POE-POF power supply transmission method and system, and storage medium

Abstract

The invention belongs to the technical field of power supply transmission, and relates to a POE-POF power supply transmission method and system, and a storage medium. The method comprises the following steps: collecting electric quantity data of POE-POF power supply transmission network nodes in real time, and extracting a dynamic power feature set; identifying an abnormal power loss event based on the feature set and extracting an abnormal mode feature; abnormal path nodes are positioned by combining a POE-POF power supply transmission network topological structure, voltage and current distribution parameters are optimized, and an optimized power distribution scheme is generated; dynamically adjusting voltage and current values of a power supply path, evaluating an energy consumption trend and generating a trend data set; and adaptively adjusting the electric quantity acquisition parameters according to the energy consumption trend, updating the acquisition configuration and circularly executing the steps. According to the invention, intelligent detection, accurate positioning and dynamic optimization of abnormal loss of the power grid are realized, the stability and energy efficiency of the power supply network are improved, and the method is suitable for fine management and energy-saving optimization of the intelligent power grid.

Description

Technical Field

[0001] This application relates to the field of power transmission technology, and in particular to a POE-POF power transmission method, system and storage medium, which is suitable for the refined management and energy-saving optimization of power supply networks. It aims to realize the intelligent detection, accurate location and dynamic optimization of abnormal power supply network losses, and improve stability and energy efficiency. Background Technology

[0002] Anomaly power loss management in smart grids is a significant technical challenge facing current power systems. With the large-scale grid connection of new energy sources and the diversification of electricity loads, traditional anomaly detection methods based on threshold alarms and manual inspections are insufficient to meet the demands of refined management and control in modern power grids. Existing technologies suffer from the following main shortcomings: first, low anomaly identification accuracy, making it difficult to distinguish between normal fluctuations and actual faults; second, slow location response speed, unable to adapt to microsecond-level dynamic changes; and third, a lack of systematic optimization and adjustment, making it difficult to balance efficiency and stability.

[0003] From a technological development perspective, power grid anomaly management has evolved from manual inspection to automated monitoring, and then to intelligent diagnostics. Early SCADA systems achieved centralized monitoring but lacked flexibility. While next-generation intelligent diagnostic technologies have incorporated machine learning algorithms, they still have significant limitations in handling complex network topologies and dynamic load changes. Especially in scenarios with a high proportion of renewable energy integration, the randomness of power fluctuations exacerbates the difficulty of anomaly management.

[0004] This application addresses the aforementioned technical bottlenecks by proposing a POE-POF power transmission method, system, and storage medium. By constructing an anomaly mode feature set, establishing a dynamic energy consumption trend assessment model, and adaptively optimizing voltage and current distribution parameters, the system solves core problems such as inaccurate anomaly identification, untimely location, and insufficient optimization, providing technical support for the safe and economical operation of smart grids. Summary of the Invention

[0005] This application provides a POE-POF power transmission method, system, and storage medium, which can realize intelligent detection, precise location, and dynamic optimization of abnormal power network losses. This method addresses core issues such as inaccurate anomaly identification, untimely location, and insufficient optimization by constructing an anomaly mode feature set, establishing a dynamic energy consumption trend assessment model, and adaptively optimizing voltage and current distribution parameters, thus providing technical support for the safe and economical operation of smart grids.

[0006] In a first aspect, this application provides a POE-POF power transmission method, which includes the following steps: S101. Obtain the real-time power data set of the power supply network from the POE-POF power transmission network node, and extract dynamic power features based on the real-time power data set to obtain a dynamic power feature set; S102. Based on the dynamic power feature set, identify abnormal power loss to obtain an abnormal power loss event set, and extract abnormal mode features based on the abnormal power loss event set to obtain an abnormal mode feature set. S103. Determine the abnormal path node set based on the abnormal mode feature set and the power transmission network topology, optimize the voltage and current allocation parameters based on the abnormal path node set, and generate an optimized power allocation scheme. S104. Adjust the voltage and current values ​​of the power supply path according to the optimized power allocation scheme to obtain a real-time power allocation parameter set, and evaluate the energy consumption trend based on the real-time power allocation parameter set to generate an energy consumption trend dataset. S105. Adjust the power consumption acquisition parameters according to the energy consumption trend dataset, generate an adaptive power consumption acquisition parameter set, update the power consumption acquisition configuration based on the adaptive power consumption acquisition parameter set, and obtain a new power consumption dataset.

[0007] Furthermore, the step of obtaining the real-time power data set from the nodes of the power supply network includes: collecting an initial power data set containing voltage, current and power values ​​using high-precision power acquisition sensors deployed at each node of the power supply network; and preprocessing the initial power data set to obtain the real-time power data set.

[0008] Furthermore, the step of extracting dynamic power features based on the real-time power dataset to obtain a dynamic power feature set includes: processing the real-time power dataset using a time-series analysis algorithm to extract the power fluctuation frequency and power fluctuation amplitude; and generating the dynamic power feature set based on the power fluctuation frequency and power fluctuation amplitude.

[0009] Furthermore, the step of identifying abnormal power loss based on the dynamic power feature set and obtaining an abnormal power loss event set includes: determining whether the power fluctuation amplitude in the dynamic power feature set exceeds a preset threshold; if so, using an anomaly detection algorithm to identify the abnormal power loss and generating the abnormal power loss event set.

[0010] Furthermore, the step of extracting abnormal pattern features based on the abnormal power loss event set to obtain an abnormal pattern feature set includes: extracting microsecond-level abnormal patterns from the abnormal power loss event set; and using a fast Fourier transform algorithm to analyze the time-domain and frequency-domain characteristics of the microsecond-level abnormal patterns to generate an abnormal pattern feature set containing time-series abnormal correlation features.

[0011] Furthermore, the step of extracting abnormal pattern features based on the abnormal power loss event set to obtain an abnormal pattern feature set includes: extracting microsecond-level abnormal patterns from the abnormal power loss event set; and using a fast Fourier transform algorithm to analyze the time-domain and frequency-domain characteristics of the microsecond-level abnormal patterns to generate the abnormal pattern feature set containing time-series abnormal correlation features.

[0012] Furthermore, determining the abnormal path node set based on the abnormal pattern feature set and the power transmission network topology includes: combining the abnormal pattern feature set and the power transmission network topology, using a graph analysis algorithm to calculate path weights and abnormal propagation paths; locating specific path nodes of the abnormal power loss based on the path weights and the abnormal propagation paths, and generating the abnormal path node set.

[0013] Furthermore, the step of optimizing voltage and current allocation parameters based on the abnormal path node set to generate an optimized power allocation scheme includes: optimizing dynamic voltage adjustment and current allocation parameters using the least squares method for the abnormal path node set; and adjusting the power supply priority of each node by combining path priority ranking and node dependency analysis to generate the optimized power allocation scheme.

[0014] Furthermore, the step of adjusting the voltage and current values ​​of the power supply path according to the optimized power allocation scheme to obtain a real-time power allocation parameter set includes: using an intelligent power adjustment algorithm to dynamically adjust the voltage and current values ​​of the power supply path through real-time feedback control and adaptive adjustment frequency; generating the real-time power allocation parameter set based on the adjustment results to suppress power fluctuation amplitude.

[0015] Secondly, this application provides a POE-POF power transmission system, the POE-POF power transmission system comprising: The data acquisition and feature extraction module is used to collect voltage, current and power data of POE-POF power transmission network nodes in real time, extract dynamic power features and generate dynamic power feature sets; The anomaly identification and feature extraction module is used to identify abnormal power loss events and extract abnormal pattern features based on the dynamic power feature set using machine learning or threshold analysis algorithms. The abnormal path location and power optimization module is used to locate abnormal path nodes, optimize voltage and current distribution parameters, and generate the optimal power distribution scheme by combining the power transmission network topology and abnormal mode characteristics. The dynamic adjustment and energy consumption assessment module is used to dynamically adjust the voltage and current values ​​of the power supply path according to the optimization scheme, and to assess the energy consumption trend in real time and generate an energy consumption trend dataset. The adaptive acquisition and control module is used to adaptively adjust the power acquisition parameters according to the energy consumption trend data, and at the same time, it adopts POE-POF power transmission technology to achieve efficient synchronous transmission of power and data.

[0016] Thirdly, this application provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the POE-POF power transmission method.

[0017] Furthermore, the storage medium is divided into the following units: a core program storage unit, used to store the power dynamic adjustment algorithm, the anomaly diagnosis program, and the optoelectronic collaborative control program, responsible for the system's operating logic and intelligent decision-making; an operating data recording unit, used to collect and store power supply parameters, anomaly event records, and energy consumption analysis data in real time; a parameter configuration unit, used to save device operating parameters and adaptive adjustment thresholds; a system management unit, which adopts an embedded architecture to realize the collaborative management of programs and data, supporting remote maintenance and program upgrades; and a security protection unit, which ensures the security of stored data and the reliability of the system through data encryption and redundant backup technologies.

[0018] This application proposes a POE-POF power transmission method, system, and storage medium, which can accurately identify, locate, and eliminate abnormal power losses in large-scale, highly dynamic power supply networks. Compared with the prior art, the beneficial effects of the technical solution of this application are at least as follows: First, the combination of autoregressive moving average and fast Fourier transform extracts dynamic power features, accurately quantifies fluctuation frequency, amplitude and harmonic peak value, and improves the sensitivity of anomaly identification.

[0019] Second, the abnormal pattern characteristics are coupled with the network topology to locate the abnormal path node, achieving millisecond-level source tracing and shortening the fault diagnosis time.

[0020] Third, the least squares method is used to optimize voltage and current distribution parameters, and the power supply path parameters are dynamically adjusted through an intelligent power regulation algorithm to effectively suppress power fluctuations and reduce line losses.

[0021] Fourth, the energy consumption trend assessment drives the adaptive adjustment and continuous cyclical optimization of power collection parameters. The sampling density is updated according to changes in operating conditions to avoid resource waste, improve overall energy efficiency and power supply stability, and ensure safety. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram of one embodiment of the POE-POF power transmission method in this application. Figure 2 This is a schematic diagram of one embodiment of the POE-POF power transmission system in this application. Figure 3 This is a schematic diagram of the structure of the computer device in the embodiments of this application. Detailed Implementation

[0024] This application provides a POE-POF power transmission method, system, and storage medium. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0025] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0026] like Figure 1 This embodiment provides a POE-POF power transmission method, which may specifically include: S101. Obtain the real-time power data set of the power supply network from the POE-POF power transmission network node, and extract dynamic power features based on the power data set to obtain a dynamic power feature set.

[0027] In one specific embodiment, the process of performing step S101 may specifically include the following steps: By deploying high-precision power acquisition sensors at various nodes of the power supply network, an initial power dataset containing voltage, current and power values ​​is collected. The initial power dataset is preprocessed to obtain the real-time power dataset.

[0028] Specifically, the POE-POF power transmission refers to a collaborative transmission system using Power over Ethernet (PoE) and power fiber optics. The PoE module provides DC power transmission, while the POF module enables electromagnetic interference-resistant optical signal transmission. Both modules achieve synchronous power and data transmission through a photoelectric conversion interface. In the data center rack of the POE-POF power network, node power acquisition sensors synchronously sample the voltage, current, and power of each port at microsecond intervals. The sampled data includes a 64-bit nanosecond timestamp and is sent to the edge gateway via a redundant link to obtain the initial power dataset. A sliding window filter within the gateway performs median filtering on each signal, using an 8-cycle fundamental frequency window to remove spikes. An adaptive IIR low-pass filter then removes switching noise above 3kHz. The filtered values ​​are mapped one-to-one with the original sampled values ​​using timestamps and written to a circular buffer queue. The queue's structures are arranged in the order of timestamp, voltage, current, and power, with fields corresponding to sensor register addresses. This ensures that transient power drops are fully preserved, eliminating sampling delays and noise superposition, thus obtaining the real-time power dataset.

[0029] In one specific embodiment, the process of performing step S101 may specifically include the following steps: The real-time power dataset is processed using a time-series analysis algorithm to extract the power fluctuation frequency and power fluctuation amplitude. The dynamic power feature set is generated based on the power fluctuation frequency and power fluctuation amplitude.

[0030] Specifically, the real-time power dataset flowing in from each node is transformed into a dynamic power feature set capable of revealing abnormal fluctuation patterns. This process revolves around time-series analysis algorithms while maintaining a consistent mapping of the data's physical meaning. First, the power dataset is continuously sent from node sensors at a sampling rate of 100 times per second. Each record contains three channels of values: timestamp, voltage, current, and power, typically controlled within a 600-point (6-second) window. Second, the entire power dataset is fed into an autoregressive moving average model to calculate the autocorrelation function of the sequence to identify periodic fluctuations. The sequence is then stabilized through differencing, and finally, model parameters are fitted to extract frequency features. Power fluctuation frequency represents the number of power changes per unit time, and power fluctuation amplitude represents the peak-to-valley difference. If the extracted power fluctuation amplitude exceeds 1.5 times the normal fluctuation amplitude within the past ten minutes, the system marks the corresponding window as a potential anomaly, requiring further refinement of feature extraction. Next, a fast Fourier transform algorithm is used to perform frequency domain transformation on the power anomaly data within the same 6-second window to extract frequency domain peak values ​​as supplementary information for the dynamic power features. The Fast Fourier Transform (FFT) algorithm decomposes discrete-time signals into different frequency components. Through optimized butterfly operations, it splits the 256-point zero-padded sequence into amplitude and phase spectra, accurately obtaining the frequency domain characteristics of the power supply network, including the dominant frequency distribution and harmonic component features. The fluctuation frequency and amplitude obtained in the time domain are concatenated with the dominant frequency and harmonic amplitude obtained in the frequency domain to form a four-element dynamic power feature vector. This vector maintains a window-by-window mapping with the original power dataset, allowing any feature value to be directly traced back to the original power waveform segment. This ensures that subsequent anomaly path localization and power allocation optimization can be performed specifically for electrical events.

[0031] In one embodiment, to extend the extraction of dynamic power features, different parameter settings are considered, such as adjusting the sampling frequency to the millisecond level. In large enterprise networks, the power dataset may contain data from thousands of nodes. After processing by time-series analysis algorithms, the dynamic power feature set can reveal propagation patterns related to the network topology. For example, the features show that power fluctuations spread from the central node to the edge, which enhances the practicality of the feature set and is beneficial for subsequent path localization. For instance, assuming the power fluctuation amplitude threshold is 5W, if the dataset shows an amplitude of 7W during the extraction process, frequency domain analysis reveals additional harmonics, effectively supporting anomaly identification. Compared with traditional static analysis, this method improves the accuracy and response speed of feature extraction.

[0032] S102. Based on the dynamic power feature set, identify abnormal power loss to obtain an abnormal power loss event set, and extract abnormal pattern features based on the abnormal power loss event set to obtain an abnormal pattern feature set.

[0033] In one specific embodiment, the process of performing step S102 may specifically include the following steps: Determine whether the power fluctuation amplitude in the dynamic power feature set exceeds a preset threshold. If so, use an anomaly detection algorithm to identify the abnormal power loss and generate the abnormal power loss event set.

[0034] Specifically, in the process of identifying abnormal power loss in a power supply network, a dynamic power feature set is used as input data and transformed into a set of constant power loss events through multi-stage processing. First, it is necessary to determine whether the real-time collected power fluctuation data exceeds a preset threshold. This preset threshold is 1.5 times the average fluctuation amplitude obtained from historical power data statistics. If it exceeds the preset threshold, it is marked as a potential anomaly. For example, in a PoE power supply network, if the power fluctuation amplitude exceeds 50 watts, it is marked as a potential anomaly. Second, for the marked potential anomalies, an isolation forest algorithm is used to isolate the abnormal data. The isolation forest algorithm constructs multiple isolation trees by randomly selecting features and randomly selecting split points between feature values. An anomaly score is calculated for each data point. Sampling points with an anomaly score higher than 0.6 are confirmed as abnormal power loss events, thus obtaining the abnormal power loss event set. The isolation forest algorithm is robust to high-dimensional data and can effectively distinguish between normal fluctuations and abnormal losses. For example, when a sudden increase in node current causes power anomalies, setting the number of isolation trees to 100 can reduce false positives and improve the identification accuracy, thereby providing a reliable event set for subsequent abnormal pattern extraction and helping to suppress overall power fluctuations.

[0035] In yet another specific embodiment, the process of executing step S102 may specifically include the following steps: Extract microsecond-level abnormal patterns from the set of abnormal power loss events; The time-domain and frequency-domain characteristics of the microsecond-level anomaly pattern are analyzed using the Fast Fourier Transform algorithm to generate a feature set of the anomaly pattern that includes time-series anomaly correlation features.

[0036] Specifically, feature extraction is performed on the abnormal power loss event set. The system extracts abnormal event window data with a 1-microsecond time resolution, obtaining a complete fluctuation sequence containing parameters such as voltage and current. A Fast Fourier Transform (FFT) algorithm is used to process these time-domain sequences. The transformation process is based on butterfly operations to decompose the discrete Fourier transform, converting time-domain power fluctuations into frequency-domain amplitude and phase spectra. In power supply network analysis, components with a dominant frequency exceeding 10 Hz are considered high-frequency anomalies. Frequency-domain analysis can effectively reveal periodic features not observable in the time domain; for example, a fixed frequency peak of 20 Hz often corresponds to a specific equipment failure mode. For example, given a power sequence containing 1000 microsecond-level sampling points, it is first zero-padded to a power of 2 length, then recursively decomposed into even and odd parts, and finally outputs the spectrum, revealing hidden periodic anomaly patterns, which is helpful in locating the source of power loss.

[0037] Furthermore, the fusion analysis of time-domain and frequency-domain features constitutes an anomaly pattern feature set. In the time domain, key parameters such as peak amplitude and fluctuation duration are extracted; in the frequency domain, important indicators such as dominant frequency components and harmonic distribution are calculated. These two types of features together comprehensively describe the characteristics of the anomaly event. Features associated with time-series anomalies are calculated; for example, the Pearson correlation coefficient is used to quantify the similarity of anomaly patterns between different nodes. Features with a correlation coefficient higher than 0.8 are integrated into the anomaly pattern feature set, indicating the existence of associated features. Power supply network topology information assists in determining the anomaly propagation path, effectively distinguishing isolated anomaly events from systemic faults. Throughout the processing flow, raw power data is transformed into a potential anomaly point set through threshold screening; the isolated forest algorithm refines it into confirmed anomaly events; and time-frequency analysis further transforms it into quantifiable and comparable feature vectors. The data form and connotation are gradually deepened with each processing stage.

[0038] During power supply network operation, when a sudden increase in node load leads to a significant increase in power fluctuation amplitude, the Fast Fourier Transform (FFT) algorithm can effectively identify specific fault characteristic frequencies. Amplitude spectrum analysis shows a significant peak at 20 Hz, and this frequency component is often directly related to the fault operation mode of specific types of electrical equipment. This frequency characteristic identification allows the system to correlate and compare frequency domain analysis results with time domain anomaly features. By establishing a complete anomaly pattern feature set containing multi-dimensional time and frequency domain features, the accuracy of power supply network operation status determination is significantly improved.

[0039] To adapt to complex operating conditions such as frequency converter anomalies in the power supply network, the system possesses adaptive parameter adjustment capabilities. When a frequency converter anomaly introduced by a variable load is detected, the feature extraction time window automatically expands to 500 microseconds. This adjustment ensures that the complete cycle of the low-frequency anomaly pattern can be captured. Simultaneously, by increasing the number of zero-fill points, the frequency resolution of the Fast Fourier Transform is improved to 0.1 Hz, enhancing the detection sensitivity of low-frequency components. The adjusted analysis parameters can more accurately capture the low-frequency characteristic components in frequency converter anomalies. These characteristics are directly related to the energy loss mechanisms in the power supply network, and targeted parameter optimization can effectively reduce the overall energy consumption of the system.

[0040] S103. Based on the abnormal mode feature set and the POE-POF power transmission network topology, determine the abnormal path node set, and optimize the voltage and current distribution parameters based on the abnormal path node set to generate an optimized power distribution scheme.

[0041] In one specific embodiment, the process of performing step S103 may specifically include the following steps: Combining the abnormal pattern feature set and the topology of the power transmission network, a graph analysis algorithm is used to calculate the path weight and abnormal propagation path; Based on the path weight and the abnormal propagation path, the specific path node of the abnormal power loss is located, and the abnormal path node set is generated.

[0042] Specifically, in the process of locating abnormal path nodes and optimizing power in the POE-POF power transmission network, the system first parses the network topology and models it as a directed graph, where nodes correspond to power supply equipment terminals and edges represent POE-POF hybrid transmission links. The edge weights of the topology graph are calculated by combining the physical path length and historical load data. The path length parameter is quantified by the actual wiring distance, and the historical load data is normalized using the sliding window average power value. The time-series correlation features provided by the abnormal mode feature set are processed through frequency domain transformation, and the spectral peak features extracted by the fast Fourier transform algorithm are used to identify abnormal propagation paths. These spectral features are mapped to the edges of the topology graph to form weighted abnormal propagation paths.

[0043] Furthermore, when processing weighted anomaly propagation paths using the improved Dijkstra's shortest path algorithm in graph analysis, the spectral peak factor is used as a dynamic adjustment coefficient. During processing, edge weights are updated using a weighted sum formula, where the new weight is the product of the initial weight and the spectral peak factor. The spectral peak factor is derived by dividing the peak amplitude of a specific frequency band by the average amplitude. Then, the path node with the highest weight is identified as the set of anomaly path nodes. In a typical PoE power supply scenario with 100 nodes, a topology graph is first constructed, with initial edge weights between nodes ranging from 1 to 10. Based on the characteristic that power fluctuations exceed a 5% threshold in the anomaly pattern feature set, the weights are updated, resulting in the anomaly path node set containing three key nodes. This localization method allows for rapid identification of loss sources and improves network stability.

[0044] In yet another specific embodiment, the process of executing step S103 may specifically include the following steps: For the set of abnormal path nodes, the least squares method is used to optimize the dynamic voltage adjustment and current distribution parameters; By combining path priority ranking and node dependency analysis, the power supply priority of each node is adjusted to generate the optimized power allocation scheme.

[0045] Specifically, node dependencies are extracted from the abnormal path node set. A dependency matrix is ​​constructed, where matrix elements represent the power supply dependency between nodes. The dependency is calculated based on historical current distribution ratios; for example, the dependency equals the upstream node's current output divided by the downstream node's average input. Then, a least squares optimization algorithm is used to establish a linear regression model of voltage and current parameters to calculate power loss. The objective function is to minimize the difference between the actual and predicted power values ​​to optimize dynamic voltage adjustment and current distribution parameters. The optimization process iteratively solves for the parameters using QR decomposition. Iteration terminates when the difference between the actual and predicted power values ​​converges to 0.1. In practical applications, for a set containing 5 abnormal path nodes, the dependency matrix is ​​first calculated, with element values ​​such as 0.8 indicating strong dependency. Then, the least squares method is applied. The initial voltage parameters are 48V and current is 1A, which are optimized to 47V and 0.9A. The system is sorted in descending order of dependency, prioritizing the adjustment of core nodes. This approach reduces power loss by 15% and improves distribution efficiency. It should be noted that this optimization process, through the iteration of the least squares method, can adapt to real-time load changes, ensuring that voltage and current parameters minimize losses at abnormal nodes. The beneficial effect is to suppress fluctuation amplitude and improve the overall network energy utilization rate.

[0046] Furthermore, a bubble sort algorithm is used to rank the paths by weight from highest to lowest, with paths having higher weights having higher priority. The weights are determined by the length and number of nodes in the anomaly propagation path, and high-priority paths are assigned optimized parameters first. Combined with the path priority ranking, nodes in the anomaly path node set are sorted in descending order of dependency, and their power supply priorities are adjusted. High-priority nodes are assigned optimized voltage and current values ​​first, resulting in an overall optimized power allocation scheme. This hierarchical optimization method effectively improves the overall power supply network's operational efficiency while ensuring the stable operation of critical equipment, achieving a balance between stability and energy efficiency.

[0047] In one specific embodiment, the optimization of dynamic voltage adjustment and current distribution parameters using the least squares method may specifically include the following steps: The voltage, current and power observation data of each node are obtained from the abnormal path node set, and the sum of the squares of the differences between the actual observation values ​​and the model prediction values ​​is constructed. The optimized dynamic voltage adjustment parameters and current distribution parameters are obtained by iteratively solving the least squares method.

[0048] Specifically, assuming for nodes i Voltage adjustment parameters x and current distribution parameters y The difference function between the actual observed value and the model prediction value is constructed as Equation (1), and then the gradient descent variant is used in the iterative process to solve for the optimal value. x and yThe parameters are updated each time until convergence. For example, the initial voltage is adjusted to 0 with a step size of 0.01. After 10 iterations, an optimized value with a 2% voltage reduction is obtained, which helps to reduce the risk of overheating at high-load nodes.

[0049] (1) in, t Sampling time, n Total sampling time P t for t The actual observed power value at time.

[0050] S104. Adjust the voltage and current values ​​of the power supply path based on the optimized power allocation scheme to obtain a real-time power allocation parameter set, and evaluate the energy consumption trend based on the real-time power allocation parameter set to generate an energy consumption trend dataset. In one specific embodiment, the process of executing step S104 may specifically include the following steps: Extract the power supply priority and path weight of each node from the optimized power allocation scheme; Based on power supply priority and path weight, an intelligent power regulation algorithm is adopted to calculate voltage adjustment and current distribution through real-time feedback control. The voltage and current values ​​of the power supply path are dynamically modified using the voltage adjustment and current allocation to suppress power fluctuations and generate a set of real-time power allocation adjustment parameters. Energy consumption data, including the adjusted power value sequence, is extracted from the real-time power allocation parameter set. The extracted energy consumption data is analyzed using a time series forecasting algorithm. Based on an autoregressive integral moving average model, the power series is input to predict the consumption trend within a future time window. Based on the prediction results, calculate indicators for minimizing energy loss, such as the average loss rate, and generate an energy consumption trend dataset.

[0051] Specifically, in the power optimization process of the power supply network, the optimized power allocation scheme serves as the input data source, storing the power supply priority and path weight parameters of each node, calculated through preliminary analysis. The power supply priority values ​​are calculated based on the criticality of each node in the network topology; core switching equipment is typically assigned a higher priority, while edge terminal equipment has a relatively lower priority. The path weight quantifies the degree of loss along the anomaly propagation path; this value is calculated by a graph-based analysis algorithm during the preprocessing stage. When the system parses these parameters, it establishes a mapping table between priority and weight, providing a decision-making basis for subsequent adjustment and control.

[0052] The intelligent power regulation algorithm achieves real-time feedback regulation based on an improved proportional-integral-derivative (PID) control principle. During the algorithm initialization phase, the target voltage and current values ​​of each node are read from the optimized scheme as reference parameters, while real-time data from sensors is collected. In the actual operating environment of the PoE power supply network, when a node's voltage deviation exceeds a preset threshold, the control algorithm immediately activates the regulation mechanism. The proportional control term responds quickly to the current deviation, the integral term gradually eliminates the system's steady-state error, and the derivative term predicts future trends and compensates in advance. The algorithm output is weighted by path weight coefficients to generate voltage and current adjustments for each node, ensuring more precise regulation for high-weight paths.

[0053] The energy consumption data extraction process focuses on the time-domain characteristic analysis of power value sequences. The system extracts power data from the real-time power allocation parameter set in timestamp order to construct a continuous time series. In the data processing stage, a statistical outlier detection algorithm is used to filter sampling noise to ensure data quality. Taking a data center power supply system as an example, a typical power sequence contains hundreds of high-precision sampling points per minute, which can accurately reflect the subtle characteristics of load fluctuations.

[0054] The time series forecasting model employs an autoregressive integral moving average algorithm to process power series data. During the model preprocessing stage, differencing is used to eliminate data non-stationarity. The autoregressive term captures the historical dependence characteristics of the series, while the moving average term effectively handles random disturbances. Parameter estimation is optimized using the maximum likelihood method. After multiple iterations and convergence, the model can accurately predict the power change trend within the next 15 minutes. The prediction results are compared with actual measurement data in real time to calculate key performance indicators such as the average loss rate.

[0055] The generation process of the energy consumption trend dataset integrates multi-dimensional analysis results. The system aligns predicted power values, actual measured values, loss rate indicators, and other data along the time dimension to construct structured data records. This dataset is provided to the upper-level management system through a standardized interface, supporting visualization and in-depth analysis functions. The entire processing flow forms a complete closed-loop optimization system, continuously improving the operational quality and service reliability of the power supply network through iterative improvements.

[0056] S105. Adjust the power consumption acquisition parameters based on the energy consumption trend dataset, generate an adaptive power consumption acquisition parameter set, update the power consumption acquisition configuration based on the adaptive power consumption acquisition parameter set, obtain a new power consumption dataset, and repeat the above steps.

[0057] In one specific embodiment, the process of executing step S105 may specifically include the following steps: Key indicators such as average power consumption rate and peak fluctuation amplitude are extracted from the energy consumption trend dataset to assess the current network load status. A proportional-integral-derivative (PID) control algorithm is adopted to calculate the sampling frequency adjustment based on the deviation between actual and target consumption, thereby achieving dynamic response. By combining dynamic network topology changes, the adjustment strategy is optimized through adjacency matrix and node degree weights; By integrating feedback control and topology weights, the sampling frequency is dynamically adjusted to improve accuracy under high load and reduce resource consumption when the system is stable. The optimized sampling frequency is sent to the sensor, which then collects data in real time at the new sampling frequency and sends it back to the edge gateway, forming a continuously optimized closed-loop control.

[0058] Specifically, the adaptive optimization method for power supply networks provided in this example achieves dynamic adjustment of power acquisition parameters through a closed-loop control mechanism. The system extracts key indicators such as average power consumption rate and peak fluctuation amplitude from the energy consumption trend dataset. This data, processed through time series analysis, accurately reflects the current network load status. Based on these indicators, an improved proportional-integral-derivative (PID) control algorithm is implemented. The algorithm handles real-time deviations through the proportional term, eliminates accumulated errors through the integral term, and predicts trend changes through the derivative term, calculating the initial sampling frequency adjustment. Simultaneously, the system continuously monitors changes in network topology, analyzes changes in node connectivity through an adjacency matrix, and incorporates node degree as a weighting factor into the adjustment scheme to ensure that the sampling strategy remains synchronized with the network structure. In practical implementation, when the energy consumption trend shows a power fluctuation amplitude exceeding 10%, the PID control algorithm calculates a deviation of 5%, the integral accumulates to 2%, and the derivative predicts a 1% change, initially adjusting the sampling frequency from 1Hz to 5Hz. Simultaneously, a topology update detects a new node with a 20% increase in degree, further adjusting the frequency to 6Hz. This approach enables timely response to anomalies, reduces data omissions, and improves monitoring accuracy. The optimized sampling frequency parameters are transmitted to each monitoring node in real time via standard network management protocols such as SNMP. The sensors immediately respond with configuration updates, collecting voltage, current, and power data at the new sampling rate. The collected high-precision data is then re-input into the system for analysis and processing, forming a complete closed-loop optimization cycle. Through this adaptive adjustment mechanism, the system can increase the sampling frequency to capture key data during high-load periods and reduce the sampling rate to save resources during stable operation, achieving intelligent optimization of power supply network monitoring.

[0059] like Figure 2 As shown, this application provides a POE-POF power transmission system 200, which includes: The data acquisition and feature extraction module 201 is used to acquire voltage, current and power data of POE-POF power transmission network nodes in real time, extract dynamic power features and generate dynamic power feature sets.

[0060] The anomaly identification and feature extraction module 202 is used to identify abnormal power loss events and extract abnormal pattern features based on the dynamic power feature set using machine learning or threshold analysis algorithms.

[0061] The abnormal path location and power optimization module 203 is used to locate abnormal path nodes, optimize voltage and current distribution parameters, and generate the optimal power distribution scheme by combining the power transmission network topology and abnormal mode characteristics.

[0062] The dynamic adjustment and energy consumption assessment module 204 is used to dynamically adjust the voltage and current values ​​of the power supply path according to the optimization scheme, and to assess the energy consumption trend in real time and generate an energy consumption trend dataset.

[0063] The adaptive acquisition and control module 205 is used to adaptively adjust the power acquisition parameters according to the energy consumption trend data, and at the same time, it adopts POE-POF power transmission technology to achieve efficient synchronous transmission of power and data.

[0064] This application provides a POE-POF power transmission system 200, which achieves intelligent monitoring and optimization of the POE-POF power supply network through multi-module collaboration. The system includes a data acquisition and feature extraction module 201, an anomaly identification and feature extraction module 202, an anomaly path location and power optimization module 203, a dynamic adjustment and energy consumption assessment module 204, and an adaptive acquisition and control module 205. These modules are connected sequentially to form a complete closed-loop control system. The data acquisition and feature extraction module 201, as the system input, collects electrical parameters of network nodes in real time and extracts dynamic features, providing a data foundation for subsequent analysis. The anomaly identification and feature extraction module 202 receives feature data, identifies abnormal events through intelligent algorithms, and extracts key features. The anomaly path location and power optimization module 203 combines network topology information to locate anomalies and optimize parameters, generating an optimal power allocation scheme. The dynamic adjustment and energy consumption assessment module 204 executes the scheme and evaluates its effectiveness, feeding back energy consumption trend data to the adaptive acquisition and control module 205. Ultimately, the adaptive acquisition and control module 205 adjusts the acquisition parameters and completes data transmission, while simultaneously achieving synchronous transmission of electrical energy and data through POE-POF technology. The collaborative work of each module forms a complete intelligent optimization closed loop, effectively solving the technical problems of slow monitoring response and low energy efficiency in traditional power supply networks.

[0065] This application also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the POE-POF power transmission method.

[0066] This application also provides an electronic device, see embodiments thereof. Figure 3 The diagram shown is a structural schematic of an electronic device 300 provided in an embodiment of this application. It includes a processor 301, a readable storage medium 302, a network interface 304, and a display screen 305 connected via a bus 303. The readable storage medium 302 can be either a non-volatile or volatile readable storage medium. It stores instructions and collected data. When the instructions are executed on a computer, the computer performs the steps of the PoE-POF power transmission method, which includes the following five units: a core program storage unit 3021, a runtime data recording unit 3022, a parameter configuration unit 3023, a system management unit 3024, and a security protection unit 3025. The core program storage unit 3021 stores the power dynamic adjustment algorithm, anomaly diagnosis program, and photoelectric collaborative control program, responsible for the system's core operating logic and intelligent decision-making functions. The operating data recording unit 3022 collects and stores power supply parameters, anomaly event records, and energy consumption analysis data in real time, providing data support for system optimization. The parameter configuration unit 3023 saves equipment operating parameters and adaptive adjustment thresholds to ensure stable system operation under different working conditions. The system management unit 3024 uses an embedded architecture to achieve collaborative management of programs and data, supporting remote maintenance and program upgrades. The security protection unit 3025 uses data encryption and redundant backup technologies to ensure the security of stored data and system reliability. Furthermore, the network interface 304 is responsible for communicating with external networks to achieve data transmission and reception. It can be a wired interface, such as an Ethernet interface, or a wireless interface, such as a WIFI or Bluetooth interface, so that the device 300 can exchange data with external devices or servers through the network, enabling remote monitoring and management. The display screen 305 displays the operating status and monitoring data of the computer device 300 for user viewing and operation.

[0067] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0068] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0069] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A POE-POF power transmission method, characterized in that, include: S101. Obtain the real-time power data set of the power supply network from the POE-POF power transmission network node, and extract dynamic power features based on the real-time power data set to obtain a dynamic power feature set; S102. Based on the dynamic power feature set, identify abnormal power loss to obtain an abnormal power loss event set, and extract abnormal mode features based on the abnormal power loss event set to obtain an abnormal mode feature set. S103. Determine the abnormal path node set based on the abnormal mode feature set and the power transmission network topology, optimize the voltage and current allocation parameters based on the abnormal path node set, and generate an optimized power allocation scheme. S104. Adjust the voltage and current values ​​of the power supply path according to the optimized power allocation scheme to obtain a real-time power allocation parameter set, and evaluate the energy consumption trend based on the real-time power allocation parameter set to generate an energy consumption trend dataset. S105. Adjust the power consumption acquisition parameters according to the energy consumption trend dataset, generate an adaptive power consumption acquisition parameter set, update the power consumption acquisition configuration based on the adaptive power consumption acquisition parameter set, and obtain a new power consumption dataset.

2. The method as described in claim 1, characterized in that, Obtain real-time power data sets from nodes in the power supply network, including: By deploying high-precision power acquisition sensors at various nodes of the power supply network, an initial power dataset containing voltage, current and power values ​​is collected. The initial power dataset is preprocessed to obtain the real-time power dataset.

3. The method as described in claim 1, characterized in that, The step of extracting dynamic power features based on the real-time power dataset to obtain a dynamic power feature set includes: The real-time power dataset is processed using a time-series analysis algorithm to extract the power fluctuation frequency and power fluctuation amplitude. The dynamic power feature set is generated based on the power fluctuation frequency and power fluctuation amplitude.

4. The method as described in claim 1, characterized in that, The process of identifying abnormal power loss based on the dynamic power feature set to obtain an abnormal power loss event set includes: Determine whether the power fluctuation amplitude in the dynamic power feature set exceeds a preset threshold. If so, use an anomaly detection algorithm to identify the abnormal power loss and generate the abnormal power loss event set.

5. The method as described in claim 1, characterized in that, The step of extracting abnormal pattern features based on the abnormal power loss event set to obtain an abnormal pattern feature set includes: Extract microsecond-level abnormal patterns from the set of abnormal power loss events; The time-domain and frequency-domain characteristics of the microsecond-level anomaly pattern are analyzed using the Fast Fourier Transform algorithm to generate a feature set of the anomaly pattern that includes time-series anomaly correlation features.

6. The method as described in claim 1, characterized in that, The step of determining the abnormal path node set based on the abnormal pattern feature set and the power transmission network topology includes: Combining the abnormal pattern feature set and the topology of the power transmission network, a graph analysis algorithm is used to calculate the path weight and abnormal propagation path; Based on the path weight and the abnormal propagation path, the specific path node of the abnormal power loss is located, and the abnormal path node set is generated.

7. The method as described in claim 1, characterized in that, The step of optimizing voltage and current allocation parameters based on the abnormal path node set to generate an optimized power allocation scheme includes: For the set of abnormal path nodes, the least squares method is used to optimize the dynamic voltage adjustment and current distribution parameters; By combining path priority ranking and node dependency analysis, the power supply priority of each node is adjusted to generate the optimized power allocation scheme.

8. The method as described in claim 1, characterized in that, The step of adjusting the voltage and current values ​​of the power supply path according to the optimized power allocation scheme to obtain a real-time power allocation parameter set includes: An intelligent power regulation algorithm is adopted to dynamically adjust the voltage and current values ​​of the power supply path through real-time feedback control and adaptive frequency adjustment. The real-time power allocation parameter set is generated based on the adjustment results to suppress power fluctuation amplitude.

9. A POE-POF power transmission system for implementing the method as described in any one of claims 1 to 8, characterized in that, include: The data acquisition and feature extraction module is used to acquire voltage, current and power data of POE-POF power transmission network nodes in real time, extract dynamic power features and generate dynamic power feature sets; The anomaly identification and feature extraction module is used to identify abnormal power loss events and extract abnormal pattern features based on the dynamic power feature set using machine learning or threshold analysis algorithms. The abnormal path location and power optimization module is used to locate abnormal path nodes, optimize voltage and current distribution parameters, and generate the optimal power distribution scheme by combining the power transmission network topology and abnormal mode characteristics. The dynamic adjustment and energy consumption assessment module is used to dynamically adjust the voltage and current values ​​of the power supply path according to the optimization scheme, and to assess the energy consumption trend in real time and generate an energy consumption trend dataset. The adaptive acquisition and control module is used to adaptively adjust the power acquisition parameters according to the energy consumption trend data, and at the same time, it adopts POE-POF power transmission technology to achieve efficient synchronous transmission of power and data.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instruction is executed by the processor, it implements a POE-POF power transmission method as described in any one of claims 1 to 7.

Images