POE power supply power control method and device, equipment and storage medium

Abstract

The invention relates to the technical field of Ethernet power supply, and discloses a POE power supply power control method, device and equipment and a storage medium. The method comprises the steps of obtaining port electrical parameter data, device temperature data and historical power supply operation data of a target POE power supply, and performing multi-dimensional calibration compensation, multi-window time-frequency decomposition, dynamic and static feature extraction and feature hierarchical mapping on the port electrical parameter data based on the device temperature data; performing classification boundary construction and load feature mapping conversion on the mapping result to obtain a load classification mapping table, and performing progressive prediction of multi-level power supply parameters and power demand integration based on historical power supply operation data to obtain a power demand prediction result; and performing multi-constraint optimization calculation on the power demand prediction result, generating a power distribution strategy, and based on the power distribution strategy, controlling the POE power supply to perform multi-power closed-loop control, and generating a power control result. According to the invention, efficient and accurate control of the POE power supply power is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of Power over Ethernet (PoE), and particularly to a method, device, equipment and storage medium for power control of a PoE power supply. Background Art

[0002] In the technical field of PoE power supply systems, dynamic power allocation is an important link in product R & D and quality control, and real-time monitoring and dynamic regulation of power supply power are the most critical and challenging stages in power allocation. Accurately managing the power allocation of PoE power supplies is crucial for network devices, smart terminals and Internet of Things (IoT) devices, which directly affects the working stability, performance and overall operation efficiency of the devices. Therefore, an effective power control method to manage these power supply parameters is crucial for ensuring the reliability and security of PoE power supply systems.

[0003] Currently, digital signal processing technology is used to analyze sampled data and establish a power allocation model to evaluate the power supply state, or artificial intelligence technology is tried to be applied to the power prediction process to better evaluate the load demand and development trend. However, these methods still face challenges in integrating multi-dimensional electrical parameters, dealing with non-linear characteristics and adapting to dynamic loads, and these control methods often ignore some unique characteristics of PoE power supply systems, such as cable attenuation, temperature drift, power supply distance, etc., and these factors may have a significant impact on power allocation and power supply quality. That is, the existing power control methods of PoE power supplies ignore multi-dimensional power supply parameters and some unique characteristics of PoE systems, resulting in a poor final power allocation effect. Summary of the Invention

[0004] The main purpose of the present invention is to solve the problem that the existing power control methods of PoE power supplies ignore multi-dimensional power supply parameters and some unique characteristics of PoE systems, resulting in a poor final power allocation effect.

[0005] The first aspect of the present invention provides a power control method for a POE power supply. The power control method for the POE power supply includes: obtaining port electrical parameter data, device temperature data, and historical power supply operation data of a target POE power supply, and based on the device temperature data, performing multi-dimensional calibration compensation on the port electrical parameter data to obtain compensated port electrical parameter data; performing time-frequency decomposition with multiple windows on the compensated port electrical parameter data to obtain port electrical decomposition data, and extracting dynamic and static features and performing feature hierarchical mapping on the port electrical decomposition data to construct a port state feature vector; constructing a classification boundary for the port state feature vector and performing mapping conversion of load features to obtain a load classification mapping table, and based on the historical power supply operation data, performing progressive prediction of multi-level power supply parameters and integrating power demand on the load classification mapping table to obtain a power demand prediction result; performing multi-constraint optimization calculation on the power demand prediction result to generate a power distribution strategy, and based on the power distribution strategy, controlling the target POE power supply to perform multiple power closed-loop controls to generate the power control result of the target POE power supply.

[0006] Optionally, in the first implementation manner of the first aspect of the present invention, the performing multi-dimensional calibration compensation on the port electrical parameter data based on the device temperature data to obtain compensated port electrical parameter data includes: calculating a temperature gradient of the device temperature data to obtain a power supply temperature field distribution, and performing window smoothing calculation and statistical calculation of parameter values on the port electrical parameter data to obtain statistically processed port electrical parameter data, and marking and deleting outliers corresponding to the statistically processed port electrical parameter data to obtain effective sampling sequence data; fitting a segmented cable attenuation coefficient to the effective sampling sequence data based on the cable length corresponding to the target POE power supply to obtain slope coefficients for each segment, and performing interpolation fitting on the slope coefficients for each segment to obtain a cable compensation coefficient table; calculating a sensitivity of corresponding electrical parameters to the power supply temperature field distribution to obtain a temperature influence coefficient, and performing weighted operation of preset key node temperatures on the temperature influence coefficient to obtain a temperature compensation matrix; performing zero-point offset calculation and gain error compensation on the port electrical parameter data to obtain primary calibration data, and based on the cable compensation coefficient table and the temperature compensation matrix, performing secondary compensation calibration on the primary calibration data to obtain compensated port electrical parameter data.

[0007] Optionally, in the second implementation manner of the first aspect of the present invention, the time-frequency decomposition of the compensated port electrical parameter data is performed in multiple windows to obtain port electrical decomposition data, and the extraction of dynamic and static features and the hierarchical mapping of features are performed on the port electrical decomposition data to construct a port state feature vector, including: performing time-series segmentation on the compensated port electrical parameter data based on a preset port sampling frequency to obtain a time-domain sampling sequence, and performing time-frequency transformation on the time-domain sampling sequence to obtain port electrical decomposition data; calculating the voltage fluctuation degree of the power supply voltage data in the port electrical decomposition data to obtain voltage characteristic parameters, and calculating the change rate and performing current dynamic response spectrum decomposition on the power supply current data in the port electrical decomposition data to obtain current characteristic parameters; performing phase detection of the synchronous sampling zero point on the voltage characteristic parameters and the current characteristic parameters to obtain a phase difference angle, and calculating the power value of the voltage characteristic parameters and the current characteristic parameters based on the phase difference angle to obtain power characteristic parameters; performing hierarchical association mapping using the voltage characteristic parameters, the current characteristic parameters, and the power characteristic parameters to obtain an initial feature structure, and performing cross-correlation calculation and weighted combination on the initial feature structure to obtain a standard power supply feature set; performing descending order arrangement and spatial mapping on the standard power supply feature set to obtain a low-dimensional feature vector, and reconstructing the feature combination of the low-dimensional feature vector to obtain a port state feature vector.

[0008] Optionally, in the third implementation manner of the first aspect of the present invention, the construction of a classification boundary and the mapping transformation of load characteristics are performed on the port state feature vector to obtain a load classification mapping table, including: clustering each feature vector in the port state feature vector to obtain a state feature distance matrix, and constructing a classification boundary for the state feature distance matrix to obtain an initial feature classification space; performing level division on the output power corresponding to the initial feature classification space based on a preset POE power supply power standard to obtain an initial power interval, and setting a boundary threshold for the initial power interval to obtain a load classification space; performing spatial combination and conditional division of spatial feature attributes on the initial feature classification space and the load classification space to obtain an initial decision space, and performing hierarchical decomposition and recursive update of the spatial node weight coefficient on the initial decision space to obtain a classification decision tree; performing adjacent sampling point difference operation and normalization on the leaf node data in the classification decision tree to obtain a load state change rate, and predicting the change trend and evaluating the confidence level of the load state change rate to obtain load dynamic characteristics; performing identification of load types and mapping update of feature associations on the load dynamic characteristics to generate a load classification mapping table.

[0009] Optionally, in the fourth implementation manner of the first aspect of the present invention, the progressive prediction of multi-level power supply parameters and the integration of power demand are performed on the load classification mapping table based on the historical power operation data to obtain a power demand prediction result, including: decomposing the historical power operation data into periodic power supply powers to obtain a historical power supply power sequence, and based on the historical power supply power sequence, performing an associated matching of load power consumption patterns and extracting pattern time series characteristics on the load classification mapping table to obtain a power supply power change pattern; performing time scale division and sampling recombination on the power supply power change pattern to obtain multi-scale power change data, and performing scale association and state space modeling on the multi-scale power change data to obtain a progressive prediction model; calculating power losses for the equivalent impedance value of the POE cable corresponding to the target POE power supply to obtain a loss distribution function, and performing transmission efficiency evaluation and temperature compensation calculation on the loss distribution function to obtain a power correction coefficient; performing a short-term moving average operation on the power correction coefficient based on the power characteristic parameters corresponding to a preset short-term time interval to obtain a short-term load prediction value, and performing pattern mapping and similarity distance calculation on the power characteristic parameters corresponding to a preset medium-term time interval and the short-term load prediction value based on the pattern similarity index in the historical power operation data to obtain a pattern matching sequence; performing weighted fusion and reconstruction extraction of trend characteristics on the pattern matching sequence to obtain a medium-term load prediction value, and performing data integration and filtering smoothing on the medium-term load prediction value based on a preset long-term time interval to obtain a long-term power trend; performing seasonal decomposition and trend extrapolation on the long-term power trend to obtain a long-term load prediction value, and performing weighted fusion and abnormal load value elimination on the short-term load prediction value, the medium-term load prediction value, and the long-term load prediction value to obtain a load prediction correction result; calculating a confidence interval and quantitatively evaluating the prediction accuracy for the load prediction correction result to obtain a power demand prediction result.

[0010] Optionally, in the fifth implementation manner of the first aspect of the present invention, the multi-constraint optimization calculation of the power demand prediction result to generate a power distribution strategy includes: based on the power threshold of the target POE power supply, matching the power demand prediction result to obtain the allocable power amount, and based on the power supply priority corresponding to each port in the target POE power supply, performing weighted operation and balancing calculation of the port power quota on the allocable power amount to obtain an initial power distribution scheme; obtaining the historical cable temperature corresponding to the power supply power in the initial power distribution scheme, and performing spatial distribution calculation based on the historical cable temperature and constructing a corresponding cable temperature rise curve; performing limit constraint and margin calculation of the safety threshold on the cable temperature rise curve to obtain temperature constraint parameters, and based on the temperature constraint parameters, performing temperature coefficient correction and loss evaluation on the cable compensation coefficient table to obtain efficiency constraint parameters; performing multi-objective adjustment solution and search and evaluation of the feasible solution space on the temperature constraint parameters and the efficiency constraint parameters to obtain an optimized distribution scheme, and performing margin risk assessment and power distribution adjustment on the optimized distribution scheme to generate a power distribution strategy.

[0011] Optionally, in the sixth implementation manner of the first aspect of the present invention, the controlling the target POE power supply to perform multiple power closed-loop controls based on the power distribution strategy to generate the power control result of the target POE power supply includes: extracting power control parameters and dividing the interval of the supply voltage for the power distribution strategy to obtain reference control target parameters, and performing voltage-current decomposition and mapping of the DC power on the reference control target parameters to obtain hierarchical control parameters; performing timing planning on the hierarchical control parameters to generate a control execution plan, and performing gain calculation and loop compensation of the voltage loop and the current loop on the control execution plan to obtain a basic power supply control amount; based on the power demand prediction result, performing power feedback compensation on the basic power supply control amount to obtain a corrected power supply control amount, and performing output amount superposition and limit constraint on the basic power supply control amount and the corrected control amount to obtain a protection power supply control amount corresponding to the power supply of the target POE power supply; determining and modulating the output priority of the basic power supply control amount, the corrected power supply control amount, and the protection power supply control amount to generate the power control result of the target POE power supply.

[0012] In a second aspect of the present invention, there is provided a power control device for a POE power supply. The power control device for the POE power supply includes: a calibration compensation module, configured to obtain port electrical parameter data, device temperature data, and historical power supply operation data of a target POE power supply, and perform multi-dimensional calibration compensation on the port electrical parameter data based on the device temperature data to obtain compensated port electrical parameter data; a feature mapping module, configured to perform time-frequency decomposition with multiple windows on the compensated port electrical parameter data to obtain port electrical decomposition data, and extract dynamic and static features and perform hierarchical mapping of the features on the port electrical decomposition data to construct a port status feature vector; a demand prediction module, configured to construct a classification boundary and perform mapping conversion of load features on the port status feature vector to obtain a load classification mapping table, and perform progressive prediction of multi-level power supply parameters and integration of power demand on the load classification mapping table based on the historical power supply operation data to obtain a power demand prediction result; a closed-loop control module, configured to perform multi-constraint optimization calculation on the power demand prediction result to generate a power distribution strategy, and control the target POE power supply to perform multiple power closed-loop controls based on the power distribution strategy to generate a power control result of the target POE power supply.

[0013] In a third aspect of the present invention, there is provided a power control device for a POE power supply, including: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the power control device for the POE power supply executes each step of the above-mentioned power control method for the POE power supply.

[0014] In a fourth aspect of the present invention, there is provided a computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, and when the instructions are run on a computer, the computer is made to execute each step of the above-mentioned power control method for the POE power supply.

[0015] The power control method, device, equipment and storage medium of the above POE power supply. In the embodiments of the present invention, by obtaining the port electrical parameter data, device temperature data and historical power supply operation data of the target POE power supply, and based on the device temperature data, multi-dimensional calibration compensation is performed on the port electrical parameter data to obtain the compensated port electrical parameter data; time-frequency decomposition of multiple windows is performed on the compensated port electrical parameter data to obtain port electrical decomposition data, and dynamic and static features are extracted from the port electrical decomposition data and feature hierarchical mapping is performed to construct a port state feature vector; a classification boundary is constructed for the port state feature vector and load feature mapping transformation is performed to obtain a load classification mapping table, and based on the historical power supply operation data, progressive prediction of multi-level power supply parameters and power demand integration are performed on the load classification mapping table to obtain a power demand prediction result; multi-constraint optimization calculation is performed on the power demand prediction result to generate a power distribution strategy, and based on the power distribution strategy, the target POE power supply is controlled to perform multiple power closed-loop controls to generate the power control result of the target POE power supply. Compared with the prior art, in this application, multi-dimensional acquisition and feature extraction are performed on the electrical parameters and temperature data of the target POE power supply, temperature compensation and calibration are combined with historical operation data to obtain high-precision port electrical characteristics, and then time-frequency decomposition and dynamic and static feature mapping are performed on these characteristics to obtain a port state feature vector; then a load classification mapping table is formed through classification boundary construction and load feature conversion, and progressive prediction of multi-level power supply parameters is performed based on historical data to achieve accurate prediction of power demand; finally, multi-constraint optimization and closed-loop control are performed based on the prediction result to output the power control result. Through hierarchical data processing and feature analysis, the problem of accurate distribution of POE power supply power control is solved, especially in aspects such as cable attenuation, temperature drift and dynamic load, fully considering the characteristics and power supply constraints of the POE power supply, effectively improving the control accuracy; and adopting a multi-level feature mapping and progressive prediction strategy, not only realizing the correlation analysis between electrical parameters, but also enhancing the reliability of power distribution; in addition, through multi-constraint optimization and closed-loop control, accurate power distribution and regulation are achieved, thus overall realizing the efficient and accurate control of the POE power supply power.

[0016] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention are achieved and obtained by the structures specifically pointed out in the specification, claims and drawings.

[0017] To make the above objectives, features and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given in conjunction with the accompanying drawings and are described in detail as follows. Description of the Drawings

[0018] Figure 1Schematic diagram of the first embodiment of the power control method for the POE power supply in the embodiments of the present invention; Figure 2 Schematic diagram of an embodiment of the power control device for the POE power supply in the embodiments of the present invention; Figure 3 Schematic diagram of an embodiment of the power control device for the POE power supply in the embodiments of the present invention. Specific embodiments

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes other steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products, or devices.

[0021] For ease of understanding of this embodiment, the specific process of the embodiments of the present invention will be described below. Please refer to Figure 1 , the first embodiment of the power control method for the POE power supply in the embodiments of the present invention includes: 101. Obtain the port electrical parameter data, device temperature data, and historical power supply operation data of the target POE power supply, and based on the device temperature data, perform multi-dimensional calibration compensation on the port electrical parameter data to obtain the compensated port electrical parameter data; Embodiments of the present application can obtain and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0022] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0023] In this embodiment, the above-mentioned target POE power supply refers to an Ethernet power supply device that needs power control, usually a POE switch or a POE power supply. It can provide both data transmission and DC power supply for network terminal devices (such as IP cameras, wireless APs, etc.) through Ethernet cables; the port electrical parameter data includes: port output voltage (usually 44V - 57V), port output current (up to 350 - 600mA at most), port output power, port impedance value, port ripple coefficient, etc., and these parameters can be obtained by real-time sampling through a high-precision sampling circuit; the device temperature data includes: the temperature of the POE power supply chip, the temperature of the power transformer, the temperature of the power output port, and the temperature of key power devices, etc., and these temperature data are collected by temperature sensors for temperature compensation and over-temperature protection; the historical power operation data includes: the records of port voltage, current, and power changes, historical load type records, historical temperature change records, historical power operation data, historical cable temperature, historical fault records, and historical power distribution schemes, etc. within a certain period of time in the past. These data are stored in the device's memory for power prediction and optimization; perform temperature gradient calculation on the device temperature data to obtain the power supply temperature field distribution, and perform window smoothing calculation and statistical calculation of parameter values on the port electrical parameter data to obtain the statistically processed port electrical parameter data, and mark and delete the outliers corresponding to the statistically processed port electrical parameter data to obtain the effective sampling sequence data; based on the cable length corresponding to the target POE power supply, fit the attenuation coefficient of each segment of the cable for the effective sampling sequence data to obtain the slope coefficient of each segment, and perform interpolation fitting on the slope coefficient of each segment to obtain the cable compensation coefficient table; perform sensitivity calculation of the corresponding electrical parameters on the power supply temperature field distribution to obtain the temperature influence coefficient, and perform weighted operation of the preset key node temperature on the temperature influence coefficient to obtain the temperature compensation matrix; perform zero-point offset calculation and gain error compensation on the port electrical parameter data to obtain the initial calibration data, and based on the cable compensation coefficient table and the temperature compensation matrix, perform secondary compensation calibration on the initial calibration data to obtain the compensated port electrical parameter data.

[0024] In practical applications, a high-precision sampling circuit is used to sample the voltage (within the voltage range corresponding to IEEE Std802.3af (POE standard), IEEE Std 802.3at (POE+ standard), and IEEE Std 802.3bt (POE++ standard)) and current (in the range of 350 - 600 mA) of the POE power supply port in real time. At the same time, temperature data of key components such as the POE power supply chip, transformer, and output port are collected, and operation record data including historical voltage, current, power changes, etc. are read from the memory. Then, the device temperature data is processed. By calculating the temperature gradient between adjacent temperature measurement points, a temperature field distribution model inside the power supply, i.e., the power supply temperature field distribution, is constructed to obtain the heat transfer and distribution inside the power supply. At the same time, the electrical parameter data such as the port voltage and current collected are smoothed using a 10 - ms sliding window to eliminate sampling noise and transient interference. Then, statistical features such as the mean and standard deviation of the data within the window are calculated, and by comparing with the preset parameter thresholds, abnormal data points such as voltage mutations and current surges are identified and marked, and these abnormal values are removed from the data sequence, retaining the valid sampling sequence data. Furthermore, considering the influence of the cable length in POE power supply on the power supply quality, according to the actual length of the network cable used for the POE power supply, the valid sampling sequence is segmented by distance, and linear fitting is performed on each segment of data to calculate the slope coefficients of the voltage and current attenuation with distance. By interpolating these segmented slope coefficients, a complete cable compensation coefficient comparison table is finally obtained. Furthermore, the influence of the power supply temperature field distribution on electrical parameters is analyzed. By calculating the sensitivity of parameters such as voltage and current to temperature changes, a coefficient representing the degree of temperature influence is obtained, and for the temperatures of key nodes such as the power supply chip and transformer, weighted calculations are performed according to their importance to form a complete temperature compensation matrix for correcting parameter deviations caused by temperature drift. Then, basic calibration is performed on the original electrical parameter data, including eliminating the zero - point offset and gain error of the sampling circuit to obtain preliminarily calibrated data, and combining the cable compensation coefficient table and the temperature compensation matrix to perform secondary compensation on the initially calibrated data. This process not only compensates for cable losses but also corrects the influence brought about by temperature changes, finally obtaining accurate and reliable port electrical parameter data, providing a high - quality data basis for subsequent power control, ensuring the accuracy of electrical parameters, and laying a foundation for the precise power control of the POE power supply. For example, for a 48V / 400W POE power supply powered through a 100 - meter CAT5e network cable at 25°C, when the port voltage is 48.5V, the actual output voltage after compensation and calibration is 47.8V, which is closer to the actual value compared to 49.2V without compensation.

[0025] 102. Perform time-frequency decomposition on the compensated port electrical parameter data in multiple windows to obtain port electrical decomposition data, extract dynamic and static features from the port electrical decomposition data, and perform hierarchical mapping of the features to construct a port state feature vector. In this embodiment, based on the preset port sampling frequency, the compensated port electrical parameter data is segmented in time series to obtain a time-domain sampling sequence, and the time-domain sampling sequence is subjected to time-frequency transformation to obtain port electrical decomposition data; the voltage fluctuation degree of the power supply voltage data in the port electrical decomposition data is calculated to obtain a voltage characteristic parameter, and the change rate of the power supply current data in the port electrical decomposition data is calculated and the current dynamic response spectrum is decomposed to obtain a current characteristic parameter; the phase detection of the synchronous sampling zero point of the voltage characteristic parameter and the current characteristic parameter is performed to obtain a phase difference angle, and based on the phase difference angle, the power value of the voltage characteristic parameter and the current characteristic parameter is calculated to obtain a power characteristic parameter; the voltage characteristic parameter, the current characteristic parameter, and the power characteristic parameter are used for hierarchical correlation mapping to obtain an initial feature structure, and the initial feature structure is subjected to cross-correlation calculation and weighted combination to obtain a standard power supply feature set; the standard power supply feature set is sorted in descending order and spatially mapped to obtain a low-dimensional feature vector, and the low-dimensional feature vector is reconstructed by feature combination to obtain a port state feature vector.

[0026] In practical applications, according to the preset port sampling frequency (such as a sampling frequency of 50 kHz), the compensated port electrical parameter data is processed by time series segmentation. Each time window contains 10 ms of sampling data to form a continuous time-domain sampling sequence. Then, the fast Fourier transform (FFT) is applied to these time-domain sequence data for time-frequency domain analysis, decomposing the time-domain signal into different frequency components to obtain the port electrical decomposition data containing fundamental wave and harmonic information. This decomposition can effectively reflect the frequency-domain characteristics of voltage and current signals. Furthermore, feature extraction is performed on voltage and current data respectively. For voltage data, voltage fluctuation index is calculated, including parameters such as maximum fluctuation amplitude, fluctuation period, and fluctuation frequency. These parameters can reflect the stability and quality characteristics of voltage. At the same time, current data is processed, the current change rate (di / dt) is calculated, and the dynamic response characteristics of current are obtained through spectrum analysis, including parameters such as transient response speed, harmonic content, and current distortion rate. These characteristics can reflect the dynamic characteristics of the load and the power consumption behavior. Then, phase analysis of synchronous sampling points of voltage and current signals is carried out. The phase difference angle between voltage and current is calculated through zero-crossing detection technology. Based on this phase difference, combined with voltage amplitude and current amplitude, accurate power characteristic parameters such as active power, reactive power, and apparent power are calculated. These parameters comprehensively reflect the power supply state of the POE port. Furthermore, a multi-level feature correlation structure is established using voltage characteristic parameters, current characteristic parameters, and power characteristic parameters, that is, these parameters are hierarchically mapped according to physical meaning and numerical characteristics to construct an initial feature structure framework. Then, the cross-correlation coefficients between different feature parameters are calculated to evaluate the degree of mutual influence between parameters, and weighted combination is performed according to the importance of parameters to form a standardized power supply feature set (this feature set contains comprehensive state information of the power supply port). Then, the parameters in the standard power supply feature set are sorted in descending order of importance, and through dimensionality reduction techniques such as principal component analysis (PCA), the high-dimensional features are mapped to a low-dimensional feature space to obtain a more compact low-dimensional feature vector. These low-dimensional features are reorganized and structured, and the features with strong correlation are combined together to form the final port state feature vector, which can not only comprehensively characterize the power supply state of the POE port but also provide an efficient feature representation form.

[0027] 103. Construct the classification boundary of the port state feature vector and perform mapping transformation of load characteristics to obtain the load classification mapping table. Based on historical power supply operation data, progressive prediction of multi-level power supply parameters and integration of power demand are performed on the load classification mapping table to obtain the power demand prediction result. In this embodiment, each feature vector in the port status feature vector is clustered to obtain a status feature distance matrix, and a classification boundary of the status feature distance matrix is constructed to obtain an initial feature classification space; based on a preset POE power standard, the output power corresponding to the initial feature classification space is classified to obtain an initial power interval, and boundary thresholds of the initial power interval are set to obtain a load classification space; the initial feature classification space and the load classification space are combined in space and the conditional division of the spatial feature attributes is performed to obtain an initial decision space, and the initial decision space is hierarchically decomposed and the weight coefficients of the spatial nodes are recursively updated to obtain a classification decision tree; the leaf node data in the classification decision tree are subjected to differential operation and normalization of adjacent sampling points to obtain a load status change rate, and the change trend of the load status change rate is predicted and the confidence level is evaluated to obtain a load dynamic feature; the load type is identified from the load dynamic feature and the mapping update of the feature association is performed to generate a load classification mapping table; the historical power supply operation data are decomposed into periodic power supply powers to obtain a historical power supply power sequence, and based on the historical power supply power sequence, the load classification mapping table is subjected to association matching of the load power consumption pattern and the extraction of the pattern time series feature to obtain a power supply power change pattern; the power supply power change pattern is divided in time scale and sampled and recombined to obtain multi-scale power change data, and the multi-scale power change data are subjected to scale association and state space modeling to obtain a progressive prediction model; the power loss of the equivalent impedance value of the POE cable corresponding to the target POE power supply is calculated to obtain a loss distribution function, and the transmission efficiency of the loss distribution function is evaluated and the temperature compensation calculation is performed to obtain a power correction coefficient; based on the power feature parameters corresponding to a preset short-term time interval, a short-term sliding average operation is performed on the power correction coefficient to obtain a short-term load prediction value, and based on the pattern similarity index in the historical power supply operation data, the power feature parameters corresponding to a preset medium-term time interval and the short-term load prediction value are subjected to pattern mapping and similarity distance calculation to obtain a pattern matching sequence; the pattern matching sequence is weighted and fused and the trend feature is reconstructed and extracted to obtain a medium-term load prediction value, and based on a preset long-term time interval, data integration and filtering and smoothing are performed on the medium-term load prediction value to obtain a long-term power trend; the long-term power trend is seasonally decomposed and trend extrapolation is performed to obtain a long-term load prediction value, and the short-term load prediction value, the medium-term load prediction value, and the long-term load prediction value are weighted and fused and abnormal load values are removed to obtain a load prediction correction result; a confidence interval calculation and a quantitative evaluation of the prediction accuracy are performed on the load prediction correction result to obtain a power demand prediction result.

[0028] In practical applications, the K-means clustering algorithm is first used to cluster the feature vectors. That is, by calculating the Euclidean distance between different feature vectors, a state feature distance matrix is constructed. This matrix reflects the similarity between different operating states. Based on this distance matrix, the density clustering algorithm is used to determine the class boundaries, and the feature vectors with high similarity are divided into the same class, thus constructing an initial feature classification space (where this space reflects the distribution characteristics of different operating states of the POE port); furthermore, in combination with the power levels specified in POE power supply standards such as IEEE 802.3af / at / bt (such as 15.4W, 30W, 60W, etc.), the power levels of each category in the initial feature classification space are divided to establish an initial power interval. The formula for the level division is: ; Where: is the membership function of the i-th power level, x is the input power value, is the central value of the i-th level, is the fuzziness parameter (for example: for an input power of 15.4W, the membership degrees to Class 3 and Class 4 can be calculated), and by setting the boundary thresholds of each power interval, such as the voltage change range, current limit, and power fluctuation range, etc., a complete load classification space is formed (this space establishes a corresponding relationship between the port state and the actual power supply capacity); furthermore, the initial feature classification space and the load classification space are combined, and conditional division is performed according to attributes such as voltage stability, current dynamic response, and power factor to construct an initial decision space. Then, hierarchical decomposition is performed on this decision space to establish a decision tree structure. Each node of the tree represents a classification judgment condition, and by analyzing the classification accuracy of different conditions in the historical operation data, the weight coefficients of each node are dynamically updated to optimize the classification performance of the decision tree, thus obtaining a classification decision tree; furthermore, the data of the leaf nodes are processed, the difference in load states at adjacent time points is calculated and normalized to obtain a standardized load state change rate. Based on this change rate, the time series analysis method is used to predict the load change trend, and by calculating the deviation between the prediction result and the actual value, the confidence level of the prediction is evaluated, thus obtaining accurate load dynamic characteristics; furthermore, the load dynamic characteristics are matched with the predefined load type templates to identify the specific type of the current load, such as continuous load, periodic load, or burst load, etc. At the same time, the association relationship between the features and the load types is continuously updated according to the new operation data to generate a dynamically updated load classification mapping table (this mapping table not only contains the recognition rules of the load types, but also records the typical feature parameters corresponding to different load types), thus realizing the accurate identification of the load types of the POE ports and having strong dynamic adaptability, and can optimize the classification effect as the power supply operates continuously.

[0029] Secondly, process the stored historical power operation data, extract the periodic components of the power supply through Fourier transform, separate the fundamental wave and each harmonic, obtain a series of historical power supply power sequences in chronological order, and match these power sequences with the load types in the load classification mapping table to analyze the power consumption characteristics of different types of loads, including power change rules, peak-to-peak values, and fluctuation periods, etc., so as to extract typical power supply power change patterns; then divide the power supply power change patterns according to the time scale, establish power change data sets at the hourly, daily, and monthly levels respectively, and through resampling and combination of these data at different time scales, form multi-scale power change data, and then use the state space method to model the multi-scale data to construct a progressive prediction model including a state equation and an observation equation (wherein, this model can describe the dynamic characteristics and observation characteristics of power change); then considering the cable loss problem in POE power supply, according to the equivalent impedance model of the POE cable, calculate the cable loss distribution under different power supply powers, that is, by analyzing the relationship between the loss and the transmission distance and current, establish a loss distribution function, and combine the influence of the cable temperature on the transmission efficiency to calculate the power correction coefficient applicable to different working conditions; then in the prediction process, first for the short-term prediction of the next 1 hour, use a 10ms sliding window to calculate the average of the power correction coefficient, and combine the current load status to obtain the short-term load prediction value. For the medium-term prediction of the next 24 hours, extract similar load patterns from the historical data, calculate the Euclidean distance between the current short-term prediction value and the historical pattern, and select several of the most similar pattern sequences for matching; then through weighted fusion of the matched pattern sequences and extraction of the trend characteristics therein, generate the medium-term load prediction value. In addition, for the long-term prediction of the next month, perform cumulative integration and low-pass filtering on the medium-term prediction value to obtain a smooth long-term power change trend, and then use the time series decomposition method to decompose the long-term trend into seasonal components and trend components, and perform trend extrapolation through the ARIMA model to obtain the long-term load prediction value (for example: in the POE power supply of an office building, the short-term prediction can predict the load change in the next 1 hour through the LSTM model, with an accuracy rate of 95%; the medium-term prediction predicts the power consumption trend in the next week through the CRF model, and the error is controlled within ±10%; the long-term prediction predicts the seasonal change through the ARIMA model, and the power supply capacity can be planned in advance).Through this multi-scale prediction method, it is possible to adapt to load changes at different time scales and provide accurate power demand predictions. Furthermore, the short-term, medium-term, and long-term prediction values are comprehensively processed. First, by setting weight coefficients for different time scales, the three prediction values are weighted and fused. At the same time, the 3σ criterion is used to identify and eliminate abnormal prediction values. Then, based on the historical prediction error distribution, the confidence interval of the prediction result is calculated, and the prediction accuracy is evaluated through indicators such as the root mean square error (RMSE) and the mean absolute percentage error (MAPE). Finally, a power demand prediction result containing the prediction value and credibility is output. It realizes taking into account the temporal characteristics of load changes while also integrating physical constraints such as cable losses and temperature effects, and can accurately predict power demands at different time scales. By fusing and correcting multiple prediction results, the accuracy and reliability of the prediction can be effectively balanced, providing a reliable decision-making basis for the power distribution of the POE power supply. At the same time, by continuously accumulating prediction experience and updating model parameters, the prediction performance is continuously optimized to ensure the long-term effectiveness of power prediction.

[0030] 104. Perform multi-constraint optimization calculations on the power demand prediction results to generate a power distribution strategy. Based on the power distribution strategy, control the target POE power supply for multiple power closed-loop controls to generate the power control result of the target POE power supply.

[0031] In this embodiment, based on the power threshold of the target POE power supply, the power demand prediction result is matched for power demand to obtain the allocable power quantity. Then, based on the power supply priority corresponding to each port in the target POE power supply, weighted operation and equilibrium calculation of port power quotas are performed on the allocable power quantity to obtain the initial power distribution scheme. The historical cable temperature corresponding to the power supply power in the initial power distribution scheme is obtained, and based on the historical cable temperature, spatial distribution calculation is carried out and the corresponding cable temperature rise curve is constructed. Limit constraint and margin calculation of the safety threshold are performed on the cable temperature rise curve to obtain the temperature constraint parameter. Based on the temperature constraint parameter, temperature coefficient correction and loss evaluation are performed on the cable compensation coefficient table to obtain the efficiency constraint parameter. Multi-objective adjustment solution and search evaluation of the feasible solution space are performed on the temperature constraint parameter and the efficiency constraint parameter to obtain the optimized distribution scheme. Then, margin risk assessment and power distribution adjustment are performed on the optimized distribution scheme to generate the power distribution strategy. Power control parameter extraction and supply voltage interval division are performed on the power distribution strategy to obtain the reference control target parameter. Voltage-current decomposition and DC power mapping are performed on the reference control target parameter to obtain the hierarchical control parameter. Timing planning is performed on the hierarchical control parameter to generate the control execution scheme. Gain calculation and loop compensation of the voltage loop and current loop are performed on the control execution scheme to obtain the basic power supply control quantity. Based on the power demand prediction result, power feedback compensation is performed on the basic power supply control quantity to obtain the corrected power supply control quantity. Output quantity superposition and limit constraint are performed on the basic power supply control quantity and the corrected control quantity to obtain the protection power supply control quantity corresponding to the power supply of the target POE power supply. Judgment and modulation of the output priority are performed on the basic power supply control quantity, the corrected power supply control quantity, and the protection power supply control quantity to generate the power control result of the target POE power supply.

[0032] In practical applications, first, according to the maximum power supply threshold specified in the POE power supply standard (such as 15.4W for Type1, 30W for Type2, 60W for Type3, etc.), matching analysis is performed on the predicted power demand to determine the total allocable power of the power supply. A non-linear programming model is used to calculate the allocable power quantity, and the calculation formula of its non-linear programming model is: ; Constraint condition: For each port , the allocated power must be within the allowable range: , the power of each port multiplied by its transmission efficiency must meet the required power: , the sum of the powers allocated to all ports cannot exceed the total available power: , Among them, is the power allocated to the i-th port, is the port priority weight, λ is the balance factor, is the total available power, is the maximum power limit of the port, is the transmission efficiency, is the required power (for example: for an 8-port POE switch with a total power of 480W, at a certain moment, the requirements of 4 high-priority ports are 60W, 55W, 45W, 40W respectively, and the requirements of 4 low-priority ports are 30W, 25W, 20W, 15W). Then, based on the pre-set power supply priorities of each POE port (such as high priority for critical service ports and low priority for ordinary service ports), the available power is weighted and allocated according to priorities, and at the same time, the load balance between ports is considered. Through the linear programming algorithm, the initial power allocation scheme for each port is calculated; furthermore, in order to ensure power supply safety, the cable temperature records corresponding to the initial power allocation scheme are extracted from the historical database, the temperature distribution of the cable at different positions is analyzed through the thermodynamic model, and a temperature rise curve of the power-temperature correspondence is established (where this temperature rise curve reflects the temperature change characteristics of the cable under different power supply powers); furthermore, according to the temperature resistance grade and safety margin requirements of the cable material, a temperature limit is set for the temperature rise curve, such as limiting the maximum working temperature to below 75°C and reserving a 15% safety margin, the temperature constraint parameters are calculated, and based on these temperature constraints, the previously established cable compensation coefficient table is corrected, considering the influence of temperature on the cable impedance, and the transmission loss at different temperatures is evaluated, so as to obtain the constraint parameters representing the power supply efficiency; furthermore, the temperature constraint and the efficiency constraint are solved as a multi-objective optimization problem, and the goal is to achieve the highest transmission efficiency on the premise of ensuring power supply safety, that is, to search for the optimal solution in the constraint space through the particle swarm optimization algorithm, obtain the optimized allocation scheme that meets all constraint conditions, and conduct a risk assessment on the optimized scheme, analyze the power supply working margin under the worst working conditions, and make fine adjustments to the power allocation according to the evaluation results, and finally form a reliable power allocation strategy, which not only ensures the safety of power supply but also realizes the efficient utilization of resources, enabling the POE power supply to operate stably and reliably.

[0033] Secondly, extract the key power control parameters from the power allocation strategy, including the target power supply power, dynamic response speed, stability requirements, etc., and divide the power supply voltage into a detection interval (2.8V - 10V), a grading interval (15.5V - 20.5V), and a normal power supply interval (44V - 57V) according to the POE power supply standard, so as to establish the reference control target parameters, and decompose these parameters according to voltage control and current control respectively. The reference control target parameters are decomposed into voltage and current components through a second-order matrix transformation to establish a power mapping relationship, and its power decomposition formula is: ; Where: and is the nominal voltage and current value, and ΔI(t) is the dynamic adjustment amount, θ(t) is the power factor angle, and t is the time variable (for example: when the POE port needs to output 30W of power, V0 = 48V and I0 = 0.625A can be set). Precise control is achieved through the dynamic adjustment amount, thereby obtaining hierarchical control parameters including voltage control parameters, current limits, and power thresholds; furthermore, based on the hierarchical control parameters, a timing control plan is carried out to generate a control execution plan including stages such as startup detection, power grading, normal power supply, and protection control. And in order to achieve precise closed-loop control, a dual-loop control structure is adopted. By calculating the proportional-integral-derivative (PID) parameters of the voltage loop and the current loop, and introducing loop compensation methods such as feedforward compensation and dead zone control, a basic power supply control amount is obtained (where this control amount can achieve stable regulation of voltage and current and precise control of power); furthermore, in combination with the previously obtained power demand prediction results, dynamic compensation is performed on the basic power supply control amount, that is, by establishing a power feedback channel, the deviation between the actual output power and the predicted power is compared in real time, the compensation control amount is calculated, and it is superimposed on the basic control amount to obtain the corrected power supply control amount. At the same time, limit constraints are set on the control amount, including protection parameters such as the maximum output voltage, maximum output current, and maximum output power, to form a safe protection power supply control amount; furthermore, a priority determination mechanism is established according to the characteristics and functions of different control amounts. Under normal working conditions, the basic power supply control amount is preferentially used for steady-state regulation; when the load changes, it switches to the corrected power supply control amount for dynamic response; and in case of abnormal situations such as overvoltage and overcurrent, the protection power supply control amount is immediately enabled for safety protection, and the selected control amount is converted into a switch drive signal through a PWM modulator, ultimately achieving precise control of the POE power supply and ensuring the stability and reliability of power supply.

[0034] In the embodiments of the present invention, by performing multi-dimensional acquisition and feature extraction on the electrical parameters and temperature data of the target POE power supply, combining historical operation data for temperature compensation and calibration to obtain high-precision port electrical characteristics, then performing time-frequency decomposition and dynamic-static feature mapping on these characteristics to obtain port status feature vectors; then forming a load classification mapping table through classification boundary construction and load feature conversion, and performing progressive prediction of multi-level power supply parameters based on historical data to achieve accurate prediction of power demand; finally, performing multi-constraint optimization and closed-loop control based on the prediction results to output power control results. Through hierarchical data processing and feature analysis, the problem of accurate allocation of POE power supply power control is solved, especially in aspects such as cable attenuation, temperature drift, and dynamic load, fully considering the characteristics and power supply constraints of the POE power supply, effectively improving the control accuracy; and adopting a multi-level feature mapping and progressive prediction strategy, which not only realizes the correlation analysis between electrical parameters but also enhances the reliability of power distribution; in addition, through multi-constraint optimization and closed-loop control, accurate power distribution and adjustment are achieved, thus overall realizing the efficient and accurate control of the POE power supply power.

[0035] The power control method of the POE power supply in the embodiments of the present invention has been described above. Next, the power control device of the POE power supply in the embodiments of the present invention will be described. Please refer to Figure 2 , an embodiment of the power control device of the POE power supply in the embodiments of the present invention includes: A calibration and compensation module 201, configured to obtain port electrical parameter data, device temperature data, and historical power supply operation data of the target POE power supply, and perform multi-dimensional calibration and compensation on the port electrical parameter data based on the device temperature data to obtain compensated port electrical parameter data; A feature mapping module 202, configured to perform time-frequency decomposition with multiple windows on the compensated port electrical parameter data to obtain port electrical decomposition data, and perform extraction of dynamic-static features and hierarchical mapping of features on the port electrical decomposition data to construct a port status feature vector; A demand prediction module 203, configured to construct a classification boundary and perform mapping conversion of load features on the port status feature vector to obtain a load classification mapping table, and perform progressive prediction of multi-level power supply parameters and integration of power demand on the load classification mapping table based on the historical power supply operation data to obtain a power demand prediction result; A closed-loop control module 204, configured to perform multi-constraint optimization calculation on the power demand prediction result to generate a power distribution strategy, and based on the power distribution strategy, control the target POE power supply to perform multiple power closed-loop controls to generate the power control result of the target POE power supply.

[0036] In the embodiments of the present invention, by performing multi-dimensional acquisition and feature extraction on the electrical parameters and temperature data of the target POE power supply, combining historical operation data for temperature compensation and calibration to obtain high-precision port electrical characteristics, then performing time-frequency decomposition and dynamic and static feature mapping on these characteristics to obtain port status feature vectors; then forming a load classification mapping table through classification boundary construction and load feature conversion, and performing progressive prediction of multi-level power supply parameters based on historical data to achieve accurate prediction of power demand; finally, performing multi-constraint optimization and closed-loop control based on the prediction results to output power control results. Through hierarchical data processing and feature analysis, the problem of accurate allocation of POE power supply power control is solved, especially in aspects such as cable attenuation, temperature drift, and dynamic load, fully considering the characteristics and power supply constraints of the POE power supply, effectively improving the control accuracy; and adopting a multi-level feature mapping and progressive prediction strategy, which not only realizes the correlation analysis between electrical parameters but also enhances the reliability of power distribution; in addition, through multi-constraint optimization and closed-loop control, accurate power distribution and adjustment are achieved, thus overall realizing the efficient and accurate control of POE power supply power.

[0037] Above Figure 2 The power control device of the POE power supply in the embodiments of the present invention is described in detail from the perspective of modular functional entities. Next, the power control device of the POE power supply in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0038] Figure 3 FIG. is a schematic structural diagram of a power control device of a POE power supply provided by an embodiment of the present invention. The power control device 300 of the POE power supply may vary greatly due to configuration or performance differences, and may include one or more processors (central processing units, CPUs) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage devices) for storing application programs 333 or data 332. Among them, the memory 320 and the storage media 330 may be transient storage or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the power control device 300 of the POE power supply. Further, the processor 310 may be configured to communicate with the storage media 330 and execute a series of instruction operations in the storage media 330 on the power control device 300 of the POE power supply.

[0039] The power control device 300 of the POE power supply may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and so on. Those skilled in the art can understand that Figure 3 The structure of the power control device of the POE power supply shown does not constitute a limitation on the power control device of the POE power supply, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0040] The present invention also provides a power control device for a POE power supply. The computer device includes a memory and a processor. When the computer-readable instructions stored in the memory are executed by the processor, the processor is caused to execute each step of the power control method for the POE power supply in the above-mentioned embodiments.

[0041] The present invention also provides a computer-readable storage medium. The computer-readable storage medium may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are run on a computer, the computer is caused to execute each step of the power control method for the POE power supply.

[0042] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0043] If the integrated unit is implemented in the form of 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 the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The foregoing storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs, etc., which can store program codes.

[0044] This application can be used in numerous general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0045] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power control method for a POE power supply, characterized in that: The power control method of the POE power supply includes: Acquire port electrical parameter data, device temperature data, and historical power supply operation data of a target POE power supply, and perform multi-dimensional calibration and compensation on the port electrical parameter data based on the device temperature data to obtain compensated port electrical parameter data; Performing multi-window time-frequency decomposition on the compensated port electrical parameter data to obtain port electrical decomposition data, and extracting dynamic and static features and mapping features hierarchically on the port electrical decomposition data to construct a port state feature vector; Constructing the classification boundary and mapping the load characteristics of the port state feature vector to obtain a load classification mapping table, and based on the historical power supply operation data, performing progressive prediction of multi-level power supply parameters and power demand integration on the load classification mapping table to obtain a power demand prediction result; A multi-constraint optimization calculation is performed on the power demand prediction result to generate a power allocation strategy, and based on the power allocation strategy, the target POE power supply is controlled to perform multiple power closed-loop controls to generate a power control result of the target POE power supply.

2. The power control method of the POE power supply according to claim 1, characterized in that: The step of performing multi-dimensional calibration compensation on the port electrical parameter data based on the device temperature data to obtain compensated port electrical parameter data includes: Performing temperature gradient calculation on the device temperature data to obtain power supply temperature field distribution, and performing window smoothing calculation and parameter value statistical calculation on the port electrical parameter data to obtain statistical port electrical parameter data, and marking and deleting abnormal values ​​corresponding to the statistical port electrical parameter data to obtain valid sampling sequence data; Based on the cable length corresponding to the target POE power supply, the effective sampling sequence data is fitted with a segmented cable attenuation coefficient to obtain a slope coefficient of each segment, and the slope coefficient of each segment is interpolated and fitted to obtain a cable compensation coefficient table; Performing sensitivity calculation of electrical parameters corresponding to the power supply temperature field distribution to obtain a temperature influence coefficient, and performing weighted calculation of preset key node temperatures on the temperature influence coefficient to obtain a temperature compensation matrix; Zero offset calculation and gain error compensation are performed on the port electrical parameter data to obtain initial calibration data, and secondary compensation calibration is performed on the initial calibration data based on the cable compensation coefficient table and the temperature compensation matrix to obtain compensated port electrical parameter data.

3. The power control method of the POE power supply according to claim 1, characterized in that: The multi-window time-frequency decomposition of the compensated port electrical parameter data is performed to obtain port electrical decomposition data, and the dynamic and static features of the port electrical decomposition data are extracted and hierarchically mapped to construct a port state feature vector, including: Based on a preset port sampling frequency, the compensated port electrical parameter data is segmented in time sequence to obtain a time domain sampling sequence, and the time domain sampling sequence is transformed in time-frequency to obtain port electrical decomposition data; Performing voltage fluctuation calculation on the power supply voltage data in the port electrical decomposition data to obtain voltage characteristic parameters, and performing change rate calculation and current dynamic response spectrum decomposition on the power supply current data in the port electrical decomposition data to obtain current characteristic parameters; Performing phase detection of a synchronous sampling zero point on the voltage characteristic parameter and the current characteristic parameter to obtain a phase difference angle, and performing power value calculation on the voltage characteristic parameter and the current characteristic parameter based on the phase difference angle to obtain a power characteristic parameter; Performing hierarchical correlation mapping using the voltage characteristic parameter, the current characteristic parameter, and the power characteristic parameter to obtain an initial characteristic structure, and performing cross-correlation calculation and weighted combination on the initial characteristic structure to obtain a standard power supply characteristic set; The standard power feature set is arranged in descending order and spatially mapped to obtain a low-dimensional feature vector, and the low-dimensional feature vector is reconstructed by feature combination to obtain a port state feature vector.

4. The power control method of the POE power supply according to claim 3, characterized in that: The construction of the classification boundary and the mapping conversion of the load characteristics of the port state feature vector to obtain a load classification mapping table includes: Clustering each feature vector in the port state feature vector to obtain a state feature distance matrix, and constructing a classification boundary for the state feature distance matrix to obtain an initial feature classification space; Based on the preset POE power supply power standard, the initial feature classification space is graded to obtain an initial power interval, and a boundary threshold is set for the initial power interval to obtain a load classification space; Performing spatial combination and conditional division of spatial feature attributes on the initial feature classification space and the load classification space to obtain an initial decision space, and performing hierarchical decomposition and recursive updating of spatial node weight coefficients on the initial decision space to obtain a classification decision tree; Performing adjacent sampling point differential operations and normalization on leaf node data in the classification decision tree to obtain a load state change rate, and performing a change trend prediction and confidence evaluation on the load state change rate to obtain a load dynamic feature; The load dynamic features are identified by load type and feature-associated mapping is updated to generate a load classification mapping table.

5. The power control method of the POE power supply according to claim 4, characterized in that: The step of performing progressive prediction of multi-level power supply parameters and power demand integration on the load classification mapping table based on the historical power supply operation data to obtain a power demand prediction result includes: Decomposing the historical power supply operation data into periodic power supply to obtain a historical power supply sequence, and based on the historical power supply sequence, performing correlation matching of load power consumption patterns and extracting pattern timing characteristics on the load classification mapping table to obtain a power supply change pattern; Performing time scale division and sampling reorganization on the power supply change pattern to obtain multi-scale power change data, and performing scale association and state space modeling on the multi-scale power change data to obtain a progressive prediction model; Performing power loss calculation on the equivalent impedance value of the POE cable corresponding to the target POE power supply to obtain a loss distribution function, and performing transmission efficiency evaluation and temperature compensation calculation on the loss distribution function to obtain a power correction coefficient; Based on the power characteristic parameters corresponding to the preset short-term time interval, a short-term sliding average operation is performed on the power correction coefficient to obtain a short-term load forecast value, and based on the pattern similarity index in the historical power supply operation data, a pattern mapping and similarity distance calculation are performed on the power characteristic parameters corresponding to the preset medium-term time interval and the short-term load forecast value to obtain a pattern matching sequence; Performing weighted fusion and reconstruction and extraction of trend features on the pattern matching sequence to obtain a medium-term load forecast value, and performing data integration and filtering and smoothing on the medium-term load forecast value based on a preset long-term time interval to obtain a long-term power trend; Perform seasonal decomposition and trend extrapolation on the long-term power trend to obtain a long-term load forecast value, and perform weighted fusion and elimination of abnormal load values ​​on the short-term load forecast value, the medium-term load forecast value, and the long-term load forecast value to obtain a load forecast correction result; The confidence interval calculation and prediction accuracy quantitative evaluation are performed on the load prediction correction result to obtain the power demand prediction result.

6. The power control method of the POE power supply according to claim 5, characterized in that: The performing multi-constraint optimization calculation on the power demand prediction result to generate a power allocation strategy includes: Based on the power power threshold of the target POE power supply, the power demand prediction result is matched with the power demand to obtain the allocable power amount, and based on the power supply priority corresponding to each port in the target POE power supply, the allocable power amount is weighted and balanced by the port power quota to obtain an initial power allocation plan; Obtaining historical cable temperatures corresponding to the power supply power in the initial power allocation scheme, and performing spatial distribution calculation based on the historical cable temperatures and constructing corresponding cable temperature rise curves; Performing limit constraints on the cable temperature rise curve and margin calculation of the safety threshold to obtain temperature constraint parameters, and based on the temperature constraint parameters, performing temperature coefficient correction and loss evaluation on the cable compensation coefficient table to obtain efficiency constraint parameters; The temperature constraint parameters and the efficiency constraint parameters are adjusted and solved for multiple objectives, and the feasible solution space is searched and evaluated to obtain an optimal allocation scheme, and the optimal allocation scheme is subjected to margin risk assessment and power allocation adjustment to generate a power allocation strategy.

7. The power control method of the POE power supply according to claim 1, characterized in that: The controlling the target POE power supply to perform multiple power closed-loop controls based on the power allocation strategy to generate a power control result of the target POE power supply includes: Extracting power control parameters and dividing supply voltage into intervals for the power allocation strategy to obtain reference control target parameters, and performing voltage and current decomposition and DC power mapping on the reference control target parameters to obtain hierarchical control parameters; Performing time sequence planning on the hierarchical control parameters, generating a control execution plan, and performing gain calculation and loop compensation on the voltage loop and the current loop on the control execution plan to obtain a basic power supply control quantity; Based on the power demand prediction result, the basic power supply control amount is subjected to power feedback compensation to obtain a modified power supply control amount, and the basic power supply control amount and the modified power supply control amount are subjected to output quantity superposition and limit constraints to obtain a protection power supply control amount of the power supply corresponding to the target POE power supply; The output priority of the basic power control amount, the modified power control amount and the protection power control amount is determined and modulated to generate a power control result of the target POE power supply.

8. A power control device for a POE power supply, characterized in that: The power control device of the POE power supply comprises: A calibration and compensation module, used to obtain port electrical parameter data, device temperature data and historical power supply operation data of a target POE power supply, and based on the device temperature data, perform multi-dimensional calibration and compensation on the port electrical parameter data to obtain compensated port electrical parameter data; A feature mapping module is used to perform multi-window time-frequency decomposition on the compensated port electrical parameter data to obtain port electrical decomposition data, extract dynamic and static features and perform feature hierarchical mapping on the port electrical decomposition data to construct a port state feature vector; A demand prediction module, used to construct classification boundaries and map and transform load characteristics of the port state feature vector to obtain a load classification mapping table, and based on the historical power supply operation data, perform progressive prediction of multi-level power supply parameters and power demand integration on the load classification mapping table to obtain a power demand prediction result; The closed-loop control module is used to perform multi-constraint optimization calculations on the power demand prediction results, generate a power allocation strategy, and based on the power allocation strategy, control the target POE power supply to perform multiple power closed-loop controls to generate a power control result of the target POE power supply.

9. A power control device for a POE power supply, characterized in that: The power control device of the POE power supply comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instruction in the memory to enable the power control device of the POE power supply to execute each step of the power control method of the POE power supply as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the power control method of the POE power supply as described in any one of claims 1-7 are implemented.

Images