A health management and resource optimization configuration method, device, medium and product for a test cabinet distributed power supply

Abstract

This application discloses a health management and resource optimization configuration method for a distributed power supply in a test cabinet. The method includes: collecting electrical and temperature parameters of multiple sub-modules of the power supply during operation; extracting feature vectors representing the degradation state of the power supply based on a sliding time window; obtaining health information of multiple sub-modules; adaptively adjusting the reporting strategy of health information; and outputting power control commands for at least one sub-module. When a sub-module's health risk indicator is detected to exceed a preset risk threshold, the sub-module is marked as a risk sub-module, and at least one type of migrateable load is migrated from the risk sub-module to a candidate sub-module. Output channel isolation control is performed on the risk sub-module, and the power supply path used for status detection is retained. The display priority of multiple sub-modules is calculated, sorted and displayed according to the display priority, and multi-level alarm prompts corresponding to the health risk indicators are triggered simultaneously.

Description

Technical Field

[0001] This application relates to the field of power management technology, specifically to a method, device, medium, and product for health management and resource optimization of distributed power supplies in a test cabinet. Background Technology

[0002] In critical tasks such as industrial automation testing or data centers, test cabinets generally adopt centralized or simple distributed power supply architectures, relying on static threshold monitoring of basic parameters such as voltage, current, and temperature to achieve fault alarms. Existing technical solutions have the following limitations: First, the power supply system lacks the ability to quantitatively analyze and predict the progressive degradation behavior of power modules, only triggering alarms when parameters exceed limits, failing to identify risks in the early stages of power performance degradation, leading to an increased probability of sudden power outages; second, under conditions of limited system communication links and human-machine interaction resources, existing solutions typically report monitoring data for all power modules at fixed intervals and priorities, resulting in a large amount of low-risk information mixed with critical risk information during transmission and display, easily causing excessive bus load and operator cognitive overload, affecting timely attention to high-risk modules; third... The current power supply resource allocation largely relies on fixed redundancy or static grouping strategies preset during the design phase. It cannot dynamically adjust based on the real-time health status, remaining lifespan, and load importance of power modules. This results in degraded modules operating under continuous high stress while healthy modules remain idle, leading to insufficient overall energy efficiency and task continuity. Furthermore, existing methods for load switching or fault isolation often rely on single threshold triggers, lacking stability constraints such as switching hysteresis and minimum dwell time. This easily causes frequent load migration between multiple power modules, leading to system oscillations. Moreover, after isolation, it is often impossible to continuously monitor module status, hindering maintenance decisions. Therefore, existing distributed power supply systems urgently need improvement in health prediction, information scheduling, dynamic resource allocation, and stability control. Summary of the Invention

[0003] This application provides a system and method for health management and resource optimization of distributed power supplies in a test cabinet, addressing issues in existing distributed power supply systems such as reliance on static threshold alarms, lack of degradation trend prediction capabilities, rigid resource allocation strategies, and information overload on the bus and human-machine interface. By introducing health prediction, risk quantification, and resource collaborative scheduling mechanisms between the distributed power supply submodules and the central management unit, the distributed power supply system is transformed from passive alarms to proactive prediction and intervention, and from static redundancy to risk-driven dynamic configuration.

[0004] Firstly, this application provides a method for health management and resource optimization of distributed power supplies in a test cabinet. The test cabinet includes a power supply and a central management unit. The power supply includes multiple sub-modules. The method for health management and resource optimization of distributed power supplies in the test cabinet includes: collecting electrical and temperature parameters of multiple sub-modules of the power supply during operation; segmenting the electrical and temperature parameters based on a sliding time window; and extracting feature vectors representing the degradation state of the power supply within each time window; inputting the feature vectors into a prediction and health management model to obtain health information of multiple sub-modules, including the remaining healthy lifespan and health risk indicators of multiple sub-modules; adaptively adjusting the reporting strategy of health information based on the health risk indicators, and reporting the health risk indicators to the central management unit according to the reporting strategy; and the central management unit, based on the health risk indicators and the power supply's... Power data indicators are used to construct a resource optimization decision model and output power regulation commands for at least one submodule. When the health risk indicators of at least one submodule exceed a preset risk threshold, and / or at least one submodule fails to meet load requirements after receiving the power regulation command, at least one submodule is marked as a risk submodule. Another submodule that meets the health risk indicators and / or meets the load requirements is selected as a candidate submodule, and at least one type of migrateable load is migrated from the risk submodule to the candidate submodule. After the load migration of the risk submodule is completed, output channel isolation control is performed on the risk submodule, and the power supply path used for status detection is reserved. Based on the health risk indicators and load requirements, the display priority of multiple submodules is calculated, sorted and displayed according to the display priority, and multi-level alarm prompts corresponding to the health risk indicators are triggered.

[0005] Optionally, the electrical parameters include at least the output voltage, output current, and input voltage; the temperature parameters include at least the power device temperature and the heat dissipation temperature; and the eigenvectors include at least one of the following: mean, variance, peak-to-peak value, rate of change, load fluctuation frequency, temperature cycle amplitude, and cycle count within the sliding time window.

[0006] Optionally, the prediction and health management model includes a remaining life prediction model based on time series learning, wherein the remaining life prediction model is a recurrent neural network structure, and the recurrent neural network structure includes a long short-term memory network structure.

[0007] Optionally, the prediction and health management model includes an uncertainty assessment model. After the prediction and health management model outputs an estimate of the remaining healthy lifespan, the expected confidence level of the estimate of the remaining healthy lifespan is assessed by the uncertainty assessment model, which includes a Gaussian process regression model.

[0008] Optionally, the formula for calculating health risk indicators is as follows:

[0009]

[0010] HRI stands for Health Risk Index. The degradation term is obtained by normalizing the remaining healthy lifespan. These are fault items obtained based on statistics of threshold exceedance events. This is the stress accumulation term obtained based on temperature cycling or load impact, and .

[0011] Optionally, the method for implementing output channel isolation control for the risk submodule includes: determining the health risk coefficient k based on health risk indicators; and setting the maximum output current of the risk submodule to: I i,max = k(HRIi)·I i,rated .

[0012] Among them, I i,max I is the maximum allowable output current of the submodule. i,rated The rated output current of the submodule, HRI i Let k represent the health risk of the submodule, and k be the risk coefficient function.

[0013] Optionally, the reporting strategy includes at least one of the following: reporting period, reporting data granularity, or data frame priority; the power regulation command includes at least one of the following: maximum output current limit, output voltage upper limit adjustment, or output channel start / stop control.

[0014] Optionally, the health management and resource optimization configuration method for the distributed power supply of the test cabinet also includes: setting load migration constraints; after completing a load migration of a risk submodule, the load migration constraints must be met before the next migration can be carried out, wherein the load migration constraints include switching hysteresis conditions, minimum dwell time, and switching cost constraints.

[0015] Optionally, the formula for calculating the display priority is as follows:

[0016] P i = w1·HRI i + w2·C i ,

[0017] Among them, P i HRI displays priority for submodules. i For the health risk indicators of the submodule, C i w1 and w2 represent the importance of the submodule load, respectively, and are weighting coefficients.

[0018] Secondly, this application also provides an electronic device, including a memory and a processor; the memory stores a computer program, and the processor runs the computer program in the memory to perform operations in the health management and resource optimization configuration method for the distributed power supply of the test cabinet provided in the first aspect.

[0019] Thirdly, this application also provides a storage medium storing multiple instructions adapted for loading by a processor to execute the steps in the health management and resource optimization configuration method for the distributed power supply of the test cabinet provided in the first aspect.

[0020] Fourthly, this application also provides a computer program product, including a computer program or instructions, which, when executed by a processor, implement the steps in the health management and resource optimization configuration method for the distributed power supply of the test cabinet provided in the first aspect.

[0021] The health management and resource optimization configuration method, equipment, media, and products for distributed power supplies in test cabinets provided in this application continuously collect electrical and temperature parameters during the operation of distributed power supply submodules. By introducing a time-series learning-based prediction and health management model, the degradation trend and remaining health life of the power supply modules are estimated online. Health risk indicators are constructed by combining fault event statistics and stress accumulation information, thereby identifying potential risks before severe performance degradation or failure of the power supply modules, reducing the probability of sudden power supply failures during critical task execution cycles. Furthermore, through an adaptive information reporting mechanism based on health risk levels, distributed power supply submodules can dynamically adjust the reporting cycle, data granularity, and priority of health information according to their own risk status. This achieves a risk-driven data scheduling strategy that prioritizes high-risk information and reduces the frequency of low-risk information, thereby ensuring real-time availability of critical risk information while reducing unnecessary data consumption, improving the utilization efficiency of the system's core data links, and enhancing the visualization and interactive interface of the monitoring terminal. To improve information readability, this application comprehensively considers the health risk indicators, remaining health lifespan, rated capacity, and load importance level of distributed power submodules in the central power management and resource allocation unit. Under the premise of meeting the continuous power supply constraints of critical loads, it dynamically generates power regulation commands such as current limiting, voltage limiting, and output channel start / stop, and performs load redistribution when necessary. This avoids the resource waste caused by traditional fixed redundancy or static grouping strategies, improves the overall resource utilization efficiency and task continuity of the distributed power system, and effectively suppresses frequent back-and-forth migration of loads between multiple distributed power submodules and avoids system oscillation during load switching and resource redistribution by introducing switching hysteresis conditions, minimum dwell time, and switching cost constraints. When the health risk of a power module reaches the emergency threshold, the output channel is isolated after the load migration is completed, while retaining the low-power monitoring power supply path. This ensures the safe operation of the system while continuously acquiring module health status information, providing a basis for maintenance decisions. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only 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 flowchart illustrating the health management and resource optimization configuration method for distributed power supplies in a test cabinet provided in an embodiment of this application.

[0024] Figure 2 This is a schematic diagram of the dynamic display, hierarchical alarms, and data recording of the monitoring terminal's visual interactive interface provided in this application embodiment. Detailed Implementation

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

[0026] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0027] "A and / or B" includes the following three combinations: A only, B only, and a combination of A and B.

[0028] The use of "applies to" or "configured to" in this application implies open and inclusive language, which does not preclude applicability to or configuration to devices performing additional tasks or steps. Furthermore, the use of "based on" implies openness and inclusivity, because processes, steps, calculations, or other actions "based on" one or more conditions or values ​​may in practice be based on additional conditions or values ​​beyond those conditions.

[0029] In this application, the term "exemplary" is used to mean "used as an example, illustration, or description." Any embodiment described as "exemplary" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes are not described in detail to avoid obscuring the description of this application with unnecessary detail. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0030] The following describes, with reference to the accompanying drawings, the health management and resource optimization configuration method, equipment, media and products of the distributed power supply for the test cabinet provided in the embodiments of this application.

[0031] Figure 1 This is a flowchart illustrating the health management and resource optimization configuration method for distributed power supplies in a test cabinet according to an embodiment of this application.

[0032] like Figure 1 As shown in the illustration, this application provides a method for health management and resource optimization configuration of a distributed power supply in a test cabinet. The test cabinet includes a power supply and a central management unit, and the power supply includes multiple sub-modules. The method includes the following steps:

[0033] S102 collects electrical and temperature parameters of multiple sub-modules of the power supply during operation, segments the electrical and temperature parameters based on a sliding time window, and extracts feature vectors representing the degradation state of the power supply within each time window.

[0034] In some embodiments of this application, step S102 includes the following sub-steps:

[0035] S102-1 collects electrical parameters from multiple sub-modules.

[0036] In some embodiments of this application, each distributed power supply submodule collects the following electrical parameters in real time through a local sampling circuit: output voltage Uout; output current Iout; and input voltage Uin. In some embodiments of this application, the sampling circuit can be implemented using shunt resistors, Hall current sensors, voltage divider networks, etc. To ensure the ability to capture transient behaviors such as load changes and current surges, the sampling period Ts can be set to 1 ms–10 ms.

[0037] Those skilled in the art will understand that although the output voltage, output current and input voltage of the submodule are collected in some embodiments of this application, this is only exemplary and not intended to limit the present embodiment. In practical applications, other suitable electrical parameters such as the output power of the submodule can also be collected.

[0038] S102-2 collects multiple temperature and environmental parameters.

[0039] In some embodiments of this application, the following temperature or environmental parameters are collected in real time for each distributed power submodule: power device junction temperature or equivalent temperature Tdev; heat sink / ambient temperature Tamb; and module housing temperature Tcase. These temperature parameters can be obtained through thermistors, digital temperature sensors, or integrated temperature monitoring chips, and are used to characterize the thermal stress level of the devices under different load conditions.

[0040] Those skilled in the art will understand that although some embodiments of this application collect the junction temperature or equivalent temperature of the power devices of the submodule, the heat sink / ambient temperature, and the module housing temperature, this is only exemplary and not intended to limit the present embodiment. In practical applications, other suitable temperature or environmental parameters of the submodule can also be collected.

[0041] S102-3 aligns the raw data cache with time.

[0042] In some embodiments of this application, the collected electrical and temperature parameters are first written into the local circular buffer inside the distributed power supply submodule, and a timestamp is uniformly added for time alignment.

[0043] In some embodiments of this application, by aligning the time of cached data for electrical and temperature parameters, time deviations caused by sampling delays of different channels can be avoided, providing accurate data for subsequent sliding window analysis and feature extraction.

[0044] S102-4, Set a sliding time window and segment the electrical and temperature parameters based on the sliding time window.

[0045] In some embodiments of this application, a sliding time window mechanism is introduced to statistically analyze the statistical behavior and dynamic degradation characteristics of the power module, which are typically reflected over a period of time. The specific settings are as follows:

[0046] Sliding window length W (e.g., 10 s–60 s);

[0047] Window sliding step size Δt (e.g., 100 ms – 1 s);

[0048] At any time t, the data within the window is represented as follows: .

[0049] S102-5 performs statistical analysis and feature extraction on electrical and temperature parameters.

[0050] Within each sliding window, the system calculates characteristics for both electrical and temperature parameters to characterize operational stability and load behavior. These characteristics can include statistical features, dynamic features, and load behavior features. Statistical features can include mean, variance, skewness, and peak-to-peak value; dynamic features can include current change rate dIout / dt and voltage change rate dUout / dt; and load behavior features can include load duty cycle and load fluctuation frequency.

[0051] S102-6 extracts thermal fatigue and stress characteristics.

[0052] In some embodiments of the application, cyclic analysis of the temperature time series is performed to extract the temperature cycle amplitude, the number of temperature cycles, and the slope of temperature rise and fall, which can be used to characterize the cumulative effect of thermal fatigue generated by power devices during long-term operation.

[0053] S102-7, the feature vector representing the power supply degradation state is used as the health state feature vector, and the health state feature vector is constructed.

[0054] The statistical characteristics, dynamic characteristics, load behavior characteristics, and thermal fatigue characteristics mentioned above are concatenated in a predetermined order to form the health status feature vector corresponding to the current time window: .

[0055] In some embodiments of this application, the electrical and temperature parameters of the power supply submodule are segmented by a sliding time window, and feature vectors characterizing the power supply degradation state are extracted in each window. This can effectively capture the gradual process of power supply performance changes and early signs of degradation from dynamic operating data, thereby providing more accurate and timely feature inputs for subsequent health prediction and risk assessment, overcoming the shortcomings of traditional static or instantaneous parameter monitoring that is insensitive to slow degradation.

[0056] S104. Input the feature vector into the prediction and health management model to obtain health information for multiple sub-modules. The health information includes the remaining healthy lifespan and health risk indicators of multiple sub-modules.

[0057] In some embodiments of this application, step S104 includes the following sub-steps:

[0058] S104-1, construct the feature sequence.

[0059] Select the health status feature vectors from the most recent N time windows and construct a feature sequence in chronological order: This serves as input to the Predictive and Health Management (PHM) model.

[0060] S104-2, Predicting Remaining Healthy Life (RUL).

[0061] The feature sequence is input into the trained PHM model, which may include a time-series learning-based remaining lifetime prediction model, and may employ a recurrent neural network structure, such as a Long Short-Term Memory (LSTM) network. The remaining lifetime prediction model outputs the remaining healthy lifetime (RUL) of the current distributed power submodule, which can be represented as remaining runtime or remaining job counts.

[0062] Those skilled in the art will understand that although some embodiments of this application show that the remaining lifetime prediction model may include a single layer of long short-term memory network, this is merely exemplary and does not limit the application. In practical applications, the remaining lifetime prediction model may also be other suitable models.

[0063] S104-3, assess the prediction uncertainty of the RUL results.

[0064] In some embodiments of this application, the prediction uncertainty of the RUL results is assessed, and a confidence interval for the RUL prediction results is given by a Gaussian process regression model, so as to improve the overall prediction accuracy.

[0065] Those skilled in the art will understand that although some embodiments of this application perform a prediction uncertainty assessment on the RUL results, this is merely exemplary and this step can be omitted in practical applications.

[0066] S104-4, Construct the degradation term of RUL based on the RUL prediction value. The specific calculation formula is as follows:

[0067]

[0068] Among them, RUL init The nominal lifespan under initial healthy conditions, and when RUL ≥ RUL init time f RUL Take 0, when RUL ≤ 0, f RUL Take 1, or truncate the calculation result to the interval 0–1.

[0069] S104-5 calculates the fault event term and stress accumulation term; by statistically analyzing the occurrence or duration of over-limit events such as overvoltage, undervoltage, overcurrent, and overtemperature within the sliding window, the fault event term f is obtained. fault A stress accumulation term f is constructed based on the characteristics of temperature cycling and load impact.stress .

[0070] S104-6, the degradation terms obtained in steps S104-4 and S104-5 are... Fault event item f fault and stress accumulation term f stress The degradation items, failure event items, and stress accumulation items are fused according to preset weights to obtain the normalized health risk index (HRI). It can be further mapped to a risk score range of 0–100 according to display or alarm requirements, and can be used for subsequent risk classification and control decisions.

[0071] In some embodiments of this application, the formulas for calculating health risk indicators are as follows:

[0072]

[0073] HRI stands for Health Risk Index. The degradation term is obtained by normalizing the remaining healthy lifespan. These are fault items obtained based on statistics of threshold exceedance events. This is the stress accumulation term obtained based on temperature cycling or load impact, and .

[0074] S106, based on health risk indicators, adaptively adjust the reporting strategy for health information, and report the health risk indicators to the central management unit according to the reporting strategy.

[0075] In some embodiments of this application, step S106 includes the following sub-steps:

[0076] S106-1, Determine the health information reporting mode based on health risk indicators.

[0077] Based on the HRI's numerical range, rate of change, and prediction uncertainty, the distributed power supply submodule is divided into a stable state, a mildly degraded state, and a rapidly degraded or high-risk state.

[0078] S106-2, Adjust the reporting strategy according to the status of the distributed power supply submodule.

[0079] If the module is determined to be in a stable state in step S106-1, the reporting period is extended, and only HRI and RUL are reported; if the module is determined to be in a high-risk state in step S106-1, the reporting period is shortened, key feature or trend information is reported, and the priority of data frames is increased.

[0080] In some embodiments of this application, the priority of the data frame can be dynamically adjusted by configuring the scheduling priority or time triggering parameters of health information messages in the core data link of the system. In some embodiments of this application, by adjusting the reporting strategy of the status of the distributed power supply submodule, a data flow control strategy of "high risk, high frequency; low risk, low frequency" can be implemented. This ensures timely delivery of critical health status while significantly reducing the average load of the system bus, avoiding information overload, and improving the information focusing capability of the monitoring terminal's visual interactive interface and the overall communication efficiency of the system.

[0081] S108, the central management unit constructs a resource optimization decision model based on health risk indicators and power data indicators of the power supply, and outputs power regulation commands for at least one sub-module.

[0082] In some embodiments of this disclosure, the central management unit may include a central power management unit and a resource configuration unit. The central power management and resource configuration unit (CPMCU) receives and caches the health summary information reported by each distributed power submodule, and synchronously maintains the module's rated power, current output power, load power requirements, and load importance level, providing a decision basis for resource optimization configuration.

[0083] S110, when at least one submodule's health risk indicator exceeds a preset risk threshold, and / or at least one submodule fails to meet load requirements after receiving a power regulation command, at least one submodule is marked as a risk submodule.

[0084] In some embodiments of this disclosure, for a marked risk submodule, the central power management and resource allocation unit issues a current-limiting or voltage-limiting command based on the health risk coefficient k (HRI) to set the maximum output current of the risk submodule to: I i,max = k(HRI i )·I i,rated This is to reduce the internal stress of the risk submodule.

[0085] S112, select another submodule that meets the health risk indicators and / or meets the load requirements as a candidate submodule, and migrate at least one type of migrateable load from the risk submodule to the candidate submodule.

[0086] In some embodiments of this application, the central power management and resource allocation unit will preferentially select modules with lower HRI and larger power margins from the set of modules that meet the voltage level and capacity conditions as candidate sub-modules. The candidate sub-modules will serve as takeover nodes to take over the transferable loads of the risky sub-modules.

[0087] In some embodiments of this application, in order to avoid repeated migrations during load switching, stability constraints are set. By introducing minimum dwell time, switching hysteresis, and switching cost constraints, system oscillations caused by multiple load switching are prevented.

[0088] S114 After completing the load migration of the risk submodule, output channel isolation control is performed on the risk submodule, and the power supply path for status detection is retained.

[0089] When the health risk of the distributed power supply submodule increases and the importance of the supplied load is low, the central power management and resource allocation unit can perform shutdown or load reduction control on some non-critical output channels to reduce the internal thermal and electrical stress of the module, while ensuring continuous power supply to critical loads.

[0090] In some embodiments of this application, when both HRI and RUL trigger emergency conditions simultaneously, the execution module is isolated, retaining only the monitoring power supply. The power supply path used for status monitoring is a low-power monitoring branch, used to maintain the operation of the sampling circuit or local processing unit to continuously acquire health status information, rather than to maintain the original load power supply.

[0091] S116 calculates the display priority of multiple sub-modules based on health risk indicators and load requirements, sorts and displays them according to display priority, and triggers multi-level alarm prompts corresponding to health risk indicators.

[0092] Figure 2 This is a schematic diagram of the dynamic display, hierarchical alarms, and data recording of the monitoring terminal's visual interactive interface provided in this application embodiment.

[0093] like Figure 2 As shown, in some embodiments of this application, the display priority can be calculated based on the HRI and load importance of the submodule, and displayed in order of priority in the monitoring terminal's visual interactive interface. The calculation formula for the display priority is as follows:

[0094] P i = w1·HRI i + w2·C i ,

[0095] Among them, P i HRI displays priority for submodules. i For the health risk indicators of the submodule, C i w1 and w2 represent the importance of the submodule load, respectively, and are weighting coefficients.

[0096] In some embodiments of this application, different levels of audible and visual alarms can be triggered based on the HRI range. Simultaneously, the central power management and resource allocation unit records historical data, such as all health assessment results, control commands, and event logs, for maintenance analysis and model updates.

[0097] On the other hand, embodiments of this application also provide an electronic device, including a memory and a processor; the memory stores a computer program, and the processor is used to run the computer program in the memory to perform the operations in the health management and resource optimization configuration method for the distributed power supply of the test cabinet provided in the first aspect.

[0098] On the other hand, embodiments of this application also provide a storage medium storing multiple instructions adapted for loading by a processor to execute the health management and resource optimization configuration method for the distributed power supply of the test cabinet as provided in the above embodiments.

[0099] On the other hand, this application also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned method for health management and resource optimization of the distributed power supply in the test cabinet.

[0100] The health management and resource optimization configuration method, equipment, media, and products for distributed power supplies in test cabinets provided in this application continuously collect electrical and temperature parameters during the operation of distributed power supply submodules. By introducing a time-series learning-based prediction and health management model, the degradation trend and remaining health life of the power supply modules are estimated online. Health risk indicators are constructed by combining fault event statistics and stress accumulation information, thereby identifying potential risks before severe performance degradation or failure of the power supply modules, reducing the probability of sudden power supply failures during critical task execution cycles. Furthermore, through an adaptive information reporting mechanism based on health risk levels, distributed power supply submodules can dynamically adjust the reporting cycle, data granularity, and priority of health information according to their own risk status. This achieves a risk-driven data scheduling strategy that prioritizes high-risk information and reduces the frequency of low-risk information, thereby ensuring real-time availability of critical risk information while reducing unnecessary data consumption, improving the utilization efficiency of the system's core data links, and enhancing the visualization and interactive interface of the monitoring terminal. To improve information readability, this application comprehensively considers the health risk indicators, remaining health lifespan, rated capacity, and load importance level of distributed power submodules in the central power management and resource allocation unit. Under the premise of meeting the continuous power supply constraints of critical loads, it dynamically generates power regulation commands such as current limiting, voltage limiting, and output channel start / stop, and performs load redistribution when necessary. This avoids the resource waste caused by traditional fixed redundancy or static grouping strategies, improves the overall resource utilization efficiency and task continuity of the distributed power system, and effectively suppresses frequent back-and-forth migration of loads between multiple distributed power submodules and avoids system oscillation during load switching and resource redistribution by introducing switching hysteresis conditions, minimum dwell time, and switching cost constraints. When the health risk of a power module reaches the emergency threshold, the output channel is isolated after the load migration is completed, while retaining the low-power monitoring power supply path. This ensures the safe operation of the system while continuously acquiring module health status information, providing a basis for maintenance decisions.

[0101] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0102] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0103] The above provides a detailed description of the health management and resource optimization configuration method, device, medium, and product of a distributed power supply for a test cabinet provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for health management and resource optimization of a distributed power supply in a test cabinet, the test cabinet comprising a power supply and a central management unit, the power supply comprising multiple sub-modules, characterized in that, include: The electrical and temperature parameters of multiple sub-modules of the power supply are collected during operation. The electrical and temperature parameters are segmented based on a sliding time window, and feature vectors characterizing the power supply degradation state are extracted within each time window. The feature vector is input into the prediction and health management model to obtain health information of multiple sub-modules, including the remaining healthy lifespan and health risk indicators of multiple sub-modules; Based on the health risk indicators, the reporting strategy for the health information is adaptively adjusted, and the health risk indicators are reported to the central management unit according to the reporting strategy. The central management unit constructs a resource optimization decision model based on the health risk indicators and the power data indicators of the power supply, and outputs power regulation commands for at least one of the sub-modules. When the health risk indicator of at least one of the sub-modules exceeds a preset risk threshold, and / or at least one of the sub-modules is unable to meet the load requirements after receiving the power regulation command, at least one of the sub-modules will be marked as a risk sub-module. Select another submodule that meets the health risk indicators and / or meets the load requirements as a candidate submodule, and migrate at least one type of migrateable load from the risk submodule to the candidate submodule; After completing the load migration of the risk submodule, output channel isolation control is performed on the risk submodule, and the power supply path for status detection is retained; as well as Based on the health risk indicators and the load requirements, the display priority of multiple sub-modules is calculated, and they are sorted and displayed according to the display priority, while triggering multi-level alarm prompts corresponding to the health risk indicators.

2. The method for health management and resource optimization of distributed power supplies in test cabinets according to claim 1, characterized in that, The electrical parameters include at least the output voltage, output current, and input voltage; The temperature parameters include at least the power device temperature and the heat dissipation temperature; The feature vector includes at least one of the following within the sliding time window: mean, variance, peak-to-peak value, rate of change, load fluctuation frequency, temperature cycle amplitude, and cycle count.

3. The method for health management and resource optimization of distributed power supplies in a test cabinet according to claim 1, characterized in that, The prediction and health management model includes a life expectancy prediction model based on time series learning, wherein the life expectancy prediction model is a recurrent neural network structure, and the recurrent neural network structure includes a long short-term memory network structure.

4. The method for health management and resource optimization of distributed power supplies in test racks according to claim 3, characterized in that, The prediction and health management model includes an uncertainty assessment model. After the prediction and health management model outputs an estimate of the remaining healthy life expectancy, the uncertainty assessment model is used to evaluate the expected confidence level of the estimate of the remaining healthy life expectancy. The uncertainty assessment model includes a Gaussian process regression model.

5. The method for health management and resource optimization of distributed power supplies in a test cabinet according to claim 1, characterized in that, The formula for calculating the health risk indicators is as follows: HRI stands for Health Risk Index. The degradation term is obtained by normalizing the remaining healthy lifespan. These are fault items obtained based on statistics of threshold exceedance events. This is the stress accumulation term obtained based on temperature cycling or load impact, and .

6. The method for health management and resource optimization of distributed power supplies in a test cabinet according to claim 5, characterized in that, The method for performing output channel isolation control on the risk submodule includes: The health risk coefficient k is determined based on the aforementioned health risk indicators; Set the maximum output current of the risk submodule to: I i,max = k(HRIi)·I i,rated . Where I i,max I is the maximum allowable output current of the submodule. i,rated The rated output current of the submodule, HRI i Let k represent the health risk of the submodule, and k be the risk coefficient function.

7. The method for health management and resource optimization of distributed power supplies in a test cabinet according to claim 1, characterized in that, The reporting strategy includes at least one of reporting period, reporting data granularity, or data frame priority; the power regulation command includes at least one of maximum output current limit, output voltage upper limit adjustment, or output channel start / stop control.

8. The method for health management and resource optimization of distributed power supplies in a test cabinet according to claim 1, characterized in that, Also includes: Set load migration constraints; After completing one load migration of the aforementioned risk submodule, the next migration can only proceed if the load migration constraints are met. The load migration constraints include switching hysteresis conditions, minimum dwell time, and switching cost constraints.

9. The method for health management and resource optimization of distributed power supplies in a test cabinet according to claim 1, characterized in that, The formula for calculating the display priority is as follows: P i = w1·HRI i + w2·C i , Among them, P i HRI displays priority for submodules. i For the health risk indicators of the submodule, C i w1 and w2 represent the importance of the submodule load, respectively, and are weighting coefficients.

10. An electronic device, characterized in that, It includes a memory and a processor; the memory stores a computer program, and the processor runs the computer program in the memory to perform the steps in the health management and resource optimization configuration method for the distributed power supply of the test cabinet according to any one of claims 1 to 9.

Images