contact, long lines, wide distribution, heterogeneous equipment, large data volumes, and high real-time requirements. The traditional centralized power consumption information collection and management model is no longer adequate for the new operational and safety needs, necessitating an upgrade towards a distributed, edge-based, and collaborative approach.

[0003] Currently, the power information management of high-speed railways mostly adopts a centralized architecture of "terminal collection - centralized upload - cloud processing": electricity meters, sensors, and data collectors are deployed at locations such as box-type substations, substations, distribution substations, and communication signal equipment rooms along the line to collect data such as voltage, current, power, energy, temperature and humidity, partial discharge, cable head temperature measurement, smoke detection, and door magnetic sensors; the collected data is transmitted to the railway bureau / section-level master station through serial port, 485, power line carrier, or private network, and is uniformly stored, analyzed, and controlled by the master station; the logic of abnormal alarms, load adjustment, fault location, and maintenance work orders is all centrally calculated and distributed by the cloud master station.

[0004] However, existing technologies have many shortcomings and deficiencies: First, they have poor real-time performance and slow response. Second, they usually use AI as an auxiliary means, but AI requires a large amount of data for training, which is easily affected by the drift of training data distribution.

[0005] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0006] This invention provides a method for edge collaborative control of electrical information in high-speed rail, including: S1: deploying edge control nodes within the high-speed rail electrical system to cover electrical equipment within a preset area; S2: Obtain raw data of the four electrical equipment through the edge control node; S3: Preprocess the raw data to obtain the optimal state estimate. ; S4: Based on the optimal state estimate Calculate theoretical electrical parameters Based on theoretical electrical parameters Generate physical residual vector Compare the physical residual vectors With dynamic adaptive threshold To generate fault identifiers ; S5: Calculate the optimal energy flow allocation reference value Area net load ; S6: Define the comprehensive control utility function Based on the physical residual vector Dynamic adaptive threshold Optimal energy flow allocation reference value Area net load Generate the final local executor output instruction vector ; S7: Define the local adjustment feasible region of node i Based on the area's net load Local adjustment feasible domain Calculate the collaborative demand vector Dynamic communication topology ; S8: Generate consistent coordination instructions ; S9: Node i's coordinated adjustment command Perform the projection operation and generate the final execution timestamp based on the operation result. To generate network-wide synchronized execution instructions. ; S10: Upload S3 to S9 to the cloud and integrate the uploaded content into operation and maintenance work orders and energy consumption statistics reports.

[0007] Furthermore, S3 also includes the following steps: The raw data received by the edge control node at discrete time k are integrated into the raw observation vector. ; Define the system state vector as ; The prediction equation is:

[0008] The observation update results are as follows: ; ; ; Calculate the dynamic observation noise covariance matrix: ; This is the state transition matrix; To control the input matrix; The control quantity from the previous moment; The Kalman gain matrix; H is the observation matrix; To estimate the error covariance matrix; For the dynamic observation noise covariance matrix; It is the identity matrix; These are adaptive coefficients.

[0009] Furthermore, S4 also includes the following steps: Define the physical mechanism equation function Based on the optimal state estimate Calculate theoretical electrical parameters for: ; Define the physical residual vector: ; Define a dynamic adaptive threshold: ; Generate fault identifiers: ; This is the threshold coefficient.

[0010] Furthermore, S5 also includes the following steps: Define the net load of the area within the coverage area of ​​the edge node: ; Construct the minimum convex objective function: ; The constraints are: ; ; Based on the Lagrange multiplier method, construct the KKT conditions and obtain the optimal energy flow allocation reference value: ; Inequality constraints; For the corresponding KKT multipliers; This represents the current actual load demand; Real-time photovoltaic power generation; This provides regenerative braking power for the train. This refers to the energy storage discharge / charge power. These are the weighting coefficients; It is in a state of energy storage and charging.

[0011] Furthermore, S6 also includes the following steps: Define the comprehensive control utility function: ; The standardized control instructions are as follows: ; For security protection instructions; Directives to optimize the economy; This is a dynamic priority weight.

[0012] Furthermore, S7 also includes the following steps: The feasible region Ωi for local adjustment is: ; When exceeding the area's net load When the local adjustment feasible region Ωi is reached, the cooperative demand vector is calculated: ; when At that time, broadcast coordination requests; Upon receiving a response, a dynamic communication topology is constructed: ; For participating nodes; Signal-to-noise ratio of the communication link; This is the collaborative dead zone threshold.

[0013] Furthermore, S8 also includes the following steps: The global optimization objective is set as minimizing the total adjustment cost across all regions: ; ; Introducing the Lagrange augmented function, the local update formula at iteration t+1 is: ; The dual variable is updated as follows: ; The iteration stopping condition is: .

[0014] This refers to the actual regulating power undertaken. The ideal expected adjustment value; For penalty parameters; Consensus dual variables; Gather for neighbors.

[0015] Furthermore, S9 also includes the following steps: The projection operation is as follows: ; Calculate the global consistency verification metric: ; when The verification passed. Based on the final execution timestamp, generate a network-wide synchronized execution command: .

[0016] This invention also discloses a high-speed rail electrical information edge collaborative control system, which applies the method described above.

[0017] Compared with the prior art, the present invention has the following advantages: This invention differs from the "probabilistic output" of black-box AI such as LSTM. Its output conforms to basic circuit principles, preventing unexpected misjudgments, making it particularly suitable for the high safety requirements of high-speed rail. Its technological advantage stems from deterministic mathematical and physical equations, unaffected by shifts in the distribution of training data.

[0018] This invention requires no GPU or NPU acceleration; a standard ARM industrial-grade chip can complete the entire calculation process within 50ms. It avoids the massive matrix multiplication and addition operations inherent in deep learning.

[0019] Traditional collaborative methods require sending detailed electrical data to a central node. This invention only exchanges marginal cost information and power deviations between neighbors; commercially / safety-sensitive data such as original load and equipment capacity remain within the local network. Its advantage stems from the "data does not move, model moves" characteristic of distributed consensus algorithms.

[0020] This invention improves state estimation accuracy by approximately 40% in scenarios with strong electromagnetic interference, such as high-speed rail phase transitions and high-power sudden changes, through a closed-loop correction mechanism driven by novel residuals.

[0021] This invention will By directly incorporating it as a negative load into the convex optimization objective function, and in conjunction with energy storage regulation, the local consumption rate of renewable energy can be increased from less than 60% in traditional schemes to over 95%.

[0022] This invention does not rely on a central coordinator. Even if some nodes leave the coordination group due to tunnel signal loss, the remaining nodes can still automatically reconstruct a new topology and converge to a suboptimal solution, making the system extremely resilient.

[0023] The unconditional safety preemption priority set by this invention ensures the determinism of the upper bound of the control delay, meeting the stringent requirements of high-speed railway relay protection. Attached Figure Description

[0024] Figure 1 Overall flowchart. Detailed Implementation

[0025] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0026] Example 1: This invention provides a method for edge-coordinated control of power information in high-speed rail's four electrical systems, referring to... Figure 1 , Figure 1 This is a schematic diagram of the process of the present invention. Specifically, it includes the following steps: S1: Deploy edge control nodes within the high-speed rail's four-electric system to cover the electrical equipment within a pre-defined area.

[0027] S2: Obtain raw data of the four electrical equipment through the edge control node.

[0028] S3: The raw data contains electromagnetic noise and random fluctuations, and cannot be directly used for control. The raw data received by the edge control nodes at discrete time k are integrated into the raw observation vector. (Includes three-phase voltage, current, and power).

[0029] Define the system state vector as (Includes actual effective value, frequency, and temperature change rate).

[0030] The prediction equation (prior estimate) is:

[0031] The observation update results (posterior estimate) are as follows: .

[0032] .

[0033] .

[0034] Calculate the dynamic observation noise covariance matrix: .

[0035] in, For the new information residual, This is an adaptive forgetting factor.

[0036] Let be the state transition matrix.

[0037] To control the input matrix.

[0038] This is the control quantity from the previous moment.

[0039] This is the Kalman gain matrix.

[0040] H is the observation matrix.

[0041] To estimate the error covariance matrix.

[0042] This is the dynamic observation noise covariance matrix.

[0043] It is an identity matrix.

[0044] These are adaptive coefficients.

[0045] Therefore, this invention improves the state estimation accuracy by about 40% in strong electromagnetic interference scenarios such as high-speed rail phase transitions and high-power sudden changes through a closed-loop correction mechanism driven by new information residuals.

[0046] S4: Define the physical mechanism equation function Based on the optimal state estimate Calculate theoretical electrical parameters for: .

[0047] For example, for the transformer substation area, the theoretical power balance equation is: .

[0048] in .

[0049] Define the physical residual vector: .

[0050] To eliminate normal fluctuations, a dynamic adaptive threshold is defined instead of a fixed dead zone: .

[0051] in, and These are the mean and standard deviation of the residuals within the sliding time window L, respectively. Generate fault identifiers: .

[0052] This is a threshold coefficient used to balance the false alarm rate and the false alarm rate. Preferably, the value ranges from 3 to 5.

[0053] In summary, this invention constructs a white-box model that is more sensitive to unknown faults and requires no training data, making it extremely suitable for high-speed rail high-safety scenarios.

[0054] S5: Define the net load within the coverage area of ​​the edge node: .

[0055] Construct a convex objective function that minimizes the "cost of electricity extraction from the upper-level power grid + energy storage lifespan loss": .

[0056] The constraints are: .

[0057] .

[0058] Based on the Lagrange multiplier method, construct the KKT conditions and obtain the optimal energy flow allocation reference value: .

[0059] This is an inequality constraint.

[0060] For the corresponding KKT multipliers.

[0061] This represents the current actual load demand.

[0062] This represents the real-time power generation of photovoltaic systems.

[0063] This provides regenerative braking feedback power for the train.

[0064] This refers to the energy storage discharge / charge power.

[0065] These are the weighting coefficients.

[0066] It is in a state of energy storage and charging.

[0067] In summary, through the aforementioned steps, the present invention will By directly incorporating it as a negative load into the convex optimization objective function, and in conjunction with energy storage regulation, the local consumption rate of renewable energy can be increased from less than 60% in traditional schemes to over 95%.

[0068] S6: Define the comprehensive control utility function: .

[0069] This is a dynamic priority weight. When... season .

[0070] The standardized control instructions are as follows: .

[0071] This is a safety protection instruction.

[0072] To optimize the output for economic purposes, an incremental PID controller is used to smooth the output. .

[0073] in, This represents the power tracking error.

[0074] These are the proportional, integral, and derivative coefficients of the PID controller.

[0075] The increment of the control quantity at the current moment.

[0076] If standardized control commands If the anomaly cannot be quelled or cross-regional power flow is involved, then subsequent collaborative steps will be executed.

[0077] In summary, through the aforementioned steps, this invention sets an unconditional safe preemption priority, ensuring the deterministic upper bound of the control delay and meeting the stringent requirements of high-speed rail relay protection.

[0078] S7: The feasible region for local adjustment Ωi is: .

[0079] When exceeding the area's net load When the local adjustment feasible region Ωi is reached, the cooperative demand vector is calculated: .

[0080] when ( When the dead zone threshold is reached, node i broadcasts a collaboration request to other nodes whose physical distance is less than the set value.

[0081] Upon receiving a response, a dynamic communication topology is constructed: .

[0082] For participating nodes.

[0083] This refers to the signal-to-noise ratio of the communication link.

[0084] This is the collaborative dead zone threshold.

[0085] S8: Set the global optimization objective to minimize the total adjustment cost across all regions. .

[0086] , Let be the adjustment cost coefficient for node i.

[0087] Global constraint: The sum of the output of each node must equal the total deficit / total surplus.

[0088] Introducing the Lagrange augmented function, the local update formula at iteration t+1 is: .

[0089] The dual variable is updated as follows: .

[0090] The iteration stopping condition is: .

[0091] This refers to the actual regulating power undertaken.

[0092] This is the ideal expected adjustment value.

[0093] This is the penalty parameter.

[0094] For consensus dual variables.

[0095] Set up a group for neighbors.

[0096] In summary, through the aforementioned steps, compared to traditional collaborative systems that require sending detailed electrical data to the central node, this invention only exchanges marginal cost information and power deviations between neighbors. Commercially / safety-sensitive data such as original loads and equipment capacity remain within the local network. Furthermore, through these steps, this invention does not rely on a central coordinator. Even if some nodes leave the collaborative group due to tunnel signal loss, the remaining nodes can automatically reconstruct a new topology and converge to a suboptimal solution, demonstrating extremely strong system resilience.

[0097] S9: Node i's coordinated adjustment command Perform a projection operation: .

[0098] Calculate the global consistency verification metric: .

[0099] when The verification was successful.

[0100] Based on the final execution timestamp, generate a network-wide synchronized execution command: .

[0101] If communication is interrupted or ADMM fails to converge, the node will automatically fall back to local safe mode. .

[0102] S10: Upload the contents of S3 to S9 to the cloud and integrate the uploaded contents into maintenance work orders and energy consumption statistics reports.

[0103] In summary, this invention differs from the "probabilistic output" of black-box AI such as LSTM. Its output conforms to basic circuit principles, preventing unexpected misjudgments, making it particularly suitable for the high safety requirements of high-speed rail. Its technological advantage stems from deterministic mathematical and physical equations, unaffected by shifts in the distribution of training data.

[0104] This invention requires no GPU or NPU acceleration; a standard ARM industrial-grade chip can complete the entire calculation process within 50ms. This avoids the massive matrix multiplication and addition operations inherent in deep learning. Therefore, it effectively improves real-time performance and accelerates response efficiency.

[0105] For ease of understanding, the following calculation examples are provided: The scenario is set as follows: a collaborative area consisting of three edge control nodes (Node-1, Node-2, and Node-3) within a power supply arm of a high-speed railway. Among them, Node-1 is a traction substation, Node-2 is an AT substation, and Node-3 is a sectioning substation.

[0106] Taking Node-1 as an example: Measured load (True power 800 kW, noise 12 kW).

[0107] Actual photovoltaic measurement (True value 100 kW, noise -2 kW).

[0108] Actual Regeneration (True power 200 kW, noise 5 kW).

[0109] Let the state variable be Observation matrix H=1. Based on the prediction and update equations: Prior estimation: (Assume the load is stable in the short term, A=1, B=0).

[0110] Calculate the new information residual: .

[0111] Adaptive adjustment of observation noise covariance: Assuming convergence but .

[0112] Final optimal estimate:

[0113] Similarly, the photovoltaic estimate can be obtained. Regeneration estimation .

[0114] The theoretical net load should be:

[0115] Measured equivalent net load (after filtering) .

[0116] Generate physical residuals: .

[0117] Check the dynamic adaptive threshold The current value is calculated based on the historical window. .because The judgment is: no hard fault, Flag_fault = 0, and the process is allowed to enter the economic optimization stage.

[0118] Node objective function construction (general):

[0119] in (In this scenario, only discharge is required, i.e.) Take the weighting coefficients. (Electricity price factor) (Energy storage lifespan penalty).

[0120] Perform analytical calculations for each node: Node-1: Net Demand Without energy storage, the upstream power draw is 512kW. With energy storage, the actual power draw... Substitute the values ​​into the objective function and take the derivative: Solving for: However, it is subject to hard constraints. Constraints, taking the boundary optimal solution: At this time, the superior unit draws power. (Lower than the grid limit of 300 kW, locally satisfied).

[0121] Node-2: Net Demand Unconstrained differentiation: Hard constraints have to: (Exceeded the grid limit of 400 kW, external support needs to be requested).

[0122] Node-3: No energy storage ( Net demand have to:

[0123] Node-1 and Node-3 did not exceed their local power extraction limits, but Node-2's calculated local power extraction of 450 kW exceeded its 400 kW limit, and its energy storage was already at maximum capacity. Node-2 reported a "shortage" in its local power input. (External assistance is needed to reduce power extraction or obtain power injection). Meanwhile, the global command issued by the grid dispatch center is that the total power extraction should be ≤800 kW. The current optimal local power extraction for the three nodes is:

[0124] The total power exceeded the global target (800 kW) by 112 kW.

[0125] Local decision: No fault signal, no safety preemption triggered; However, Node-2 itself exceeds the limit, and the global condition is not met, so subsequent coordination steps must be triggered; (Node-1 and Node-3 will remain unchanged for the time being, while Node-2 will be temporarily limited to 400 kW of power intake, but a shortfall of 50 kW will be calculated).

[0126] Calculate the adjustment feasible region of each node.

[0127] Node-1 and Node-3 receive a coordination request and build a communication topology. .

[0128] Define the node adjustment cost function .

[0129] in This represents the additional reduction value actually borne by node i (a positive number indicates a reduction in power consumption). The cost coefficient represents the node's local expected reduction value (local ideals are all 0, meaning it doesn't want to exert any effort). The smaller the value, the more willing the employee is to exert effort (flexible adjustment). (Has the ability to send counter-transfers at a low cost) (Completely unadjustable, extremely costly, and forcing it to bear the load would lead to the loss of the main traction load, which is impractical) (The severable portion has a minor load and moderate cost) Global coupling constraints: .

[0130] Using ADMM iteration, the iteration formula (local update of node i) is as follows:

[0131] The final iteration result is as follows: Node-1 handles 60.5 kW (the reverse power supply is close to the limit of 50 kW, slightly exceeding it, which is achieved by slightly adjusting the station service transformer), and Node-3 handles 50.3 kW (cutting off non-critical loads). The sum of the three precisely meets the global constraint of 112 kW.

[0132] Node-1: Calculated power is 60.5 kW, but the actual reverse transmission capacity is at most 50 kW. Perform projection operation:

[0133] Node-2: Calculated result 1.2 kW, within the allowable range (0~50), maintain .

[0134] Node-3: The calculated result is 50.3 kW, but the maximum non-critical load is 50 kW. Projection yields:

[0135] Total reduction after projection:

[0136] The power output did not reach 112 kW, leaving a residual of 10.8 kW. At this point, a secondary local fine-tuning was triggered (Node-1 over-generated 0.8 kW, Node-3 over-cut 0.8 kW, and Node-2 took over 9.2 kW by adjusting the UPS charging sequence—in actual engineering, this is made up by PID tracking at the second level; this is just for illustrative purposes and the processing is simplified).

[0137] Assuming the final execution value after two fine-tuning steps is:

[0138] Consistency verification metrics:

[0139] Synchronous execution: Each node executes according to the PTP precision clock. The final command is issued simultaneously at all times: Node-1: The inverter switches from discharging 300 kW to discharging 351 kW (equivalent to feeding back 51 kW).

[0140] Node-2: Reduce the power of the station's air conditioning by 10 kW.

[0141] Node-3: Short-term shelving of non-critical loads of 51 kW.

[0142] After implementation, the entire network will actually draw power from its parent organization: .

[0143] Example 2: An edge collaborative control system for power consumption information of high-speed railway's four electrical systems (electrical, electrical, and electronic) uses the method described in Example 1. Specifically, edge control nodes are deployed along the high-speed railway's four electrical systems, including transformer substations, power distribution stations, communication signal rooms, and traction power supply stations. Each edge control node covers the four electrical equipment within a preset area. All edge control nodes establish communication connections with the cloud management platform through an industrial-grade wireless mesh network, and peer-to-peer communication links are established between each edge control node.

[0144] Among them, the edge control node adopts an industrial-grade edge computing gateway, which has the characteristics of electromagnetic interference prevention, waterproof and dustproof, and high and low temperature resistance, and is suitable for the complex outdoor environment along the high-speed rail line; the industrial-grade wireless mesh network adopts 5G+LoRa dual-mode communication to ensure the stability of data transmission in weak network and no network environments, with a transmission latency of ≤100ms; the division of the preset area is determined according to the distribution of stations, equipment density and communication distance along the high-speed rail line, to ensure that each four-electrical equipment can be covered by the edge control node and avoid blind spots in data collection.

[0145] Each edge control node collects raw data of the four electrical equipment within its coverage area in real time through its built-in acquisition module.

[0146] Each edge control node compresses and uploads its locally pre-processed effective data, local decision results, collaborative decision results, and control execution logs to the cloud management platform at preset intervals (e.g., 10 minutes, 1 hour). The cloud management platform performs global analysis, model training, and strategy optimization on the aggregated data, and distributes the optimized control strategies and AI model parameters to each edge control node to achieve collaborative iteration between the "end-edge-cloud".

[0147] The data upload process employs a breakpoint resume mechanism. When the network is interrupted, the edge control node temporarily stores the data to be uploaded to its local storage unit and resumes uploading once the network is restored, ensuring that the data is not lost. The model training on the cloud management platform adopts a federated learning mode, iterating the model based on the anonymized data uploaded by each edge node (removing sensitive information such as device number and specific location), avoiding the leakage of original data and protecting data privacy. The cloud-optimized control strategy and AI model parameters are distributed to each edge control node through an industrial-grade wireless mesh network. The edge nodes automatically update their local strategies and models without manual intervention.

[0148] On the other hand, the cloud management platform generates equipment maintenance work orders and energy consumption statistics reports based on the data uploaded by each edge node, and monitors the operating status of edge nodes and electrical equipment in real time. According to the changes in high-speed rail operating load, equipment aging and "dual carbon" energy saving requirements, the control strategy and coordination rules of edge control nodes are dynamically adjusted to ensure the long-term stable operation of the system.

[0149] Maintenance work orders include information such as the location of the faulty equipment, the type of fault, handling suggestions, and the responsible person. They are pushed to maintenance personnel via mobile devices to enable rapid fault repair. Energy consumption statistics reports are generated daily, weekly, and monthly, intuitively displaying the energy consumption of each area and each piece of equipment, providing data support for energy-saving optimization. During dynamic adjustments, the cloud management platform optimizes the coverage, collaboration rules, and control thresholds of edge nodes through global analysis to adapt to the dynamic changes in high-speed rail operation requirements.

[0150] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0151] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0152] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0153] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

[0154] It is understood that the system provided in the embodiments of the present invention corresponds to the method provided in the embodiments of the present invention, and the explanation, examples and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

Claims

1. A method for edge-coordinated control of electrical information in high-speed rail systems, characterized in that, Including: S1: Deploying edge control nodes within the high-speed rail's four electrical systems to cover the electrical equipment within a pre-defined area; S2: Obtain the raw data of the four electrical equipment through the edge control node; S3: Preprocess the raw data to obtain the optimal state estimate. ; S4: Based on the aforementioned optimal state estimate Calculate theoretical electrical parameters Based on the aforementioned theoretical electrical parameters Generate physical residual vector Compare the physical residual vector With dynamic adaptive threshold To generate fault identifiers ; S5: Calculate the optimal energy flow allocation reference value Area net load ; S6: Define the comprehensive control utility function Based on the physical residual vector The dynamic adaptive threshold The optimal energy flow allocation reference value The net load of the area Generate the final local executor output instruction vector ; S7: Define the local adjustment feasible region of node i Based on the net load of the area The local adjustment feasible domain Calculate the collaborative demand vector Dynamic communication topology ; S8: Generate consistent coordination instructions ; S9: Node i's coordinated adjustment command Perform the projection operation and generate the final execution timestamp based on the operation result. To generate network-wide synchronized execution instructions. ; S10: Upload the contents of S3 to S9 to the cloud and integrate the uploaded contents into maintenance work orders and energy consumption statistics reports.

2. The edge collaborative control method for power information consumption in high-speed rail as described in claim 1, characterized in that, S3 further includes the following steps: The raw data received by the edge control node at discrete time k is integrated into the raw observation vector. ; Define the system state vector as ; The prediction equation is: The observation update results are as follows: ； ； ； Calculate the dynamic observation noise covariance matrix: ； The This is the state transition matrix; The To control the input matrix; The The control quantity from the previous moment; The The Kalman gain matrix; H is the observation matrix; The To estimate the error covariance matrix; The For the dynamic observation noise covariance matrix; The It is the identity matrix; The These are adaptive coefficients.

3. The edge collaborative control method for power consumption information in high-speed rail as described in claim 1, characterized in that, S4 further includes the following steps: Define the physical mechanism equation function Based on the optimal state estimate Calculate theoretical electrical parameters for: ； Define the physical residual vector as follows: ； Define a dynamic adaptive threshold: ； Generate fault identifiers: ； The This is the threshold coefficient.

4. The edge collaborative control method for power information consumption in high-speed rail as described in claim 1, characterized in that, S5 further includes the following steps: Define the net load of the area within the coverage area of ​​the edge node: ； Construct the minimum convex objective function: ； The constraints are: ； ； Based on the Lagrange multiplier method, the KKT conditions are constructed, and the optimal energy flow allocation reference value is obtained: ； The Inequality constraints; The For the corresponding KKT multipliers; The This represents the current actual load demand; The Real-time photovoltaic power generation; The This provides regenerative braking power for the train. The This refers to the energy storage discharge / charge power. The These are the weighting coefficients; The It is in a state of energy storage and charging.

5. The edge collaborative control method for power information of high-speed rail four-electric systems as described in claim 1, characterized in that, S6 further includes the following steps: Define the comprehensive control utility function: ； The standardized control instructions are as follows: ； The For security protection instructions; The Directives to optimize the economy; The This is a dynamic priority weight.

6. The edge collaborative control method for power information of high-speed rail four-electric systems as described in claim 1, characterized in that, S7 further includes the following steps: The local adjustment feasible region Ωi is: ； When the net load of the area exceeds the limit When the local adjustment feasible region Ωi is reached, the cooperative demand vector is calculated: ； when At that time, broadcast coordination requests; Upon receiving a response, the dynamic communication topology is constructed as follows: ； The For participating nodes; The Signal-to-noise ratio of the communication link; The This is the collaborative dead zone threshold.

7. The edge collaborative control method for power information consumption in high-speed rail as described in claim 1, characterized in that, S8 further includes the following steps: The global optimization objective is set as minimizing the total adjustment cost across all regions: ； The ; Introducing the Lagrange augmented function, the local update formula at iteration t+1 is: ； The dual variable is updated as follows: ； The iteration stopping condition is: 。 The This refers to the actual regulating power undertaken. The The ideal expected adjustment value; The For penalty parameters; The Consensus dual variables; The Gather for neighbors.

8. The edge collaborative control method for power information of high-speed rail four-electric systems as described in claim 1, characterized in that, S9 further includes the following steps: The projection operation is as follows: ； Calculate the global consistency verification metric: ； when The verification passed. Based on the final execution timestamp, the network-wide synchronized execution instruction is generated: 。 9. A high-speed rail electrical, electronic, and electronic information edge collaborative control system, characterized in that: The method described in any one of claims 1 to 8 is applied.

Images