Server power supply detection method and device, storage medium and electronic equipment

Abstract

The present application discloses a server power supply detection method and device, a storage medium, and an electronic device, which relate to the field of computer technology. The method comprises the following steps: obtaining multidimensional operation data of the server power supply; performing feature extraction on the multidimensional operation data to obtain a power consumption feature of the server, wherein the power consumption feature is used to indicate a temporal variation relationship of the server's demand for electric energy; generating a decay feature of the server power supply under the power consumption feature based on the multidimensional operation data; and detecting the remaining service life of the server power supply at a target time after the current time based on the decay feature and the power consumption feature. The method solves the problem of low detection efficiency of the server power supply in related technologies and achieves the technical effect of improving the detection efficiency of the server power supply.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method and device for detecting a server power supply, a storage medium, and an electronic device. Background Art

[0002] The server power supply is essential for the normal operation of the server. It provides electrical energy to the server, thereby ensuring its normal operation. To better ensure the operational quality of the server power supply, power module life prediction technology has been proposed in the current field. Power module life prediction technology aims to predict the remaining useful life (RUL) of the power module by monitoring and analyzing its operating status, thereby providing a basis for maintenance planning and fault prevention.

[0003] Current power module life prediction technology usually predicts the remaining service life of the power supply based on the number of times the power supply is used. This prediction method is relatively one-sided, and the number of times the power supply is used cannot truly reflect the actual usage status of the server power supply, which leads to inaccurate prediction results. Summary of the Invention

[0004] The present application provides a method and device for detecting a server power supply, a storage medium, and an electronic device, so as to at least solve the problem of low detection efficiency of a server power supply in the related art.

[0005] The present application provides a method for detecting a server power supply, comprising: obtaining multidimensional operating data of the server power supply, wherein the multidimensional operating data is used to characterize, from multiple feature dimensions, the power supply status of the server power supply to the server within a reference time period before a current moment; performing feature extraction on the multidimensional operating data to obtain a power consumption characteristic of the server, wherein the power consumption characteristic is used to indicate a temporal variation relationship of the server's demand for electric energy; generating, based on the multidimensional operating data, a decay characteristic of the server power supply under the power consumption characteristic, wherein the decay characteristic is used to indicate an influence relationship of the power consumption characteristic on the decay of the server power supply; and detecting, based on the decay characteristic and the power consumption characteristic, a remaining service life of the server power supply at a target moment after the current moment.

[0006] The present application also provides a detection device for a server power supply, comprising: an acquisition module for acquiring multidimensional operation data of the server power supply, wherein the multidimensional operation data is used to characterize the power supply status of the server power supply to the server within a reference time period before the current moment from multiple feature dimensions; an extraction module for performing feature extraction on the multidimensional operation data to obtain the power consumption characteristics of the server, wherein the power consumption characteristics are used to indicate the temporal change relationship of the server's demand for electric energy; a generation module for generating, based on the multidimensional operation data, a decay characteristic of the server power supply under the power consumption characteristics, wherein the decay characteristic is used to indicate the influence relationship of the power consumption characteristics on the decay of the server power supply; a first detection module for detecting the remaining service life of the server power supply at a target moment after the current moment based on the decay characteristics and the power consumption characteristics.

[0007] The present application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for implementing the steps of any of the above-mentioned server power supply detection methods when executing the computer program.

[0008] The present application also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned server power supply detection methods are implemented.

[0009] The present application also provides a computer program product, including a computer program, which implements the steps of any of the above-mentioned server power supply detection methods when executed by a processor.

[0010] Through the present application, by extracting features from multidimensional operating data used to characterize the power supply status of the server power supply to the server from multiple dimensions, the power consumption characteristics of the server are obtained to indicate the time-series change relationship of the server's power demand, thereby realizing the reflection of the server's actual power usage demand for the server power supply based on the multidimensional operating data, thereby comprehensively reflecting the actual usage status of the server power supply, and then generating the decay characteristics of the server power supply under the power consumption characteristics based on the multidimensional operating data, detecting the remaining service life of the server based on the decay characteristics and the power consumption characteristics, and realizing the prediction of the remaining service life of the server power supply based on the actual usage status of the server power supply and the decay of the server power supply under the actual usage status. Therefore, the technical problem of low detection efficiency of the server power supply in the related technology can be solved, and the technical effect of improving the detection efficiency of the server power supply can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0012] Figure 1 This is a hardware structure block diagram of a server power supply detection method according to an embodiment of the present application;

[0013] Figure 2 is a flow chart of a method for detecting a server power supply according to an embodiment of the present application;

[0014] Figure 3 is a flowchart of power module life prediction according to an embodiment of the present application;

[0015] Figure 4 This is a schematic diagram of an optional edge computing power module life prediction system according to an embodiment of the present application;

[0016] Figure 5 This is a structural block diagram of a server power supply detection device according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0018] It should be noted that, in the description of this application, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. The terms "first," "second," etc., in this application are used to distinguish similar objects, and are not used to describe a particular order or sequence.

[0019] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0020] In conjunction with the specific application environment architecture or specific hardware architecture on which the execution of the server power detection method depends, the specific application environment architecture or specific hardware architecture is described herein.

[0021] The method embodiments provided in the embodiments of the present application can be executed in a server device or a similar computing device. Taking running on a server device as an example, Figure 1 FIG. 1 is a hardware structure diagram of a method for detecting a server power supply according to an embodiment of the present application. Figure 1 As shown, the server device may include one or more ( Figure 1 Only one is shown) a processor 102 (the processor 102 may include but is not limited to a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data. The server device may also include a transmission device 106 and an input / output device 108 for communication functions. It will be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above server device. Figure 1 More or fewer components than shown, or with Figure 1 Different configurations shown.

[0022] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the startup method of the operating system in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, implementing the above-mentioned method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely located relative to the processor 102, and these remote memories can be connected to the server device via a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0023] Transmission device 106 is used to receive or transmit data via a network. A specific example of the aforementioned network may include a wireless network provided by a communication provider of the server device. In one embodiment, transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to enable communication with the Internet. In another embodiment, transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0024] An embodiment of the present application provides a method for detecting a server power supply. The method is described in detail in conjunction with the execution flow of the method for detecting a server power supply.

[0025] The following is an explanation of the professional terms that appear in this application:

[0026] RUL: Remaining Useful Life, which refers to the estimated time that a power module can continue to operate normally under its current conditions.

[0027] FFT: Fast Fourier Transform, an algorithm used to convert a time-domain signal into a frequency-domain representation, helps extract the main frequency components of the signal.

[0028] LSTM: Long Short-Term Memory, a special type of recurrent neural network (RNN) suitable for processing and predicting long-term dependencies in time series data.

[0029] ML: Machine Learning is a type of technology that uses algorithms to analyze data, learn from it, and make predictions or decisions about unknown data based on the learned knowledge.

[0030] IoT: Internet of Things, refers to the communication network between various physical devices connected by the Internet.

[0031] Edge computing: A distributed computing framework that brings computation and data storage closer to the devices where data is collected, rather than relying on remote data centers. This reduces latency and improves efficiency.

[0032] Lightweight machine learning models: These are machine learning models that are optimized to reduce computing resource requirements and are suitable for deployment in resource-constrained environments, such as mobile devices or IoT edge nodes.

[0033] Arrhenius equation: An empirical formula that describes the relationship between chemical reaction rate and temperature, often used in the field of material aging or life prediction, especially when considering the effects of thermal stress.

[0034] Federated learning: A privacy-preserving technique that allows multiple participants to jointly train a shared global model without directly exchanging local datasets, thereby protecting data privacy.

[0035] Pruning optimization: A model compression technique that reduces model size and computational complexity by removing unimportant weights or neurons in a neural network while maintaining model performance as much as possible.

[0036] Wavelet transform: A mathematical tool used for signal processing and analysis, particularly suitable for decomposing and reconstructing non-stationary signals and can effectively remove noise.

[0037] Isolation Forest Algorithm: An ensemble learning-based anomaly detection method, particularly suitable for high-dimensional data sets, that identifies anomalies by constructing multiple isolation trees.

[0038] Homomorphic encryption: A form of encryption that allows certain types of computation to be performed on ciphertext, resulting in a result that, after decryption, is identical to the result of performing the same operation directly on the plaintext. It is widely used in privacy-preserving data analysis.

[0039] In this embodiment, a method for detecting a server power supply is provided. Figure 2 FIG. 1 is a flow chart of a method for detecting a server power supply according to an embodiment of the present application, as shown in FIG. Figure 2 As shown, the method includes the following steps:

[0040] Step S202: Acquire multi-dimensional operation data of the server power supply, wherein the multi-dimensional operation data is used to characterize the power supply status of the server power supply to the server within a reference time period before the current moment from multiple characteristic dimensions;

[0041] Step S204: extracting features from the multi-dimensional operation data to obtain power consumption features of the server, wherein the power consumption features are used to indicate a temporal variation relationship of the power demand of the server;

[0042] Step S206: generating a decay characteristic of the server power supply under the power consumption characteristic according to the multi-dimensional operation data, wherein the decay characteristic is used to indicate an influence relationship of the power consumption characteristic on the decay of the server power supply;

[0043] Step S208 : detecting the remaining service life of the server power supply at a target time after the current time according to the decay characteristics and the power consumption characteristics.

[0044] Through the above steps, by extracting features from the multidimensional operating data used to characterize the power supply status of the server power supply to the server from multiple dimensions, the power consumption characteristics of the server are obtained to indicate the time-series change relationship of the server's power demand, thereby realizing the reflection of the server's actual power usage demand for the server power supply according to the multidimensional operating data, thereby comprehensively reflecting the actual usage status of the server power supply, and then generating the decay characteristics of the server power supply under the power consumption characteristics according to the multidimensional operating data, detecting the remaining service life of the server according to the decay characteristics and the power consumption characteristics, and realizing the prediction of the remaining service life of the server power supply according to the actual usage status of the server power supply and the decay of the server power supply under the actual usage status. Therefore, the technical problem of low detection efficiency of the server power supply in the related technology can be solved, and the technical effect of improving the detection efficiency of the server power supply can be achieved.

[0045] The server power supply detection method provided in this application can be applied to, but is not limited to, a device having a function of detecting the service life of a server power supply, such as a server, or a power supply detection device configured separately for a server, and this solution does not limit this.

[0046] In the embodiment provided in step S202, the server power supply is used to supply power to the server. The multi-dimensional operating data of the server power supply may be, but is not limited to, operating parameters that can reflect the delivery status of the server power supply to the server. The operating data may include, but is not limited to, the current output value of the server power supply, the voltage output value of the server power supply, the power supply temperature of the server power supply, etc. This solution does not limit this.

[0047] Optionally, in an embodiment of the present application, the operating data of the server power supply can be, but is not limited to, obtained by detecting the operating status of the server power supply through a data detection device deployed on the server power supply, for example: using a DS18B20 digital temperature sensor to monitor the internal temperature of the server power supply, obtaining the temperature changes of the server power supply during operation, and obtaining the original temperature data; using an ACS712 Hall effect current sensor to monitor the current intensity flowing through the server power supply, capturing the current fluctuations when the server power supply is working, and then obtaining the original current data.

[0048] In the embodiment provided in step S204, the power consumption characteristics are used to indicate the actual usage status of the server power supply by the server, and feature extraction is performed on the multi-dimensional operation data to obtain the power consumption characteristics of the server. The implementation method may include, but is not limited to: feature extraction of the operation data of each feature dimension to obtain the operation characteristics corresponding to each feature dimension, wherein the operation characteristics are used to indicate the power operation status of the server power supply in the corresponding feature dimension; feature splicing of multiple operation characteristics, and inputting the obtained spliced ​​characteristics into the target prediction model to obtain the power consumption characteristics of the server, wherein the target prediction model records the conversion relationship between the operation characteristics of the server power supply and the power consumption characteristics.

[0049] Optionally, in an embodiment of the present application, feature extraction is performed on the multi-dimensional operating data to obtain the power consumption characteristics of the server. The implementation method can also be: feature extraction is performed on the operating data of each feature dimension to obtain the operating characteristics corresponding to each feature dimension, wherein the operating characteristics are used to indicate the power operating status of the server power supply in the corresponding feature dimension; feature splicing is performed on multiple operating characteristics, and the spliced ​​characteristics obtained by splicing are determined as power consumption characteristics.

[0050] In the embodiment provided in step S206, the method of generating the decay characteristics of the server power supply under the power consumption characteristics based on the multidimensional operation data may be, but is not limited to: extracting target operation data of the target feature dimension from the multidimensional operation data, wherein the target feature dimension is used to characterize the power performance of the server power supply, and the target operation parameters of the target feature dimension may be, but are not limited to, including the maximum energy storage capacity of the power supply, the maximum power supply of the power supply, the peak power output of the power supply, etc.; converting the power decay information of the server power supply according to the target operation data, wherein the power decay information is used to characterize the time series change relationship of the decay amount of the server power supply; performing feature extraction on the power decay information to obtain the decay characteristics of the server power supply corresponding to the power consumption characteristics. In this embodiment, the method for converting the power decay information of the server power supply based on the target operating data can be: determining a reference power life corresponding to the target operating data from the corresponding operating data and power life; calculating the power life decay value of the server power supply at the corresponding moment using the reference power life and the initial power life of the power supply; and constructing a power life decay curve of the power supply according to the power life decay values ​​of the server power supply at each moment within the reference time period, wherein the power decay information includes the power life decay curve. Through the above content, by performing feature processing using operating data in the multidimensional operating data that can reflect the power supply performance of the power supply, the decay characteristics of the server power supply are converted when the server obtains power from the server power supply according to the power consumption method indicated by the power consumption characteristics, thereby making the characteristic dimensions of the power consumption characteristics and the decay characteristics more compatible and improving the characteristic correlation between the power consumption characteristics and the decay characteristics.

[0051] Optionally, in an embodiment of the present application, a method for generating the decay characteristics of the server power supply under the power consumption characteristics based on the multi-dimensional operation data may be, but is not limited to, extracting temperature data of the server power supply within a reference time period from the multi-dimensional operation data, and calculating the decay parameters of the server power supply using the following calculation method for the temperature data at each moment: ,in, represents the ideal gas constant, is the temperature in degrees Celsius, is the activation energy, is the frequency factor, is the thermal stress coefficient; the decay parameters are then sorted according to the time series relationship to obtain a decay sequence of the server power supply, wherein the decay sequence is used to indicate the relationship between the decay amount of the server power supply and time within a reference time period; and feature extraction is performed on the decay sequence of the server power supply to obtain the decay feature of the server power supply corresponding to the power consumption feature.

[0052] In the embodiment provided in step S208, the method of detecting the remaining service life of the server power supply at a target moment after the current moment based on the decay characteristics and the power consumption characteristics can be: inputting the decay characteristics and the power consumption characteristics into a target prediction model to obtain the remaining service life output by the target prediction model, wherein the target prediction model maintains the correlation between the decay characteristics, the power consumption characteristics and the remaining service life, and this solution does not limit this.

[0053] Optionally, in an embodiment of the present application, there is a corresponding relationship between the decay characteristics and the power consumption characteristics, that is, the decay characteristics reflect the decay characteristics of the server power supply when the server power supply obtains electric energy from the server power supply in the power consumption manner indicated by the power consumption characteristics. Therefore, the method of detecting the remaining service life of the server power supply at a target moment after the current moment based on the decay characteristics and the power consumption characteristics can also be: feature fusion of the decay characteristics and the power consumption characteristics to obtain a fusion feature of the server power supply, which reflects the correlation between the power supply characteristics of the server power supply to the server and the decay characteristics, and then the fusion characteristics are input into the target prediction model to obtain the remaining service life output by the target prediction model, wherein the target prediction model is for the correlation between the fusion characteristics and the remaining service life.

[0054] Optionally, in an embodiment of the present application, after detecting the remaining service life of the server power supply at a target moment after the current moment based on the decay characteristics and the power consumption characteristics, power supply maintenance operations can be performed on the server power supply based on the detected remaining service life, thereby ensuring the reliable operation of the server and avoiding the impact of the service life and temperature of the server power supply on business processing during the normal operation of the server.

[0055] As an optional implementation manner, extracting features from the multi-dimensional operation data to obtain the power consumption features of the server includes:

[0056] performing feature extraction on the operation data of each feature dimension in the multi-dimensional operation data to obtain an initial power supply feature, wherein the initial power supply feature is used to indicate an operating state of the server power supply within the reference time period;

[0057] Feature conversion is performed on the multiple initial power supply features corresponding to the multi-dimensional operation data to obtain the power consumption features.

[0058] Optionally, in an embodiment of the present application, the initial power supply characteristics represent the power supply status of the server power supply to the server, and then by performing feature conversion on multiple initial power supply characteristics corresponding to multiple operating data, a power consumption characteristic representing the server's usage status of the server power supply is obtained, thereby realizing the characterization of the server's usage status of the server power supply through the multi-dimensional operating data of the server power supply.

[0059] Optionally, in an embodiment of the present application, a method for performing feature conversion on multiple initial power supply features to obtain power consumption features may be: obtaining a reference feature weight for each initial power supply feature, wherein each reference feature weight is used to characterize the degree of influence of the initial power supply feature of the corresponding dimension on the power demand of the server, and then using the reference feature weight to perform weighted fusion on the multiple initial power supply features to obtain the power consumption feature. The reference feature weight may be a pre-set fixed weight, or may be obtained by weighting multi-dimensional operating data, which is not limited in this solution.

[0060] Through the above content, by performing feature conversion on the initial power supply characteristics that characterize the operating status of the server power supply, it is possible to reflect the actual usage requirements of the server for the server power supply based on the multi-dimensional operating data of the server power supply, thereby improving the accuracy of the remaining service life prediction results of the server power supply.

[0061] As an optional implementation manner, performing feature conversion on the multiple initial power supply features corresponding to the multi-dimensional operation data to obtain the power consumption feature includes:

[0062] Sorting the initial power supply characteristics of each characteristic dimension within the reference time period in chronological order to obtain a characteristic sequence of the operating data of each characteristic dimension;

[0063] The multiple feature sequences corresponding to the multidimensional operating data are input into the target conversion model to obtain the power consumption characteristics output by the target conversion model, wherein the target conversion model is used to identify the time dependency between the initial power supply characteristics included in the multiple feature sequences and convert the corresponding power consumption characteristics.

[0064] Optionally, in an embodiment of the present application, the target conversion model is used to extract the time dependency between any features in multiple feature sequences, thereby reflecting the actual power consumption requirements of the server. In an embodiment of the present application, the target conversion model is an algorithm model with the function of extracting feature time dependency, which can be but is not limited to LSTM (Long Short-Term Memory).

[0065] Optionally, in an embodiment of the present application, before multiple feature sequences are input into the target conversion model, the features of the multiple feature sequences can be spliced ​​first, and the spliced ​​features can be input into the target conversion model, so that the target conversion model can extract the time dependency between any features in the multiple sequence features, thereby ensuring the accuracy of the obtained power consumption features.

[0066] Through the above content, by arranging the initial power supply features in chronological order and using the target conversion model to extract the time dependency of multiple feature sequences, the obtained power consumption features are made more accurate and reliable.

[0067] As an optional implementation manner, extracting features from the operation data of each feature dimension in the multi-dimensional operation data to obtain initial power supply features includes:

[0068] Using a wavelet transform algorithm to filter the noise signal carried in the operating data to obtain first data;

[0069] Performing outlier filtering on the first data using an isolation forest algorithm to obtain second data;

[0070] Feature extraction is performed on the second data to obtain a spectrum feature of the second data, wherein the spectrum feature indicates a fluctuation of a data signal of the corresponding feature dimension, and the initial power supply feature includes the spectrum feature.

[0071] Through the above content, by using the wavelet transform algorithm to filter the noise signal of the operating data, and using the isolation forest algorithm to filter outliers, the accuracy of the second data is guaranteed, and then by extracting the memory spectrum characteristics of the second data, the data periodic transformation state of the corresponding operating data is reflected through the spectrum characteristics.

[0072] As an optional implementation manner, extracting features from the second data to obtain spectral features of the second data includes:

[0073] When the second data is temperature data of the server power supply during operation, Fourier transform is performed on the temperature data to obtain a temperature spectrum feature in a frequency domain space, wherein the temperature spectrum feature is used to indicate a periodic fluctuation state of the operating temperature of the server power supply, and the spectrum feature includes the temperature spectrum feature;

[0074] In the case where the second data is current data of the output current of the server power supply, the current data is subjected to the Fourier transform to obtain a current spectrum feature in the frequency domain space, wherein the current spectrum feature is used to indicate the periodic fluctuation state of the output current value of the server power supply, and the spectrum feature includes the current spectrum feature.

[0075] As an optional implementation manner, detecting the remaining service life of the server power supply at a target time after the current time according to the decay characteristic and the power consumption characteristic includes:

[0076] Merging the decay feature and the power consumption feature to obtain a merged feature;

[0077] The combined features are input into a target prediction model to obtain the remaining service life of the server power supply output by the target prediction model, wherein the target prediction model records the correlation between the power consumption features of the server, the decay features of the server power supply and the remaining service life of the server power supply.

[0078] Optionally, in an embodiment of the present application, the decay feature and the power consumption feature can be merged by directly concatenating the power consumption feature and the decay feature to obtain a merged feature. Alternatively, feature weights of the power consumption feature and the decay feature can be obtained, where the feature weights indicate the impact of the corresponding feature on the server power supply life prediction. The decay feature and the power consumption feature can then be weighted and fused using the feature weights to obtain a merged feature.

[0079] Optionally, in an embodiment of the present application, the target prediction model can be, but is not limited to, obtained by training the initial prediction model using reference merged features marked with the remaining useful life as sample data, and this solution does not impose any limitation on this.

[0080] Through the above content, by merging the decay characteristics and power consumption characteristics, and using the target prediction model to predict the merged characteristics, the actual remaining life of the server in the future can be predicted based on the server's power usage demand characteristics and the server's decay characteristics under the usage demand.

[0081] As an optional implementation manner, merging the decay feature and the power consumption feature to obtain a merged feature includes:

[0082] Obtaining a first feature weight of the decay feature and a second feature weight of the power consumption feature, wherein the first feature weight is used to indicate the degree of influence of the decay feature on the detection result of the remaining useful life of the power supply, and the second feature weight is used to indicate the degree of influence of the power consumption feature on the detection result of the remaining useful life of the power supply;

[0083] The decay feature is weightedly calculated using the first feature weight, and the power consumption feature is weightedly calculated using the second feature weight, and the weighted calculation results are summed to obtain the combined feature.

[0084] Optionally, in an embodiment of the present application, the first feature weight and the second feature weight can be, but are not limited to, pre-set fixed values, or can also be weight values ​​obtained by calculating the importance of the first feature weight and the second feature weight. This solution does not limit this.

[0085] As an optional implementation manner, after inputting the combined features into a target prediction model to obtain the remaining useful life of the server power supply output by the target prediction model, the method further includes:

[0086] Detecting the actual service life of the server power supply at the target time;

[0087] Calculating the difference between the actual useful life and the remaining useful life;

[0088] The model parameters of the target prediction model are updated according to the difference value to obtain the updated target prediction model.

[0089] Optionally, in an embodiment of the present application, after using the target prediction model to predict the remaining service life of the server power supply at the target time, the actual service life of the server power supply at the target time can also be collected, and the model parameters of the target prediction model are self-updated based on the actual service life, thereby ensuring the accuracy of the target prediction model. In an embodiment of the present application, the method for updating the model parameters of the target prediction model may be, but is not limited to: inputting the comprehensive feature vector into the initially deployed model, performing a complete forward propagation process to obtain the initial remaining service life prediction value of the power module; based on the difference between the initial prediction result and the actual aging state monitored in real time, calculating the model output error, and calculating the gradient of each layer weight through the back propagation algorithm, the expression is: ;in, is the learning rate, is the partial derivative of the error with respect to the weight; the adjusted model parameters are obtained ; Generate a new comprehensive feature vector using the latest temperature raw data and current raw data, and combine it with the adjusted model parameters They are used together in the model retraining process to obtain an updated model; the new comprehensive feature vector is input into the updated model, and the forward propagation process is performed to obtain a more accurate prediction value of the remaining service life of the power module.

[0090] As an optional implementation manner, updating the model parameters of the target prediction model according to the difference value to obtain a reference prediction model includes:

[0091] Calculating the difference value by a back propagation algorithm to obtain parameter adjustment information for the model parameters of the target prediction model, wherein the parameter adjustment information is used to indicate an adjustment method for the model parameters;

[0092] The model parameters of the target prediction model are adjusted according to the parameter adjustment method indicated by the parameter adjustment information to obtain the updated target prediction model.

[0093] Through the above content, the model parameters of the target prediction model are adjusted by using the back propagation algorithm, thereby realizing the dynamic update of the model parameters of the target prediction model, thereby ensuring the reliability of the output results of the target prediction model.

[0094] As an optional implementation manner, the server system includes a plurality of servers, each server is correspondingly configured with one target prediction model, and the method further includes:

[0095] Obtaining current target model parameters of each target prediction model, wherein the target model parameters are used to indicate a correlation between the power consumption characteristics of the server, the decay characteristics of the server power supply, and the remaining service life of the server power supply;

[0096] Converting candidate model parameters of the target prediction model according to the plurality of target model parameters;

[0097] The candidate model parameters are sent to each of the target prediction models, wherein the target prediction model is used to update the model parameters currently used by the target prediction model using the candidate model parameters.

[0098] Optionally, in an embodiment of the present application, the above-mentioned server power supply detection method can be applied to a server system, which includes multiple servers. By deploying a target prediction model for each server, the remaining service life of the corresponding server can be predicted, and the model parameters of multiple target prediction models are dynamically updated in the server system through federated learning.

[0099] Optionally, in an embodiment of the present application, the above-mentioned server power supply detection method can be applied to a server system, which includes multiple servers. By deploying a target prediction model for each server, the remaining service life of the corresponding server can be predicted. In order to better execute the above-mentioned server power supply detection method, an edge device for executing the above-mentioned server power supply detection method can be deployed for each server, and the target prediction model can be run through the edge device to achieve dynamic prediction of the remaining service life of the server power deployed by the corresponding server. Furthermore, a cloud server can be configured for the edge nodes of all servers in the server system to implement the update and maintenance of the model parameters of the target prediction model on each edge node. The target prediction model deployed on the edge node can be, but is not limited to, a lightweight machine learning model that has been pruned and optimized, and is updated and corrected using the initial prediction results. The model parameters are adjusted based on the latest sensor data to obtain an accurate RUL prediction value; the model parameters on multiple edge devices are aggregated to the cloud through the federated learning mechanism, and global model optimization is performed on the cloud and fed back to each edge device. The specific steps are: running a federated learning client program on each edge device; the federated learning client program is responsible for collecting the model parameters of the local lightweight machine learning model after multiple updates to obtain a local model parameter set. ; Apply homomorphic encryption technology to the model parameters on each edge device Encrypt and generate encrypted model parameters , get the encrypted model parameter set ; The cloud server receives the encrypted model parameters sent by all edge devices , and restore the original model parameters through the corresponding decryption algorithm ; Use the weighted average method to summarize all model parameters to obtain the global model parameters, which are expressed as: ;in, Representative The importance weight of each edge device, The global model parameters are summarized; the global optimization algorithm is run on the cloud server to optimize the global model parameters. Fine-tune to obtain optimized global model parameters ; The cloud server will optimize the global model parameters Transmitted back to each edge device, the edge device receives After that, it replaces the current local model parameters , thereby updating the local model to a new local model; using a feedback mechanism to regularly check and update the global model parameters, and continuously provide accurate power module remaining service life prediction, to obtain the final accurate RUL prediction value.

[0100] Through the above content, by configuring a target prediction model for each of multiple servers in the server system, and by summarizing and updating the model parameters of multiple target prediction models, it is possible to achieve self-update of model parameters of multiple target prediction models on the one hand, and after the target prediction model self-updates the model parameters, the model parameters of multiple target prediction models can be obtained and summarized and updated, thereby ensuring the accuracy and reliability of the model parameters of multiple target prediction models in the server system and avoiding extreme updates of the model parameters of the models.

[0101] As an optional implementation manner, converting candidate model parameters of the target prediction model according to the multiple target model parameters includes:

[0102] Performing a weighted summation on the plurality of target model parameters using the model weight of each target prediction model to obtain a model parameter sum value, wherein the model weight is used to indicate the importance of the model parameter of the corresponding target prediction model;

[0103] Calculating a quotient of the model parameter and value and the number of models of the target prediction model to obtain an average value of the model parameters;

[0104] The model parameter average value is sent to each of the target prediction models, wherein the target prediction model is used to update the model parameters of the target prediction model using the model parameter average value.

[0105] Optionally, in an embodiment of the present application, when sending the average value of model parameters to the target prediction model, the average value of model parameters can be encrypted and transmitted using homomorphic encryption, and then after receiving the encrypted average value of model parameters, the target prediction model can restore the encrypted model parameters to obtain the average value of model parameters.

[0106] As an optional implementation manner, generating the decay characteristics of the server power supply under the power consumption characteristics based on the multi-dimensional operation data includes:

[0107] The decay characteristics are obtained by calculating the temperature data included in the multi-dimensional operation data using the following formula:

[0108] ;

[0109] in, is the gas constant, is the temperature in degrees Celsius, is the activation energy, is the frequency factor, is a thermal stress coefficient, and the decay characteristics include the thermal stress coefficient.

[0110] As an optional embodiment, the present application provides a power module life prediction method based on edge computing, Figure 3 The flowchart of the power module life prediction according to an embodiment of the present application includes:

[0111] Use sensors to monitor the working status of the power module and obtain original sensor data;

[0112] Preprocessing the original sensor data to obtain preprocessed sensor data;

[0113] Fast Fourier transform (FFT) and long short-term memory (LSTM) are used to extract features from the preprocessed sensor data, capturing the spectral and time series characteristics of the signal and obtaining key feature vectors.

[0114] A prediction model is built based on the Arrhenius equation and combined with lightweight machine learning. Key feature vectors are input into the prediction model to output the initial prediction results.

[0115] Deploy a lightweight machine learning model that has undergone pruning optimization on the edge device and use the initial prediction results for update correction. Adjust the model parameters based on the latest sensor data to obtain accurate RUL prediction values.

[0116] Through the federated learning mechanism, the model parameters on multiple edge devices are aggregated to the cloud, and global model optimization is performed in the cloud and fed back to each edge device.

[0117] As a preferred solution of the power module life prediction method based on edge computing described in this application, wherein: the sensor is used to monitor the working status of the power module to obtain the original sensor data, and the specific steps are as follows:

[0118] The DS18B20 digital temperature sensor is used to monitor the internal temperature of the power module, obtain the temperature changes of the power module during operation, and obtain the original temperature data;

[0119] The ACS712 Hall effect current sensor is used to monitor the current intensity flowing through the power module, capture the current fluctuations when the power module is working, and then obtain the original current data;

[0120] The temperature raw data and current raw data are aggregated into complete raw sensing data.

[0121] As a preferred solution of the power module life prediction method based on edge computing described in this application, wherein: the raw sensor data is preprocessed to obtain preprocessed sensor data, and the specific steps are:

[0122] The original sensor data set is denoised using wavelet transform. By selecting the wavelet basis function db4 to decompose the signal and filter out the high-frequency noise components, the denoised sensor data is obtained.

[0123] The isolation forest algorithm is used to detect anomalies in the denoised sensor data, and the anomaly score threshold is set to identify and eliminate abnormal data points to ensure the quality of the data set and obtain the preprocessed sensor data.

[0124] As a preferred solution of the power module life prediction method based on edge computing described in this application, wherein: the fast Fourier transform FFT and long short-term memory network LSTM are used to extract features of the preprocessed sensor data, capture the spectral characteristics and time series characteristics of the signal, and obtain key feature vectors. The specific steps are as follows:

[0125] Apply fast Fourier transform to the preprocessed temperature data, convert it into frequency domain representation, extract the main frequency components as features, and obtain the temperature spectrum features;

[0126] Apply fast Fourier transform to the preprocessed current data, convert it into frequency domain representation, extract the main frequency components as features, and obtain the current spectrum features;

[0127] The long short-term memory network (LSTM) is used to perform time series analysis on the temperature spectrum features and current spectrum features to capture the trends and patterns of data changes over time and obtain key feature vectors containing time series information.

[0128] As a preferred solution of the power module life prediction method based on edge computing described in this application, wherein: the prediction model is constructed based on the Arrhenius equation and combined with lightweight machine learning, key feature vectors are input into the prediction model, and the initial prediction results are output. The specific steps are as follows:

[0129] The Arrhenius equation is used to calculate the thermal stress coefficient of the preprocessed temperature data. The activation energy and frequency factor are adjusted according to the ambient temperature to obtain the thermal stress coefficient. The expression is:

[0130] ;

[0131] in, represents the ideal gas constant, is the temperature in degrees Celsius, is the activation energy, is the frequency factor, is the thermal stress coefficient;

[0132] Input the key feature vector into the pre-trained LSTM model;

[0133] The LSTM model is used to identify long-term dependencies in the data and extract patterns related to power module aging to obtain a time series feature representation;

[0134] The thermal stress coefficient and time series feature representation are input into the lightweight machine learning model after pruning optimization. The two features are integrated by weighted average to obtain a comprehensive feature vector, which is expressed as:

[0135] ;

[0136] in, is the time series feature, is the thermal stress coefficient, and is the weight coefficient;

[0137] The comprehensive feature vector The data is input into the lightweight machine learning model, and by applying and adjusting the internal parameters of the model, the initial remaining service life prediction value of the power module is output to obtain the initial prediction result.

[0138] As a preferred solution of the power module life prediction method based on edge computing described in the application, wherein: the lightweight machine learning model that has been pruned and optimized is deployed on the edge device, and the initial prediction results are used for update correction, and the model parameters are adjusted based on the latest sensor data to obtain an accurate RUL prediction value. The specific steps are as follows:

[0139] The lightweight machine learning model MobileNet is optimized through pruning technology and deployed on edge devices to obtain a preliminary deployment model;

[0140] The comprehensive feature vector is input into the initially deployed model and a full forward propagation process is performed to obtain the initial remaining useful life prediction value of the power module;

[0141] Based on the difference between the initial prediction result and the actual aging state monitored in real time, the model output error is calculated, and the gradient of each layer weight is calculated by the back propagation algorithm. The expression is:

[0142] ;

[0143] in, is the learning rate, is the partial derivative of the error with respect to the weight;

[0144] Get the adjusted model parameters ;

[0145] Generate a new comprehensive feature vector using the latest temperature raw data and current raw data, and combine it with the adjusted model parameters Used together in the model retraining process to obtain an updated model;

[0146] The new comprehensive feature vector is input into the updated model, and a forward propagation process is performed to obtain a more accurate prediction value of the remaining useful life of the power module.

[0147] As a preferred solution of the power module life prediction method based on edge computing described in the present invention, wherein: the model parameters on multiple edge devices are aggregated to the cloud through the federated learning mechanism, and global model optimization is performed in the cloud and fed back to each edge device. The specific steps are as follows:

[0148] Run the federated learning client program on each edge device;

[0149] The federated learning client program is responsible for collecting the model parameters of the local lightweight machine learning model after multiple updates to obtain the local model parameter set ;

[0150] Apply homomorphic encryption technology to the model parameters on each edge device Encrypt and generate encrypted model parameters , get the encrypted model parameter set ;

[0151] The cloud server receives the encrypted model parameters sent by all edge devices , and restore the original model parameters through the corresponding decryption algorithm ;

[0152] All model parameters are summarized using the weighted average method to obtain the global model parameters, which are expressed as:

[0153] ;

[0154] in, Representative The importance weight of each edge device, are the summarized global model parameters;

[0155] Run the global optimization algorithm on the cloud server to optimize the global model parameters Fine-tune to obtain optimized global model parameters ;

[0156] The cloud server will optimize the global model parameters Transmitted back to each edge device, the edge device receives After that, it replaces the current local model parameters , thereby updating the local model to a new local model;

[0157] A feedback mechanism is used to regularly check and update global model parameters, and continuously provide accurate predictions of the remaining useful life of the power module to obtain the final accurate RUL prediction value.

[0158] This application provides a power module life prediction system based on edge computing, including:

[0159] Data acquisition module, data preprocessing module, feature extraction module, model building module, model update module and federated learning module;

[0160] The data acquisition module is used to monitor the temperature changes inside the power module using a DS18B20 digital temperature sensor to obtain raw temperature data, and to monitor the current intensity flowing through the power module using an ACS712 Hall effect current sensor to capture current fluctuations and summarize them into complete raw sensor data;

[0161] The data preprocessing module is used to perform denoising on the original sensor data set using wavelet transform, decompose the signal by selecting the wavelet basis function db4, filter out high-frequency noise components, and then use the isolation forest algorithm to identify and eliminate abnormal data points to ensure data quality and obtain preprocessed sensor data;

[0162] The feature extraction module is used to apply fast Fourier transform to the preprocessed temperature data and current data, convert them into frequency domain representation, extract the main frequency components as features, and then use long short-term memory network to perform time series analysis on the temperature spectrum features and current spectrum features to capture trends and patterns that change over time and obtain key feature vectors containing time series information;

[0163] The model building module is used to calculate the thermal stress coefficient according to the Arrhenius equation, and to build a prediction model in combination with a lightweight machine learning model, input key feature vectors into the model, and output initial prediction results;

[0164] The model update module is used to deploy a pruned and optimized lightweight machine learning model on the edge device, and use the initial prediction results for update correction, adjust the model parameters based on the latest sensor data, and perform a forward propagation process to obtain a more accurate prediction value of the remaining useful life of the power module;

[0165] The federated learning module is used to run the federated learning client program on each edge device, collect the model parameters after local training, and encrypt them using homomorphic encryption technology before sending them to the cloud. The cloud server receives the encrypted model parameters, decrypts them, and uses the weighted average method to summarize all model parameters to obtain the global model parameters.

[0166] Figure 4is a schematic diagram of an optional edge computing power module life prediction system according to an embodiment of the present application, such as Figure 4 As shown, the following steps are included:

[0167] S1. Use sensors to monitor the working status of the power module and obtain raw sensor data;

[0168] Furthermore, a DS18B20 digital temperature sensor is used to monitor the internal temperature of the power module, obtain the temperature changes of the power module during operation, and obtain the original temperature data;

[0169] The ACS712 Hall effect current sensor is used to monitor the current intensity flowing through the power module, capture the current fluctuations when the power module is working, and then obtain the original current data;

[0170] Aggregate the temperature raw data and current raw data into complete raw sensing data;

[0171] It should be noted that by selecting the high-precision DS18B20 digital temperature sensor and ACS712 Hall-effect current sensor, the collected data can be ensured to have high accuracy and reliability. The high-precision data provides a solid foundation for subsequent preprocessing and feature extraction, which helps to improve the accuracy of the final life prediction.

[0172] S2. Preprocessing the original sensor data to obtain preprocessed sensor data;

[0173] Furthermore, wavelet transform is used to denoise the original sensor data set. By selecting the wavelet basis function db4 to decompose the signal and filter out the high-frequency noise components, the denoised sensor data is obtained;

[0174] The isolation forest algorithm is used to detect anomalies in the denoised sensor data. The anomaly score threshold is set to identify and remove abnormal data points to ensure the quality of the data set and obtain the preprocessed sensor data.

[0175] It should be noted that the application of wavelet transform and isolation forest algorithm not only effectively removes noise and outliers, but also retains the key information in the data. This step is crucial for subsequent feature extraction and model training because it directly affects the quality of model input and thus affects the accuracy of prediction results.

[0176] S3, using fast Fourier transform (FFT) and long short-term memory (LSTM) network to extract features from the pre-processed sensor data, capturing the spectral characteristics and time series characteristics of the signal, and obtaining key feature vectors;

[0177] Furthermore, the preprocessed temperature data is converted into a frequency domain representation by applying fast Fourier transform, and the main frequency components are extracted as features to obtain the temperature spectrum features;

[0178] Apply fast Fourier transform to the preprocessed current data, convert it into frequency domain representation, extract the main frequency components as features, and obtain the current spectrum features;

[0179] Use the long short-term memory network (LSTM) to perform time series analysis on the temperature spectrum characteristics and current spectrum characteristics, capture the trends and patterns of data changes over time, and obtain key feature vectors containing time series information;

[0180] It should be noted that the combined use of FFT and LSTM can comprehensively capture the changing characteristics of the working state of the power module, taking into account both the frequency domain characteristics of the signal and its timing characteristics. This method not only improves the comprehensiveness of feature extraction, but also enhances the model's ability to understand the state of the power module under different working conditions, thereby improving the accuracy and reliability of the prediction.

[0181] S4. Build a prediction model based on the Arrhenius equation and combined with lightweight machine learning, input key feature vectors into the prediction model, and output the initial prediction results;

[0182] Furthermore, the Arrhenius equation is used to calculate the thermal stress coefficient of the preprocessed temperature data. The activation energy and frequency factor are adjusted according to the ambient temperature to obtain the thermal stress coefficient, which is expressed as:

[0183] ;

[0184] in, represents the ideal gas constant, is the temperature in degrees Celsius, is the activation energy, is the frequency factor, is the thermal stress coefficient;

[0185] Input the key feature vector into the pre-trained LSTM model;

[0186] The LSTM model is used to identify long-term dependencies in the data and extract patterns related to power module aging to obtain a time series feature representation;

[0187] The thermal stress coefficient and time series feature representation are input into the lightweight machine learning model after pruning optimization. The two features are integrated by weighted average to obtain a comprehensive feature vector, which is expressed as:

[0188] ;

[0189] in, is the time series feature, is the thermal stress coefficient, and is the weight coefficient;

[0190] The comprehensive feature vector The data is input into a lightweight machine learning model, and by applying and adjusting the model's internal parameters, the initial remaining useful life prediction value of the power module is output to obtain the initial prediction result.

[0191] It should be noted that the combination of the Arrhenius equation and the lightweight machine learning model fully utilizes the advantages of physical knowledge and data-driven methods. In this way, it can not only explain the impact of ambient temperature on the aging of power modules, but also dynamically adjust the model parameters to adapt to different working conditions, thereby providing a more scientific and accurate remaining service life prediction.

[0192] S5. Deploy a lightweight machine learning model that has undergone pruning optimization on the edge device and use the initial prediction results for update correction. Adjust the model parameters based on the latest sensor data to obtain an accurate RUL prediction value.

[0193] Furthermore, the lightweight machine learning model MobileNet was optimized through pruning technology and deployed on edge devices to obtain a preliminary deployment model;

[0194] The comprehensive feature vector is input into the initially deployed model and a full forward propagation process is performed to obtain the initial remaining useful life prediction value of the power module;

[0195] Based on the difference between the initial prediction result and the actual aging state monitored in real time, the model output error is calculated, and the gradient of each layer weight is calculated by the back propagation algorithm. The expression is:

[0196] ;

[0197] in, is the learning rate, is the partial derivative of the error with respect to the weight;

[0198] Get the adjusted model parameters ;

[0199] Generate a new comprehensive feature vector using the latest temperature raw data and current raw data, and combine it with the adjusted model parameters Used together in the model retraining process to obtain an updated model;

[0200] The new comprehensive feature vector is input into the updated model and a forward propagation process is performed to obtain a more accurate prediction value of the remaining useful life of the power module;

[0201] It should be noted that the lightweight machine learning model after pruning optimization can run efficiently on resource-constrained edge devices while maintaining high prediction accuracy. By continuously updating and correcting model parameters, the model can better adapt to changes in actual working conditions, thereby continuously providing accurate remaining service life prediction values.

[0202] S6. Aggregate model parameters on multiple edge devices to the cloud through a federated learning mechanism, perform global model optimization in the cloud, and provide feedback to each edge device.

[0203] Furthermore, the federated learning client program is run on each edge device;

[0204] The federated learning client program is responsible for collecting the model parameters of the local lightweight machine learning model after multiple updates to obtain the local model parameter set. ;

[0205] Apply homomorphic encryption technology to the model parameters on each edge device Encrypt and generate encrypted model parameters , get the encrypted model parameter set ;

[0206] The cloud server receives the encrypted model parameters sent by all edge devices , and restore the original model parameters through the corresponding decryption algorithm ;

[0207] All model parameters are summarized using the weighted average method to obtain the global model parameters, which are expressed as:

[0208] ;

[0209] in, Representative The importance weight of each edge device, are the summarized global model parameters;

[0210] Run the global optimization algorithm on the cloud server to optimize the global model parameters Fine-tune to obtain optimized global model parameters ;

[0211] The cloud server will optimize the global model parameters Transmitted back to each edge device, the edge device receives After that, it replaces the current local model parameters , thereby updating the local model to a new local model;

[0212] A feedback mechanism is used to regularly check and update global model parameters, and continuously provide accurate predictions of the remaining useful life of the power modules, obtaining the final accurate RUL prediction value;

[0213] It should be noted that the federated learning mechanism not only protects the data privacy of each edge device, but also significantly improves the generalization ability and prediction accuracy of the model through global model optimization. Regular model updates and feedback mechanisms ensure that all edge devices can obtain the latest global model parameters in a timely manner, thereby achieving continuous and accurate prediction of the remaining service life of the power module. While improving the overall performance of the system, this method also provides strong support for maintenance and management in practical applications.

[0214] This embodiment also provides a power module life prediction system based on edge computing, including:

[0215] Data acquisition module, data preprocessing module, feature extraction module, model building module, model update module and federated learning module;

[0216] The data acquisition module uses a DS18B20 digital temperature sensor to monitor temperature changes inside the power module and obtain raw temperature data. It also uses an ACS712 Hall-effect current sensor to monitor the current intensity flowing through the power module, capture current fluctuations, and summarize them into complete raw sensor data.

[0217] The data preprocessing module is used to denoise the original sensor data set using wavelet transform. By selecting the wavelet basis function db4 to decompose the signal, high-frequency noise components are filtered out. Then, the isolation forest algorithm is used to identify and eliminate abnormal data points to ensure data quality and obtain preprocessed sensor data.

[0218] The feature extraction module applies fast Fourier transform to the preprocessed temperature and current data, converts them into frequency domain representation, extracts the main frequency components as features, and then uses a long short-term memory network to perform time series analysis on the temperature and current spectrum features to capture trends and patterns over time and obtain key feature vectors containing time series information;

[0219] The model building module is used to calculate the thermal stress coefficient according to the Arrhenius equation and build a prediction model in combination with a lightweight machine learning model. The key feature vectors are input into the model and the initial prediction results are output.

[0220] The model update module deploys a pruned and optimized lightweight machine learning model on the edge device and uses the initial prediction results for update correction. It adjusts the model parameters based on the latest sensor data and performs a forward propagation process to obtain a more accurate prediction of the remaining useful life of the power module.

[0221] The federated learning module is used to run the federated learning client program on each edge device, collect the locally trained model parameters, encrypt them using homomorphic encryption technology, and send them to the cloud. The cloud server receives the encrypted model parameters, decrypts them, and uses the weighted average method to summarize all model parameters to obtain the global model parameters.

[0222] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0223] The embodiment of the present application also provides a server power supply detection device, Figure 5 This is a structural block diagram of a server power supply detection device according to an embodiment of the present application. Figure 5 As shown, the device includes:

[0224] an acquisition module, configured to acquire multi-dimensional operating data of a server power supply, wherein the multi-dimensional operating data is used to characterize, from multiple characteristic dimensions, a power supply status of the server power supply to the server within a reference time period before a current moment;

[0225] an extraction module, configured to perform feature extraction on the multi-dimensional operation data to obtain power consumption features of the server, wherein the power consumption features are used to indicate a temporal variation relationship of the power demand of the server;

[0226] a generating module, configured to generate, based on the multi-dimensional operating data, a decay characteristic of the server power supply under the power consumption characteristic, wherein the decay characteristic is used to indicate an influence relationship of the power consumption characteristic on the decay of the server power supply;

[0227] The first detection module is configured to detect a remaining service life of the server power supply at a target time after a current time according to the decay characteristics and the power consumption characteristics.

[0228] Through the above device, by extracting features from the multidimensional operating data used to characterize the power supply status of the server power supply to the server from multiple dimensions, the power consumption characteristics of the server are obtained to indicate the time-series change relationship of the server's power demand, thereby realizing the reflection of the server's actual power usage demand for the server power supply according to the multidimensional operating data, thereby comprehensively reflecting the actual usage status of the server power supply, and then generating the decay characteristics of the server power supply under the power consumption characteristics according to the multidimensional operating data, detecting the remaining service life of the server according to the decay characteristics and the power consumption characteristics, and realizing the prediction of the remaining service life of the server power supply according to the actual usage status of the server power supply and the decay of the server power supply under the actual usage status. Therefore, the technical problem of low detection efficiency of the server power supply in the related technology can be solved, and the technical effect of improving the detection efficiency of the server power supply can be achieved.

[0229] Optionally, the extraction module includes:

[0230] an extraction unit, configured to perform feature extraction on the operation data of each feature dimension in the multi-dimensional operation data to obtain an initial power supply feature, wherein the initial power supply feature is used to indicate an operating state of the server power supply within the reference time period;

[0231] The conversion unit is used to perform feature conversion on the multiple initial power supply features corresponding to the multi-dimensional operation data to obtain the power consumption feature.

[0232] Optionally, the conversion unit is configured to:

[0233] Sorting the initial power supply characteristics of each characteristic dimension within the reference time period in chronological order to obtain a characteristic sequence of the operating data of each characteristic dimension;

[0234] The multiple feature sequences corresponding to the multidimensional operating data are input into the target conversion model to obtain the power consumption characteristics output by the target conversion model, wherein the target conversion model is used to identify the time dependency between the initial power supply characteristics included in the multiple feature sequences and convert the corresponding power consumption characteristics.

[0235] Optionally, the extraction unit is used to:

[0236] Using a wavelet transform algorithm to filter the noise signal carried in the operating data to obtain first data;

[0237] Performing outlier filtering on the first data using an isolation forest algorithm to obtain second data;

[0238] Feature extraction is performed on the second data to obtain a spectrum feature of the second data, wherein the spectrum feature indicates a fluctuation of a data signal of the corresponding feature dimension, and the initial power supply feature includes the spectrum feature.

[0239] Optionally, the extraction unit is used to:

[0240] When the second data is temperature data of the server power supply during operation, Fourier transform is performed on the temperature data to obtain a temperature spectrum feature in a frequency domain space, wherein the temperature spectrum feature is used to indicate a periodic fluctuation state of the operating temperature of the server power supply, and the spectrum feature includes the temperature spectrum feature;

[0241] In the case where the second data is current data of the output current of the server power supply, the current data is subjected to the Fourier transform to obtain a current spectrum feature in the frequency domain space, wherein the current spectrum feature is used to indicate the periodic fluctuation state of the output current value of the server power supply, and the spectrum feature includes the current spectrum feature.

[0242] Optionally, the first detection module includes:

[0243] a merging unit, configured to merge the decay feature and the power consumption feature to obtain a merged feature;

[0244] An input unit is used to input the combined features into a target prediction model to obtain the remaining service life of the server power supply output by the target prediction model, wherein the target prediction model records the correlation between the power consumption characteristics of the server, the decay characteristics of the server power supply and the remaining service life of the server power supply.

[0245] Optionally, the merging unit is used to:

[0246] Obtaining a first feature weight of the decay feature and a second feature weight of the power consumption feature, wherein the first feature weight is used to indicate the degree of influence of the decay feature on the detection result of the remaining useful life of the power supply, and the second feature weight is used to indicate the degree of influence of the power consumption feature on the detection result of the remaining useful life of the power supply;

[0247] The decay feature is weightedly calculated using the first feature weight, and the power consumption feature is weightedly calculated using the second feature weight, and the weighted calculation results are summed to obtain the combined feature.

[0248] Optionally, the device further includes:

[0249] a second detection module, configured to detect the actual service life of the server power supply at the target moment after inputting the combined feature into a target prediction model to obtain the remaining service life of the server power supply output by the target prediction model;

[0250] a calculation module, configured to calculate a difference between the actual useful life and the remaining useful life;

[0251] An updating module is used to update the model parameters of the target prediction model according to the difference value to obtain the updated target prediction model.

[0252] Optionally, the update module includes:

[0253] a calculation unit, configured to calculate the difference value by a back propagation algorithm to obtain parameter adjustment information for the model parameters of the target prediction model, wherein the parameter adjustment information is used to indicate an adjustment method for the model parameters;

[0254] An adjustment unit is used to adjust the model parameters of the target prediction model according to the parameter adjustment method indicated by the parameter adjustment information to obtain the updated target prediction model.

[0255] Optionally, the server system includes a plurality of servers, each server is configured with a corresponding target prediction model, and the device further includes:

[0256] an acquisition module, configured to acquire current target model parameters of each target prediction model, wherein the target model parameters are used to indicate a correlation between the power consumption characteristics of the server, the decay characteristics of the server power supply, and the remaining service life of the server power supply;

[0257] A conversion module, configured to convert candidate model parameters of the target prediction model according to the plurality of target model parameters;

[0258] A sending module is used to send the candidate model parameters to each of the target prediction models, wherein the target prediction model is used to use the candidate model parameters to update the model parameters currently used by the target prediction model.

[0259] Optionally, the conversion module includes:

[0260] a processing unit, configured to perform weighted summation on the plurality of target model parameters using a model weight of each target prediction model to obtain a model parameter sum value, wherein the model weight is used to indicate the importance of the model parameter of the corresponding target prediction model;

[0261] a calculation unit, configured to calculate a quotient of the model parameter and value and the number of models of the target prediction model to obtain an average value of the model parameters;

[0262] A sending unit is used to send the average value of the model parameters to each of the target prediction models, wherein the target prediction model is used to update the model parameters of the target prediction model using the average value of the model parameters.

[0263] Optionally, the generating module includes:

[0264] A calculation unit is configured to calculate the temperature data included in the multidimensional operation data using the following formula to obtain the decay characteristic:

[0265] ;

[0266] in, is the gas constant, is the temperature in degrees Celsius, is the activation energy, is the frequency factor, is a thermal stress coefficient, and the decay characteristics include the thermal stress coefficient.

[0267] For the description of the features in the embodiment corresponding to the detection device of the server power supply, reference can be made to the relevant description of the embodiment corresponding to the detection method of the server power supply, which will not be repeated here.

[0268] An embodiment of the present application further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any of the above-mentioned server power supply detection method embodiments.

[0269] An embodiment of the present application further provides a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps of any of the above-mentioned server power supply detection method embodiments when running.

[0270] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0271] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of any of the above-mentioned server power supply detection method embodiments are implemented.

[0272] An embodiment of the present application further provides another computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any of the above-mentioned server power supply detection method embodiments are implemented.

[0273] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0274] The above describes in detail the server power supply detection method and device, storage medium, and electronic device provided by this application. This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only intended to help understand the method and core concept of this application. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of this application, several improvements and modifications can be made to this application, and these improvements and modifications also fall within the scope of protection of the claims of this application.

Claims

1. A method for detecting a server power supply, characterized in that: include: Acquire multi-dimensional operation data of a server power supply, wherein the multi-dimensional operation data is used to characterize, from multiple characteristic dimensions, a power supply status of the server power supply to the server within a reference time period before a current moment; Extracting features from the multi-dimensional operation data to obtain power consumption features of the server, wherein the power consumption features are used to indicate a temporal variation relationship of power demand of the server; generating, based on the multi-dimensional operating data, a decay characteristic of the server power supply under the power consumption characteristic, wherein the decay characteristic is used to indicate an influence relationship of the power consumption characteristic on the decay of the server power supply; detecting a remaining service life of the server power supply at a target time after the current time according to the decay characteristics and the power consumption characteristics; The detecting the remaining service life of the server power supply at a target moment after the current moment based on the decay characteristics and the power consumption characteristics includes: merging the decay characteristics and the power consumption characteristics to obtain a merged characteristic; inputting the merged characteristic into a target prediction model to obtain the remaining service life of the server power supply output by the target prediction model, wherein the target prediction model records the correlation between the power consumption characteristics of the server, the decay characteristics of the server power supply, and the remaining service life of the server power supply; Generating the decay characteristics of the server power supply under the power consumption characteristics according to the multi-dimensional operation data includes: The decay characteristics are obtained by calculating the temperature data included in the multi-dimensional operation data using the following formula: ； in, is the gas constant, is the temperature in degrees Celsius, is the activation energy, is the frequency factor, is a thermal stress coefficient, and the decay characteristics include the thermal stress coefficient.

2. The method according to claim 1, characterized in that The extracting features of the multi-dimensional operation data to obtain the power consumption features of the server includes: performing feature extraction on the operation data of each feature dimension in the multi-dimensional operation data to obtain an initial power supply feature, wherein the initial power supply feature is used to indicate an operating state of the server power supply within the reference time period; Feature conversion is performed on the multiple initial power supply features corresponding to the multi-dimensional operation data to obtain the power consumption features.

3. The method according to claim 2, characterized in that The performing feature conversion on the plurality of initial power supply features corresponding to the multi-dimensional operation data to obtain the power consumption feature includes: Sorting the initial power supply characteristics of each characteristic dimension within the reference time period in chronological order to obtain a characteristic sequence of the operating data of each characteristic dimension; The multiple feature sequences corresponding to the multidimensional operating data are input into the target conversion model to obtain the power consumption characteristics output by the target conversion model, wherein the target conversion model is used to identify the time dependency between the initial power supply characteristics included in the multiple feature sequences and convert the corresponding power consumption characteristics.

4. The method according to claim 2, characterized in that The extracting features of the operation data of each feature dimension in the multi-dimensional operation data to obtain initial power supply features includes: Using a wavelet transform algorithm to filter the noise signal carried in the operating data to obtain first data; Performing outlier filtering on the first data using an isolation forest algorithm to obtain second data; Feature extraction is performed on the second data to obtain a spectrum feature of the second data, wherein the spectrum feature indicates a fluctuation of a data signal of the corresponding feature dimension, and the initial power supply feature includes the spectrum feature.

5. The method according to claim 4, characterized in that The extracting features of the second data to obtain spectral features of the second data includes: When the second data is temperature data of the server power supply during operation, Fourier transform is performed on the temperature data to obtain a temperature spectrum feature in a frequency domain space, wherein the temperature spectrum feature is used to indicate a periodic fluctuation state of the operating temperature of the server power supply, and the spectrum feature includes the temperature spectrum feature; In the case where the second data is current data of the output current of the server power supply, the current data is subjected to the Fourier transform to obtain a current spectrum feature in the frequency domain space, wherein the current spectrum feature is used to indicate the periodic fluctuation state of the output current value of the server power supply, and the spectrum feature includes the current spectrum feature.

6. The method according to claim 1, characterized in that The merging of the decay feature and the power consumption feature to obtain a merged feature includes: Obtaining a first feature weight of the decay feature and a second feature weight of the power consumption feature, wherein the first feature weight is used to indicate the degree of influence of the decay feature on the detection result of the remaining useful life of the power supply, and the second feature weight is used to indicate the degree of influence of the power consumption feature on the detection result of the remaining useful life of the power supply; The decay feature is weightedly calculated using the first feature weight, and the power consumption feature is weightedly calculated using the second feature weight, and the weighted calculation results are summed to obtain the combined feature.

7. The method according to claim 1, characterized in that After inputting the combined features into a target prediction model to obtain the remaining useful life of the server power supply output by the target prediction model, the method further includes: Detecting the actual service life of the server power supply at the target time; Calculating the difference between the actual useful life and the remaining useful life; The model parameters of the target prediction model are updated according to the difference value to obtain the updated target prediction model.

8. The method according to claim 7, characterized in that The updating of the model parameters of the target prediction model according to the difference value to obtain a reference prediction model includes: Calculating the difference value by a back propagation algorithm to obtain parameter adjustment information for the model parameters of the target prediction model, wherein the parameter adjustment information is used to indicate an adjustment method for the model parameters; The model parameters of the target prediction model are adjusted according to the parameter adjustment method indicated by the parameter adjustment information to obtain the updated target prediction model.

9. The method according to claim 7, characterized in that The server system includes a plurality of servers, each server is configured with a corresponding target prediction model, and the method further includes: Obtaining current target model parameters of each target prediction model, wherein the target model parameters are used to indicate a correlation between the power consumption characteristics of the server, the decay characteristics of the server power supply, and the remaining service life of the server power supply; Converting candidate model parameters of the target prediction model according to the plurality of target model parameters; The candidate model parameters are sent to each of the target prediction models, wherein the target prediction model is used to update the model parameters currently used by the target prediction model using the candidate model parameters.

10. The method according to claim 9, characterized in that The converting of candidate model parameters of the target prediction model according to the plurality of target model parameters includes: Performing a weighted summation on the plurality of target model parameters using the model weight of each target prediction model to obtain a model parameter sum value, wherein the model weight is used to indicate the importance of the model parameter of the corresponding target prediction model; Calculating a quotient of the model parameter and value and the number of models of the target prediction model to obtain an average value of the model parameters; The model parameter average value is sent to each of the target prediction models, wherein the target prediction model is used to update the model parameters of the target prediction model using the model parameter average value.

11. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the server power supply detection method according to any one of claims 1 to 10 when executing the computer program.

12. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the server power supply detection method according to any one of claims 1 to 10 are implemented.

13. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the server power supply detection method according to any one of claims 1 to 10 are implemented.

Images