A switching power supply control system

By integrating data acquisition, feature extraction and fault prediction models in the switching power supply control system, early detection and early warning of faults are realized, and regulatory strategies are generated, which solves the problems of untimely fault prediction and lagging regulation response in the existing technology, and improves the problem of The stability and reliability of the system are achieved.

CN118763874BActive Publication Date: 2024-11-15SHENZHEN DINGSHUO TONGBANG TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411255128.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-09
Publication Date
2024-11-15
Estimated Expiration
2044-09-09

AI Technical Summary

Technical Problem

The existing switching power control system cannot be pre-warned before a fault occurs, resulting in the system being affected during the fault, which may lead to shutdown or equipment damage.

Method used

A switching power supply control system including data acquisition module, data processing and feature extraction module, fault prediction model construction module, fault trend analysis module, early warning and response module, and feedback and iterative optimization module were designed. By collecting operation parameters in real time, the Data preprocessing and feature extraction, build fault prediction models, conduct trend analysis and early warning, generate regulatory strategies, and continuously improve system stability through feedback and iterative optimization.

Benefits of technology

It effectively improves the accuracy and response speed of fault prediction, reduces the risk of system downtime or damage caused by sudden problems, ensures the continuous and stable operation of the equipment, and extends the service life of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118763874B_ABST
    Figure CN118763874B_ABST
Patent Text Reader

Abstract

The present invention discloses a switching power supply control system, which relates to the field of automation control technology. When the system is running, the stability and reliability of the system are continuously improved through feedback and iterative optimization, and the key operating parameters are collected in real time by using an integrated sensor group. Through data processing and feature extraction, the changes of the system are accurately captured to form a feature vector FTR, and a machine learning algorithm is further used to build a fault prediction model, generate a fault prediction index in real time and perform trend analysis, which can effectively evaluate the current operating status, and generate and execute corresponding control strategies based on early warning. In addition, after executing the control, the system ensures the control effect and further optimizes the system status through fixed-cycle feedback and iterative evaluation. The fault prediction accuracy and response speed of the switching power supply system are greatly improved, and the risk of system shutdown or damage caused by sudden problems is reduced, thereby ensuring the continuous and stable operation of the equipment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of automatic control, in particular to a switch power supply control system. Background Art

[0002] Switching power supply control systems are mainly used to convert, regulate and control electrical energy. In the field of power electronics, switching power supplies are widely used in various electronic devices and industrial control systems due to their efficient and reliable power conversion capabilities. Specifically in industrial automation control systems, switching power supplies are usually used to provide a stable power supply for key equipment such as controllers, sensors, and actuators.

[0003] In industrial automation control systems, although switching power supplies have high efficiency and reliability, they may still encounter various faults in long-term operation, including overvoltage, overcurrent and overheating. These faults are usually caused by equipment aging, environmental changes or load fluctuations, and are often difficult to predict in advance. In existing control systems, passive protection measures such as overcurrent protection and overvoltage protection are usually relied on to prevent further expansion of faults. However, this passive protection method cannot provide early warning before the fault occurs, resulting in the system being affected when the fault occurs, which may cause downtime or equipment damage. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a switch power supply control system, which solves the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a switching power supply control system, including a data acquisition module, a data processing and feature extraction module, a fault prediction model construction module, a fault trend analysis module, an early warning and response module and a feedback and iterative optimization module;

[0006] The data acquisition module collects the operating parameters of the switching power supply in real time through a sensor group integrated in the operating area of ​​the switching power supply to form a collection data group PRM;

[0007] The data processing and feature extraction module performs preprocessing on the collected data group PRM, including denoising preprocessing and normalization preprocessing, and then performs feature extraction on the preprocessed collected data group PRM to obtain temperature gradient feature DT, current fluctuation rate feature DC, voltage fluctuation rate feature DV, electromagnetic interference feature DI, vibration feature VB and power loss feature PL to form a feature vector FTR;

[0008] The fault prediction model building module uses a machine learning algorithm to build a fault prediction model based on the acquired feature vector FTR, obtains a fault prediction index PE through training and analysis of the built fault prediction model, and matches it with a preset fault prediction evaluation threshold GZ to obtain a fault prediction evaluation result;

[0009] The fault trend analysis module analyzes the trend of the triggering time series according to the fault prediction evaluation result, evaluates the system operation status, and obtains the fault trend accumulation factor TH;

[0010] The early warning and response module performs fitting according to the fault prediction index PE and the fault trend accumulation factor TH, obtains the system operation status evaluation coefficient HS, matches it with the preset system operation status evaluation threshold PZ, obtains the system operation status evaluation result, and generates a control system operation status strategy plan according to the system operation status evaluation result;

[0011] The feedback and iterative optimization module records the fixed-cycle response state of the system that executes the control system operation status strategy program content, obtains the iterative evaluation index DZ, and matches it with the preset system iterative control execution threshold TZ to obtain the system operation status result after control.

[0012] Preferably, the data acquisition module collects the operating parameters of the switching power supply in real time through a sensor group integrated in the working area of ​​the switching power supply operation device to form a collection data group PRM, including a temperature value TMP, a current value CUR, a voltage value VOL, an electromagnetic interference intensity value EMI, a vibration frequency value VIB, an input power value PWRin and an output power value PWRou;

[0013] Among them, the sensor group includes a temperature sensor, a current sensor, a voltage sensor, an electromagnetic interference sensor, a vibration sensor and a power sensor.

[0014] Preferably, the data processing and feature extraction module includes a data preprocessing unit and a feature extraction unit;

[0015] The data preprocessing unit performs preprocessing on the collected data group PRM, including denoising preprocessing and normalization preprocessing, to form a preprocessed collected data group PRMy;

[0016] Among them, the denoising preprocessing includes using a low-pass filter to filter the noise of the collected data group PRM; the normalization preprocessing includes using MAX and MIN normalization processing to process the noise-filtered collected data group PRM dimensionlessly;

[0017] The feature extraction unit performs feature extraction on the preprocessed collected data group PRMy to obtain temperature gradient features DT, current fluctuation rate features DC, voltage fluctuation rate features DV, electromagnetic interference features DI, vibration features VB and power loss features PL to form a feature vector FTR, specifically FTR={DT, DC, DV, DI, VB, PL}.

[0018] Preferably, the temperature gradient feature DT is obtained by the following feature extraction formula:

[0019] ;

[0020] Where N represents the total number of sampling points in a fixed time window, TMPmax represents the peak temperature value among the total number of sampling points N, and TMPmin represents the valley temperature value among the number of sampling points N;

[0021] The current fluctuation rate feature DC is obtained by the following feature extraction formula:

[0022] ;

[0023] Where CURi represents the current value sampled at the i-th time. It represents the average value of the CUR current values ​​obtained from the total number of sampling points N;

[0024] The voltage fluctuation rate feature DV is obtained by the following feature extraction formula:

[0025] ;

[0026] In the formula, It represents the voltage value obtained from the total number of sampling points N. It represents the average value of the VOL voltage value obtained from the total number of sampling points N;

[0027] The electromagnetic interference feature DI is obtained by the following feature extraction formula:

[0028] ;

[0029] Where EMIpeak represents the peak electromagnetic interference intensity value EMI obtained from the total number of sampling points N. It represents the average value of the EMI electromagnetic interference intensity values ​​obtained from the total number of sampling points N;

[0030] The vibration feature VB is obtained by the following feature extraction formula:

[0031] ;

[0032] Where VIBpeak represents the peak vibration frequency value VIB obtained from the total number of sampling points N. It represents the average value of the vibration frequency value VIB obtained from the total number of sampling points N;

[0033] The power loss feature PL is obtained by the following feature extraction formula:

[0034] ;

[0035] Wherein, PWRin represents the input power value, and PWRout represents the output power value.

[0036] Preferably, the fault prediction model building module includes a modeling unit and a fault matching unit;

[0037] The modeling unit uses a machine learning algorithm to construct a fault prediction model according to the acquired feature vector FTR, and obtains a fault prediction index PE by training and analyzing the constructed fault prediction model;

[0038] The fault matching unit matches the fault prediction index PE with the preset fault prediction evaluation threshold GZ. The matching method obtains the fault prediction evaluation results, and triggers the fault trend analysis strategy according to the fault prediction evaluation results;

[0039] When the return signal is 1, the fault prediction evaluation result is abnormal, the fault trend analysis strategy is triggered, and the fault prediction index PE is synchronously sent to the fault trend analysis module for trend analysis;

[0040] When the return signal is 0, the fault prediction evaluation result is normal and the fault trend analysis strategy is not triggered.

[0041] Preferably, the fault prediction index PE is obtained by the following calculation formula:

[0042] ;

[0043] Wherein, tanh() represents the hyperbolic tangent function, specifically represents nonlinear mapping, and compresses the mapping result to the range of [-1, 1]. DT*DC represents the association between the temperature gradient feature DT and the current fluctuation rate feature DC, thereby reflecting the mutual influence. ln(1+DV) represents the logarithmic processing of the voltage fluctuation rate feature DV, which is specifically used to adjust the influence of the voltage fluctuation rate feature DV on the control system. Indicates that the absolute value of the electromagnetic interference characteristic DI takes the cube root, which is specifically used to adjust the nonlinear influence of the electromagnetic interference characteristic DI. Sin represents the sine function. It means that the vibration feature VB is processed by sine function and the processing result is mapped to the range of [0, 1]. cos represents the cosine function. It means that the power loss feature PL is processed by the cosine function and the processing result is mapped into the range of [0, 1].

[0044] Preferably, the fault trend analysis module performs a trigger time series trend analysis according to the fault prediction evaluation result, triggers the evaluation mechanism when the fault prediction evaluation result returns a signal of 1, evaluates the system operation status, and obtains the fault trend accumulation factor TH;

[0045] The fault trend accumulation factor TH is obtained by the following calculation formula:

[0046] ;

[0047] Where T represents the total number of fixed period detections, and j represents the jth detection within the fixed period.

[0048] Preferably, the early warning and response module includes a state assessment unit and a strategy generation unit;

[0049] The state evaluation unit performs fitting according to the fault prediction index PE and the fault trend accumulation factor TH to obtain the system operation state evaluation coefficient HS;

[0050] The system operation status evaluation coefficient HS is obtained by the following calculation formula:

[0051] ;

[0052] The strategy generation unit matches the preset system operation status evaluation threshold value PZ with the system operation status evaluation coefficient HS to obtain the system operation status evaluation result, and generates a control system operation status strategy scheme according to the system operation status evaluation result.

[0053] Preferably, the system operation status evaluation result is obtained by The matching method is obtained, and the strategy scheme for regulating the operation status of the system is generated according to the return signal of the system operation status evaluation result;

[0054] When the return signal of the system operation status evaluation result is 1, the system operation status evaluation result is obtained as abnormal, and a strategy scheme for regulating the system operation status is generated, including adjusting the frequency of the control switch and triggering the switch protection mechanism;

[0055] When the system operation status evaluation result return signal is 0, the system operation status evaluation result is obtained to be qualified, no control system operation status strategy is generated, and no control switch is regulated.

[0056] Preferably, the feedback and iterative optimization module includes a response state evaluation unit and an iterative matching unit;

[0057] The response state evaluation unit records the fixed-period response state of the system that executes the content of the control system operation state strategy plan, and obtains the iterative evaluation index DZ;

[0058] The iterative evaluation index DZ is obtained by the following calculation formula:

[0059] ;

[0060] Where, PEpost and PEprev represent the fault prediction index PE recorded after and before the execution of the control system operation status strategy plan, THpost and THprev represent the fault trend accumulation factor TH recorded after and before the execution of the control system operation status strategy plan, respectively;

[0061] The iterative matching unit performs the matching through the preset system iterative control execution threshold TZ and the iterative evaluation index DZ. Mode matching, obtain the system operation status result after regulation;

[0062] When the return match return signal is 1, the result of obtaining the regulated system operation status is unqualified, triggering the operation mechanism of iterative execution of the fault trend analysis module and the early warning and response module;

[0063] When the return match return signal is 0, the result of obtaining the regulated system operation status is qualified, and the iteration mechanism is not triggered.

[0064] The present invention provides a switching power supply control system, which has the following beneficial effects:

[0065] (1) When the system is running, the stability and reliability of the system are continuously improved through feedback and iterative optimization. The integrated sensor group is used to collect key operating parameters in real time. Through data processing and feature extraction, the changes in the system's temperature, current, voltage, electromagnetic interference, vibration and power loss are accurately captured to form a feature vector FTR. The machine learning algorithm is further used to build a fault prediction model, generate a fault prediction index in real time and perform trend analysis. It can effectively evaluate the current operating status and generate and execute corresponding control strategies based on early warning. In addition, after executing the control, the system ensures the control effect and further optimizes the system status through fixed-cycle feedback and iterative evaluation. The fault prediction accuracy and response speed of the switching power supply system are greatly improved, and the risk of system shutdown or damage caused by sudden problems is reduced, thereby ensuring the continuous and stable operation of the equipment and effectively extending the service life of the system. It effectively solves the problem that the traditional system is not timely enough in fault prediction and the control response is delayed.

[0066] (2) Based on the feature vector FTR, the fault prediction model is accurately constructed. Through nonlinear mapping, multiple factors such as temperature, current, voltage, electromagnetic interference, vibration and power loss are comprehensively considered to obtain a more sensitive fault prediction index PE. In addition, accurate matching is performed through the preset fault prediction evaluation threshold GZ to quickly determine whether there is an abnormal state. When an abnormal signal is detected, the system immediately triggers the fault trend analysis, and further performs time series analysis on the fault trend to obtain the fault trend accumulation factor TH, thereby providing sufficient data support for the system's early warning and subsequent regulation. It improves the accuracy and response speed of fault detection, and can effectively capture early signs of faults, so that the system can take appropriate preventive measures when the fault risk first appears, reducing the impact of sudden faults on the system.

[0067] (3) Through the state evaluation unit, the current operating state can be accurately evaluated, the system operating state evaluation coefficient HS can be calculated, and targeted control strategies can be generated under abnormal conditions, including adjusting the switching frequency or triggering the protection mechanism. Subsequently, the feedback and iterative optimization module periodically evaluates the control effect, calculates the iterative evaluation index DZ, and decides whether the control mechanism needs to be triggered again based on the evaluation results. The closed-loop feedback mechanism not only ensures the effectiveness of the control strategy, but also enables the system to gradually adapt to the complex operating environment through continuous iterative optimization, thereby improving its long-term stability and self-healing ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 The present invention is a schematic diagram of a switch power supply control system block diagram. DETAILED DESCRIPTION

[0069] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0070] Example 1

[0071] The present invention provides a switch power supply control system, please refer to Figure 1 , including data acquisition module, data processing and feature extraction module, fault prediction model building module, fault trend analysis module, early warning and response module and feedback and iterative optimization module;

[0072] The data acquisition module collects the operating parameters of the switching power supply in real time through the sensor group integrated in the switching power supply operation area to form a collection data group PRM;

[0073] The data processing and feature extraction module preprocesses the collected data group PRM, including denoising preprocessing and normalization preprocessing, and then extracts features from the preprocessed collected data group PRM to obtain temperature gradient features DT, current fluctuation rate features DC, voltage fluctuation rate features DV, electromagnetic interference features DI, vibration features VB and power loss features PL to form a feature vector FTR;

[0074] The fault prediction model building module uses a machine learning algorithm to build a fault prediction model based on the acquired feature vector FTR. By training and analyzing the built fault prediction model, the fault prediction index PE is obtained and matched with the preset fault prediction evaluation threshold GZ to obtain the fault prediction evaluation result.

[0075] The fault trend analysis module analyzes the trend of the triggering time series according to the fault prediction evaluation results, evaluates the system operation status, and obtains the fault trend accumulation factor TH;

[0076] The early warning and response module performs fitting based on the fault prediction index PE and the fault trend accumulation factor TH, obtains the system operation status evaluation coefficient HS, matches it with the preset system operation status evaluation threshold PZ, obtains the system operation status evaluation result, and generates a control system operation status strategy plan based on the system operation status evaluation result;

[0077] The feedback and iterative optimization module records the fixed-cycle response status of the system that executes the control system operation status strategy plan, obtains the iterative evaluation index DZ, and matches it with the preset system iterative control execution threshold TZ to obtain the system operation status result after control.

[0078] In this embodiment, the stability and reliability of the system are continuously improved through feedback and iterative optimization, and the key operating parameters are collected in real time by using an integrated sensor group. Through data processing and feature extraction, the changes in the system's temperature, current, voltage, electromagnetic interference, vibration and power loss are accurately captured to form a feature vector FTR. The machine learning algorithm is further used to build a fault prediction model, and the fault prediction index is generated in real time and trend analysis is performed. It can effectively evaluate the current operating status and generate and execute corresponding control strategies based on early warning. In addition, after executing the control, the system ensures the control effect and further optimizes the system status through fixed-cycle feedback and iterative evaluation. The fault prediction accuracy and response speed of the switching power supply system are greatly improved, and the risk of system shutdown or damage caused by sudden problems is reduced, thereby ensuring the continuous and stable operation of the equipment and effectively extending the service life of the system, effectively solving the problem that the traditional system is not timely enough in fault prediction and the control response is delayed.

[0079] Example 2

[0080] This embodiment is explained in Example 1, please refer to Figure 1 Specifically: the data acquisition module collects the operating parameters of the switching power supply in real time through the sensor group integrated in the working area of ​​the switching power supply operation equipment to form a collection data group PRM, including temperature value TMP, current value CUR, voltage value VOL, electromagnetic interference intensity value EMI, vibration frequency value VIB, input power value PWRin and output power value PWRou;

[0081] Among them, the sensor group includes a temperature sensor, a current sensor, a voltage sensor, an electromagnetic interference sensor, a vibration sensor and a power sensor.

[0082] The data processing and feature extraction module includes a data preprocessing unit and a feature extraction unit;

[0083] The data preprocessing unit preprocesses the collected data group PRM, including denoising preprocessing and normalization preprocessing, to form a preprocessed collected data group PRMy;

[0084] Among them, the denoising preprocessing includes using a low-pass filter to filter the noise of the collected data group PRM; the normalization preprocessing includes using MAX and MIN normalization processing to process the noise-filtered collected data group PRM dimensionlessly;

[0085] The feature extraction unit performs feature extraction on the preprocessed collected data group PRMy to obtain the temperature gradient feature DT, the current fluctuation rate feature DC, the voltage fluctuation rate feature DV, the electromagnetic interference feature DI, the vibration feature VB and the power loss feature PL to form a feature vector FTR, specifically FTR={DT, DC, DV, DI, VB, PL}.

[0086] The temperature gradient feature DT is obtained by the following feature extraction formula:

[0087] ;

[0088] Where N represents the total number of sampling points in a fixed time window, TMPmax represents the peak temperature value among the total number of sampling points N, and TMPmin represents the valley temperature value among the number of sampling points N;

[0089] The current fluctuation rate feature DC is obtained by the following feature extraction formula:

[0090] ;

[0091] Where CURi represents the current value sampled at the i-th time. It represents the average value of the CUR current values ​​obtained from the total number of sampling points N;

[0092] The voltage fluctuation rate feature DV is obtained by the following feature extraction formula:

[0093] ;

[0094] In the formula, It represents the voltage value obtained from the total number of sampling points N. It represents the average value of the VOL voltage value obtained from the total number of sampling points N;

[0095] The electromagnetic interference feature DI is obtained by the following feature extraction formula:

[0096] ;

[0097] Where EMIpeak represents the peak electromagnetic interference intensity value EMI obtained from the total number of sampling points N. It represents the average value of the EMI electromagnetic interference intensity values ​​obtained from the total number of sampling points N;

[0098] The vibration feature VB is obtained by the following feature extraction formula:

[0099] ;

[0100] Where VIBpeak represents the peak vibration frequency value VIB obtained from the total number of sampling points N. It represents the average value of the vibration frequency value VIB obtained from the total number of sampling points N;

[0101] The power loss feature PL is obtained by the following feature extraction formula:

[0102] ;

[0103] Wherein, PWRin represents the input power value, and PWRout represents the output power value.

[0104] Example 3

[0105] This embodiment is explained in Example 2. Please refer to Figure 1 ,Specifically: the fault prediction model building module includes a ,modeling unit and a fault matching unit;

[0106] The modeling unit uses a machine learning algorithm to build a fault prediction model based on the acquired feature vector FTR, and obtains a fault prediction index PE through training and analysis of the built fault prediction model;

[0107] The fault matching unit matches the fault prediction evaluation threshold GZ with the fault prediction index PE. The matching method obtains the fault prediction evaluation results, and triggers the fault trend analysis strategy according to the fault prediction evaluation results;

[0108] When the return signal is 1, the fault prediction evaluation result is abnormal, the fault trend analysis strategy is triggered, and the fault prediction index PE is synchronously sent to the fault trend analysis module for trend analysis;

[0109] When the return signal is 0, the fault prediction evaluation result is normal and the fault trend analysis strategy is not triggered.

[0110] The fault prediction index PE is obtained by the following calculation formula:

[0111] ;

[0112] Wherein, tanh() represents the hyperbolic tangent function, specifically represents nonlinear mapping, and compresses the mapping result to the range of [-1, 1]. DT*DC represents the association between the temperature gradient feature DT and the current fluctuation rate feature DC, thereby reflecting the mutual influence. ln(1+DV) represents the logarithmic processing of the voltage fluctuation rate feature DV, which is specifically used to adjust the influence of the voltage fluctuation rate feature DV on the control system. Indicates that the absolute value of the electromagnetic interference characteristic DI takes the cube root, which is specifically used to adjust the nonlinear influence of the electromagnetic interference characteristic DI. Sin represents the sine function. It means that the vibration feature VB is processed by sine function and the processing result is mapped to the range of [0, 1]. cos represents the cosine function. It means that the power loss feature PL is processed by the cosine function and the processing result is mapped into the range of [0, 1].

[0113] The fault trend analysis module triggers the time series trend analysis according to the fault prediction evaluation result. When the fault prediction evaluation result returns a signal of 1, the evaluation mechanism is triggered to evaluate the system operation status and obtain the fault trend accumulation factor TH.

[0114] The fault trend accumulation factor TH is obtained by the following calculation formula:

[0115] ;

[0116] Where T represents the total number of fixed period detections, and j represents the jth detection within the fixed period.

[0117] In this embodiment, a fault prediction model is accurately constructed based on the feature vector FTR, and multiple factors such as temperature, current, voltage, electromagnetic interference, vibration and power loss are comprehensively considered through nonlinear mapping to obtain a more sensitive fault prediction index PE. In addition, accurate matching is performed through the preset fault prediction evaluation threshold GZ to quickly determine whether there is an abnormal state. When an abnormal signal is detected, the system immediately triggers a fault trend analysis, further performs a time series analysis on the fault trend, and obtains the fault trend accumulation factor TH, thereby providing sufficient data support for the system's early warning and subsequent regulation. The accuracy and response speed of fault detection are improved, and early signs of faults can be effectively captured, so that the system can take appropriate preventive measures when the risk of faults first appears, reducing the impact of sudden faults on the system.

[0118] Example 4

[0119] This embodiment is explained in Example 3, please refer to Figure 1 ,Specifically: the early warning and response module includes a ,status evaluation unit and a strategy generation unit;

[0120] The state evaluation unit performs fitting based on the fault prediction index PE and the fault trend accumulation factor TH to obtain the system operation state evaluation coefficient HS;

[0121] The system operation status evaluation coefficient HS is obtained by the following calculation formula:

[0122] ;

[0123] The strategy generation unit matches the preset system operation status evaluation threshold PZ with the system operation status evaluation coefficient HS to obtain the system operation status evaluation result, and generates a control system operation status strategy scheme according to the system operation status evaluation result.

[0124] The system operation status assessment results passed The matching method is obtained, and the strategy scheme for regulating the operation status of the system is generated according to the return signal of the system operation status evaluation result;

[0125] When the return signal of the system operation status evaluation result is 1, the system operation status evaluation result is obtained as abnormal, and a strategy scheme for regulating the system operation status is generated, including adjusting the frequency of the control switch and triggering the switch protection mechanism;

[0126] When the system operation status evaluation result return signal is 0, the system operation status evaluation result is obtained to be qualified, no control system operation status strategy is generated, and no control switch is regulated.

[0127] The feedback and iterative optimization module includes a response state evaluation unit and an iterative matching unit;

[0128] The response state evaluation unit records the fixed-cycle response state of the system that executes the content of the control system operation state strategy plan and obtains the iterative evaluation index DZ;

[0129] The iterative evaluation index DZ is obtained by the following calculation formula:

[0130] ;

[0131] Where, PEpost and PEprev represent the fault prediction index PE recorded after and before the execution of the control system operation status strategy plan, THpost and THprev represent the fault trend accumulation factor TH recorded after and before the execution of the control system operation status strategy plan, respectively;

[0132] The iterative matching unit performs the iterative matching through the preset system iterative control execution threshold TZ and the iterative evaluation index DZ. Mode matching, obtain the system operation status result after regulation;

[0133] When the return match return signal is 1, the result of obtaining the regulated system operation status is unqualified, triggering the operation mechanism of iterative execution of the fault trend analysis module and the early warning and response module;

[0134] When the return match return signal is 0, the result of obtaining the regulated system operation status is qualified, and the iteration mechanism is not triggered.

[0135] In this embodiment, the state evaluation unit can accurately evaluate the current operating state, calculate the system operating state evaluation coefficient HS, and generate targeted control strategies in abnormal states, including adjusting the switching frequency or triggering the protection mechanism. Subsequently, the feedback and iterative optimization module periodically evaluates the control effect, calculates the iterative evaluation index DZ, and decides whether to trigger the control mechanism again based on the evaluation results. The closed-loop feedback mechanism not only ensures the effectiveness of the control strategy, but also enables the system to gradually adapt to the complex operating environment through continuous iterative optimization, thereby improving its long-term stability and self-healing ability.

[0136] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A switching power supply control system, characterized in that: It includes data acquisition module, data processing and feature extraction module, fault prediction model building module, fault trend analysis module, early warning and response module and feedback and iterative optimization module; The data acquisition module collects the operating parameters of the switching power supply in real time through a sensor group integrated in the operating area of ​​the switching power supply to form a collection data group PRM; The data processing and feature extraction module performs preprocessing on the collected data group PRM, including denoising preprocessing and normalization preprocessing, and then performs feature extraction on the preprocessed collected data group PRM to obtain temperature gradient feature DT, current fluctuation rate feature DC, voltage fluctuation rate feature DV, electromagnetic interference feature DI, vibration feature VB and power loss feature PL. Composition feature vector FTR; The fault prediction model building module uses a machine learning algorithm to build a fault prediction model based on the acquired feature vector FTR, obtains a fault prediction index PE through training and analysis of the built fault prediction model, and matches it with a preset fault prediction evaluation threshold GZ to obtain a fault prediction evaluation result; The fault prediction model building module includes a modeling unit and a fault matching unit; The modeling unit uses a machine learning algorithm to construct a fault prediction model according to the acquired feature vector FTR, and obtains a fault prediction index PE by training and analyzing the constructed fault prediction model; The fault matching unit matches the fault prediction index PE with the preset fault prediction evaluation threshold GZ. The matching method obtains the fault prediction evaluation results, and triggers the fault trend analysis strategy according to the fault prediction evaluation results; When the return signal is 1, the fault prediction evaluation result is abnormal, the fault trend analysis strategy is triggered, and the fault prediction index PE is synchronously sent to the fault trend analysis module for trend analysis; When the return signal is 0, the fault prediction evaluation result is normal and the fault trend analysis strategy is not triggered; The fault prediction index PE is obtained by the following calculation formula: Wherein, tanh() represents the hyperbolic tangent function, specifically represents nonlinear mapping, and compresses the mapping result to the range of [-1, 1]. DT*DC represents the association between the temperature gradient feature DT and the current fluctuation rate feature DC, thereby reflecting the mutual influence. ln(1+DV) represents the logarithmic processing of the voltage fluctuation rate feature DV, which is specifically used to adjust the influence of the voltage fluctuation rate feature DV on the control system. Indicates that the absolute value of the electromagnetic interference characteristic DI takes the cube root, which is specifically used to adjust the nonlinear influence of the electromagnetic interference characteristic DI. Sin represents the sine function. Indicates that the vibration feature VB is processed by a sine function and the processing result is mapped to the range of [0, 1]. cos represents the cosine function. Indicates that the power loss feature PL is processed by the cosine function and the processing result is mapped to the range of [0, 1]; The fault trend analysis module analyzes the trend of the triggering time series according to the fault prediction evaluation result, evaluates the system operation status, and obtains the fault trend accumulation factor TH; The fault trend analysis module triggers the time series trend analysis according to the fault prediction evaluation result, triggers the evaluation mechanism when the fault prediction evaluation result returns a signal of 1, evaluates the system operation status, and obtains the fault trend accumulation factor TH; The fault trend accumulation factor TH is obtained by the following calculation formula: In the formula, T represents the total number of fixed period detections, and j represents the jth detection in the fixed period; The early warning and response module performs fitting according to the fault prediction index PE and the fault trend accumulation factor TH, obtains the system operation status evaluation coefficient HS, matches it with the preset system operation status evaluation threshold PZ, obtains the system operation status evaluation result, and generates a control system operation status strategy plan according to the system operation status evaluation result; The feedback and iterative optimization module records the fixed-cycle response state of the system that executes the control system operation status strategy program content, obtains the iterative evaluation index DZ, and matches it with the preset system iterative control execution threshold TZ to obtain the system operation status result after control.

2. The switching power supply control system according to claim 1, characterized in that: The data acquisition module collects the operating parameters of the switching power supply in real time through the sensor group integrated in the working area of ​​the switching power supply operation equipment to form a collection data group PRM, including temperature value TMP, current value CUR, voltage value VOL, electromagnetic interference intensity value EMI, vibration frequency value VIB, input power value PWRin and output power value PWRou; Among them, the sensor group includes a temperature sensor, a current sensor, a voltage sensor, an electromagnetic interference sensor, a vibration sensor and a power sensor.

3. The switching power supply control system according to claim 2, characterized in that: The data processing and feature extraction module includes a data preprocessing unit and a feature extraction unit; The data preprocessing unit performs preprocessing on the collected data group PRM, including denoising preprocessing and normalization preprocessing, to form a preprocessed collected data group PRMy; Among them, the denoising preprocessing includes using a low-pass filter to filter the noise of the collected data group PRM; the normalization preprocessing includes using MAX and MIN normalization processing to process the noise-filtered collected data group PRM dimensionlessly; The feature extraction unit performs feature extraction on the preprocessed collected data group PRMy, obtains temperature gradient feature DT, current fluctuation rate feature DC, voltage fluctuation rate feature DV, electromagnetic interference feature DI, vibration feature VB and power loss feature PL, and forms a feature vector FTR, specifically FTR = {DT, DC, DV, DI, VB, PL}.

4. The switching power supply control system according to claim 3, characterized in that: The temperature gradient feature DT is obtained by the following feature extraction formula: Where N represents the total number of sampling points in a fixed time window, TMPmax represents the peak temperature value among the total number of sampling points N, and TMPmin represents the valley temperature value among the number of sampling points N; The current fluctuation rate feature DC is obtained by the following feature extraction formula: Where CURi represents the current value sampled at the i-th time. It represents the average value of the current value CUR obtained by statistics in the total number of sampling points N; The voltage fluctuation rate feature DV is obtained by the following feature extraction formula: Where σVOL represents the voltage value obtained from the total number of sampling points N. Represents the average value of the voltage value VOL obtained by statistics in the total number of sampling points N; The electromagnetic interference feature DI is obtained by the following feature extraction formula: Where EMIpeak represents the peak electromagnetic interference intensity value EMI obtained from the total number of sampling points N. It represents the average value of the electromagnetic interference intensity value EMI obtained by statistics from the total number of sampling points N; The vibration feature VB is obtained by the following feature extraction formula: Where VIBpeak represents the peak vibration frequency value VIB obtained from the total number of sampling points N. It represents the average value of the vibration frequency value VIB obtained from the total number of sampling points N; The power loss feature PL is obtained by the following feature extraction formula: Wherein, PWRin represents the input power value, and PWRout represents the output power value.

5. The switching power supply control system according to claim 1, characterized in that: The early warning and response module includes a state assessment unit and a strategy generation unit; The state evaluation unit performs fitting according to the fault prediction index PE and the fault trend accumulation factor TH to obtain the system operation state evaluation coefficient HS; The system operation status evaluation coefficient HS is obtained by the following calculation formula: The strategy generation unit matches the preset system operation status evaluation threshold value PZ with the system operation status evaluation coefficient HS to obtain the system operation status evaluation result, and generates a control system operation status strategy scheme according to the system operation status evaluation result.

6. The switching power supply control system according to claim 5, characterized in that: The system operation status evaluation result is The matching method is obtained, and the strategy scheme for regulating the operating status of the system is generated according to the return signal of the system operating status evaluation result; When the return signal of the system operation status evaluation result is 1, the system operation status evaluation result is obtained as abnormal, and a strategy scheme for regulating the system operation status is generated, including adjusting the frequency of the control switch and triggering the switch protection mechanism; When the system operation status evaluation result return signal is 0, the system operation status evaluation result is obtained to be qualified, no control system operation status strategy is generated, and the control switch is not regulated.

7. The switching power supply control system according to claim 1, characterized in that: The feedback and iterative optimization module includes a response state evaluation unit and an iterative matching unit; The response state evaluation unit records the fixed-period response state of the system that executes the content of the control system operation state strategy plan, and obtains the iterative evaluation index DZ; The iterative evaluation index DZ is obtained by the following calculation formula: Where, PEpost and PEprev represent the fault prediction index PE recorded after and before the execution of the control system operation status strategy plan, THpost and THprev represent the fault trend accumulation factor TH recorded after and before the execution of the control system operation status strategy plan, respectively; The iterative matching unit performs the matching through the preset system iterative control execution threshold TZ and the iterative evaluation index DZ. Mode matching, obtain the system operation status result after regulation; When the return match return signal is 1, the result of obtaining the regulated system operation status is unqualified, triggering the operation mechanism of iterative execution of the fault trend analysis module and the early warning and response module; When the return match return signal is 0, the result of obtaining the system operation status after regulation is qualified, and the iteration mechanism is not triggered.