Method and system for detecting emergency starting power supply

Abstract

The invention relates to the technical field of power supplies, and provides an emergency starting power supply detection method and system, and the method comprises the steps: obtaining continuously inputted input voltage data, continuously outputted output current data and temperature change data of an emergency starting power supply in a preset time period; inputting the input voltage data and the output current data into a preset detection model to obtain power supply state information, comparing the temperature change data with a preset temperature threshold to obtain temperature comparison data, and performing health state evaluation on the emergency starting power supply in combination with the power supply state information to obtain a power supply health state. And performing fault diagnosis on the emergency starting power supply according to the power supply health state to obtain a diagnosis result, and generating a detection report of the emergency starting power supply based on the diagnosis result. By comprehensively acquiring and analyzing the input voltage data, the output current data and the temperature change data, the problem that the health state and potential faults of the emergency starting power supply cannot be diagnosed and evaluated under the condition that the environment temperature change is large is solved.

Description

Technical Field

[0001] The present application relates to the technical field of power supplies, and in particular to a detection method and system for an emergency starting power supply. Background Art

[0002] Emergency starting power supplies play a vital role in power systems and electronic equipment, providing timely power support when power is interrupted or insufficient. With the advancement of technology and the improvement of people's requirements for power safety, more and more industries and fields have begun to rely on emergency starting power supplies to ensure the continuous operation of critical systems and equipment.

[0003] In related technical means, the working status of the emergency starting power supply is determined by monitoring its input voltage and output current. These methods include using voltage and current sensors to record the operating data of the power supply and using these data for analysis and fault diagnosis. This method can detect abnormal conditions of the power supply in a timely manner, such as unstable voltage or excessive current, so that corresponding measures can be taken for maintenance and repair.

[0004] For the above technical solutions, although the basic evaluation of the power supply status can be achieved through voltage and current monitoring, there is a problem of insufficient monitoring data in practical applications. Specifically, relying solely on voltage and current data cannot fully reflect the health status and service life of the emergency starting power supply, especially when the ambient temperature changes greatly, it is impossible to diagnose and evaluate the health status and potential faults of the emergency starting power supply. Summary of the invention

[0005] In order to improve the problem that the health status and potential faults of the emergency starting power supply cannot be diagnosed and evaluated when the ambient temperature changes greatly, the present application provides a detection method and system for the emergency starting power supply.

[0006] The present invention provides a detection method for an emergency starting power supply, comprising: obtaining input voltage data continuously input and output current data continuously output by the emergency starting power supply within a preset time period, and recording the temperature change within the time period to obtain temperature change data; inputting the input voltage data and the output current data into a preset detection model to obtain power supply status information, and comparing the temperature change data with a preset temperature threshold to obtain temperature comparison data; performing a health status assessment on the emergency starting power supply based on the power supply status information and the temperature comparison data to obtain a power supply health status, performing a fault diagnosis on the emergency starting power supply according to the power supply health status to obtain a diagnosis result, and generating a detection report of the emergency starting power supply based on the diagnosis result.

[0007] As a preferred solution, the step of obtaining the input voltage data and the output current data of the emergency starting power supply continuously input within a preset time period, and recording the temperature change within the time period to obtain the temperature change data includes: performing real-time voltage measurement on the emergency starting power supply within a preset time period to obtain battery voltage values ​​at multiple time points, inputting the battery voltage values ​​at all time points into a preset Fourier transform algorithm to generate a voltage change curve, and denoising the voltage change curve through wavelet transform to obtain input voltage data; performing real-time current measurement on the output end of the emergency starting power supply to obtain output current values ​​at multiple time points, inputting the output current values ​​at all time points into a preset Kalman filter algorithm to generate a current change curve, and smoothing the current change curve using a Savitzky-Golay smoothing filter to obtain output current data; and recording ambient temperature changes through a temperature sensor within the preset time period to generate temperature change data.

[0008] As a preferred solution, the steps of inputting the battery voltage values ​​at all time points into a preset Fourier transform algorithm to generate a voltage change curve, and performing denoising processing on the voltage change curve by wavelet transform to obtain input voltage data include: performing Fourier transform on the battery voltage values ​​at all time points to obtain frequency domain data, performing high-frequency noise filtering on the frequency domain data by a high-pass filter, and performing inverse Fourier transform on the filtered data to obtain a voltage change curve; performing wavelet transform on the voltage change curve to obtain change curve data after wavelet transform, and performing denoising processing on the change curve data by a threshold method to obtain denoised data. The voltage change data after the voltage change is reconstructed in time domain by statistical time domain analysis to obtain the reconstructed mean, standard deviation and peak value; the mean is standardized by logarithmic transformation to obtain a standardized mean, the standardized mean and the standard deviation are input into a multivariate linear regression model for regression analysis to obtain a regression analysis result; the peak value is smoothed by an exponential smoothing method to obtain a smoothed peak value, the smoothed peak value and the regression analysis result are input into a neural network model for feature learning to obtain a voltage feature vector, and the voltage feature vector is processed by a support vector machine to obtain input voltage data.

[0009] As a preferred solution, the step of inputting the output current values ​​at all time points into a preset Kalman filter algorithm to generate a current change curve, and smoothing the current change curve using a Savitzky-Golay smoothing filter to obtain output current data includes: inputting the output current values ​​at all time points into a Kalman filter for filtering to obtain filtered data; optimizing the filtered data using a recursive least squares algorithm to obtain optimized data, performing curve fitting on the optimized data to generate a current change curve; inputting the current change curve into a Savitzky-Golay filter for smoothing, performing polynomial fitting on the smoothed data to obtain a smooth change curve; and resampling the smooth change curve by a downsampling technique to obtain output current data.

[0010] As a preferred embodiment, the step of inputting the input voltage data and the output current data into a preset detection model to obtain the power supply status information includes: inputting the input voltage data and the output current data into a preset detection model, in the detection model, performing feature extraction on the input voltage data and the output current data by a convolutional neural network to obtain time series features, frequency domain features and energy distribution features, analyzing the time series features by linear discriminant analysis to obtain noise level and signal stability; fusing the noise level with the frequency domain features to obtain denoised frequency domain features, fusing the energy distribution features with the signal stability to obtain dynamic response features; and inputting the denoised frequency domain features and the dynamic response features into a logistic regression model to obtain power supply status information.

[0011] As a preferred scheme, the step of performing a health status assessment on the emergency starting power supply based on the power status information and the temperature comparison data to obtain a power health status includes: normalizing the power status information to obtain normalized status data, performing a health assessment on the normalized status data through a Bayesian network to obtain a health assessment probability; performing weighted averaging processing on the health assessment probability and the temperature comparison data based on weighted logistic regression to obtain a comprehensive health index; classifying the comprehensive health index using a support vector machine to obtain a power health status classification result, cross-validating the power health status classification result, and generating a power health status.

[0012] As a preferred scheme, the steps of performing fault diagnosis on the emergency starting power supply according to the power supply health status to obtain a diagnosis result, and generating a detection report of the emergency starting power supply based on the diagnosis result include: analyzing the power supply health status through a fuzzy neural network to obtain a health status feature vector and a predicted failure probability, classifying the health status feature vector using a decision tree to obtain a health level, and verifying the predicted failure probability using trend analysis to obtain a verified failure risk; inputting the health level and the verified failure risk into a preset fault diagnosis model, in which a multilayer perceptron network is used to perform fault diagnosis on the emergency starting power supply based on the health level and the verified failure risk, and outputting a diagnosis result; performing root cause analysis on the fault information in the diagnosis result to obtain a detection report of the emergency starting power supply.

[0013] The present application also provides a detection system for an emergency starting power supply, comprising: an acquisition unit, used to acquire input voltage data continuously input and output current data continuously output by the emergency starting power supply within a preset time period, and record the temperature change within the time period to obtain temperature change data; a comparison unit, used to input the input voltage data and the output current data into a preset detection model to obtain power supply status information, and compare the temperature change data with a preset temperature threshold to obtain temperature comparison data; a generation unit, used to perform a health status assessment on the emergency starting power supply based on the power supply status information and the temperature comparison data to obtain a power supply health status, perform a fault diagnosis on the emergency starting power supply according to the power supply health status to obtain a diagnosis result, and generate a detection report of the emergency starting power supply based on the diagnosis result.

[0014] Compared with the prior art, this application has the following beneficial effects: rapid detection and strong applicability. Through data analysis algorithms, abnormal conditions in power supply operation can be captured in a timely manner, providing more accurate power supply status information. Combined with the monitoring and comparison of temperature change data, it can better identify the risk of failures caused by ambient temperature changes, thereby improving the reliability and safety of emergency starting power supplies; and through the generated test report, it is possible to quickly understand the health status and potential problems of the power supply, formulate scientific maintenance and repair strategies, ensure that the emergency starting power supply can provide stable and reliable power support at critical moments, and improve the problem of being unable to diagnose and evaluate the health status and potential failures of the emergency starting power supply when the ambient temperature changes greatly. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0016] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0017] Figure 1 It is a flow chart of a method for detecting an emergency starting power supply provided by an embodiment of the present invention; Figure 2 The figure is a schematic block diagram of the structure of a detection system for an emergency starting power supply provided by an embodiment of the present invention.

[0018] Description of reference numerals: 10. Emergency starting power supply detection system; 11. Acquisition unit; 12. Comparison unit; 13. Generation unit. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0021] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0022] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0023] The technical solution of the present invention is further described below with reference to the accompanying drawings and through specific implementation methods.

[0024] Embodiment 1: like Figure 1 As shown, the emergency starting power supply detection method provided in the embodiment of the present application includes steps S100 to S300.

[0025] Step S100: obtaining input voltage data and output current data of the emergency starting power supply continuously input within a preset time period, and recording the temperature change within the time period to obtain temperature change data.

[0026] In this step, high-precision voltage and current sensors are used to obtain the continuous input voltage data and continuous output current data of the emergency starting power supply within a preset time period, and a temperature sensor is used to record the temperature changes within the time period; specifically, the data acquisition module will transmit the acquired data to the data processing unit in real time for recording and analysis.

[0027] For example, by setting 1 hour as the preset time period, input voltage data, output current data and temperature change data are collected every 1 minute, which can ensure the integrity and continuity of the data.

[0028] Step S200: input the input voltage data and the output current data into a preset detection model to obtain power supply status information, and compare the temperature change data with a preset temperature threshold to obtain temperature comparison data.

[0029] In this step, the input voltage data and output current data are first preprocessed and input into a preset detection model to generate power status information, and the recorded temperature change data is compared with the temperature threshold set by the system to generate temperature comparison data; specifically, the detection model uses a machine learning algorithm to perform a comprehensive analysis based on historical data and current data to obtain the power status information.

[0030] For example, the detection model can use regression analysis or classification algorithms to process data, obtain more accurate power status information through big data analysis, and compare temperature changes with preset temperature thresholds (such as 40°C) to identify abnormal temperature changes.

[0031] Step S300: perform health status assessment on the emergency starting power supply based on the power status information and the temperature comparison data to obtain the power health status, perform fault diagnosis on the emergency starting power supply according to the power health status to obtain the diagnosis result, and generate a detection report of the emergency starting power supply based on the diagnosis result.

[0032] In this step, the system integrates the power status information and temperature comparison data to comprehensively evaluate the health status of the emergency starting power supply, and the evaluation results are used for further fault diagnosis; specifically, the health status assessment module determines the working status, abnormal conditions and potential faults of the power supply by analyzing the power status information and temperature comparison data, and generates a detailed health status report.

[0033] For example, if the health status assessment results show that the power supply voltage fluctuates greatly and the temperature exceeds the threshold, the system will determine that the power supply has an overheating problem or voltage instability problem, and record the specific fault type and recommended maintenance measures in the detection report.

[0034] In this embodiment, the temperature change data is obtained by obtaining the input voltage data and the output current data of the emergency starting power supply continuously input within a preset time period, and recording the temperature change within the time period. Then, the input voltage data and the output current data are input into the preset detection model to generate the power state information, and the temperature change data is compared with the preset temperature threshold to obtain the temperature comparison data. Finally, the health state of the emergency starting power supply is evaluated based on the power state information and the temperature comparison data to obtain the power health state, and the emergency starting power supply is fault diagnosed according to the power health state to generate a detection report. By combining the input voltage data, the output current data and the temperature change data for comprehensive evaluation, the health state and service life of the emergency starting power supply can be more accurately reflected, the accuracy of fault diagnosis is effectively improved, and the problem of being unable to diagnose and evaluate the health state and potential faults of the emergency starting power supply when the ambient temperature changes greatly is improved. In addition, by monitoring the temperature change data, the abnormal conditions caused by the ambient temperature change can be discovered in time, thereby further ensuring the stable operation of the emergency starting power supply, which not only improves the reliability and safety of the emergency starting power supply, but also provides more scientific data support for its maintenance and management.

[0035] Embodiment 2: In step S100, the emergency starting power supply is subjected to real-time voltage measurement within a preset time period to obtain battery voltage values ​​at multiple time points, the battery voltage values ​​at all time points are input into a preset Fourier transform algorithm to generate a voltage change curve, and the voltage change curve is denoised by wavelet transform to obtain input voltage data.

[0036] The emergency starting power supply is measured in real time by a high-precision voltage sensor, and the battery voltage values ​​at multiple time points are recorded. Specifically, these battery voltage values ​​are input into a preset Fourier transform algorithm to obtain frequency domain data, and then the frequency domain data is filtered for high-frequency noise by a high-pass filter. Finally, the filtered data is inversely Fourier transformed to generate a voltage change curve. Then, the voltage change curve is denoised by wavelet transform to obtain input voltage data.

[0037] For example, the battery voltage value is recorded every 30 seconds, and these data are input into the Fourier transform algorithm. After high-pass filtering and inverse transformation, the voltage change curve is obtained, and then the noise is removed by wavelet transform to obtain smoother input voltage data.

[0038] The output end of the emergency starting power supply is measured in real time to obtain the output current values ​​at multiple time points. The output current values ​​at all time points are input into the preset Kalman filter algorithm to generate a current change curve. The current change curve is smoothed using the Savitzky-Golay smoothing filter to obtain the output current data.

[0039] The output end of the emergency starting power supply is measured in real time by a high-precision current sensor, and the output current values ​​at multiple time points are recorded; specifically, these output current values ​​are input into a preset Kalman filter algorithm for filtering to generate a current change curve. Then, the current change curve is smoothed using a Savitzky-Golay smoothing filter to obtain output current data.

[0040] For example, the output current value is recorded every 30 seconds, and these data are input into the Kalman filter for filtering to generate a current change curve, which is then smoothed by the Savitzky-Golay smoothing filter to obtain smoother output current data.

[0041] During a preset time period, the temperature sensor records the changes in ambient temperature and generates temperature change data.

[0042] A high-precision temperature sensor is used to record the ambient temperature changes in real time within a preset time period to generate temperature change data; specifically, the temperature sensor records the ambient temperature every 30 seconds to form a continuous temperature change data sequence.

[0043] For example, the ambient temperature is recorded every 30 seconds, and the data obtained by the sensor constitutes a temperature change curve, which can reflect in detail the changes in ambient temperature within a preset time period.

[0044] The steps of inputting the battery voltage values ​​at all time points into a preset Fourier transform algorithm to generate a voltage change curve, and performing denoising on the voltage change curve by wavelet transform to obtain input voltage data include: The battery voltage values ​​at all time points are Fourier transformed to obtain frequency domain data, and the frequency domain data are filtered for high-frequency noise using a high-pass filter. The filtered data are inverse Fourier transformed to obtain a voltage change curve.

[0045] Frequency domain data is obtained by performing fast Fourier transform on the battery voltage values ​​at all time points. Specifically, a high-pass filter is used to filter high-frequency noise on the frequency domain data, and then the filtered frequency domain data is converted back to the time domain through inverse Fourier transform to generate a voltage change curve.

[0046] For example, the frequency domain data obtained by fast Fourier transform can effectively remove high-frequency noise and obtain a more stable voltage change curve.

[0047] The voltage change curve is subjected to wavelet transform to obtain the change curve data after wavelet transform. The change curve data is denoised using the threshold method to obtain the denoised voltage change data. The voltage change data is reconstructed in the time domain through statistical time domain analysis to obtain the reconstructed mean, standard deviation and peak value.

[0048] By performing wavelet transform on the voltage change curve, specifically, using the threshold method to denoise the wavelet transformed data, the denoised voltage change data is obtained, and then the data is reconstructed in the time domain through statistical time domain analysis to obtain the mean, standard deviation and peak value of the voltage data.

[0049] For example, most high-frequency noise can be eliminated through wavelet transform and threshold denoising, and more stable and reliable data indicators such as mean, standard deviation and peak value can be obtained after time domain reconstruction of voltage data.

[0050] The mean was standardized using logarithmic transformation to obtain the standardized mean, and the standardized mean and standard deviation were input into the multiple linear regression model for regression analysis to obtain the regression analysis results.

[0051] The mean of the voltage data is standardized by using logarithmic transformation. Specifically, the standardized mean and standard deviation are input into a multiple linear regression model to perform regression analysis to obtain regression analysis results.

[0052] For example, logarithmic transformation can reduce the data range and enhance the stability and prediction effect of the model.

[0053] The peak value is smoothed by exponential smoothing method to obtain a smoothed peak value, the smoothed peak value and the regression analysis result are input into the neural network model for feature learning to obtain a voltage feature vector, and the voltage feature vector is processed by a support vector machine to obtain input voltage data.

[0054] The peak value of the voltage data is smoothed by the exponential smoothing method. Specifically, the smoothed peak value and the regression analysis result are input into the neural network model for feature learning to generate a voltage feature vector. The voltage feature vector is then classified and processed by a support vector machine to obtain the final input voltage data.

[0055] For example, the exponential smoothing method can effectively smooth the peak value of voltage data, thereby improving the accuracy of learning features of the neural network model, and the support vector machine can further optimize the data classification effect.

[0056] The steps of inputting the output current values ​​at all time points into a preset Kalman filter algorithm to generate a current change curve, smoothing the current change curve using a Savitzky-Golay smoothing filter, and obtaining output current data include: The output current values ​​at all time points are input into the Kalman filter for filtering to obtain filtered data.

[0057] The output current values ​​at all time points are input into the Kalman filter for filtering. Specifically, smooth current change data is obtained through filtering.

[0058] For example, the Kalman filter algorithm can effectively filter out the noise in the current data and improve the accuracy and stability of the data.

[0059] The recursive least squares algorithm is used to optimize the filtered data to obtain the optimized data, and the optimized data is curve fitted to generate the current change curve.

[0060] The filtered data is optimized by using a recursive least squares algorithm, specifically, optimized data is generated, and then curve fitting is performed to obtain a current variation curve.

[0061] For example, the recursive least squares algorithm can gradually improve the fitting effect of the current variation curve by continuously optimizing the data.

[0062] The current variation curve is input into the Savitzky-Golay filter for smoothing, and a polynomial fitting is performed on the smoothed data to obtain a smooth variation curve.

[0063] The current variation curve is input into a Savitzky-Golay filter for smoothing. Specifically, a polynomial fitting is performed on the smoothed data to obtain a smooth variation curve.

[0064] For example, the Savitzky-Golay filter can better smooth the current data and reduce the impact of noise on the results, while the polynomial fit provides higher fitting accuracy.

[0065] The smooth changing curve is resampled by down sampling technology to obtain the output current data.

[0066] The smoothly changing curve is resampled by a down-sampling technique, specifically, to generate the final output current data.

[0067] For example, downsampling can reduce the amount of data and improve processing efficiency while retaining the key features of the data.

[0068] In step S200, the step of inputting the input voltage data and the output current data into a preset detection model to obtain the power supply status information includes: The input voltage data and output current data are input into the preset detection model. In the detection model, the input voltage data and output current data are featured by convolutional neural network to obtain time series features, frequency domain features and energy distribution features. The time series features are analyzed by linear discriminant analysis to obtain the noise level and signal stability.

[0069] By inputting the input voltage data and the output current data into the preset detection model, specifically, using a convolutional neural network to extract features of these data, generating time series features, frequency domain features, and energy distribution features, and then using linear discriminant analysis to analyze the time series features to obtain the noise level and signal stability.

[0070] For example, convolutional neural networks can extract complex features from data, while linear discriminant analysis can effectively distinguish between noise and signals in data.

[0071] The noise level and frequency domain features are fused to obtain denoised frequency domain features, the noise level and frequency domain features are fused to obtain denoised frequency domain features, and the energy distribution features and signal stability are fused to obtain dynamic response features. By fusing the noise level and frequency domain features, specifically, the denoised frequency domain features are combined with the original frequency domain data using an algorithm to obtain a purer denoised frequency domain feature. At the same time, the energy distribution features and signal stability are fused to generate dynamic response features to better reflect the dynamic response capability of the power supply under different working conditions. For example, by combining frequency domain features with noise levels, redundant noise can be effectively filtered out to obtain more accurate frequency domain features. The fusion of energy distribution features and signal stability can reflect the stability of the power supply under different working conditions.

[0072] The denoised frequency domain features and dynamic response features are input into the logistic regression model to obtain the power supply status information.

[0073] The denoised frequency domain features and dynamic response features are input into a logistic regression model for analysis. Specifically, the model generates power supply status information through parameter training and optimization.

[0074] For example, the logistic regression model can accurately determine the working status and potential problems of the power supply by analyzing the characteristic data, providing a basis for subsequent health status assessment and fault diagnosis.

[0075] In step S300, the health status of the emergency starting power supply is evaluated based on the power status information and the temperature comparison data to obtain the health status of the power supply, including: The power supply status information is normalized to obtain normalized status data, and the health assessment of the normalized status data is performed through a Bayesian network to obtain a health assessment probability.

[0076] The power supply status information is normalized, specifically, the power supply status information is converted into data of a uniform scale using data normalization technology, and input into a Bayesian network for health assessment to obtain the health assessment probability of the power supply.

[0077] For example, normalization can eliminate differences between different dimensions, making data analysis more accurate and reliable, while Bayesian networks use statistical methods to evaluate health status.

[0078] Based on weighted logistic regression, the health assessment probability and temperature comparison data were weighted averaged to obtain a comprehensive health index.

[0079] The weighted logistic regression model is used to comprehensively analyze the health assessment probability and temperature comparison data. Specifically, different indicators are assigned different weights, and weighted average processing is performed to generate a comprehensive health index.

[0080] For example, through weighted logistic regression, various influencing factors can be taken into consideration to obtain a comprehensive index that reflects the overall health status of the power supply.

[0081] The comprehensive health index is classified by support vector machine to obtain the power health status classification result. The power health status classification result is cross-validated to generate the power health status.

[0082] The comprehensive health index is input into a support vector machine for classification. Specifically, the classification result is cross-validated to ensure its accuracy and generate the final power health status.

[0083] For example, the support vector machine can effectively distinguish the health status level of the power supply by classifying the health index, and cross-validation further improves the reliability of the classification results.

[0084] In step S300, the emergency starting power supply is diagnosed for a fault according to the power health status to obtain a diagnosis result, and the step of generating a detection report of the emergency starting power supply based on the diagnosis result includes: The health status of the power supply is analyzed through fuzzy neural network to obtain the health status feature vector and predicted failure probability. The health status feature vector is classified using decision tree to obtain the health level. The predicted failure probability is verified using trend analysis to obtain the verified failure risk.

[0085] The power supply health status data is deeply analyzed through fuzzy neural networks. Specifically, the health status feature vector is extracted and the failure probability is predicted. The feature vector is classified using a decision tree model to obtain the power supply health level. The predicted failure probability is then verified through trend analysis to generate a verified failure risk.

[0086] For example, fuzzy neural networks can perform effective analysis under conditions of high uncertainty, decision trees can accurately classify health levels, and trend analysis is used to verify the reliability of failure probabilities.

[0087] The health level and the verification failure risk are input into a preset fault diagnosis model. In the fault diagnosis model, a multilayer perceptron network is used to perform fault diagnosis on the emergency starting power supply based on the health level and the verification failure risk, and the diagnosis result is output.

[0088] By inputting the health level and verification failure risk into the multilayer perceptron network, specifically, the network diagnoses the fault type of the emergency starting power supply through parameter training and optimization, and generates a diagnosis result.

[0089] For example, a multilayer perceptron network can be trained with large amounts of data and has strong fault diagnosis capabilities, and can accurately identify different types of fault conditions.

[0090] Perform root cause analysis on the fault information in the diagnosis results to obtain a test report on the emergency starting power supply.

[0091] By performing root cause analysis on the fault information in the diagnosis results, specifically, using the fault tree analysis method to trace the root cause of the fault, a detailed detection report is generated.

[0092] For example, fault tree analysis can effectively identify the key factors that cause failures and provide a basis for the repair and maintenance of emergency starting power supplies.

[0093] In this embodiment, the voltage and current of the emergency starting power supply are measured in real time within a preset time period, and the data are processed using the Fourier transform algorithm and the Kalman filter algorithm to generate voltage and current change curves, and then the curves are denoised and smoothed by wavelet transform and Savitzky-Golay smoothing filter to obtain more accurate input voltage data and output current data. The system uses a temperature sensor to record the change of ambient temperature, combines the input voltage data and the output current data, and inputs the preset detection model for comprehensive analysis to obtain the power supply status information. Based on the power supply status information and temperature comparison data, health status assessment and fault diagnosis are performed, and a variety of algorithms such as Bayesian network, weighted logistic regression, support vector machine and fuzzy neural network are used to comprehensively evaluate the health status and fault conditions of the power supply, generate a detailed detection report, and realize accurate monitoring, status assessment and fault diagnosis of the emergency starting power supply, with high reliability and accuracy, providing important technical support for the maintenance and management of the power supply.

[0094] Embodiment 3: like Figure 2 As shown, the present application also provides an emergency starting power supply detection system 10, including an acquisition unit 11, a comparison unit 12 and a generation unit 13.

[0095] The acquisition unit 11 is mainly used to acquire the input voltage data and the output current data of the emergency starting power supply continuously input within a preset time period, and to record the temperature change within the time period to obtain the temperature change data.

[0096] The comparison unit 12 is mainly used to input the input voltage data and the output current data into a preset detection model to obtain power supply status information, and compare the temperature change data with a preset temperature threshold to obtain temperature comparison data.

[0097] The generation unit 13 is mainly used to evaluate the health status of the emergency starting power supply based on the power status information and the temperature comparison data, obtain the power health status, perform fault diagnosis on the emergency starting power supply according to the power health status, obtain the diagnosis result, and generate a detection report of the emergency starting power supply based on the diagnosis result.

[0098] In this embodiment, the input voltage data and output current data of the emergency starting power supply within a preset time period are collected in real time by the high-precision sensor of the acquisition unit 11, and the ambient temperature changes are recorded by the temperature sensor to generate comprehensive temperature change data. After being collected, these data are transmitted to the comparison unit 12 for processing. The comparison unit 12 inputs the input voltage data and the output current data into a preset detection model, and uses a convolutional neural network to extract features from the data to generate detailed power supply status information. At the same time, the comparison unit 12 compares the temperature change data with the preset temperature threshold to generate temperature comparison data. In the generation unit 13, the system integrates the power supply status information and the temperature comparison data, evaluates the health status of the emergency starting power supply, performs fault diagnosis based on the evaluation results, and generates detailed diagnosis results and a test report of the emergency starting power supply.

[0099] It should be noted that technicians in the relevant technical field can clearly understand that for the convenience and simplicity of description, the specific working process of the system and each unit described above can refer to the corresponding process in the aforementioned emergency starting power supply detection method embodiment, and will not be repeated here.

[0100] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting an emergency starting power supply, characterized in that: include: Acquire input voltage data and output current data of the emergency starting power supply continuously input within a preset time period, and record the temperature change within the time period to obtain temperature change data; Inputting the input voltage data and the output current data into a preset detection model to obtain power supply status information, and comparing the temperature change data with a preset temperature threshold to obtain temperature comparison data; Based on the power status information and the temperature comparison data, the health status of the emergency starting power supply is evaluated to obtain the power health status, and according to the power health status, a fault diagnosis is performed on the emergency starting power supply to obtain a diagnosis result, and a detection report of the emergency starting power supply is generated based on the diagnosis result.

2. The method for detecting an emergency starting power supply according to claim 1, characterized in that: The step of obtaining the input voltage data and the output current data of the emergency starting power supply continuously input within a preset time period, and recording the temperature change within the time period to obtain the temperature change data includes: Performing real-time voltage measurement on the emergency starting power supply within a preset time period to obtain battery voltage values ​​at multiple time points, inputting the battery voltage values ​​at all time points into a preset Fourier transform algorithm to generate a voltage change curve, and performing denoising processing on the voltage change curve by wavelet transform to obtain input voltage data; Performing real-time current measurement on the output end of the emergency starting power supply to obtain output current values ​​at multiple time points, inputting the output current values ​​at all time points into a preset Kalman filter algorithm to generate a current change curve, and smoothing the current change curve using a Savitzky-Golay smoothing filter to obtain output current data; During the preset time period, the ambient temperature change is recorded by the temperature sensor to generate temperature change data.

3. The method for detecting an emergency starting power supply according to claim 2, characterized in that: The step of inputting the battery voltage values ​​at all time points into a preset Fourier transform algorithm to generate a voltage change curve, and performing denoising processing on the voltage change curve by wavelet transform to obtain input voltage data comprises: Performing Fourier transformation on the battery voltage values ​​at all time points to obtain frequency domain data, filtering the frequency domain data for high-frequency noise using a high-pass filter, and performing inverse Fourier transformation on the filtered data to obtain a voltage change curve; Performing wavelet transformation on the voltage change curve to obtain change curve data after wavelet transformation, performing denoising on the change curve data using a threshold method to obtain denoised voltage change data, and performing time domain reconstruction on the voltage change data through statistical time domain analysis to obtain a reconstructed mean, standard deviation and peak value; The mean is standardized by logarithmic transformation to obtain a standardized mean, and the standardized mean and the standard deviation are input into a multiple linear regression model for regression analysis to obtain a regression analysis result; The peak value is smoothed by an exponential smoothing method to obtain a smoothed peak value, the smoothed peak value and the regression analysis result are input into a neural network model for feature learning to obtain a voltage feature vector, and the voltage feature vector is processed by a support vector machine to obtain input voltage data.

4. The method for detecting an emergency starting power supply according to claim 2, characterized in that: The step of inputting the output current values ​​at all time points into a preset Kalman filter algorithm to generate a current change curve, and smoothing the current change curve using a Savitzky-Golay smoothing filter to obtain output current data includes: Inputting the output current values ​​at all time points into a Kalman filter for filtering to obtain filtered data; Optimizing the filtered data using a recursive least squares algorithm to obtain optimized data, performing curve fitting on the optimized data to generate a current change curve; The current change curve is input into a Savitzky-Golay filter for smoothing, and a polynomial fitting is performed on the smoothed data to obtain a smooth change curve; The smooth change curve is resampled by down sampling technology to obtain output current data.

5. The method for detecting an emergency starting power supply according to claim 1, characterized in that: The step of inputting the input voltage data and the output current data into a preset detection model to obtain power supply status information includes: The input voltage data and the output current data are input into a preset detection model, in which a convolutional neural network is used to extract features of the input voltage data and the output current data to obtain time series features, frequency domain features, and energy distribution features, and a linear discriminant analysis is used to analyze the time series features to obtain noise level and signal stability; The noise level is fused with the frequency domain feature to obtain a denoised frequency domain feature, and the energy distribution feature is fused with the signal stability to obtain a dynamic response feature; The denoised frequency domain features and the dynamic response features are input into a logistic regression model to obtain power supply status information.

6. The method for detecting an emergency starting power supply according to claim 1, characterized in that: The step of evaluating the health status of the emergency starting power supply based on the power status information and the temperature comparison data to obtain the health status of the power supply includes: Normalizing the power state information to obtain normalized state data, and performing health assessment on the normalized state data through a Bayesian network to obtain a health assessment probability; Performing weighted average processing on the health assessment probability and the temperature comparison data based on weighted logistic regression to obtain a comprehensive health index; The comprehensive health index is classified by using a support vector machine to obtain a power health status classification result, and the power health status classification result is cross-validated to generate a power health status.

7. The method for detecting an emergency starting power supply according to claim 1, characterized in that: The step of performing fault diagnosis on the emergency starting power supply according to the power supply health status to obtain a diagnosis result, and generating a detection report of the emergency starting power supply based on the diagnosis result includes: Analyze the health status of the power supply by using a fuzzy neural network to obtain a health status feature vector and a predicted failure probability, classify the health status feature vector by using a decision tree to obtain a health level, and verify the predicted failure probability by using trend analysis to obtain a verified failure risk; Inputting the health level and the verification failure risk into a preset fault diagnosis model, in which a multilayer perceptron network is used to perform fault diagnosis on the emergency starting power supply based on the health level and the verification failure risk, and outputting a diagnosis result; Perform root cause analysis on the fault information in the diagnosis result to obtain a detection report of the emergency starting power supply.

8. A detection system for an emergency starting power supply, characterized in that: include: An acquisition unit, used to acquire input voltage data and output current data of the emergency starting power supply continuously input within a preset time period, and record the temperature change within the time period to obtain temperature change data; A comparison unit, used for inputting the input voltage data and the output current data into a preset detection model to obtain power supply state information, and comparing the temperature change data with a preset temperature threshold to obtain temperature comparison data; A generating unit is used to evaluate the health status of the emergency starting power supply based on the power status information and the temperature comparison data to obtain the power health status, perform fault diagnosis on the emergency starting power supply according to the power health status to obtain a diagnosis result, and generate a detection report of the emergency starting power supply based on the diagnosis result.

Images