Crankshaft contour error preprocessing method based on robust Gaussian regression filter
By preprocessing crankshaft profile errors using a robust Gaussian regression filter, identifying and replacing outliers, and performing low-pass filtering, the problem of insufficient robustness in existing technologies is solved, thereby improving the accuracy and stability of the data and simplifying the subsequent processing procedures.
Patent Information
- Application Number
- CN202411424108.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-12
- Publication Date
- 2025-01-10
AI Technical Summary
Existing Gaussian filters are not robust enough when dealing with crankshaft profile errors, and cannot effectively remove noise and outliers from the data, resulting in poor filtering performance and affecting the accuracy and stability of subsequent processing.
A robust Gaussian regression filter is used to preprocess the crankshaft profile error. Outliers are identified and replaced by difference judgment, and low-pass filtering is performed to obtain the final profile error, thereby improving the accuracy and stability of the data.
It effectively removes noise and outliers from crankshaft profile errors, improves data accuracy and stability, simplifies subsequent processing, and enhances the precision and efficiency of error compensation and quality control.
Smart Images
Figure CN119293405A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of crankshaft profile error analysis, and more specifically to a crankshaft profile error preprocessing method based on a robust Gaussian regression filter. Background Art
[0002] Contour error is the difference between the actual contour of the workpiece and the standard contour, so the filtering method of contour error is the filtering method of engineering contour or surface texture. The earliest filter used for contour error signal filtering was the 2RC filter, but due to its nonlinear phase disadvantage, its phase was distorted. WhiteHouse proposed a phase-corrected 2RC filter to solve this problem. Although the phase-corrected 2RC filter solves the phase distortion problem, it cannot separate the roughness and waviness in the contour error. Therefore, ISO has deleted the filter from the standard, but ASME B46.1 still has a 2RC filter. Later, the emergence of the Gaussian filter not only solved the shortcomings of the 2RC filter, but also was easy to implement, so it entered the ASME and ISO standards respectively. However, the Gaussian filter has poor robustness and is powerless against marginal effects and abnormal contours (such as high peaks or deep valleys). In order to solve the disadvantage of poor robustness of the Gaussian filter, ISO recommends the use of the Rk filter. However, this filtering method is not always robust to very high peaks or very deep valleys.
[0003] Therefore, how to solve the above problems is urgently needed to be studied by those skilled in the art. Summary of the invention
[0004] In view of this, the present invention provides a crankshaft profile error preprocessing method based on a robust Gaussian regression filter to solve the problems in the background technology.
[0005] In order to achieve the above object, the present invention adopts the following technical solution:
[0006] A crankshaft profile error preprocessing method based on a robust Gaussian regression filter comprises the following steps:
[0007] The original contour error is filtered using a robust Gaussian regression filter to obtain a filtering result;
[0008] Subtract the original contour error from the filtering result;
[0009] If the absolute value of the difference is greater than the set threshold, the filtering result is used as the first contour error to replace the original contour error;
[0010] The first contour error is low-pass filtered using the robust Gaussian regression filter to obtain a final contour error.
[0011] Optionally, it is characterized in that the formula of the robust Gaussian regression filter is as follows:
[0012]
[0013] Where z is the contour error; w is the error mean; n is the number of data points; k is the center position index of the weight function; i is the contour error point index; ω c is the cut-off frequency; S is the weight function.
[0014] Optionally, the robust Gaussian regression filter can also be expressed as:
[0015]
[0016] Where δ(p,i) is the additional vertical weight, p is the number of iterations, z is the contour error, w is the error mean, n is the number of data points, k is the center position index of the weight function, i is the contour error point index, ω is the contour error point index, and c is the cut-off frequency; S is the weight function.
[0017] Optionally, the formula of the weight function is as follows:
[0018]
[0019] Optionally, in the first iteration of the robust Gaussian regression filter, δ(p,i) is set to 1, and in subsequent iterations, δ(p,i) is set according to the following formula:
[0020]
[0021] In the formula, r(p,i)=z(i)-w(i), C B =4.4×Median(|r(p,i)|).
[0022] It can be seen from the above technical solution that, compared with the prior art, the present invention discloses a crankshaft profile error preprocessing method based on a robust Gaussian regression filter. The technical solution realizes effective preprocessing of the crankshaft profile error through the application of a robust Gaussian regression filter, improves the accuracy and stability of the data, and provides a strong guarantee for subsequent error compensation and quality control. At the same time, the technical solution also has the advantages of simple operation and strong adaptability, and has high practical value and promotion significance, as follows:
[0023] 1. Improve the accuracy of error data
[0024] Excellent filtering effect: The robust Gaussian regression filter can effectively filter the crankshaft profile error data, remove noise and outliers in the data, and obtain smoother and more accurate filtering results.
[0025] Accurate difference judgment: By calculating the difference between the original contour error and the filtered result and setting a reasonable threshold, abnormal points or noise points in the original data can be accurately identified.
[0026] Reasonable replacement of outliers: For points whose absolute difference value is greater than the set threshold, the filtering result is used for replacement. This processing method not only retains most of the information of the original data, but also effectively removes the influence of outliers, making the first contour error closer to the true value.
[0027] 2. Enhance data stability and reliability
[0028] Strong robustness: The robust Gaussian regression filter itself has strong robustness and can adapt to data processing requirements in different situations. Even when there is large noise or outliers in the data, stable filtering results can be obtained.
[0029] Multiple filtering optimization: The first contour error is low-pass filtered again to further smooth the data and reduce high-frequency noise and fluctuations in the data, thereby obtaining a more stable and reliable final contour error.
[0030] 3. Improve the efficiency and accuracy of subsequent processing
[0031] Preprocessing simplifies subsequent work: By preprocessing the crankshaft profile error, the subsequent data processing and analysis work can be simplified, reducing the calculation complexity and time cost.
[0032] Improve the accuracy of subsequent processing: The pre-processed data is more accurate and stable, and can provide more reliable data support for subsequent error compensation, quality control and other links, thereby improving the accuracy and efficiency of the entire processing process. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0034] Figure 1 It is a schematic diagram of the process of the present invention;
[0035] Figure 2 The abnormal contour error input in the embodiment of the present invention and the filtering results obtained by different filtering methods are shown. DETAILED DESCRIPTION
[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0037] The embodiment of the present invention discloses a crankshaft profile error preprocessing method based on a robust Gaussian regression filter, such as Figure 1 As shown, the following steps are included:
[0038] The original contour error is filtered using a robust Gaussian regression filter to obtain a filtering result;
[0039] Subtract the original contour error from the filtering result;
[0040] If the absolute value of the difference is greater than the set threshold, the filtering result is used as the first contour error to replace the original contour error;
[0041] The first contour error is low-pass filtered using a robust Gaussian regression filter to obtain the final contour error.
[0042] The specific steps include:
[0043] Filter selection and setting: Select the robust Gaussian regression filter as the main filtering tool. According to the characteristics and processing requirements of the crankshaft profile error data, set the relevant parameters of the filter, such as window size, number of iterations, etc.
[0044] Threshold setting: According to historical data or experience, set a reasonable difference absolute value threshold. This threshold is used to determine whether the difference between the original contour error and the filtering result is within the abnormal range.
[0045] Filtering of original profile error: Input the original crankshaft profile error data into the robust Gaussian regression filter. Through the filter processing, a smoothed filtering result is obtained.
[0046] Calculate the difference: Compare the original contour error with the filtered result point by point, calculate the difference between them, and record the absolute value of each difference.
[0047] Outlier point identification and replacement: Traverse all the absolute values of the difference and find the points that are greater than the set threshold. For these outliers, replace the corresponding values in the original contour error with the filtering results to form the first contour error.
[0048] Low-pass filtering: The first contour error is input into the robust Gaussian regression filter again for low-pass filtering. Through low-pass filtering, the data is further smoothed to remove high-frequency noise and fluctuations.
[0049] Get the final crankshaft profile error data.
[0050] Data quality check: Perform a quality check on the final contour error data to ensure that the data has no abnormal fluctuations or noise. This can be verified by drawing data curves, calculating statistics, etc.
[0051] Result verification: Compare the processed data with the original data to verify whether the processing effect meets expectations. The processing effect can be evaluated by calculating indicators such as error rate and correlation coefficient.
[0052] The Gaussian regression filter is an improvement on the Gaussian filter and can be defined by the following regression process:
[0053]
[0054] Where z is the contour error; w is the error mean; n is the number of data points; k is the center position index of the weight function; i is the contour error point index; ω c is the cut-off frequency; S is the weight function, which can be expressed by the following formula:
[0055]
[0056] In a zero-order Gaussian regression filter:
[0057]
[0058] Taking partial derivatives with respect to C:
[0059]
[0060] therefore:
[0061]
[0062] The zero-order Gaussian regression filter is equivalent to:
[0063]
[0064] Among them, S′ is the modified weight function,
[0065] The second-order Gaussian regression filter is implemented by minimizing the deviation between the input data and the second-order polynomial, thereby obtaining the mean value of the contour error at each point. At this point, the robust Gaussian regression filter becomes:
[0066]
[0067] Here, x(k,i) = (ki) / n, i = 1:n, so the problem becomes to determine the following binomial by minimizing E(k):
[0068] w(k,i)=Ax(k,i) 2 +Bx(k,i)+C;
[0069] Take partial derivatives of equations (5-9) and set the results equal to zero, and we get:
[0070]
[0071]
[0072]
[0073] The robust Gaussian regression filter is expressed as follows:
[0074]
[0075] Where δ(p,i) is the additional vertical weight and p is the number of iterations. In the first iteration of the filter, δ(p,i) is set to 1. In subsequent iterations, δ(p,i) is set according to the following formula:
[0076]
[0077] In the formula, r(p,i)=z(i)-w(i), C B =4.4×Median(|r(p,i)|).
[0078] Will Figure 2 The abnormal contour error is used as the original contour error input and robust Gaussian regression filtering is performed. The filtering result is as follows: Figure 2 In addition, by comparing the results of several other filtering methods, it is proved that the robust Gaussian regression filter has the best effect on suppressing or correcting abnormal contours.
[0079] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0080] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A crankshaft profile error preprocessing method based on a robust Gaussian regression filter, characterized in that: The following steps are involved: The original contour error is filtered using a robust Gaussian regression filter to obtain a filtering result; Subtract the original contour error from the filtering result; If the absolute value of the difference is greater than the set threshold, the filtering result is used as the first contour error to replace the original contour error; The first contour error is low-pass filtered using the robust Gaussian regression filter to obtain a final contour error.
2. The crankshaft profile error preprocessing method based on robust Gaussian regression filter according to claim 1 is characterized in that: The formula of the robust Gaussian regression filter is as follows: Where z is the contour error; w is the error mean; n is the number of data points; k is the center position index of the weight function; i is the contour error point index; ω c is the cut-off frequency; S is the weight function.
3. The crankshaft profile error preprocessing method based on robust Gaussian regression filter according to claim 1 is characterized in that: The robust Gaussian regression filter can also be expressed as: Where δ(p,i) is the additional vertical weight, p is the number of iterations, z is the contour error, w is the error mean, n is the number of data points, k is the center position index of the weight function, i is the contour error point index, ω is the contour error point index, and c is the cut-off frequency; S is the weight function.
4. A crankshaft profile error preprocessing method based on a robust Gaussian regression filter according to claim 2 or 3, characterized in that: The formula of the weight function is as follows:
5. The crankshaft profile error preprocessing method based on robust Gaussian regression filter according to claim 3 is characterized in that: In the first iteration of the robust Gaussian regression filter, δ(p,i) is set to 1. In subsequent iterations, δ(p,i) is set according to the following formula: where r(p,i) = z(i) - w(i), C B = 4.4 × Median(|r(p,i)|).