Turning signal noise reduction method based on improved wavelet threshold

Abstract

The invention discloses a turning signal noise reduction method based on an improved wavelet threshold, and the method comprises the steps: carrying out the multi-layer wavelet decomposition of a noise-containing turning signal, obtaining the low-frequency and high-frequency wavelet coefficients of each layer, and processing the high-frequency wavelet coefficients of each layer through an improved threshold function, thereby achieving the noise reduction of the turning signal. And reconstructing a signal through wavelet inverse transformation of the processed wavelet coefficient to realize denoising. The optimal number of decomposition layers required by a sampling signal is determined according to noise power, and finally, an optimal wavelet basis is determined by taking a signal-to-noise ratio, a mean square error and the like as evaluation indexes. By means of the improved wavelet threshold function, on one hand, the signal denoising effect can be achieved, and on the other hand, the situation that details are damaged due to excessive noise reduction can be avoided. Experimental data show that the improved threshold function effectively suppresses random errors, the signal-to-noise ratio is improved, the root-mean-square error is reduced, and the denoising effect is good.

Description

Technical Field

[0001] The invention relates to a turning processing signal denoising method based on improved wavelet threshold, belonging to the technical field of signal processing. Background Art

[0002] With the continuous development of automation technology, turning has become one of the most basic, widespread and important process methods in the machinery manufacturing industry. In the turning system, the tool is the component that is most susceptible to wear and failure. The vibration of the tool during the cutting process will change with the change of wear condition. However, due to the influence of external interference sources, such as external noise and lathe noise itself, the collected tool vibration signal will become more complex, the signal-to-noise ratio will be reduced, and the subsequent feature extraction and pattern recognition research will be affected.

[0003] In order to solve this problem, it is crucial to denoise the collected original signal. Denoising technology can effectively improve the quality of the signal, making subsequent signal processing and analysis more accurate and reliable. Commonly used denoising methods include wavelet transform, wavelet packet transform, etc. These methods can effectively filter the noise in the signal and extract effective vibration signals, which can help the subsequent feature extraction and wear status monitoring.

[0004] At present, wavelet transform methods are widely used in the field of signal denoising. Wavelet threshold denoising methods have become the most common signal processing methods in practical engineering applications due to their advantages such as easy implementation and ability to fully preserve the characteristics of the original signal. The threshold function is an important part of wavelet threshold denoising, and the appropriate threshold reduction strategy plays a key role in the denoising results. The threshold function specifies how to process wavelet coefficients at different levels, including strategies for retaining, reducing or setting to zero. Different threshold functions have different focuses and characteristics, but the basic idea is to suppress low-amplitude wavelet coefficients containing noise energy and protect the signal characteristics reflected by high-amplitude wavelet coefficients.

[0005] The traditional hard threshold function jumps at the threshold, which may produce breakpoints or discontinuity points, resulting in additional oscillations in the signal; the soft threshold function will be over-smoothed during the processing process, which seriously affects the approximation degree between the reconstructed signal and the real signal, and there is a large deviation. Therefore, it is necessary to design a new threshold function, which can achieve the effect of signal denoising on the one hand, and avoid the destruction of details by excessive reduction of noise on the other hand. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a turning processing signal denoising method based on improved wavelet threshold, which effectively suppresses random errors, improves the signal-to-noise ratio, reduces the root mean square error, and has a good denoising effect.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] A turning signal denoising method based on improved wavelet threshold comprises the following steps:

[0009] Step 1, set the adjustment factor, construct an improved threshold function, use the signal-to-noise ratio and root mean square error as evaluation indicators, and determine the optimal decomposition layer number and optimal wavelet basis of wavelet transform;

[0010] Step 2, according to the optimal decomposition layer number and the optimal wavelet basis, multi-layer wavelet decomposition is performed on the noisy turning processing signal to extract the low-frequency and high-frequency wavelet coefficients of each layer;

[0011] Step 3, using the improved threshold function to process the high-frequency wavelet coefficients of each layer to obtain new high-frequency wavelet coefficients;

[0012] Step 4: Reconstruct the low-frequency wavelet coefficients and new high-frequency wavelet coefficients of each layer using inverse wavelet transform to obtain a reconstructed signal, calculate the signal-to-noise ratio and root mean square error of the reconstructed signal, and determine the denoising effect of the reconstructed signal;

[0013] Step 5, adjust the adjustment factor in the improved threshold function, and return to step 3;

[0014] Step 6, repeat steps 3 to 5, and select the reconstructed signal corresponding to the best denoising effect as the denoised turning processing signal.

[0015] As a preferred embodiment of the present invention, the specific process of step 1 is as follows:

[0016] Step 1.1, fix the wavelet basis, select different decomposition levels, use wavelet transform to perform wavelet decomposition on the noisy turning signal at each decomposition level, and obtain the low-frequency wavelet coefficients and high-frequency wavelet coefficients corresponding to each decomposition level at each decomposition level;

[0017] Step 1.2, using the constructed improved threshold function to process the high-frequency wavelet coefficients corresponding to each decomposition layer under each decomposition layer number, to obtain new high-frequency wavelet coefficients;

[0018] Step 1.3, using inverse wavelet transform to reconstruct the low-frequency wavelet coefficients and new high-frequency wavelet coefficients corresponding to each decomposition layer at each decomposition layer number, and obtain the turning processing signal after denoising at each decomposition layer number;

[0019] Step 1.4, calculate the signal-to-noise ratio and root mean square error of the turning processing signal after denoising at each decomposition level, and select the decomposition level with the best denoising effect;

[0020] Step 1.5, when the optimal number of decomposition layers is determined, select different wavelet bases, and perform the same operations as steps 1.1-1.4, and select the wavelet base with the best denoising effect as the optimal wavelet base.

[0021] As a preferred solution of the present invention, the improved threshold function in step 1 is specifically as follows:

[0022]

[0023] Among them, w j,k , are the high-frequency wavelet coefficients before and after denoising, λ is the threshold, σ is the standard deviation of the noise, N is the length of the noisy signal, α and β are adjustment factors, α≥0, m≥2.

[0024] As a preferred solution of the present invention, the calculation formulas of the signal-to-noise ratio and the root mean square error are as follows:

[0025]

[0026] Where SNR and RSME represent the signal-to-noise ratio and root mean square error respectively, x(t) is the noisy signal, d(t) is the signal after denoising, and N is the length of the noisy signal.

[0027] Based on the turning signal denoising method based on improved wavelet threshold, the turning tool wear state monitoring method is specifically as follows:

[0028] According to the turning signal denoising method based on improved wavelet threshold, the denoised turning signal and the corresponding tool wear state label are obtained;

[0029] Perform feature extraction on the turning processing signal after noise reduction to extract features that characterize the tool wear state;

[0030] A wear state prediction model is constructed using a machine learning algorithm, and the wear state prediction model is trained using the denoised turning processing signal and the corresponding tool wear state label to obtain a trained prediction model;

[0031] The trained prediction model is used to predict the denoised turning processing signal acquired in real time to obtain the tool wear status.

[0032] A computer device comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the computer program, the steps of the turning processing signal denoising method based on improved wavelet threshold are implemented.

[0033] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the turning signal denoising method based on improved wavelet threshold are implemented.

[0034] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0035] 1. The present invention can achieve the effect of signal denoising through the improved wavelet threshold function, and can avoid the destruction of details by excessive reduction of noise. Simulation data show that the improved threshold function effectively suppresses random errors, improves the signal-to-noise ratio, reduces the root mean square error, and has a good denoising effect.

[0036] 2. The present invention optimizes the wavelet basis function, the threshold function, and introduces the adjustment factors α and β to adapt to different signals and expand the application field. This method helps to achieve accurate noise reduction calculation of the signal and avoid signal distortion and inaccuracy.

[0037] 3. During the turning process, the present invention can more accurately monitor the wear state of the tool, judge the wear degree and remaining life of the tool by extracting and analyzing the features of the denoised vibration signal, thereby monitoring the wear state of the tool in real time, replacing the tool in time, improving the processing quality and efficiency, and reducing production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a flow chart of a turning processing signal denoising method based on improved wavelet threshold of the present invention;

[0039] Figure 2 is the noisy signal image used in the present invention;

[0040] Figure 3 It is the signal image after noise reduction of the present invention. DETAILED DESCRIPTION

[0041] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0042] The present invention proposes a turning processing signal denoising method based on improved wavelet threshold. The low-frequency and high-frequency wavelet coefficients of each layer are obtained by performing multi-layer wavelet decomposition on the noisy tool head vibration signal. The low-frequency coefficients are retained by analyzing their amplitudes and the high-frequency coefficients are processed by applying an improved threshold function. Finally, the signal is reconstructed using inverse wavelet transform to obtain a denoised signal. This method produces almost no burrs during the processing process, making the signal closer to the ideal state while retaining the integrity of the information, which significantly improves the denoising effect.

[0043] like Figure 1 As shown, the specific steps are as follows:

[0044] 1) Apply wavelet transform to process the original noisy cutter vibration signal, so as to determine the wavelet basis function and the number of decomposition layers used for decomposition. By decomposing the signal through several layers of wavelets, the corresponding coefficients of low-frequency and high-frequency wavelets of each layer can be obtained. Corresponding to the amplitude of the wavelet coefficients, the useful signal is higher and the noise is lower.

[0045] The optimal number of decomposition layers and the optimal wavelet basis are determined based on the signal-to-noise ratio and the root mean square error as evaluation indicators. The evaluation indicators of the denoising effect include the signal-to-noise ratio (SNR) and the root mean square error (RSME). The larger the signal-to-noise ratio, the smaller the root mean square error, and the better the denoising effect. The formulas of the signal-to-noise ratio (SNR) and the root mean square error (RSME) are respectively expressed as:

[0046]

[0047] Where x(t) is the noisy signal; d(t) is the denoised signal; and N is the length of the noisy signal.

[0048] When determining the selection of wavelet basis, the characteristics of the specific signal need to be considered. During the turning process, the vibration of the cutter head is periodic, including characteristics such as different frequency components, waveform changes, pulse characteristics and noise components. Therefore, the present invention selects the coif5 wavelet as the wavelet basis function. The coif5 wavelet basis has tight support, smoothness, symmetry and multi-scale properties, and is suitable for processing these complex signal characteristics.

[0049] When determining the number of decomposition layers, it is necessary to consider that increasing the number of decomposition layers will make the difference between the useful signal and the noise more obvious. However, too many decomposition layers will increase the offset between the signal after the inverse wavelet transform and the original signal, resulting in serious signal distortion. In wavelet denoising, it is generally recommended to control the number of wavelet decomposition layers to 2-6 layers. According to the characteristics of the present invention, the number of decomposition layers is selected to be set to 2 layers.

[0050] Table 1 is a comparison of the denoising results of the noisy signal using different decomposition levels and threshold functions under the wavelet basis coif5; Table 2 is a comparison of the denoising results of the noisy signal using multiple wavelet basis functions and different threshold functions when the decomposition level is 2. The threshold 1 in the table is the existing threshold function, and its expression is:

[0051]

[0052] Table 1

[0053]

[0054] Table 2

[0055]

[0056]

[0057] 2) For the input signal, the corresponding threshold function needs to be used to process the wavelet coefficients. The specific operation is to use the improved threshold function to process each layer of high-frequency wavelet coefficients.

[0058] A new improved threshold function is proposed, as shown in the following formula:

[0059]

[0060] in, α and β are regulatory factors, α≥0, w j,k , They refer to the wavelet transform coefficients before and after denoising, λ is the threshold, σ is the standard deviation of the noise, N is the length of the noisy signal, and sgn(·) is the sign function.

[0061] Analyzing formula (1), we can get:

[0062] 1. hour,

[0063] When j,k →λ,

[0064]

[0065] When j,k →-λ,

[0066]

[0067] Formula (1) in w j,k = ±λ, and thus equation (1) is continuous at (-∞, +∞).

[0068] 2. When w j,k →+∞,

[0069]

[0070] When j,k →-∞,

[0071]

[0072] When j,k →∞,

[0073]

[0074] In summary, formula (1) is is the asymptote.

[0075] The construction method in formula (1) is similar to the general threshold function, and its main purpose is to convert |w j,k |<λ part is reset to zero, only |w is processed j,k |≥λ. From equations (2) and (3), we can see that the function is continuous, unlike the hard threshold function which has serious oscillation. From equations (4), (5) and (6), we can see that the function is continuous. is an asymptote. When the wavelet coefficient |w j,k |→∞, let and w j,k The distance is close to zero. This method can solve the constant deviation problem in the soft threshold. By adjusting the parameters α and β, the threshold function can be adapted to different signals, making its application range wider.

[0076] 3) Use the inverse wavelet transform to perform wavelet reconstruction on the wavelet coefficients retained in step 2) and output the desired cutter vibration signal after filtering. When the signal-to-noise ratio and root mean square error of the denoised signal do not meet the requirements, adjust the adaptive parameters α and β, reset the corresponding threshold function, reduce the wavelet coefficients of the signal after wavelet decomposition, and perform an inverse wavelet transform based on the reduced wavelet coefficients to reconstruct the denoised signal again until the signal-to-noise ratio and root mean square error meet the requirements.

[0077] The coif5 wavelet is selected, the number of decomposition layers is set to 2, the adjustment factors are selected to be α=50, β=30, the threshold function of the present invention is applied for processing, and then wavelet reconstruction is performed to obtain the output signal. Figure 2 and Figure 3 They are the noisy turning signal image and the turning signal image after denoising respectively.

[0078] By extracting and analyzing the features of the denoised vibration signal, the wear state of the tool can be monitored more accurately, and the degree of wear and remaining life of the tool can be determined. This allows the tool to be monitored in real time, replaced in a timely manner, and the processing quality and efficiency can be improved, reducing production costs.

[0079] Based on the same inventive concept, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the aforementioned turning processing signal denoising method based on improved wavelet threshold are implemented.

[0080] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the aforementioned turning signal denoising method based on improved wavelet threshold are implemented.

[0081] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0082] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0083] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0084] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0085] The above embodiments are only for illustrating the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A turning signal denoising method based on improved wavelet threshold, characterized in that: The steps include: Step 1, set the adjustment factor, construct an improved threshold function, use the signal-to-noise ratio and root mean square error as evaluation indicators, and determine the optimal decomposition layer number and optimal wavelet basis of wavelet transform; Step 2, according to the optimal decomposition layer number and the optimal wavelet basis, multi-layer wavelet decomposition is performed on the noisy turning processing signal to extract the low-frequency and high-frequency wavelet coefficients of each layer; Step 3, using the improved threshold function to process the high-frequency wavelet coefficients of each layer to obtain new high-frequency wavelet coefficients; Step 4: Reconstruct the low-frequency wavelet coefficients and new high-frequency wavelet coefficients of each layer using inverse wavelet transform to obtain a reconstructed signal, calculate the signal-to-noise ratio and root mean square error of the reconstructed signal, and determine the denoising effect of the reconstructed signal; Step 5, adjust the adjustment factor in the improved threshold function, and return to step 3; Step 6, repeat steps 3 to 5, and select the reconstructed signal corresponding to the best denoising effect as the denoised turning processing signal.

2. The turning signal denoising method based on improved wavelet threshold according to claim 1 is characterized in that: The specific process of step 1 is as follows: Step 1.1, fix the wavelet basis, select different decomposition levels, use wavelet transform to perform wavelet decomposition on the noisy turning signal at each decomposition level, and obtain the low-frequency wavelet coefficients and high-frequency wavelet coefficients corresponding to each decomposition level at each decomposition level; Step 1.2, using the constructed improved threshold function to process the high-frequency wavelet coefficients corresponding to each decomposition layer under each decomposition layer number, to obtain new high-frequency wavelet coefficients; Step 1.3, using inverse wavelet transform to reconstruct the low-frequency wavelet coefficients and new high-frequency wavelet coefficients corresponding to each decomposition layer at each decomposition layer number, and obtain the turning processing signal after denoising at each decomposition layer number; Step 1.4, calculate the signal-to-noise ratio and root mean square error of the turning processing signal after denoising at each decomposition level, and select the decomposition level with the best denoising effect; Step 1.5, when the optimal number of decomposition layers is determined, select different wavelet bases, and perform the same operations as steps 1.1-1.4, and select the wavelet base with the best denoising effect as the optimal wavelet base.

3. The turning signal denoising method based on improved wavelet threshold according to claim 1 is characterized in that: The improved threshold function described in step 1 is specifically as follows: Among them, w j,k , are the high-frequency wavelet coefficients before and after denoising, λ is the threshold, σ is the standard deviation of the noise, N is the length of the noisy signal, α and β are adjustment factors, α≥0, m≥2.

4. The turning signal denoising method based on improved wavelet threshold according to claim 1 is characterized in that: The calculation formulas of the signal-to-noise ratio and the root mean square error are as follows: Where SNR and RSME represent the signal-to-noise ratio and root mean square error respectively, x(t) is the noisy signal, d(t) is the signal after denoising, and N is the length of the noisy signal.

5. A turning tool wear state monitoring method based on the turning signal denoising method based on improved wavelet threshold according to any one of claims 1 to 4, characterized in that: The monitoring method is specifically as follows: According to the turning signal denoising method based on improved wavelet threshold, the denoised turning signal and the corresponding tool wear state label are obtained; Perform feature extraction on the turning processing signal after noise reduction to extract features that characterize the tool wear state; A wear state prediction model is constructed using a machine learning algorithm, and the wear state prediction model is trained using the denoised turning processing signal and the corresponding tool wear state label to obtain a trained prediction model; The trained prediction model is used to predict the denoised turning processing signal acquired in real time to obtain the tool wear status.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the computer program, the steps of the turning signal denoising method based on improved wavelet threshold as described in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the turning signal denoising method based on improved wavelet threshold are implemented as claimed in any one of claims 1 to 4.

Images