Comprehensive test method, device and equipment for direct-current power supply and storage medium

Abstract

The invention discloses a comprehensive testing method, device and equipment for a direct-current power supply and a storage medium, and the method comprises the steps: obtaining dynamic parameters and static parameters of the direct-current power supply, and constructing a multi-dimensional parameter set; performing feature extraction according to the multi-dimensional parameter set to obtain an independent variable and a dependent variable; based on the independent variables and the dependent variables, calculating coupling coefficients among the variables to obtain a coupling coefficient matrix; performing error analysis on the multi-dimensional parameter set to obtain dynamic distribution characteristics of accumulated errors; performing error prediction and compensation according to the dynamic distribution characteristics to obtain an error correction sequence; and according to the multi-dimensional parameter set, the coupling coefficient matrix and the error correction sequence, calculating to obtain a comprehensive score of the direct-current power supply. The accuracy of performance analysis of the direct-current power supply can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power supply testing, and in particular to a comprehensive testing method, device, equipment and storage medium for a direct current power supply. Background Art

[0002] As the core power supply unit of modern electronic devices, the performance of DC power supplies is directly related to system stability and reliability. In fields such as industrial control, communications equipment, and new energy systems, the accuracy and efficiency of DC power supplies have become important indicators of technological progress in these fields. Current DC power supply testing typically involves collecting key parameters from the power supply's specifications, such as using a digital multimeter to measure voltage or an oscilloscope to observe ripple. Simple statistical analysis is then performed on the acquired data to generate test results.

[0003] However, with the increasing complexity of application scenarios and technological advancements, the demand for DC power supply performance evaluation is also increasing. Traditional testing methods have limitations in evaluating the comprehensive performance of DC power supplies, often focusing on static characteristics or dynamic response under specific conditions. Faced with increasingly complex usage scenarios, existing testing methods struggle to provide sufficiently accurate data support, limiting the understanding and evaluation of the power supply's true operating conditions and resulting in low accuracy in DC power supply performance analysis. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a comprehensive testing method, device, equipment and storage medium for a DC power supply, which can effectively improve the accuracy of DC power supply testing.

[0005] An embodiment of the present invention provides a comprehensive testing method for a DC power supply, comprising:

[0006] Obtaining dynamic parameters and static parameters of the DC power supply, and constructing a multidimensional parameter set based on the dynamic parameters and the static parameters; the static parameters are parameters that do not change over time under stable operating conditions of the DC power supply, and the dynamic parameters are parameters for evaluating the performance of the DC power supply under transient conditions;

[0007] Performing feature extraction based on the multidimensional parameter set to obtain independent variables and dependent variables; the independent variables are parameters with low mutual correlation, and the dependent variables are parameters whose value changes are affected by the independent variables;

[0008] Based on the independent variables and the dependent variables, the coupling coefficients between the variables are calculated to obtain a coupling coefficient matrix;

[0009] Performing error analysis on the multidimensional parameter set to obtain dynamic distribution characteristics of the cumulative error;

[0010] Perform error prediction and compensation according to the dynamic distribution characteristics to obtain an error correction sequence;

[0011] A comprehensive score of the DC power supply is calculated based on the multidimensional parameter set, the coupling coefficient matrix, and the error correction sequence.

[0012] As an improvement to the above solution, the feature extraction is performed based on the multidimensional parameter set to obtain independent variables and dependent variables, including:

[0013] performing standardization processing on the multidimensional parameter set to obtain a standardized parameter matrix;

[0014] Performing covariance calculation based on the standardized parameter matrix to obtain a covariance matrix;

[0015] Extracting the eigenvalues and eigenvectors of the covariance matrix, and selecting several eigenvectors with the largest eigenvalues to construct a principal component matrix;

[0016] Performing a matrix multiplication operation on the principal component matrix and the standardized parameter matrix to obtain a dimension reduction matrix;

[0017] A weight analysis is performed based on the dimension reduction matrix and the principal component matrix to obtain independent variables and dependent variables.

[0018] As an improvement to the above solution, the coupling coefficients between the variables are calculated based on the independent variables and the dependent variables to obtain a coupling coefficient matrix, including:

[0019] Calculating the correlation between the independent variable and the dependent variable, and calculating the coupling coefficient between the variables based on the correlation;

[0020] The independent variables and the dependent variables are used as row and column indices of a matrix, and a coupling coefficient matrix is constructed based on the coupling coefficients between the variables.

[0021] As an improvement to the above solution, performing error analysis on the multi-dimensional parameter set to obtain dynamic distribution characteristics of the cumulative error includes:

[0022] Dividing the multidimensional parameter set into a number of time windows according to a preset time length;

[0023] Calculating the difference between the parameter value in the time window and the preset reference value to obtain an error sequence;

[0024] Extracting statistical features of the error sequence to obtain error statistical features;

[0025] A time series analysis is performed on the error statistical characteristics to obtain a dynamic distribution characteristic of the accumulated error.

[0026] As an improvement to the above solution, performing error prediction and compensation according to the dynamic distribution characteristics to obtain an error correction sequence includes:

[0027] Inputting the dynamic distribution features into a preset error prediction model to obtain an error prediction sequence; the error prediction model is obtained by training a long short-term memory network based on historical data;

[0028] Performing reverse offset calculation on the error prediction sequence to obtain an error correction sequence;

[0029] Performing application simulation based on the error correction sequence to generate a convergence indicator;

[0030] Determine whether the convergence index is less than a preset convergence threshold; if so, output an error correction sequence; if not, optimize the parameters of the error prediction model, reacquire the error prediction sequence, error correction sequence and convergence index until the convergence index is less than the preset convergence threshold, and output the error correction sequence.

[0031] As an improvement to the above solution, the comprehensive score of the DC power supply is calculated based on the multi-dimensional parameter set, the coupling coefficient matrix and the error correction sequence, including:

[0032] Calculating a standardized parameter score based on the multidimensional parameter set and preset parameter weights;

[0033] Calculating a coupling influence value according to the coupling coefficient matrix and a preset coupling coefficient weight;

[0034] Calculating an error index value based on the error correction sequence and a preset time weight;

[0035] A weighted sum is performed on the standardized parameter score, the coupling influence value, and the error index value to obtain a comprehensive score of the DC power supply.

[0036] As an improvement to the above solution, after obtaining the comprehensive score of the DC power supply through calculation, the method further includes:

[0037] If the comprehensive score is less than the preset qualified threshold, the random forest algorithm is used to perform anomaly tracing analysis based on the multidimensional parameter set, the coupling coefficient matrix and the error correction sequence, to determine the key anomaly features, and generate a performance deviation distribution report based on the key anomaly features.

[0038] An embodiment of the present invention further provides a comprehensive test device for a DC power supply, comprising:

[0039] a data acquisition module, configured to acquire dynamic and static parameters of the DC power supply, and construct a multidimensional parameter set based on the dynamic and static parameters; the static parameters are parameters that do not change over time under stable operating conditions, and the dynamic parameters are parameters used to evaluate the performance of the DC power supply under transient conditions;

[0040] a feature extraction module for performing feature extraction based on the multidimensional parameter set to obtain independent variables and dependent variables; the independent variables are parameters with low mutual correlation, and the dependent variables are parameters whose value changes are affected by the independent variables;

[0041] A coupling calculation module, configured to calculate the coupling coefficients between the variables based on the independent variables and the dependent variables to obtain a coupling coefficient matrix;

[0042] An error analysis module, configured to perform error analysis on the multidimensional parameter set to obtain dynamic distribution characteristics of the accumulated error;

[0043] An error correction module, configured to perform error prediction and compensation according to the dynamic distribution characteristics to obtain an error correction sequence;

[0044] A power supply scoring module is used to calculate a comprehensive score of the DC power supply based on the multi-dimensional parameter set, the coupling coefficient matrix and the error correction sequence.

[0045] An embodiment of the present invention further provides a computer device comprising a processor and a memory, wherein a computer program is stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the comprehensive testing method for a DC power supply described in any one of the above items is implemented.

[0046] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program, wherein when the computer program is executed, the device containing the computer-readable storage medium is controlled to execute any one of the above-mentioned comprehensive testing methods for DC power supplies.

[0047] Compared with the prior art, the comprehensive testing method, device, equipment, and storage medium for a DC power supply provided by the embodiments of the present invention have the following beneficial effects:

[0048] By acquiring the dynamic parameters and static parameters of the DC power supply to construct a multidimensional parameter set, the comprehensiveness and accuracy of parameter acquisition are improved, and a rich data basis is provided for subsequent analysis; feature extraction is performed based on the multidimensional parameter set to obtain independent variables and dependent variables, which is conducive to simplifying complex parameter relationships, clarifying the relationship between variables, and improving the accuracy of analysis; based on the independent variables and the dependent variables, the coupling coefficients between each variable are calculated to obtain a coupling coefficient matrix, which not only reveals the deep relationship between the variables, but also can further guide the system optimization direction, which is conducive to improving system performance and stability; error analysis is performed on the multidimensional parameter set to obtain the dynamic distribution characteristics of the cumulative error, and error prediction and compensation are performed based on the dynamic distribution characteristics to obtain an error correction sequence, which can reduce the uncertainty caused by the error and further improve accuracy and reliability; based on the multidimensional parameter set, the comprehensive score of the DC power supply is calculated, which fully considers the influence of various factors, thereby improving the comprehensiveness and accuracy of the DC current test. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a flow chart of a comprehensive testing method for a DC power supply provided by an embodiment of the present invention;

[0050] Figure 2 1 is a schematic structural diagram of a comprehensive test device for a DC power supply provided by an embodiment of the present invention;

[0051] Figure 3 It is a structural diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0053] See also Figure 1 , Figure 1 1 is a flow chart of a comprehensive test method for a DC power supply provided by an embodiment of the present invention. The comprehensive test method for a DC power supply includes:

[0054] S1: Acquire dynamic parameters and static parameters of a DC power supply, and construct a multidimensional parameter set based on the dynamic parameters and the static parameters; the static parameters are parameters that do not change over time under stable operating conditions of the DC power supply, and the dynamic parameters are parameters for evaluating the performance of the DC power supply under transient conditions;

[0055] Specifically, static parameters are parameters that do not change or change very little over time under stable operating conditions. Static parameters reflect the basic performance indicators of DC power supplies, including but not limited to input voltage range, output voltage accuracy (such as nominal value ± deviation), ripple factor (the ratio of the AC component to the DC component in the output voltage), efficiency (the ratio of output power to input power), load regulation (the change in output voltage when the load changes) and temperature stability (the impact of ambient temperature changes on the output voltage).

[0056] Dynamic parameters indicate the performance of a DC power supply under sudden load changes or other transient conditions, including but not limited to transient response time (the time required for the output voltage to stabilize after a load change), overshoot / undershoot amplitude (the deviation of the output voltage peak caused by the load change), recovery time (the time required for the output voltage to return to the specified range after the deviation), voltage drop amplitude under dynamic load, and frequency response characteristics (the ability to regulate voltage at different frequencies).

[0057] In an embodiment of the present invention, static parameters and dynamic parameters of a DC power supply under different working conditions are first collected in real time through sensors and measuring equipment to ensure that the data covers the entire working condition range; then, the collected parameter data is preliminarily processed, including filtering, denoising, and smoothing, to eliminate accidental errors and improve data quality; the preliminarily processed parameter data is formatted according to a predefined data structure, and static parameters and dynamic parameters are marked separately, and assigned unique identifiers to facilitate subsequent processing and query; finally, the static parameters and dynamic parameters are integrated, and all parameters are stored in a structured database or file system in a tabular or binary format, where each row represents a sampling point, and each column corresponds to a different parameter type and its value. Timestamp information is also attached to record the time sequence of data collection, thereby forming a complete multidimensional parameter set, providing a data basis for subsequent analysis, and comprehensively evaluating the performance of the DC power supply under various working conditions.

[0058] S2: performing feature extraction based on the multidimensional parameter set to obtain independent variables and dependent variables; the independent variables are parameters with low mutual correlation, and the dependent variables are parameters whose value changes are affected by the independent variables;

[0059] As one of the optional embodiments, the feature extraction based on the multidimensional parameter set to obtain independent variables and dependent variables includes:

[0060] performing standardization processing on the multidimensional parameter set to obtain a standardized parameter matrix;

[0061] Performing covariance calculation based on the standardized parameter matrix to obtain a covariance matrix;

[0062] Extracting the eigenvalues and eigenvectors of the covariance matrix, and selecting several eigenvectors with the largest eigenvalues to construct a principal component matrix;

[0063] Performing a matrix multiplication operation on the principal component matrix and the standardized parameter matrix to obtain a dimension reduction matrix;

[0064] A weight analysis is performed based on the dimension reduction matrix and the principal component matrix to obtain independent variables and dependent variables.

[0065] Specifically, each parameter in the multidimensional parameter set is standardized. For example, each parameter is subtracted from its mean value and divided by its standard deviation to ensure that the mean value of all parameters is 0 and the standard deviation is 1, thereby forming a standardized parameter matrix.

[0066] Then, principal component analysis (PCA) is performed on the standardized parameter matrix. The covariance matrix is calculated based on the standardized parameter matrix, quantifying the mutual influence of each parameter by evaluating the strength of the linear relationship between them. Each element of the covariance matrix represents the covariance value between two parameters. Eigenvalues are extracted from the covariance matrix to identify the principal directions and their corresponding eigenvectors that reflect the degree of data variation. These eigenvectors are then sorted in descending order of eigenvalue, and the top several eigenvectors with the highest contribution are selected to construct the principal component matrix. A matrix multiplication is then performed on the principal component matrix and the standardized parameter matrix to transform the original data into a new coordinate system, generating a reduced dimension matrix that simplifies the data structure while retaining the key information. Finally, a weight analysis is performed based on the reduced dimension matrix and the PCA matrix. By analyzing the weight distribution of each original parameter along each principal component direction, the independent variables are identified, i.e., those that have a significant impact on system performance and have low correlation with each other, such as input voltage. The dependent variables are those that are directly affected by the independent variables and have a strong dependence on their own changes, such as output voltage sag. This weight analysis helps clarify the relationship between the parameters and lays the foundation for further coupling strength analysis.

[0067] In a specific example, the multidimensional parameter set of a DC power supply includes: input voltage, output voltage, load current, ambient temperature, output ripple, and voltage drop. First, the standardized parameters are reduced in dimension to obtain two principal components (PC1 and PC2), of which PC1 explains 70% of the variance and PC2 explains 20% of the variance. In the principal component matrix, the weight distribution of PC1 shows that the weight of input voltage is 0.6, temperature is 0.5, load current is 0.3, while the weight of output voltage and ripple is only 0.1; the weight distribution of PC2 shows that the weight of voltage drop is 0.7, output voltage is 0.6, while the weights of input voltage and temperature drop to 0.1 and 0.2. By analyzing these weights, it is clear that the input voltage and ambient temperature are independent variables because they occupy the dominant weight in PC1, indicating that they are the core factors driving the performance of the system and have low correlation with each other; while the voltage drop amplitude is a dependent variable and is highly correlated with the output voltage in PC2, and its changes are obviously affected by the input voltage and temperature (for example, input voltage fluctuations will cause the voltage drop amplitude to increase). This analysis not only simplifies the original six parameters into two principal components, but also clearly distinguishes between "driving" parameters (such as input voltage) and "response" parameters (such as voltage drop amplitude). For example, changes in input voltage as an independent variable will directly trigger a chain reaction of dependent variables, thereby providing a clear basis for subsequent analysis of the coupling relationship between parameters. In this way, engineers can prioritize the optimization of independent variables, while evaluating system stability based on changes in dependent variables, and ultimately improve the overall performance of the DC power supply.

[0068] S3: Based on the independent variables and the dependent variables, calculating the coupling coefficients between the variables to obtain a coupling coefficient matrix;

[0069] As one of the optional embodiments, the coupling coefficients between the variables are calculated based on the independent variables and the dependent variables to obtain a coupling coefficient matrix, including:

[0070] Calculating the correlation between the independent variable and the dependent variable, and calculating the coupling coefficient between the variables based on the correlation;

[0071] The independent variables and the dependent variables are used as row and column indices of a matrix, and a coupling coefficient matrix is constructed based on the coupling coefficients between the variables.

[0072] Specifically, the coupling strength analysis is performed based on the independent and dependent variables. First, the correlation between the independent and dependent variables is calculated to assess the strength of the linear relationship between the two variables. For example, the correlation is calculated using the Pearson correlation method, which measures the correlation by measuring the ratio of the degree of covariation of the two variables to the product of their standard deviations. The value of a perfect positive correlation is 1, the value of a perfect negative correlation is -1, and the value of no correlation is 0.

[0073] Then, the coupling coefficient is calculated based on the calculated correlation. The specific formula is as follows:

[0074]

[0075] Among them, C xy represents the coupling coefficient between variables X and Y, r xy represents the correlation, k represents the proportional coefficient, and its value is 2.

[0076] The coupling coefficient calculation function considers both the absolute value and direction of the correlation to quantify the degree of mutual influence between variables. When the correlation is less than a preset threshold, the relationship is considered weakly coupled, and the corresponding coupling coefficient is set to 0 to simplify subsequent analysis.

[0077] Furthermore, a matrix framework is constructed using independent variables as row indices and dependent variables as column indices, and the corresponding positions in the matrix are filled according to the coupling coefficients between each pair of variables to form a final coupling coefficient matrix. This matrix intuitively shows the coupling strength between each independent variable and the dependent variable, helps identify the key coupling paths in the system, and provides a basis for further optimization. The embodiment of the present invention not only reveals the intrinsic connection between parameters, but also removes unnecessary weak coupling relationships by setting thresholds, so that the analysis is more focused on significant influencing factors.

[0078] S4: performing error analysis on the multidimensional parameter set to obtain dynamic distribution characteristics of the cumulative error;

[0079] As one of the optional embodiments, performing error analysis on the multidimensional parameter set to obtain dynamic distribution characteristics of cumulative errors includes:

[0080] Dividing the multidimensional parameter set into a number of time windows according to a preset time length;

[0081] Calculating the difference between the parameter value in the time window and the preset reference value to obtain an error sequence;

[0082] Extracting statistical features of the error sequence to obtain error statistical features;

[0083] A time series analysis is performed on the error statistical characteristics to obtain a dynamic distribution characteristic of the accumulated error.

[0084] Specifically, the data in the multidimensional parameter set is divided into multiple continuous time windows according to preset time intervals (such as every 10 seconds or every minute), and each window contains all parameter values within the time period. For the parameter values in each time window, the difference between them and the preset benchmark value is calculated one by one to generate the error sequence within the window. Subsequently, statistical features are extracted from the error sequence of each window, including the mean value of the error, the fluctuation range of the maximum and minimum values, the standard deviation, the skewness and the kurtosis. Finally, the statistical features of the different time windows are arranged in chronological order, and the dynamic distribution characteristics of the cumulative error are obtained through time series analysis methods such as trend line fitting, periodicity detection or fluctuation pattern recognition, which explains the law of error change over time, such as whether the error shows a gradually increasing trend, whether there are periodic fluctuations or sudden drastic changes. These dynamic distribution features, such as the rising trend of the error mean and the periodic changes in the fluctuation amplitude, are used to characterize the pattern of error evolution over time and provide a basis for subsequent error prediction and compensation. The embodiment of the present application converts the original parameter error into quantifiable dynamic features through window analysis and statistical modeling, thereby capturing the potential law of system performance change over time.

[0085] In a specific example, if the multi-dimensional parameter set of the DC power supply includes input voltage (nominal value 24V), output voltage (nominal value 12V), output ripple (nominal value ≤50mV) and load current (nominal value 10A). First, the monitoring data for one hour is divided into six time windows of 10 minutes each. In the first window (0-10 minutes), if the average value of the actual input voltage is 23.8V, the error with the reference value of 24V is -0.2V; the average value of the output ripple is 60mV, and the error is +10mV. The error values of all parameters in each window are calculated in a similar way to form an error sequence. Statistical features are then extracted from the error sequence for each window. For example, in the third window (20-30 minutes), the input voltage error has a mean value of -0.5V, fluctuates between -0.3V and -0.7V, and has a standard deviation of 0.15V, indicating a small but persistent error fluctuation. Meanwhile, the output ripple error's mean value suddenly rises to +20mV, and its standard deviation increases to 4mV, suggesting the presence of transient interference. Finally, a time series analysis of the error statistical features reveals that the mean value of the input voltage error gradually increases over time (from -0.2V to -0.5V), indicating device aging or temperature drift. The output ripple error exhibits periodic fluctuations (every two minutes) in the fifth window (40-50 minutes), synchronized with the periodic variation of the load current. These dynamic features (such as error trends and periodic patterns) are integrated into dynamic distribution features for subsequent prediction of the spatiotemporal evolution of the error.

[0086] S5: performing error prediction and compensation according to the dynamic distribution characteristics to obtain an error correction sequence;

[0087] As one of the optional embodiments, performing error prediction and compensation according to the dynamic distribution characteristics to obtain an error correction sequence includes:

[0088] Inputting the dynamic distribution features into a preset error prediction model to obtain an error prediction sequence; the error prediction model is obtained by training a long short-term memory network based on historical data;

[0089] Performing reverse offset calculation on the error prediction sequence to obtain an error correction sequence;

[0090] Performing application simulation based on the error correction sequence to generate a convergence indicator;

[0091] Determine whether the convergence index is less than a preset convergence threshold; if so, output an error correction sequence; if not, optimize the parameters of the error prediction model, reacquire the error prediction sequence, error correction sequence and convergence index until the convergence index is less than the preset convergence threshold, and output the error correction sequence.

[0092] The training process of the error prediction model includes:

[0093] The historical dynamic distribution features are used as input variables and the historical error values are used as labels to construct a training sample set.

[0094] The training sample set is used to perform model training, and the difference between the predicted error and the actual error is calculated using a loss function;

[0095] Based on gradient descent optimization, adjust the model parameters to minimize the loss function;

[0096] When the prediction error of the error prediction model on the validation set is lower than the preset training threshold, the model training is considered complete.

[0097] Specifically, the error prediction model uses an LSTM (Long Short-Term Memory) network as its core architecture. Its structure consists of an input gate, a forget gate, and an output gate, enabling it to capture long-term dependencies in time series data. The specific steps are as follows: First, historical dynamic distribution features are used as the input sequence, and the corresponding historical actual error values are used as labels to construct a training sample set. This training sample set is then fed into the LSTM network. The network processes the data moment by moment, dynamically updating its internal memory cells to learn the mapping between input features and errors. In each round of training, the model predicts error values based on the current parameters, calculates the difference between the predicted and actual values, and optimizes the parameters using a backpropagation algorithm combined with gradient descent to gradually reduce the prediction error. This process is repeated until the model's prediction error on the validation dataset falls below a preset threshold, for example, a mean error of less than 0.1%. At this point, the model is deemed to have sufficient predictive power and training is complete. The LSTM's time series modeling capabilities enable it to effectively handle complex changes in dynamic distribution features, providing a reliable basis for subsequent error correction.

[0098] After obtaining the error prediction sequence based on the error prediction model, an error correction sequence is first generated through reverse offset calculation. For example, if the error predicted at the next moment is +2V (the output voltage is too high), the error correction sequence will generate a compensation value of -2V, and the output voltage will be adjusted in real time through the control circuit to offset the prediction error.

[0099] The error correction sequence is then applied to a simulation or actual system to simulate the system's operating state after compensation. Key metrics, such as the standard deviation of the residual error, fluctuation amplitude, and number of deviations, are extracted as convergence indicators. If the convergence indicator falls below a preset threshold, the error correction sequence is considered valid and output for actual compensation. If the convergence indicator fails to meet the threshold, it indicates a deviation in the model prediction, requiring a return to the training phase to optimize the parameters of the error prediction model, regenerate the prediction sequence, and revalidate it. This iterative process continues until the correction effect meets the requirements, and the error correction sequence is output to ensure the accuracy and stability of error compensation.

[0100] S6: Calculate a comprehensive score of the DC power supply based on the multi-dimensional parameter set, the coupling coefficient matrix, and the error correction sequence.

[0101] As one of the optional embodiments, the calculating of the comprehensive score of the DC power supply according to the multidimensional parameter set, the coupling coefficient matrix and the error correction sequence includes:

[0102] Calculating a standardized parameter score based on the multidimensional parameter set and preset parameter weights;

[0103] Calculating a coupling influence value according to the coupling coefficient matrix and a preset coupling coefficient weight;

[0104] Calculating an error index value based on the error correction sequence and a preset time weight;

[0105] A weighted sum is performed on the standardized parameter score, the coupling influence value, and the error index value to obtain a comprehensive score of the DC power supply.

[0106] Specifically, the normalized parameter score is calculated by calculating the normalized degree of deviation of each parameter from its mean (i.e., the difference between the parameter value and the mean divided by the standard deviation), multiplying it by the weight of the parameter, and finally taking the average value. For example, if the weight of the input voltage is high and its actual value deviates greatly from the mean, it will significantly affect the score. This process quantifies the severity of each parameter's deviation from the normal range and its weighted impact on the overall performance. The normalized parameter score is calculated using the following formula:

[0107]

[0108] Among them, S p represents the standardized parameter score, N represents the total number of parameters, and x i represents the parameter value of the i-th parameter in the multidimensional parameter set, μ represents the mean of the multidimensional parameter set, σ represents the standard deviation of the multidimensional parameter set, and w p,i Represents the weight of the i-th parameter.

[0109] The coupling impact value is calculated by multiplying the absolute value of each element in the coupling coefficient matrix by the corresponding weight and then averaging it. This value reflects the impact of the interaction between parameters on system stability. For example, if the coupling coefficient between input voltage and output ripple is large and the weight is high, it indicates that the strong correlation between the two will cause chain reactions and require close monitoring. The coupling impact value is calculated using the following formula:

[0110]

[0111] Among them, S c represents the coupling influence value, M represents the total number of coupling coefficient matrix rows, L represents the total number of coupling coefficient matrix columns, m represents the number of coupling coefficient matrix rows, l represents the number of coupling coefficient matrix columns, C m,l The coupling coefficient of the mth row and lth column is w c,m,l Represents the weight of the coupling coefficient in the mth row and lth column.

[0112] The error index is a weighted sum of the error values at each time point in the error correction sequence to evaluate the overall effect of the error correction. For example, if the error correction effect is poor during a sudden load change, its high weight will lower the error index. The error index is calculated using the following formula:

[0113]

[0114] Among them, S e represents the error index, T represents the total number of timestamps, t represents the timestamp number, e t represents the error correction value of timestamp t, w e,t Represents the time weight of timestamp t.

[0115] Finally, a comprehensive score for the DC power supply is obtained by weightedly summing the standardized parameter scores, coupling influence values, and error index values. If the comprehensive score of the DC power supply is greater than or equal to the preset pass threshold, the DC power supply is considered qualified; if the comprehensive score of the DC power supply is less than the preset pass threshold, the DC power supply is considered unqualified.

[0116] As one of the optional embodiments, after calculating and obtaining the comprehensive score of the DC power supply, the method further includes:

[0117] If the comprehensive score is less than the preset qualified threshold, the random forest algorithm is used to perform anomaly tracing analysis based on the multidimensional parameter set, the coupling coefficient matrix and the error correction sequence, to determine the key anomaly features, and generate a performance deviation distribution report based on the key anomaly features.

[0118] Specifically, the qualified threshold of the DC power supply is set to 85 points. If the comprehensive score calculation result of a device is 78 points, the abnormality tracing process is triggered, and the random forest algorithm is used to trace the abnormality of the DC power supply and generate a performance deviation distribution report.

[0119] Taking an industrial DC power supply as an example, its multidimensional parameter set includes input voltage, output ripple, load current, etc. Its coupling coefficient matrix shows that the coupling coefficient between input voltage and output ripple is high (such as 0.8), but the actual input voltage fluctuations frequently exceed the normal range.

[0120] The specific process for tracing DC power supply anomalies using the random forest algorithm is as follows: First, the DC power supply's multi-dimensional parameter set, coupling coefficient matrix, and error correction sequence are used as feature data, combined with historically labeled data of passing and failing equipment to construct a training set. The random forest model uses a bagging strategy to generate multiple decision trees, each trained on a randomly selected subset of features. Ultimately, voting determines the key anomaly features. Feature importance analysis reveals that "input voltage fluctuation amplitude" and "input voltage-output ripple coupling coefficient" have the highest weights (e.g., 40% and 30% of the total importance), indicating that input voltage instability and its strong coupling with output ripple are the primary causes of performance deviation. Based on this, the system generates a performance deviation distribution report that lists key anomaly parameters (e.g., input voltage standard deviation exceeding the standard), coupling relationships (e.g., a strong correlation between input voltage and ripple), and identifies high-frequency time windows where the problem occurs (e.g., error correction failure during a sudden load increase). This provides targeted repair or parameter adjustment options (e.g., inspecting the input filter circuit or optimizing the compensation algorithm for the coupling path). The embodiments of the present invention automatically locate the root cause of the problem through machine learning, avoiding the inefficiency of manual troubleshooting.

[0121] The embodiment of the present invention constructs a multidimensional parameter set by acquiring the dynamic parameters and static parameters of the DC power supply, thereby improving the comprehensiveness and accuracy of parameter acquisition and providing a rich data basis for subsequent analysis; feature extraction is performed based on the multidimensional parameter set to obtain independent variables and dependent variables, which is conducive to simplifying complex parameter relationships, clarifying the relationship between variables, and improving the accuracy of analysis; the coupling coefficients between each variable are calculated based on the independent variables and the dependent variables to obtain a coupling coefficient matrix, which not only reveals the deep relationship between the variables, but also can further guide the system optimization direction, which is conducive to improving system performance and stability; error analysis is performed on the multidimensional parameter set to obtain the dynamic distribution characteristics of the cumulative error, and error prediction and compensation are performed based on the dynamic distribution characteristics to obtain an error correction sequence, which can reduce the uncertainty caused by the error and further improve accuracy and reliability; based on the multidimensional parameter set, the comprehensive score of the DC power supply is calculated, which fully considers the influence of various factors, thereby improving the comprehensiveness and accuracy of the DC current test.

[0122] Correspondingly, the present invention also provides a comprehensive testing device for a DC power supply, which can implement all the processes of the comprehensive testing method for a DC power supply in the above embodiment.

[0123] See also Figure 2 , Figure 2 1 is a schematic diagram of the structure of a comprehensive test device for a DC power supply provided by an embodiment of the present invention. The comprehensive test device for a DC power supply comprises:

[0124] The data acquisition module 201 is configured to acquire dynamic parameters and static parameters of the DC power supply and construct a multidimensional parameter set based on the dynamic parameters and the static parameters; the static parameters are parameters that do not change over time under stable operating conditions of the DC power supply, and the dynamic parameters are parameters used to evaluate the performance of the DC power supply under transient conditions;

[0125] A feature extraction module 202 is configured to extract features based on the multidimensional parameter set to obtain independent variables and dependent variables; the independent variables are parameters with low correlation, and the dependent variables are parameters whose value changes are affected by the independent variables;

[0126] A coupling calculation module 203 is configured to calculate the coupling coefficients between the variables based on the independent variables and the dependent variables to obtain a coupling coefficient matrix;

[0127] An error analysis module 204 is configured to perform error analysis on the multi-dimensional parameter set to obtain dynamic distribution characteristics of the accumulated error;

[0128] An error correction module 205 is configured to perform error prediction and compensation based on the dynamic distribution characteristics to obtain an error correction sequence;

[0129] The power supply scoring module 206 is configured to calculate a comprehensive score of the DC power supply based on the multi-dimensional parameter set, the coupling coefficient matrix, and the error correction sequence.

[0130] Preferably, the feature extraction module 202 is specifically used to:

[0131] performing standardization processing on the multidimensional parameter set to obtain a standardized parameter matrix;

[0132] Performing covariance calculation based on the standardized parameter matrix to obtain a covariance matrix;

[0133] Extracting the eigenvalues and eigenvectors of the covariance matrix, and selecting several eigenvectors with the largest eigenvalues to construct a principal component matrix;

[0134] Performing a matrix multiplication operation on the principal component matrix and the standardized parameter matrix to obtain a dimension reduction matrix;

[0135] A weight analysis is performed based on the dimension reduction matrix and the principal component matrix to obtain independent variables and dependent variables.

[0136] Preferably, the coupling calculation module 203 is specifically used to:

[0137] Calculating the correlation between the independent variable and the dependent variable, and calculating the coupling coefficient between the variables based on the correlation;

[0138] The independent variables and the dependent variables are used as row and column indices of a matrix, and a coupling coefficient matrix is constructed based on the coupling coefficients between the variables.

[0139] Preferably, the error analysis module 204 is specifically used to:

[0140] Dividing the multidimensional parameter set into a number of time windows according to a preset time length;

[0141] Calculating the difference between the parameter value in the time window and the preset reference value to obtain an error sequence;

[0142] Extracting statistical features of the error sequence to obtain error statistical features;

[0143] A time series analysis is performed on the error statistical characteristics to obtain a dynamic distribution characteristic of the accumulated error.

[0144] Preferably, the error correction module 205 is specifically configured to:

[0145] Inputting the dynamic distribution features into a preset error prediction model to obtain an error prediction sequence; the error prediction model is obtained by training a long short-term memory network based on historical data;

[0146] Performing reverse offset calculation on the error prediction sequence to obtain an error correction sequence;

[0147] Performing application simulation based on the error correction sequence to generate a convergence indicator;

[0148] Determine whether the convergence index is less than a preset convergence threshold; if so, output an error correction sequence; if not, optimize the parameters of the error prediction model, reacquire the error prediction sequence, error correction sequence and convergence index until the convergence index is less than the preset convergence threshold, and output the error correction sequence.

[0149] Preferably, the power scoring module 206 is specifically configured to:

[0150] Calculating a standardized parameter score based on the multidimensional parameter set and preset parameter weights;

[0151] Calculating a coupling influence value according to the coupling coefficient matrix and a preset coupling coefficient weight;

[0152] Calculating an error index value based on the error correction sequence and a preset time weight;

[0153] A weighted sum is performed on the standardized parameter score, the coupling influence value, and the error index value to obtain a comprehensive score of the DC power supply.

[0154] Preferably, the comprehensive testing device for the DC power supply is further used for:

[0155] If the comprehensive score is less than the preset qualified threshold, the random forest algorithm is used to perform anomaly tracing analysis based on the multidimensional parameter set, the coupling coefficient matrix and the error correction sequence, to determine the key anomaly features, and generate a performance deviation distribution report based on the key anomaly features.

[0156] In specific implementation, the working principle, control process and technical effects achieved by the comprehensive testing device for a DC power supply provided in an embodiment of the present invention are the same as those of the comprehensive testing method for a DC power supply in the above embodiment, and will not be repeated here.

[0157] See also Figure 3 , Figure 3 This is a block diagram of a computer device provided in an embodiment of the present invention. The computer device includes a processor 301, a memory 302, and a computer program stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program, the steps of the aforementioned embodiment of the comprehensive testing method for a DC power supply are implemented. Alternatively, when the processor 301 executes the computer program, the functions of the modules / units in the aforementioned device embodiments are implemented.

[0158] For example, the computer program may be divided into one or more modules / units, which are stored in the memory 302 and executed by the processor 301 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the computer device.

[0159] The computer device may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will appreciate that the schematic diagram is merely an example of a computer device and does not limit the computer device. The computer device may include more or fewer components than shown, or a combination of certain components, or different components. For example, the computer device may also include input and output devices, network access devices, buses, and the like.

[0160] The processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor. The processor 301 is the control center of the computer device, connecting various parts of the entire computer device using various interfaces and lines.

[0161] The memory 302 can be used to store the computer programs and / or modules. The processor 301 implements various functions of the computer device by running or executing the computer programs and / or modules stored in the memory 302 and calling the data stored in the memory 302. The memory 302 can mainly include a program storage area and a data storage area. The program storage area can store an operating system and at least one application required for a function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created based on the use of the mobile phone (such as audio data, a phone book, etc.). In addition, the memory 302 can include high-speed random access memory and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0162] Wherein, if the module / unit integrated in the computer device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor 301, it can implement the steps of the above-mentioned various method embodiments. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0163] An embodiment of the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the comprehensive testing method for the DC power supply described in any of the above embodiments.

[0164] Embodiments of the present invention provide a comprehensive testing method, apparatus, device, and storage medium for a DC power supply, which have the following beneficial effects: by acquiring dynamic parameters and static parameters of the DC power supply to construct a multidimensional parameter set, the comprehensiveness and accuracy of parameter acquisition are improved, and a rich data basis is provided for subsequent analysis; feature extraction is performed based on the multidimensional parameter set to obtain independent variables and dependent variables, which is conducive to simplifying complex parameter relationships, clarifying the relationship between variables, and improving the accuracy of analysis; coupling coefficients between each variable are calculated based on the independent variables and the dependent variables to obtain a coupling coefficient matrix, which not only reveals the deep relationship between the variables but also further guides the direction of system optimization, which is conducive to improving system performance and stability; error analysis is performed on the multidimensional parameter set to obtain dynamic distribution characteristics of the cumulative error, and error prediction and compensation are performed based on the dynamic distribution characteristics to obtain an error correction sequence, which can reduce the uncertainty caused by the error and further improve accuracy and reliability; a comprehensive score of the DC power supply is calculated based on the multidimensional parameter set, the coupling coefficient matrix, and the error correction sequence, which fully considers the influence of various factors, thereby improving the comprehensiveness and accuracy of the DC current test.

[0165] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A comprehensive test method for a DC power supply, characterized in that: include: Acquiring dynamic parameters and static parameters of a DC power supply, and constructing a multidimensional parameter set based on the dynamic parameters and the static parameters; The static parameters are parameters that do not change over time under stable operating conditions of the DC power supply, and the dynamic parameters are parameters that evaluate the performance of the DC power supply under transient conditions; Perform feature extraction based on the multidimensional parameter set to obtain independent variables and dependent variables; The independent variables are parameters with low mutual correlation, and the dependent variables are parameters whose value changes are affected by the independent variables; Based on the independent variables and the dependent variables, the coupling coefficients between the variables are calculated to obtain a coupling coefficient matrix; Performing error analysis on the multidimensional parameter set to obtain dynamic distribution characteristics of cumulative errors; Perform error prediction and compensation according to the dynamic distribution characteristics to obtain an error correction sequence; A comprehensive score of the DC power supply is calculated based on the multidimensional parameter set, the coupling coefficient matrix, and the error correction sequence.

2. The comprehensive testing method for a DC power supply according to claim 1, wherein: The feature extraction is performed based on the multidimensional parameter set to obtain independent variables and dependent variables, including: performing standardization processing on the multidimensional parameter set to obtain a standardized parameter matrix; Performing covariance calculation based on the standardized parameter matrix to obtain a covariance matrix; Extracting the eigenvalues and eigenvectors of the covariance matrix, and selecting several eigenvectors with the largest eigenvalues to construct a principal component matrix; Performing a matrix multiplication operation on the principal component matrix and the standardized parameter matrix to obtain a dimension reduction matrix; A weight analysis is performed based on the dimension reduction matrix and the principal component matrix to obtain independent variables and dependent variables.

3. The comprehensive testing method for a DC power supply according to claim 1, wherein: The step of calculating the coupling coefficients between the variables based on the independent variables and the dependent variables to obtain a coupling coefficient matrix includes: Calculating the correlation between the independent variable and the dependent variable, and calculating the coupling coefficient between the variables based on the correlation; The independent variables and the dependent variables are used as row and column indices of a matrix, and a coupling coefficient matrix is constructed based on the coupling coefficients between the variables.

4. The comprehensive testing method for a DC power supply according to claim 1, wherein: The performing error analysis on the multi-dimensional parameter set to obtain dynamic distribution characteristics of cumulative errors includes: Dividing the multidimensional parameter set into a number of time windows according to a preset time length; Calculating the difference between the parameter value in the time window and the preset reference value to obtain an error sequence; Extracting statistical features of the error sequence to obtain error statistical features; A time series analysis is performed on the error statistical characteristics to obtain a dynamic distribution characteristic of the accumulated error.

5. The comprehensive testing method for a DC power supply according to claim 1, wherein: The error prediction and compensation are performed according to the dynamic distribution characteristics to obtain an error correction sequence, including: Inputting the dynamic distribution features into a preset error prediction model to obtain an error prediction sequence; the error prediction model is obtained by training a long short-term memory network based on historical data; Performing reverse offset calculation on the error prediction sequence to obtain an error correction sequence; Performing application simulation based on the error correction sequence to generate a convergence indicator; Determine whether the convergence index is less than a preset convergence threshold; if so, output an error correction sequence; if not, optimize the parameters of the error prediction model, reacquire the error prediction sequence, error correction sequence and convergence index until the convergence index is less than the preset convergence threshold, and output the error correction sequence.

6. The comprehensive testing method for a DC power supply according to claim 1, wherein: The calculating, based on the multidimensional parameter set, the coupling coefficient matrix, and the error correction sequence, to obtain a comprehensive score of the DC power supply includes: Calculating a standardized parameter score based on the multidimensional parameter set and preset parameter weights; Calculating a coupling influence value according to the coupling coefficient matrix and a preset coupling coefficient weight; Calculating an error index value based on the error correction sequence and a preset time weight; A weighted sum is performed on the standardized parameter score, the coupling influence value, and the error index value to obtain a comprehensive score of the DC power supply.

7. The comprehensive testing method for a DC power supply according to claim 1, wherein: After obtaining the comprehensive score of the DC power supply through calculation, the method further includes: If the comprehensive score is less than the preset qualified threshold, the random forest algorithm is used to perform anomaly tracing analysis based on the multidimensional parameter set, the coupling coefficient matrix and the error correction sequence, to determine the key anomaly features, and generate a performance deviation distribution report based on the key anomaly features.

8. A comprehensive test device for a DC power supply, characterized in that: include: A data acquisition module, configured to acquire dynamic parameters and static parameters of the DC power supply, and construct a multidimensional parameter set based on the dynamic parameters and the static parameters; The static parameters are parameters that do not change over time under stable operating conditions of the DC power supply, and the dynamic parameters are parameters that evaluate the performance of the DC power supply under transient conditions; A feature extraction module, configured to extract features based on the multidimensional parameter set to obtain independent variables and dependent variables; The independent variables are parameters with low mutual correlation, and the dependent variables are parameters whose value changes are affected by the independent variables; A coupling calculation module, configured to calculate the coupling coefficients between the variables based on the independent variables and the dependent variables to obtain a coupling coefficient matrix; An error analysis module, configured to perform error analysis on the multidimensional parameter set to obtain dynamic distribution characteristics of the accumulated error; An error correction module, configured to perform error prediction and compensation according to the dynamic distribution characteristics to obtain an error correction sequence; A power supply scoring module is used to calculate a comprehensive score of the DC power supply based on the multi-dimensional parameter set, the coupling coefficient matrix and the error correction sequence.

9. A computer device, characterized in that: The device comprises a processor and a memory, wherein a computer program is stored in the memory and the computer program is configured to be executed by the processor, and when the processor executes the computer program, the comprehensive testing method for a DC power supply according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the device where the computer-readable storage medium is located executes the computer program, the comprehensive testing method for a DC power supply according to any one of claims 1 to 7 is implemented.

Images