Raman spectrum detection and correction method, system and equipment for transformer oil and medium
By combining Pearson correlation coefficient and variance analysis with Bayesian optimization algorithm and BP neural network, a correction model for Raman spectroscopy detection of transformer oil was constructed, which solved the problem of the influence of environmental factors on the detection results and achieved high accuracy and comprehensive detection in complex environments.
Patent Information
- Application Number
- CN202511758552.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-27
- Publication Date
- 2026-02-17
AI Technical Summary
Existing Raman spectroscopy detection methods fail to effectively consider the influence of environmental factors, resulting in insufficient accuracy and reliability of the detection results, making it difficult to fully reflect the true state of transformer oil.
By using Pearson correlation coefficient analysis and variance analysis, the correlation between single operating condition parameters such as temperature, humidity, and vibration and spectral characteristic parameters is quantified. A multi-factor correction model is constructed, and combined with Bayesian optimization algorithm and BP neural network, spectral data correction under complex environments is achieved.
It improves the comprehensiveness and accuracy of detection, and can fully consider the synergistic interference of multiple factors in complex environments, thereby improving the reliability of transformer oil condition monitoring.
Smart Images
Figure CN121540689A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of Raman spectroscopy detection technology, specifically to a method, system, equipment, and medium for correcting Raman spectroscopy detection of transformer oil. Background Technology
[0002] In power systems, transformers are critical equipment, and their operating status directly affects the safety and stability of the power grid. Transformer oil, as an important medium for insulation and cooling within the transformer, often reflects the transformer's health condition through changes in its properties. Raman spectroscopy, as a non-destructive, rapid, and sensitive analytical method, is widely used in transformer oil condition monitoring. However, in practical applications, the operating environment of transformer oil is complex and variable; factors such as temperature, humidity, and vibration can all interfere with Raman spectroscopy results, affecting the accuracy and reliability of the data.
[0003] Traditional Raman spectroscopy methods often neglect the influence of environmental factors on the detection results, or only consider the effect of a single environmental factor, making it difficult to comprehensively and accurately reflect the true state of transformer oil. Therefore, how to effectively correct for the influence of environmental factors on the Raman spectroscopy results of transformer oil and improve the accuracy and reliability of the detection has become an urgent problem to be solved.
[0004] Based on this, the present invention aims to provide a method, system, device and medium for correcting the Raman spectroscopy detection of transformer oil, so as to solve the related problems mentioned above. Summary of the Invention
[0005] The technical problem this invention aims to solve is that existing Raman spectroscopy detection methods neglect the influence of the operating environment on the detection results, making it difficult to accurately reflect the true state of transformer oil. The purpose is to provide a correction method, system, equipment, and medium for transformer oil Raman spectroscopy detection. Through Pearson correlation coefficient analysis, the linear correlation between single operating condition parameters such as temperature, humidity, and vibration and spectral characteristic parameters such as peak position and peak intensity is accurately quantified. Through variance analysis, the main effects and interaction effects of temperature, humidity, and vibration are decomposed, clarifying the synergistic interference laws of multiple factors under complex environments. This solves the problem of existing technologies neglecting the synergistic effects of multiple factors and failing to quantify the influence of complex environments, enabling the correction process to fully consider the complex interferences under typical or extreme environments of urban transformers, thus improving the comprehensiveness of the detection.
[0006] This invention is achieved through the following technical solution:
[0007] A method for correcting the Raman spectroscopy detection of transformer oil, the method comprising:
[0008] Data feature parameters were extracted from Raman spectral data of different oil sample types under a single operating condition, and Pearson correlation coefficients between data feature parameters and corresponding operating condition parameters were calculated to construct the correlation between data feature parameters and single operating condition parameters.
[0009] Data characteristic parameters of Raman spectral data of different oil sample types under multiple operating conditions were extracted. The parameter effects of the multiple operating conditions parameters were analyzed using the analysis of variance method to determine the influence weight between the data characteristic parameters and the multiple operating conditions parameters.
[0010] Multiple single-factor and multi-factor correction sub-models are constructed, and a Bayesian optimization algorithm is used to weight and fuse these sub-models to construct a globally coupled correction model. The single-factor correction sub-models are constructed by combining the correlation between data feature parameters and single operating condition parameters with a multiple linear regression model. The multi-factor correction model is obtained by training a backpropagation neural network by inputting the influence weights between data feature parameters and multiple operating condition parameters.
[0011] The spectral data of the transformer oil to be tested and the operating parameters of the current operating conditions are obtained. The operating parameters are input into the global coupled correction model to obtain the parameter correction weights. The spectral data is corrected using the parameter correction weights to obtain the corrected spectral data.
[0012] Furthermore, the single operating environment includes temperature operating environment, humidity operating environment, and vibration operating environment; the operating parameters include temperature, humidity, and vibration frequency; the different oil sample types include new oil, lightly aged oil, moderately aged oil, and heavily aged oil; and the data characteristic parameters include peak position, peak intensity, full width at half maximum (FWHM), and peak area.
[0013] Furthermore, data feature parameters of Raman spectral data for different oil sample types under a single operating condition were extracted, and the Pearson correlation coefficient between the data feature parameters and the corresponding operating condition parameters was calculated. The correlation between the data feature parameters and the single operating condition parameters was then constructed, specifically as follows:
[0014] Multiple data feature parameters were extracted from the Raman spectral data of four oil sample types under different single operating conditions, and four environment-feature parameter pairing datasets corresponding to the oil sample types were constructed. Each environment-feature parameter pairing dataset contains multiple data matrices, and each data matrix represents the correlation between an operating condition parameter and a data feature parameter.
[0015] For each data matrix, the Pearson correlation coefficient is calculated using the Pearson correlation coefficient calculation function. Based on the Pearson correlation coefficient of the data matrix, the correlation of the data matrix is filtered to obtain the filtered environment-feature parameter paired dataset.
[0016] Furthermore, data characteristic parameters of Raman spectral data of different oil sample types under multiple operating conditions were extracted. Analysis of variance was used to analyze the parameter effects of these multiple operating condition parameters, determining the influence weights between the data characteristic parameters and the multiple operating condition parameters. Specifically:
[0017] Multiple data feature parameters are extracted from Raman spectral data of different oil sample types under multiple operating conditions, and feature parameter matrix sets corresponding to the four data feature parameters are constructed. Each feature parameter matrix set contains multiple three-dimensional data matrices constructed based on oil sample type, multiple operating condition parameters, and data feature parameters.
[0018] Based on the correlation between operating condition parameters and data feature parameters in the feature parameter matrix, an ANOVA model is constructed; based on the ANOVA model, the effects of each operating condition parameter are decomposed to obtain the sum of squares of each effect and the net sum of squares of the effects.
[0019] Based on the proportion of each operating condition parameter effect in the net sum of squared effects, the effect weight of each operating condition parameter effect is calculated. By combining the effect weight of each operating condition parameter effect with the corresponding data feature parameters, an effect weight set representing the influence weight between data feature parameters and multiple operating condition parameters is constructed.
[0020] Furthermore, after constructing four paired environmental-feature parameter datasets corresponding to oil sample types, the method also includes performing normality and linearity checks on each data matrix to obtain paired environmental-feature parameter datasets that pass both checks.
[0021] Furthermore, after calculating the sum of squares of the effects of each operating condition parameter and the net sum of squares of the effects, the method also includes:
[0022] Obtain the degrees of freedom for the effect of each operating condition parameter, and calculate the F-value for each operating condition parameter effect using the sum of the degrees of freedom and squares of each operating condition parameter effect; based on the F-value, filter the significance of each operating condition parameter effect to obtain a set of significant effects with higher significance.
[0023] The present invention also provides a transformer oil Raman spectroscopy detection correction system, which is used in the transformer oil Raman spectroscopy detection correction method described in any one of the above claims, the system comprising:
[0024] The single-operation condition analysis module is used to extract the data feature parameters of Raman spectral data of different oil sample types under a single operating condition, calculate the Pearson correlation coefficient between the data feature parameters and the corresponding operating condition parameters, and construct the correlation between the data feature parameters and the single operating condition parameters.
[0025] The multi-operating-condition analysis module is used to extract the data characteristic parameters of Raman spectral data of different oil sample types under multiple operating conditions. The parameter effect of the multi-operating-condition parameters is analyzed using the variance analysis method to determine the influence weight between the data characteristic parameters and the multi-operating-condition parameters.
[0026] The correction model construction module is used to construct multiple single-factor correction sub-models and multi-factor correction sub-models. A Bayesian optimization algorithm is used to weight and fuse these sub-models to construct a globally coupled correction model. The single-factor correction sub-models are constructed by combining the correlation between data feature parameters and single operating condition parameters with a multiple linear regression model. The multi-factor correction model is obtained by training a BP neural network by inputting the influence weights between data feature parameters and multiple operating condition parameters.
[0027] The spectral data correction module is used to acquire the spectral data of the transformer oil to be tested and the operating condition parameters of the current operating condition. The operating condition parameters are input into the global coupled correction model to obtain the parameter correction weights. The spectral data is corrected using the parameter correction weights to obtain the corrected spectral data.
[0028] The present invention also provides a computer device, including a system memory and a processor, wherein the system memory stores a computer program, and the processor executes the computer program to implement the steps of any of the methods described above.
[0029] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods described above.
[0030] The present invention also provides a computer program product containing instructions that, when executed by a cluster of computer devices, cause the cluster of computer devices to perform the method described in any of the preceding claims.
[0031] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0032] In this invention, Pearson correlation coefficient analysis is used to accurately quantify the linear correlation between single operating condition parameters such as temperature, humidity, and vibration and spectral characteristic parameters such as peak position and peak intensity. Variance analysis is used to decompose the main effects and interaction effects of temperature, humidity, and vibration, clarifying the synergistic interference laws of multiple factors in a complex environment. This solves the problem that existing technologies ignore the synergy of multiple factors and cannot quantify the impact of complex environments, enabling the correction process to fully consider the complex interferences in typical or extreme environments of urban transformers and improve the comprehensiveness of detection. Attached Figure Description
[0033] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:
[0034] Figure 1 This is a schematic diagram of the method flow for a transformer oil Raman spectroscopy detection correction method in this embodiment;
[0035] Figure 2 This is a schematic diagram of the module connection of a transformer oil Raman spectroscopy detection and correction system in this embodiment;
[0036] Figure 3 This is a schematic diagram of the structure of a computer device in this embodiment. Detailed Implementation
[0037] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.
[0038] In this disclosure, unless otherwise stated, the use of terms such as "first," "second," etc., to describe various elements is not intended to limit the positional, temporal, or importance relationships of these elements; such terms are merely used to distinguish one element from another. In some examples, the first element and the second element may refer to the same instance of that element, while in other cases, based on the context, they may refer to different instances.
[0039] The terminology used in the description of the various examples in this disclosure is for the purpose of describing particular examples only and is not intended to be limiting. Unless the context explicitly indicates otherwise, an element may be one or more unless the number of elements is specifically limited. Furthermore, the term "and / or" as used in this disclosure covers any one of the listed items and all possible combinations thereof.
[0040] Example 1
[0041] See Figure 1 , Figure 1 A schematic flowchart of a method for correcting Raman spectroscopy detection of transformer oil is shown, wherein the method includes:
[0042] S1: Extract the data feature parameters of Raman spectral data of different oil sample types under single operating conditions, calculate the Pearson correlation coefficient between the data feature parameters and the corresponding operating condition parameters, and construct the correlation between the data feature parameters and the single operating condition parameters;
[0043] It should be noted that, in this embodiment, the operating condition parameters include temperature, humidity, and vibration frequency. Temperature data is obtained by setting a temperature sensor, humidity data is obtained by setting a humidity sensor, and vibration frequency data is obtained by setting an acceleration sensor. The different oil sample types include new oil, lightly aged oil, moderately aged oil, and heavily aged oil. The data characteristic parameters include peak position, peak intensity, full width at half maximum (FWHM), and peak area. In other embodiments, other parameter settings may be used, and no further restrictions are imposed here.
[0044] It should also be noted that, in this embodiment, the single operating conditions include temperature operating conditions, humidity operating conditions, and vibration operating conditions. The temperature operating conditions refer to operating conditions where all parameters except temperature are standard parameters, and the other single operating conditions are similar to the temperature operating conditions. Therefore, in this embodiment, the temperature parameters of the temperature operating conditions range from -30℃ to 70℃, with a 5℃ interval between different temperature operating conditions, totaling 19 temperature characteristic parameters. The standard humidity parameters are 50%RH and the vibration frequency. The standard parameter value is 10Hz; the humidity parameter value range for the humidity operating environment is [20%RH, 90%RH], with a 10%RH interval between different humidity operating environments, for a total of 8 humidity characteristic parameters; the standard temperature parameter value is 25℃; and the standard vibration frequency parameter value is 10Hz; the vibration frequency value range for the vibration operating environment is [5Hz, 100Hz], with a 5Hz interval between different vibration operating environments, for a total of 20 vibration frequency characteristic parameters; the standard temperature parameter value is 25℃; and the standard humidity parameter value is 50%RH.
[0045] Specifically, in this embodiment, Raman spectroscopy was first performed on transformer oil samples of four different types under three single-operation conditions to obtain multiple Raman spectral data.
[0046] Then, multiple data feature parameters were extracted from the Raman spectral data of four oil sample types under different single operating conditions, and four environment-feature parameter pairing datasets corresponding to oil sample types were constructed. Each environment-feature parameter pairing dataset contains multiple data matrices, and each data matrix represents the correlation between an operating condition parameter and a data feature parameter.
[0047] For example, in this embodiment, Table 1 below shows the environment-feature parameter pairing dataset containing partial data matrices for transformer oil with the oil sample type being new oil under three single operating conditions; the same applies to other oil sample types.
[0048] Table 1. Partial examples of the environmental-feature parameter paired dataset for oil samples of the new oil type.
[0049]
[0050] For each data matrix, the Pearson correlation coefficient is calculated using the Pearson correlation coefficient calculation function. Based on the Pearson correlation coefficient, the correlation of the data matrices is filtered to obtain a filtered environment-feature parameter paired dataset. Specifically, a Shapiro-Wilk test is performed to determine whether the data feature parameters in each data matrix follow a normal distribution, with a significance level of 0.05. The test statistic W and corresponding p-value are calculated. If the p-value is greater than or equal to 0.05, the data feature parameters conform to a normal distribution; if the p-value is less than 0.05, a logarithmic transformation is used. The data feature parameters were corrected using a different method, and the verification was repeated until the normality requirement was met. Then, a scatter plot was plotted for each data matrix, with the operating condition parameters on the x-axis and the data feature parameters on the y-axis. The scatter plot was used to determine the linear trend. If the scatter points showed a clear linear clustering distribution, a linear relationship was determined. If the linear trend was not obvious, a linear fitting was used to calculate the goodness of fit. If the goodness of fit was greater than or equal to 0.6, a weak linear relationship was determined; if it was less than 0.6, a non-linear relationship was determined, and the data matrix was discarded. Finally, a paired dataset of environment and feature parameters that passed the double verification was obtained.
[0051] Then, for each data matrix in the double-validated environment-feature parameter paired dataset, the Pearson correlation coefficient is calculated using the Pearson correlation coefficient calculation function, where the Pearson correlation coefficient calculation function is: In the formula, This indicates the number of operating condition parameters. ; Indicates the first Each operating condition parameter value; This represents the average value of operating parameters; Indicates the value of data feature parameters; This represents the mean of the data's characteristic parameters.
[0052] Finally, the correlation of the data matrices was filtered based on the Pearson correlation coefficient, resulting in a filtered environment-feature parameter paired dataset. When the absolute value of the Pearson correlation coefficient is greater than or equal to 0.8, the data matrix is considered extremely strongly correlated; when the absolute value of the Pearson correlation coefficient is between 0.7 and 0.8, the data matrix is considered strongly correlated; when the absolute value of the Pearson correlation coefficient is between 0.6 and 0.7, the data matrix is considered moderately correlated; and when the absolute value of the Pearson correlation coefficient is less than 0.6, the data matrix is considered weakly correlated. The weakly correlated data matrices were removed to obtain the filtered environment-feature parameter paired dataset.
[0053] S2: Extract the data characteristic parameters of Raman spectral data of different oil sample types under multiple operating conditions, and use the variance analysis method to analyze the parameter effect of the multiple operating conditions parameters to determine the influence weight between the data characteristic parameters and the multiple operating conditions parameters.
[0054] Specifically, in this embodiment, multiple data feature parameters are extracted from Raman spectral data of different oil sample types under multiple operating conditions, and feature parameter matrix sets corresponding to the four data feature parameters are constructed respectively;
[0055] It should be noted that, in this embodiment, the multi-operating-condition environment refers to the operating environment under different temperatures, different humidity and different vibration frequencies, and each of the feature parameter matrix sets contains multiple three-dimensional data matrices constructed based on oil sample type, multi-operating-condition parameters and data feature parameters.
[0056] For example, in this embodiment, Table 2 below shows the feature parameter matrix set where the data feature parameter is the peak position; the same applies to other data feature parameters.
[0057] Table 2. Examples of feature parameter matrices for peak positions.
[0058]
[0059] Based on the correlation between operating condition parameters and data feature parameters in the feature parameter matrix, an ANOVA model is constructed. Specifically, the constructed ANOVA model is as follows: In the formula, Indicates by the first Oil sample type, first The temperature value, the first The humidity value, the first Measurement values of data characteristic parameters of a three-dimensional data matrix composed of vibration frequencies. This represents the mean of the data feature parameters in the three-dimensional data matrix. Indicates the first Oil-like sample type Indicates the first A temperature value, Indicates the first A humidity value, Indicates the first One vibration frequency, Represents the random error term;
[0060] Based on the ANOVA model, the effects of each working condition parameter are decomposed to obtain the sum of squares of each effect and the net sum of squares of the effects, as detailed below:
[0061] Total sum of squares: In the formula, This represents the sum of squares of all three-dimensional data matrices in the feature parameter matrix set. This represents the mean of the data feature parameters of all three-dimensional data matrices in the feature parameter matrix set;
[0062] Sum of squares of oil sample type variables: In the formula, Indicates the first The number of data feature parameters corresponding to the oil sample type. Indicates the first Mean values of data characteristic parameters corresponding to oil sample types;
[0063] Main effect sum of squares: In the formula, Represents the sum of squares of the temperature effect. Indicates the first The number of data feature parameters corresponding to each temperature value Indicates the first The mean of the data characteristic parameters corresponding to each temperature value; In the formula, Represents the sum of squares of the humidity effect. Indicates the first The number of data feature parameters corresponding to each humidity value Indicates the first The mean of the data characteristic parameters corresponding to each humidity value; In the formula, Represents the sum of squares of the vibration effects. Indicates the first The number of data feature parameters corresponding to each vibration frequency Indicates the first The mean value of the data characteristic parameters corresponding to each vibration frequency;
[0064] Sum of squares of interaction effects: In the formula, This represents the sum of squares of the temperature-humidity interaction effect. This represents the sum of squares of the combined effects of temperature and humidity. , Indicates the first The temperature value and the first The number of data feature parameters corresponding to each humidity value Indicates the first The temperature value and the first Mean values of data characteristic parameters corresponding to each humidity value; sum of squares of humidity-vibration interaction effect. Sum of squares of temperature-vibration interaction effects and the sum of squares of temperature-humidity-vibration interaction effects Similarly, this will not be elaborated upon further here;
[0065] Sum of squared errors: ;
[0066] Net sum of squares of effects: .
[0067] It should be noted that, in this embodiment, the method further includes: obtaining the degrees of freedom of each operating condition parameter effect, and calculating the F value of each operating condition parameter effect using the sum of the degrees of freedom and squares of each operating condition parameter effect; it should be noted that, in this embodiment, the degrees of freedom of each operating condition parameter effect is calculated by subtracting 1 from the number of operating condition parameters. For example, the degrees of freedom of the temperature effect is 18, the degrees of freedom of the humidity effect is 8, and the degrees of freedom of the interaction effect is the product of the degrees of freedom of each main effect.
[0068] The F-value of each working condition parameter effect is calculated using the degrees of freedom and sum of squares of the effect. Specifically, the mean square of each working condition parameter effect is first calculated. , To represent the sum of squares, Represent the degrees of freedom; then calculate the F-value of the effect of each working condition parameter. , This represents the ratio of the effect of the operating condition parameter to the error effect. Finally, based on the F-value, the significance of each operating condition parameter effect is screened to obtain a set of significant effects with higher significance. That is, when the F-value of the operating condition parameter effect is less than the set F-value threshold, the effect is determined to be insignificant and the operating condition parameter effect is removed.
[0069] It should be noted that in this embodiment, the F-value threshold is set to 0.05. Other thresholds can be set in other embodiments, and no further restrictions are imposed here.
[0070] Based on the proportion of each operating condition parameter effect in the net sum of squared effects, the effect weight of each operating condition parameter effect is calculated. By combining the effect weight of each operating condition parameter effect with the corresponding data feature parameters, an effect weight set representing the influence weight between data feature parameters and multiple operating condition parameters is constructed.
[0071] S3: Construct multiple single-factor and multi-factor correction sub-models, and use a Bayesian optimization algorithm to weightedly fuse these sub-models to build a globally coupled correction model. The single-factor correction sub-model is constructed by combining the correlation between data feature parameters and single operating condition parameters with a multiple linear regression model. The multi-factor correction model is obtained by training a BP neural network by inputting the influence weights between data feature parameters and multiple operating condition parameters.
[0072] It should be noted that in this embodiment, the single-factor correction sub-model includes a temperature correction sub-model, a humidity correction sub-model, and a vibration correction sub-model, and their construction and training processes are the same. In this embodiment, the temperature correction sub-model is taken as an example, specifically as follows: First, based on four paired environmental-feature parameter datasets corresponding to the oil sample type, a linear relationship is fitted to each data matrix using a linear regression equation. In the formula, This represents the mean of the data's characteristic parameters. Indicates temperature parameter, For regression coefficients, The intercept is used; then the calculation is performed. , The specific value, It should be noted that, in this embodiment, when the data feature parameter is the peak position, The value is 0.02 when the data characteristic parameter is peak intensity. The value is -0.015 when the data feature parameter is half-width and half-height. The value is 0.008 when the data characteristic parameter is the peak area. The value is -0.012; based on the standard operating condition parameters, calculate the theoretical values of the data characteristic parameters under the standard operating conditions. Define the change in temperature parameter. and the change in data feature parameters In the formula, This represents the actual temperature measurement value. The data feature parameters corresponding to the actual temperature measurements are substituted into the linear regression equation to construct a quantization function of the change in temperature parameters minus the change in feature parameters. The temperature correction sub-model is derived based on the quantization function of the change in temperature parameter minus the change in characteristic parameter, as follows: The same logic applies to other modified sub-models.
[0073] The multi-factor correction sub-model is obtained by training a BP neural network by inputting the influence weights between data feature parameters and multiple operating condition parameters. Specifically, the data feature parameters, operating condition parameters, and influence weights are input into the BP neural network for model training, while the data feature parameters under standard operating conditions are used as the model output labels. The BP neural network includes an input layer, three hidden layers, and an output layer. The BP neural network is iteratively trained until the root mean square error loss function converges to obtain the trained multi-factor correction sub-model.
[0074] Finally, a Bayesian optimization algorithm is used to weightedly fuse multiple single-factor correction sub-models and multi-factor correction sub-models to construct a globally coupled correction model. Specifically, initial coefficients are set for the correction output values of multiple single-factor correction sub-models and the correction output values of multi-factor correction sub-models. , , , ; And define an initial global coupling correction model (taking the peak position as an example): The Bayesian optimization parameters are set, and the Gaussian process model is used as the surrogate model, specifically as follows: In the formula, It is a mean function. Represents the covariance kernel function. Represent a Gaussian process; use the desired improvement function as the acquisition function, specifically the current optimal performance. Candidate weights The expected improvement value is: ,in, This represents the number of samples; the number of optimization iterations is set to 100, and the data feature parameters under standard working conditions are used as the model validation set; the optimal coefficients of the model are searched using the Bayesian optimization algorithm, and the optimal coefficients are finally fused with the initial global coupling correction model to obtain the final global coupling correction model; it should be noted that in this embodiment, the final fusion weights are 0.3, 0.1, 0.2, and 0.4 respectively.
[0075] S4: Obtain the spectral data of the transformer oil to be tested and the operating parameters of the current operating condition. Input the operating parameters into the global coupled correction model to obtain the parameter correction weights. Use the parameter correction weights to correct the spectral data to obtain the corrected spectral data.
[0076] Specifically, in this embodiment, Pearson correlation coefficient analysis is used to accurately quantify the linear correlation between single operating condition parameters such as temperature, humidity, and vibration and spectral characteristic parameters such as peak position and peak intensity; variance analysis is used to decompose the main effects and interaction effects weights of temperature, humidity, and vibration, clarifying the synergistic interference laws of multiple factors under composite operating conditions; this solves the problem that existing technologies ignore the synergy of multiple factors and cannot quantify the influence of composite operating conditions, enabling the correction process to fully consider the complex interference under typical or extreme operating conditions of urban transformers, and improving the comprehensiveness of detection.
[0077] Example 2
[0078] See Figure 2The present invention also provides a transformer oil Raman spectroscopy detection correction system, which is used in the transformer oil Raman spectroscopy detection correction method described in any one of the above claims, the system comprising:
[0079] The single-operation condition analysis module 100 is used to extract the data feature parameters of Raman spectral data of different oil sample types under a single-operation condition environment, calculate the Pearson correlation coefficient between the data feature parameters and the corresponding operating condition parameters, and construct the correlation between the data feature parameters and the single-operation condition parameters.
[0080] The multi-operating-condition analysis module 200 is used to extract the data characteristic parameters of Raman spectral data of different oil sample types under multiple operating conditions, and to analyze the parameter effects of the multi-operating-condition parameters using the variance analysis method to determine the influence weight between the data characteristic parameters and the multi-operating-condition parameters.
[0081] The correction model construction module 300 is used to construct multiple single-factor correction sub-models and multi-factor correction sub-models. A Bayesian optimization algorithm is used to weight and fuse these multiple single-factor and multi-factor correction sub-models to construct a globally coupled correction model. The single-factor correction sub-models are constructed by combining the correlation between data feature parameters and single operating condition parameters with a multiple linear regression model. The multi-factor correction model is obtained by training a BP neural network by inputting the influence weights between data feature parameters and multiple operating condition parameters.
[0082] The spectral data correction module 400 is used to acquire the spectral data of the transformer oil to be tested and the operating condition parameters of the current operating condition. The operating condition parameters are input into the global coupling correction model to obtain the parameter correction weights. The spectral data is corrected using the parameter correction weights to obtain the corrected spectral data.
[0083] It should be noted that the modules in the system of Embodiment 2 correspond to the steps in the method of Embodiment 1. The steps in the method of Embodiment 1 have been described in detail in Embodiment 1, and the module content in the system will not be described in detail in this Embodiment 2.
[0084] Example 3
[0085] See Figure 3 This embodiment also provides a computer device, including a system memory 1005 and a processor 1001. The system memory 1005 stores a computer program, and the processor 1001 executes the computer program to implement the steps of any of the methods described above.
[0086] It should be noted that the processor 1001 is used to execute the steps in the above method embodiments according to the instructions in the program code. Alternatively, when the processor 1001 executes the computer program, it implements the functions of each module / unit in the above system / device embodiments.
[0087] Specifically, in this embodiment, the computer program can be divided into one or more modules / units. One or more modules / units are stored in the system memory 1005 and executed by the processor 1001 to complete this application. One or more modules / units can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the terminal device.
[0088] The terminal device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor 1001 and a system memory 1005. Those skilled in the art will understand that this does not constitute a limitation on the terminal device, which may include more or fewer components than shown, or a combination of certain components, or different components. For example, the terminal device may also include an input / output device 1003, a network access device 1002, a bus 1006, etc.
[0089] The processor 1001 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.
[0090] System memory 1005 can be an internal storage unit of the terminal device, such as a hard drive or RAM. System memory 1005 can also be a storage device 1004 of the terminal device, such as an external hard drive, SmartMedia Card (SMC), Secure Digital (SD) card, or FlashCard. Furthermore, system memory 1005 can include both internal storage units and storage device 1004. System memory 1005 is used to store computer programs and other programs and data required by the terminal device. System memory 1005 can also be used to temporarily store data that has been output or will be output.
[0091] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0092] Example 4
[0093] This embodiment provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of any of the methods described above.
[0094] The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), registers, hard disks, optical fibers, compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof, or any other form of computer-readable storage medium in the art.
[0095] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside within an application-specific integrated circuit (ASIC). In embodiments of the invention, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device.
[0096] Example 5
[0097] This embodiment also provides a computer program product containing instructions that, when executed by a cluster of computer devices, cause the cluster of computer devices to perform the method described in Embodiment 1.
[0098] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for correcting the Raman spectroscopy detection of transformer oil, characterized in that the method... include: Data feature parameters were extracted from Raman spectral data of different oil sample types under a single operating condition, and Pearson correlation coefficients between data feature parameters and corresponding operating condition parameters were calculated to construct the correlation between data feature parameters and single operating condition parameters. Data characteristic parameters of Raman spectral data of different oil sample types under multiple operating conditions were extracted. The parameter effects of the multiple operating conditions parameters were analyzed using the analysis of variance method to determine the influence weight between the data characteristic parameters and the multiple operating conditions parameters. Multiple single-factor and multi-factor correction sub-models are constructed, and a Bayesian optimization algorithm is used to weight and fuse these sub-models to construct a globally coupled correction model. The single-factor correction sub-models are constructed by combining the correlation between data feature parameters and single operating condition parameters with a multiple linear regression model. The multi-factor correction model is obtained by training a backpropagation neural network by inputting the influence weights between data feature parameters and multiple operating condition parameters. The spectral data of the transformer oil to be tested and the operating parameters of the current operating conditions are obtained. The operating parameters are input into the global coupled correction model to obtain the parameter correction weights. The spectral data is corrected using the parameter correction weights to obtain the corrected spectral data.
2. The method for correcting Raman spectroscopy detection of transformer oil according to claim 1, characterized in that, The single operating environment includes temperature operating environment, humidity operating environment, and vibration operating environment; the operating parameters include temperature, humidity, and vibration frequency; the different oil sample types include new oil, lightly aged oil, moderately aged oil, and heavily aged oil; the data characteristic parameters include peak position, peak intensity, full width at half maximum (FWHM), and peak area.
3. The method for correcting Raman spectroscopy detection of transformer oil according to claim 2, characterized in that, Data characteristic parameters were extracted from Raman spectral data of different oil sample types under a single operating condition, and Pearson correlation coefficients were calculated between the data characteristic parameters and the corresponding operating condition parameters. The correlation between the data characteristic parameters and the single operating condition parameters was then constructed, specifically as follows: Multiple data feature parameters were extracted from the Raman spectral data of four oil sample types under different single operating conditions, and four environment-feature parameter pairing datasets corresponding to the oil sample types were constructed. Each environment-feature parameter pairing dataset contains multiple data matrices, and each data matrix represents the correlation between an operating condition parameter and a data feature parameter. For each data matrix, the Pearson correlation coefficient is calculated using the Pearson correlation coefficient calculation function. The correlation of the data matrix is filtered based on the Pearson correlation coefficient of the data matrix to obtain the filtered environment-feature parameter paired dataset.
4. The method for correcting Raman spectroscopy detection of transformer oil according to claim 1, characterized in that, Data characteristic parameters were extracted from Raman spectral data of different oil sample types under multiple operating conditions. Analysis of variance was used to analyze the parameter effects of these multi-operating-condition parameters, determining the influence weights between the data characteristic parameters and the multi-operating-condition parameters. Specifically: Multiple data feature parameters are extracted from Raman spectral data of different oil sample types under multiple operating conditions, and feature parameter matrix sets corresponding to the four data feature parameters are constructed. Each feature parameter matrix set contains multiple three-dimensional data matrices constructed based on oil sample type, multiple operating condition parameters, and data feature parameters. Based on the correlation between operating condition parameters and data feature parameters in the feature parameter matrix, an ANOVA model is constructed; based on the ANOVA model, the effects of each operating condition parameter are decomposed to obtain the sum of squares of each effect and the net sum of squares of the effects. Based on the proportion of each operating condition parameter effect in the net sum of squared effects, the effect weight of each operating condition parameter effect is calculated. By combining the effect weight of each operating condition parameter effect with the corresponding data feature parameters, an effect weight set representing the influence weight between data feature parameters and multiple operating condition parameters is constructed.
5. The method for correcting Raman spectroscopy detection of transformer oil according to claim 3, characterized in that, After constructing four paired environmental-feature parameter datasets corresponding to oil sample types, the method also includes performing normality and linearity checks on each data matrix to obtain paired environmental-feature parameter datasets that pass both checks.
6. The method for correcting Raman spectroscopy detection of transformer oil according to claim 4, characterized in that, After calculating the sum of squares of the effects of each operating condition parameter and the net sum of squares of the effects, the method also includes: Obtain the degrees of freedom for the effect of each operating condition parameter, and calculate the F-value for each operating condition parameter effect using the sum of the degrees of freedom and squares of each operating condition parameter effect; based on the F-value, filter the significance of each operating condition parameter effect to obtain a set of significant effects with higher significance.
7. A transformer oil Raman spectroscopy detection and correction system, characterized in that, This system is used in the transformer oil Raman spectroscopy detection correction method according to any one of claims 1-6, the system comprising: The single-operation condition analysis module is used to extract the data feature parameters of Raman spectral data of different oil sample types under a single operating condition, calculate the Pearson correlation coefficient between the data feature parameters and the corresponding operating condition parameters, and construct the correlation between the data feature parameters and the single operating condition parameters. The multi-operating-condition analysis module is used to extract the data characteristic parameters of Raman spectral data of different oil sample types under multiple operating conditions. The parameter effect of the multi-operating-condition parameters is analyzed using the variance analysis method to determine the influence weight between the data characteristic parameters and the multi-operating-condition parameters. The correction model construction module is used to construct multiple single-factor correction sub-models and multi-factor correction sub-models. A Bayesian optimization algorithm is used to weight and fuse these sub-models to construct a globally coupled correction model. The single-factor correction sub-models are constructed by combining the correlation between data feature parameters and single operating condition parameters with a multiple linear regression model. The multi-factor correction model is obtained by training a BP neural network by inputting the influence weights between data feature parameters and multiple operating condition parameters. The spectral data correction module is used to acquire the spectral data of the transformer oil to be tested and the operating condition parameters of the current operating condition. The operating condition parameters are input into the global coupled correction model to obtain the parameter correction weights. The spectral data is corrected using the parameter correction weights to obtain the corrected spectral data.
8. A computer device comprising a system memory and a processor, wherein the system memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method described in any one of claims 1 to 6.
10. A computer program product containing instructions, characterized in that, When the instructions are executed by a cluster of computer devices, the cluster of computer devices causes the cluster of computer devices to perform the method as described in any one of claims 1 to 6.