Power topology adaptive reconfiguration and energy consumption optimization control method for HPLC communication

Abstract

The application relates to the technical field of smart grid communication and discloses a power topology adaptive reconstruction and energy consumption optimization control method for HPLC communication, which comprises the following steps: collecting HPLC communication characteristic data of power network nodes in real time; constructing a dynamic space-time topology graph and predicting a future topology state based on the communication characteristic data; combining the predicted future topology state and real-time energy consumption parameters, and generating a communication power distribution and routing path joint optimization strategy by using a multi-objective reinforcement learning algorithm; verifying the electromagnetic compatibility and protocol compliance of the optimization strategy by using a digital twin model, and generating device executable control instructions; and dynamically adjusting the transmission power and communication routing of the HPLC node by using the control instructions. The application improves the communication quality stability and energy efficiency level, realizes topology dynamic perception, multi-objective collaborative optimization and strategy safety verification, and guarantees the efficient and reliable operation of HPLC communication in a complex power grid environment.

Description

Technical Field

[0001] This invention relates to the field of smart grid communication technology, and in particular to a power topology adaptive reconfiguration and energy consumption optimization control method for HPLC communication. Background Technology

[0002] In the real-time monitoring and control system of smart grids, a reliable and efficient communication network is the core foundation for ensuring the stable operation of the power system. Traditional power line carrier (PLC) technology transmits data through power lines, which has the advantage of eliminating wiring. However, its communication quality is easily affected by factors such as changes in line impedance, load fluctuations, and electromagnetic interference. Especially in the scenario of dynamic grid topology reconfiguration, problems such as channel parameter drift and surges in communication latency often occur. With the popularization of high-precision power equipment and wide-area IoT terminals, high-precision positioning and time synchronization (HPLC) communication technology has gradually become a new paradigm for smart grid communication due to its high stability and low latency characteristics. However, in the complex and ever-changing power network environment, how to achieve adaptive topology reconfiguration and energy consumption optimization of HPLC communication systems still faces significant technical challenges.

[0003] Current mainstream HPLC communication control schemes mostly employ pre-configured static routing strategies and fixed power allocation mechanisms, setting communication paths and transmission power thresholds based on historical operating data. While these methods can maintain basic communication needs in scenarios with relatively stable power grid structures, they often lack real-time sensing and dynamic adjustment capabilities when facing dynamic topology changes caused by sudden load shifts or distributed power source integration. This leads to frequent problems such as communication congestion at critical nodes and offline issues with edge devices, severely restricting the real-time nature of power grid status sensing.

[0004] Existing technologies have significant limitations in adaptability to dynamic environments: power control strategies based on empirical thresholds are difficult to accurately match time-varying channel conditions, resulting in communication quality fluctuations and energy waste; static routing configuration dynamically optimizes path selection based on link status, leading to a decrease in network throughput and an increase in end-to-end latency; at the same time, the lack of a systematic verification mechanism for electromagnetic interference and protocol constraints can easily cause communication conflicts and equipment overload.

[0005] Therefore, this invention proposes a power topology adaptive reconstruction and energy consumption optimization control method for HPLC communication to address the shortcomings of existing technologies. Summary of the Invention

[0006] The purpose of this invention is to provide a power topology adaptive reconfiguration and energy consumption optimization control method for HPLC communication, which solves the problems of difficult adaptive reconfiguration, low energy efficiency and insufficient policy compliance in dynamic power topologies of HPLC communication.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a power topology adaptive reconstruction and energy consumption optimization control method for HPLC communication, comprising the following steps:

[0008] Real-time acquisition of HPLC communication characteristic data of power grid nodes;

[0009] Based on the aforementioned communication feature data, a dynamic spatiotemporal topology graph is constructed and the future topology state is predicted.

[0010] By combining the predicted future topology and real-time energy consumption parameters, a multi-objective reinforcement learning algorithm is used to generate a joint optimization strategy for communication power allocation and routing paths.

[0011] The electromagnetic compatibility and protocol compliance of the optimization strategy are verified through a digital twin model, and control commands that can be executed by the device are generated.

[0012] The transmit power and communication routing of the HPLC node are dynamically adjusted by control commands.

[0013] Preferably, the HPLC communication characteristic data includes:

[0014] Current phase angle at the node This is used to reflect the dynamic impact of electrical load on the characteristics of communication channels;

[0015] Frequency domain impedance characteristics of nodes ;

[0016] in, Indicates frequency The complex impedance measured below; Let be the real part of the impedance, representing the frequency at which the impedance is at a given frequency. The resistive component below; The imaginary part of the impedance is represented by the frequency. The reactance component below; The imaginary unit;

[0017] The performance parameters of the communication link should include at least the signal-to-noise ratio. Packet loss rate and average delay .

[0018] Preferably, the step of constructing a dynamic spatiotemporal topology map based on the communication feature data and predicting future topology states includes:

[0019] Based on the acquired HPLC communication characteristic data, a spatiotemporal topology map is constructed:

[0020] ;

[0021] in, For the set of power network nodes, The total number of nodes; Let it be the set of edges; The node feature matrix;

[0022] edge set In, any edge The weights are defined as follows:

[0023] ;

[0024] in, For a moment node To node Frequency domain impedance; This represents the signal-to-noise ratio of the corresponding link; This is the impedance weighting coefficient;

[0025] Node feature matrix Defined as:

[0026] ;

[0027] in, The HPLC center frequency; For the first The current phase angle of each node; This represents the packet loss rate of that node.

[0028] Based on history Topological sequence at time 1 Through spatiotemporal graph attention network Predicting the future The topological state at each time step:

[0029] ;

[0030] in, Indicates the time of prediction The topology graph, .

[0031] Preferably, the spatiotemporal graph attention network It contains the following sub-modules:

[0032] The spatial attention coefficient calculation module calculates the node using the following formula. with neighboring nodes Attention coefficient between:

[0033] ;

[0034] in, and They are nodes with neighboring nodes eigenvectors; This is a trainable weight matrix; This is a trainable attention vector; This indicates vector concatenation;

[0035] The spatiotemporal feature aggregation module fuses the features of neighboring nodes with the node's own time-series features:

[0036] ;

[0037] in, This is the normalized attention coefficient; Indicates a node Perform temporal convolution operations on the feature sequences; It is a non-linear activation function;

[0038] The multi-step prediction output module predicts future topological states based on feature aggregation results.

[0039] ;

[0040] in, For layer normalization processing; It is a multilayer perceptron; For layer normalization operation; .

[0041] Preferably, the step of generating a joint optimization strategy for communication power allocation and routing paths using a multi-objective reinforcement learning algorithm, combining the predicted future topology state and real-time energy consumption parameters, includes:

[0042] Predicted future topological state With real-time energy consumption parameters Perform feature fusion to generate joint input features:

[0043] ;

[0044] in, Flattening operation for topological graph sequences; For feature splicing operations;

[0045] Define a multi-objective reinforcement learning model:

[0046] State space:

[0047] ;

[0048] in, The set of node packet loss rates;

[0049] Action space:

[0050] ;

[0051] in, Represents a node The transmit power control quantity; For routing selection indication;

[0052] Reward function:

[0053] ;

[0054] in, This represents the average transmission power across the entire network. The average signal-to-noise ratio; Maximum packet loss rate; , , These are the weighting coefficients;

[0055] The dual-delay deep deterministic policy gradient algorithm, a component of multi-objective reinforcement learning algorithms, is used to train and optimize the policy network, and solve for the joint optimization policy. The format is:

[0056] ;

[0057] in, Discount factor; Mathematical expectation operator; To optimize the time window length; For the future The reward value for each step.

[0058] Preferably, the step of verifying the electromagnetic compatibility and protocol compliance of the optimization strategy through a digital twin model and generating control commands executable by the device includes:

[0059] strategy The generated transmit power set Input the digital twin model to calculate the electromagnetic interference intensity of the entire network:

[0060] ;

[0061] in, For nodes arrive Antenna gain coefficient; This refers to the node spacing; The path loss exponent is... If so, it is determined that the electromagnetic compatibility requirements are met. The preset electromagnetic interference threshold;

[0062] Verify routing policy satisfy:

[0063] ;

[0064] in, This represents the maximum number of connections allowed per node by the protocol. This refers to the link transmission delay; This represents the maximum permissible delay for the protocol.

[0065] Generate control instructions that the device can execute:

[0066] ;

[0067] in, and This provides standardized device control interface functions.

[0068] Preferably, the path loss index The value is dynamically adjusted based on the network deployment environment, specifically including:

[0069] Free space environment Take the first preset value;

[0070] Urban environment Take the second preset value;

[0071] Dense obstacle environment Take the third preset value;

[0072] Electromagnetic interference threshold The values ​​of satisfy:

[0073] ;

[0074] in, For regulatory standard limits; This is the device's maximum transmission power; This represents the maximum antenna gain. This represents the minimum distance between nodes in the network.

[0075] Preferably, the step of dynamically adjusting the transmit power and communication routing of the HPLC node through control commands includes:

[0076] Based on real-time channel quality parameters and Generate nodes Transmit power adjustment command:

[0077] ;

[0078] in, Represents a node Signal-to-noise ratio; Represents a node Bit error rate; SN The preset signal-to-noise ratio threshold; Adjust the power step size; It is a symbolic function;

[0079] Based on link quality indicators and Generate routing table update instructions:

[0080] ;

[0081] in, Represents a node arrive Transmission delay Represents a node arrive The packet loss rate; The preset delay threshold; This is the preset packet loss rate threshold.

[0082] Preferably, the power adjustment step size satisfy:

[0083] ;

[0084] in, are adaptive coefficients and ; This is the maximum allowable transmission power of the equipment; This is the minimum permissible transmit power of the device; This represents the average signal-to-noise ratio for the previous time period.

[0085] This invention also provides a power topology adaptive reconfiguration and energy consumption optimization control system for HPLC communication, including:

[0086] The data acquisition module is used to acquire HPLC communication characteristic data of power grid nodes in real time;

[0087] The topology analysis module is used to construct a dynamic spatiotemporal topology map based on communication characteristic data and predict the future network topology state.

[0088] The strategy generation module is used to generate a joint optimization strategy for communication power allocation and routing paths based on the predicted future topology state and real-time energy consumption parameters using a multi-objective reinforcement learning algorithm.

[0089] The verification execution module is used to verify the electromagnetic compatibility and protocol compliance of the optimization strategy through a digital twin model, and to generate control commands that can be executed by the device.

[0090] The power routing control module is used to dynamically adjust the transmit power and communication routing configuration of the HPLC node according to control commands.

[0091] In summary, the present invention has at least one of the following beneficial technical effects:

[0092] 1. This invention, by real-time acquisition of HPLC communication characteristic data such as current phase angle and frequency domain impedance, combined with a dynamic construction and prediction mechanism of spatiotemporal topology maps, can accurately perceive changes in power network load and fluctuations in channel status. Based on the future topology prediction technology of ST-GAT networks, the system possesses forward-looking adjustment capabilities, effectively overcoming the problem of delayed response to topology changes in traditional methods, and ensuring that communication quality remains stable even when the power grid structure dynamically changes.

[0093] 2. This invention employs a multi-objective reinforcement learning algorithm to jointly optimize power allocation and routing paths. Through the design of a composite reward function, it reduces the average transmit power of the entire network while improving the signal-to-noise ratio and suppressing the maximum packet loss rate. This mechanism overcomes the limitations of traditional fixed strategies, achieving a dynamic balance between communication quality, energy efficiency, and network reliability, and fully leveraging the high-speed characteristics of HPLC communication.

[0094] 3. This invention introduces a digital twin model for electromagnetic compatibility verification and protocol constraint validation. By pre-calculating the electromagnetic interference intensity of the entire network and the compliance of routing parameters, it avoids communication interruptions or equipment damage caused by policy violations. This verification mechanism provides dual assurance for the physical feasibility of control commands, significantly improving the security of system deployment.

[0095] 4. This invention utilizes an adaptive power adjustment algorithm based on channel state and a link quality-driven routing update mechanism to achieve fine-grained control of transmit power and real-time optimization of communication paths. Through dynamic step size adjustment and environmental adaptive coefficient design, it minimizes ineffective energy consumption while ensuring communication reliability, thus achieving the green and energy-saving goals of smart grid communication systems. Attached Figure Description

[0096] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0097] Figure 2 This is a schematic diagram of the system architecture of the present invention. Detailed Implementation

[0098] The following is in conjunction with the appendix Figure 1 - Appendix Figure 2 The present invention will be further described in detail below.

[0099] This invention provides a method for adaptive power topology reconstruction and energy consumption optimization control for HPLC communication, comprising the following steps:

[0100] S1. Real-time acquisition of HPLC communication characteristic data of power network nodes;

[0101] S2. Construct a dynamic spatiotemporal topology map based on the communication feature data and predict future topology states;

[0102] S3. Combining the predicted future topology and real-time energy consumption parameters, a multi-objective reinforcement learning algorithm is used to generate a joint optimization strategy for communication power allocation and routing paths.

[0103] S4. Verify the electromagnetic compatibility and protocol compliance of the optimization strategy through a digital twin model, and generate control commands that can be executed by the device;

[0104] S5. Dynamically adjust the emission power and communication routing of the HPLC node through control commands.

[0105] For step S1, in this embodiment, in order to realize the identification of power network communication characteristics for high-frequency power line communication (HPLC) and subsequent topology reconstruction and control strategy optimization, it is first necessary to collect HPLC communication characteristic data in real time from each communication node of the power network.

[0106] The aforementioned HPLC communication characteristic data includes, but is not limited to, the following core parameters:

[0107] First, at the node level, the current phase angle of the current node is collected and denoted as... ,in Represents the first in the power network Each node. This current phase angle parameter reflects the dynamic impact of the power load on the communication channel characteristics and is an important physical quantity for judging the degree of disturbance to channel conditions caused by changes in node load. Since the characteristics of power line communication channels are significantly affected by current distribution and phase, a phase angle parameter is introduced. It can be used to sense changes in the power grid topology and potential interference of load conditions on the transmission performance of HPLC channels.

[0108] Secondly, to further characterize the electrical properties of the nodes in the frequency domain, this embodiment collects the frequency domain impedance characteristics of the nodes, which can be expressed by the following formula:

[0109] ;

[0110] in, Indicates frequency The complex impedance measured below is a concentrated manifestation of the influence of a node on signal propagation at a specific frequency. Let be the real part of the impedance, representing the frequency at which the impedance is at a given frequency. The resistance component below mainly reflects the energy dissipation characteristics at the nodes; The imaginary part of the impedance is represented by the frequency. The reactance component is further distinguished into inductive and capacitive effects; The unit is imaginary. Frequency domain impedance parameters not only reveal the electrical characteristics of the node itself, but can also be used to reflect the harmonic distribution and crosstalk trend in adjacent communication links.

[0111] Furthermore, at the communication link level, in order to comprehensively evaluate the current performance status of the HPLC link, this embodiment introduces the collection of multiple communication performance parameters, including but not limited to:

[0112] Signal-to-noise ratio (SNR) SNR (Signal-to-Noise Ratio) represents the energy ratio between the effective communication signal and the background noise, used to measure the signal quality of a link during actual transmission. A higher SNR value generally means more stable link communication.

[0113] Packet Loss Rate PLR indicates the proportion of data packets that failed to reach their destination within a certain time or data volume. PLR can reveal the stability of the link or the degree of noise impact under high load conditions.

[0114] Average Delay ): This represents the average delay time experienced by a data packet from the source node to the destination node, reflecting the transmission efficiency of the network under the current topology and scheduling strategy.

[0115] During data acquisition, a periodic sampling mechanism combined with an asynchronous reporting strategy is preferred to ensure the real-time performance of key feature data while reducing communication load. For each of the above data items, it is necessary to ensure the sampling time accuracy, frequency resolution, and signal integrity to avoid data offset affecting subsequent processing.

[0116] The feature data acquisition in this embodiment not only provides basic data support for subsequent topology reconstruction and optimization control, but also improves the system's adaptability to power grid disturbances and load changes. In particular, it has good robustness and real-time performance when facing heterogeneous interference and multipath effects in complex power communication environments.

[0117] In summary, this embodiment achieves real-time, efficient, and complete acquisition of HPLC communication characteristic data of power network nodes through structured design and multi-dimensional data acquisition methods.

[0118] For step S2, in this embodiment, in order to realize the dynamic modeling of the topology and prediction of the future state of the power network under HPLC communication conditions, a dynamic spatiotemporal topology graph construction method based on communication feature data is proposed, and a multi-step topology evolution prediction is realized by combining the Spatio-Temporal Graph Attention Network (ST-GAT).

[0119] In this embodiment, a multi-time topology sequence is first constructed based on real-time acquired HPLC communication characteristic data. The current system state time is set as... Then at every moment The constructed graph structure is represented as follows:

[0120] ;

[0121] in, For the set of power network nodes, The total number of nodes; Let be the set of edges, representing the communication links that exist between any two nodes; Let be the node feature matrix, representing the node at time t. The multidimensional feature data possessed by each node at that time;

[0122] In the definition of edges, this embodiment uses a weighted approach that combines impedance characteristics and link communication quality to define arbitrary edges. weight This weight is used to comprehensively reflect the communication capability between nodes and the status of the power physical connection. Its calculation formula is as follows:

[0123] ;

[0124] in, For a moment node To node The frequency domain impedance reflects the degree of physical coupling of the power line; This represents the signal-to-noise ratio of the corresponding link; is the impedance weighting coefficient, which is adjusted in the (0,1) interval to balance the contribution of physical connection characteristics and communication performance characteristics in the edge weight.

[0125] Node feature matrix The combined feature coding method is defined as follows:

[0126] ;

[0127] in, The HPLC center frequency; For the first The current phase angle of each node; This represents the packet loss rate of the node; the above features reflect the electrical load characteristics and communication reliability, and serve as the input feature dimensions for each node in the graph neural network.

[0128] To predict future topological states, this embodiment employs the Spatiotemporal Graph Attention Network (ST-GAT) as the core model for both modeling and prediction. This network uses historical data as its foundation. The topological graph sequence at each time step is used as input, i.e.:

[0129] ;

[0130] Network output is the future The predicted topology sequence at each time step:

[0131] ;

[0132] in, Indicates the time of prediction The topology graph, .

[0133] In the ST-GAT structure, the connection relationship between each node and its neighbors is first modeled using a spatial attention coefficient calculation module, with the following formula:

[0134] ;

[0135] in, and They are nodes with neighboring nodes eigenvectors; This is a trainable weight matrix; This is a trainable attention vector; This represents vector concatenation. This module is used to automatically learn the importance scores of neighboring nodes to the central node, adapting to situations with uneven weight distribution across multiple links in complex power communication environments.

[0136] After calculating the attention coefficient, the spatiotemporal feature aggregation module is used to fuse the spatiotemporal features of the nodes. The processing method is as follows:

[0137] ;

[0138] in, Represents a node fused feature vectors; Represents a node The set of neighboring nodes; This is the normalized attention coefficient; Indicates a node The feature sequences are subjected to temporal convolution operations to extract their temporal evolution features; This is a non-linear activation function used to enhance the network's expressive power. This module considers both the spatial structure and the intrinsic correlation between time series data.

[0139] Finally, in the multi-step prediction output module, the topological state prediction output of the fused feature vector is performed in the following form:

[0140] ;

[0141] in, Layer normalization is performed to eliminate the influence of feature distribution offset; It is a multilayer perceptron used to perform regression prediction of edge weights and node attributes in graph structures; For layer normalization operation; , which is the time step index for future predictions; This represents the fusion of feature vectors from all nodes in the network. The resulting fusion feature matrix.

[0142] In summary, this embodiment effectively realizes the modeling and prediction of power network topology in the spatiotemporal dimension by constructing a weighted graph model based on HPLC communication features and introducing a multi-level graph attention and temporal convolution mechanism.

[0143] For step S3, in this embodiment, the predicted future topological state sequence is first... Real-time energy consumption parameters of each node Feature fusion processing is performed. Specifically, joint input features are constructed through the following steps:

[0144] ;

[0145] in, The topological graph sequence is flattened while preserving node features and edge weights at each time step. The feature concatenation operation forms a joint feature vector that includes spatiotemporal evolution features and energy state information.

[0146] Based on the above features, the state space of the reinforcement learning model is constructed and defined as follows:

[0147] ;

[0148] in, This represents the set of packet loss rates for each node. The state space design fully considers future network topology evolution trends, real-time node energy consumption levels, and communication reliability indicators, providing comprehensive environmental awareness for policy generation.

[0149] Regarding the definition of the action space, this embodiment adopts a joint optimization framework to unify the modeling of transmit power adjustment and routing decision-making:

[0150] ;

[0151] in, Represents a node The transmit power control quantity is subject to physical limitations of the equipment. For routing selection indication, when This indicates that data packets are allowed to pass through the link. transmission.

[0152] This motion space design achieves coupled optimization of power control and routing decisions.

[0153] To balance multiple objectives such as communication quality, energy efficiency, and network reliability, this embodiment designs a composite reward function:

[0154] ;

[0155] in, This represents the average transmit power across the entire network, reflecting the overall energy consumption level of the system. The average signal-to-noise ratio (SNR) characterizes communication quality. Maximum packet loss rate; , , The weighting coefficient is used to adjust the relative importance of each optimization objective, and its specific value can be configured according to actual deployment needs.

[0156] At the algorithm implementation level, this embodiment preferably employs the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm for policy optimization. This algorithm solves for the optimal policy in the following form. :

[0157] ;

[0158] in, This is a discount factor used to balance immediate rewards and long-term benefits; Mathematical expectation operator; To optimize the time window length, the forward-looking range of the control strategy is adjusted; For the future The algorithm updates the policy network and value network alternately using the Actor-Critic framework. The Critic network employs a dual Q-network structure to suppress value overestimation, and a policy update delay mechanism is introduced to improve training stability.

[0159] During training, the policy network receives the current state. As input, the output is a continuous power control quantity. Discrete routing To address the characteristics of the mixed action space, a branch network structure is preferred for handling different types of actions. The power control branch outputs a normalized power value, and the routing branch implements differentiable discrete decision-making through the Gumbel-Softmax technique.

[0160] For step S4, this embodiment proposes a virtual verification method based on a digital twin model to address the electromagnetic compatibility verification and protocol compliance verification requirements of the optimization strategy. This method is used when a multi-objective reinforcement learning algorithm generates a joint optimization strategy. Then, the transmit power set output by the strategy will be... With routing strategy set Input the data into the digital twin simulation environment to perform electromagnetic interference intensity calculation and protocol constraint verification.

[0161] First, the electromagnetic interference intensity of the entire network is calculated based on the electromagnetic field propagation model, and its expression is as follows:

[0162] ;

[0163] in, For nodes arrive Antenna gain coefficient; This refers to the node spacing; The preset electromagnetic interference threshold; This is the path loss index. This index is dynamically adjusted based on the network deployment environment: in a free-space propagation environment, a first preset value is used. The urban multipath environment adopts the second preset value. The dense obstacle environment adopts the third preset value. This is to accurately reflect the signal attenuation patterns under different scenarios.

[0164] Electromagnetic Interference Threshold The determination follows the principle of dual constraints:

[0165] ;

[0166] in, For regulatory standard limits; This is the device's maximum transmission power; This represents the maximum antenna gain. This represents the minimum distance between nodes in the network.

[0167] During the protocol compliance verification phase, the routing policy undergoes a double constraint check:

[0168] Single-node connection number constraint To prevent node overload;

[0169] Link delay constraints To ensure real-time requirements.

[0170] in, This represents the maximum number of connections allowed per node by the protocol. This refers to the link transmission delay; This represents the maximum allowable delay for the protocol.

[0171] If and only if the electromagnetic interference intensity Furthermore, when the routing policy passes protocol verification, the digital twin model converts the optimization parameters into standardized control commands:

[0172] ;

[0173] in, The interface function writes the target power value to the node power controller; The function sends the binary routing matrix to the network switching device. In cases where verification fails, a policy rollback mechanism is triggered, and the violation parameters are recorded as negative sample inputs for subsequent reinforcement learning training, thereby improving the robustness and compliance of the policy generation module.

[0174] For step S5, in this embodiment, an adaptive power and routing adjustment mechanism based on channel state awareness is proposed to meet the real-time control requirements of the HPLC node. After the digital twin model is verified, the system sends the generated control commands to each network node to perform dynamic adjustment of transmit power and communication routing.

[0175] Based on real-time channel quality parameters and Generate nodes Transmit power adjustment command:

[0176] Regarding transmit power adjustment, nodes Based on real-time channel quality parameters (Signal-to-noise ratio) and (Bit error rate) is used to calculate the power correction amount. The generation of power adjustment commands follows the formula:

[0177] ;

[0178] in, Represents a node Signal-to-noise ratio; Represents a node Bit error rate; SN The preset signal-to-noise ratio threshold; Adjust the power step size; It is a symbolic function.

[0179] When measured When it is below this threshold, A positive function output triggers a power increase; conversely, a negative output reduces the transmission power.

[0180] The calculation method for the power adjustment step size is as follows:

[0181] ;

[0182] in, are adaptive coefficients and This is used to control the balance between the adjustment range and the convergence speed; This is the maximum allowable transmission power of the equipment; This is the minimum permissible transmit power of the device; The average signal-to-noise ratio of the previous period is used to introduce a logarithmic term to achieve a smooth response to channel quality and avoid power oscillations.

[0183] Regarding route updates, nodes According to link quality indicators (Transmission delay) and (Packet loss rate) Generate routing table update command:

[0184] ;

[0185] in, Represents a node arrive Transmission delay Represents a node arrive The packet loss rate; The preset delay threshold; This is the preset packet loss rate threshold.

[0186] This instruction iterates through all potential links. Filter out those that simultaneously meet the time delay threshold and packet loss rate threshold For qualified links, update the routing table of the node. Perform thorough cleaning, synchronizing the routing update cycle with the power adjustment cycle to ensure the spatiotemporal consistency of the control policy.

[0187] To achieve precise control, the power adjustment step size The design incorporates an environment-adaptive mechanism: when At higher levels, The term increases, making Reduce power to avoid over-adjusting power under good channel conditions; when When the value is low, the adjustment step size is increased to quickly improve communication quality. Coefficient It can be dynamically adjusted according to network load, using a smaller value to suppress fluctuations during high-load periods and a larger value to accelerate convergence during low-load periods.

[0188] The above control commands are transmitted through standardized interface functions. and The commands are sent to the node devices. Power control commands are written to the node's power amplifier control circuit in analog form, while route update commands are configured in the network switching chip's route lookup table in binary matrix form, achieving efficient hardware-level execution.

[0189] This invention also provides a power topology adaptive reconfiguration and energy consumption optimization control system for HPLC communication, including:

[0190] The data acquisition module is used to acquire HPLC communication characteristic data of power grid nodes in real time;

[0191] The topology analysis module is used to construct a dynamic spatiotemporal topology map based on communication characteristic data and predict the future network topology state.

[0192] The strategy generation module is used to generate a joint optimization strategy for communication power allocation and routing paths based on the predicted future topology state and real-time energy consumption parameters using a multi-objective reinforcement learning algorithm.

[0193] The verification execution module is used to verify the electromagnetic compatibility and protocol compliance of the optimization strategy through a digital twin model, and to generate control commands that can be executed by the device.

[0194] The power routing control module is used to dynamically adjust the transmit power and communication routing configuration of the HPLC node according to control commands.

[0195] The system in this embodiment can be used to execute the above method embodiments, and its principle and technical effect are similar, so they will not be described again here.

[0196] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A power topology adaptive reconstruction and energy consumption optimization control method for HPLC communication, characterized in that, Includes the following steps: Real-time acquisition of HPLC communication characteristic data of power network nodes; Based on the communication feature data, a dynamic spatiotemporal topology map is constructed and the future topology state is predicted; By combining the predicted future topology and real-time energy consumption parameters, a multi-objective reinforcement learning algorithm is used to generate a joint optimization strategy for communication power allocation and routing paths. The electromagnetic compatibility and protocol compliance of the optimization strategy are verified through a digital twin model, and control commands that can be executed by the device are generated. The emission power and communication routing of the HPLC node are dynamically adjusted by control commands; The step of generating a joint optimization strategy for communication power allocation and routing paths using a multi-objective reinforcement learning algorithm, which combines the predicted future topology state and real-time energy consumption parameters, includes: Predicted future topological state With real-time energy consumption parameters Perform feature fusion to generate joint input features: ； in, Flattening operation for topological graph sequences; For feature splicing operations; Define a multi-objective reinforcement learning model: State space: ； in, The set of node packet loss rates; Action space: ； in, Represents a node The transmit power control quantity; For routing selection indication; Reward function: ； in, This represents the average transmission power across the entire network. The average signal-to-noise ratio; Maximum packet loss rate; , , These are the weighting coefficients; The dual-delay deep deterministic policy gradient algorithm, a multi-objective reinforcement learning algorithm, is used to train and optimize the policy network, and solve for the joint optimization policy. The format is: ； in, Discount factor; Mathematical expectation operator; To optimize the time window length; For the future The reward value for each step.

2. The power topology adaptive reconfiguration and energy consumption optimization control method for HPLC communication according to claim 1, characterized in that, The HPLC communication feature data includes: Current phase angle at the node This is used to reflect the dynamic impact of power load on the characteristics of communication channels; Frequency domain impedance characteristics of nodes ; in, Indicates frequency The complex impedance measured below; Let be the real part of the impedance, representing the frequency at which the impedance is at a given frequency. The resistive component below; The imaginary part of the impedance is represented by the frequency. The reactance component below; The imaginary unit; The performance parameters of the communication link should include at least the signal-to-noise ratio. Packet loss rate and average latency .

3. The power topology adaptive reconstruction and energy consumption optimization control method for HPLC communication according to claim 1, characterized in that, The step of constructing a dynamic spatiotemporal topology map based on the communication feature data and predicting the future topology state includes: Based on the collected HPLC communication characteristic data, a spatiotemporal topology map is constructed: ； in, For the set of power network nodes, The total number of nodes; Let it be the set of edges; The node feature matrix; edge set In, any edge The weights are defined as follows: ； in, For a moment node To node Frequency domain impedance; This represents the signal-to-noise ratio of the corresponding link; This is the impedance weighting coefficient; Node feature matrix Defined as: ； in, The HPLC center frequency; For the first The current phase angle of each node; This represents the packet loss rate of that node. Based on history Topological sequence at time 1 Through spatiotemporal graph attention network Predicting the future The topological state at each time step: ； in, Indicates the time of prediction The topology graph, .

4. The power topology adaptive reconstruction and energy consumption optimization control method for HPLC communication according to claim 3, characterized in that, The spatiotemporal graph attention network It contains the following sub-modules: The spatial attention coefficient calculation module calculates the node using the following formula. with neighboring nodes Attention coefficient between: ； in, and They are nodes with neighboring nodes eigenvectors; is a trainable weight matrix; This is a trainable attention vector; This indicates vector concatenation; The spatiotemporal feature aggregation module fuses the features of neighboring nodes with the node's own time-series features: ； in, Represents a node fused feature vectors; This is the normalized attention coefficient; Indicates a node Perform temporal convolution operations on the feature sequences; It is a non-linear activation function; Represents a node The set of neighboring nodes; The multi-step prediction output module predicts future topological states based on feature aggregation results. ； in, For layer normalization processing; It is a multilayer perceptron; For layer normalization operation; ; This represents the fusion of feature vectors from all nodes in the network. The resulting fusion feature matrix.

5. The power topology adaptive reconfiguration and energy consumption optimization control method for HPLC communication according to claim 1, characterized in that, The steps of verifying the electromagnetic compatibility and protocol compliance of the optimization strategy through a digital twin model and generating control commands executable by the device include: strategy The generated transmit power set Input the digital twin model and calculate the electromagnetic interference intensity of the entire network: ； in, For nodes arrive Antenna gain coefficient; This refers to the node spacing; The path loss exponent is... If so, it is determined that the electromagnetic compatibility requirements are met. The preset electromagnetic interference threshold is used; Verify routing policy satisfy: ； in, This represents the maximum number of connections allowed per node by the protocol. This refers to the link transmission delay; This represents the maximum permissible delay for the protocol. Generate control instructions that the device can execute: ； in, and This provides standardized device control interface functions.

6. The power topology adaptive reconfiguration and energy consumption optimization control method for HPLC communication according to claim 5, characterized in that, The path loss index The value is dynamically adjusted based on the network deployment environment, specifically including: Free space environment Take the first preset value; Urban environment Take the second preset value; Dense obstacle environment Take the third preset value; Electromagnetic interference threshold The values ​​of satisfy: ； in, For regulatory standard limits; This is the device's maximum transmission power; This represents the maximum antenna gain. This represents the minimum distance between nodes in the network.

7. The power topology adaptive reconstruction and energy consumption optimization control method for HPLC communication according to claim 1, characterized in that, The steps of dynamically adjusting the emission power and communication routing of the HPLC node through control commands include: Based on real-time channel quality parameters and Generate nodes Transmit power adjustment command: ； in, Represents a node Signal-to-noise ratio; Represents a node Bit error rate; SN The preset signal-to-noise ratio threshold; Adjust the power step size; It is a symbolic function; Based on link quality indicators and Generate routing table update instructions: ； in, Represents a node arrive Transmission delay Represents a node arrive The packet loss rate; The preset delay threshold; This is the preset packet loss rate threshold.

8. The power topology adaptive reconfiguration and energy consumption optimization control method for HPLC communication according to claim 7, characterized in that, The power adjustment step size satisfy: ； in, are adaptive coefficients and ; This refers to the maximum allowable transmission power of the equipment. This is the minimum permissible transmit power of the device; This represents the average signal-to-noise ratio for the previous time period.

9. A power topology adaptive reconfiguration and energy consumption optimization control system for HPLC communication, applied to the power topology adaptive reconfiguration and energy consumption optimization control method for HPLC communication as described in any one of claims 1-8, characterized in that, include: The data acquisition module is used to acquire HPLC communication characteristic data of power grid nodes in real time; The topology analysis module is used to construct a dynamic spatiotemporal topology map based on communication characteristic data and predict the future network topology state. The strategy generation module is used to generate a joint optimization strategy for communication power allocation and routing paths based on the predicted future topology state and real-time energy consumption parameters using a multi-objective reinforcement learning algorithm. The verification execution module is used to verify the electromagnetic compatibility and protocol compliance of the optimization strategy through a digital twin model, and to generate control commands that can be executed by the device. The power routing control module is used to dynamically adjust the transmit power and communication routing configuration of the HPLC node according to control commands.

Images