Tool signal denoising method based on IWOA optimized wavelet parameters

Abstract

The application discloses a tool signal denoising method based on IWOA optimization of wavelet parameters, and the wavelet parameters are optimized through the IWOA algorithm, so that different types of tool signals can be better adapted. Specifically, the application is based on the traditional WOA, and the population diversity is easy to be insufficient, the algorithm is easy to fall into a local optimal solution, and the exploration ability of the algorithm is insufficient. In order to solve the problems, firstly, the population generated by Circle mapping is used to replace the initial population randomly generated in the original algorithm, so that the population diversity is realized; secondly, the pinhole imaging reverse learning is combined with the optimal worst reverse learning strategy, so that the ability of the algorithm to jump out of the local optimum is improved; finally, a chaotic adaptive inertia weight is proposed, and the exploration ability of the algorithm is enhanced by using the method; in summary, through the improvement of the whale optimization algorithm, the wavelet parameter combination for the tool signal is more optimal, and the precision and reliability of the tool signal processing are improved, so that the application has important practical application value.

Description

Technical Field

[0001] This invention belongs to the field of signal processing technology, specifically a tool signal noise reduction method based on IWOA-optimized wavelet parameters. Background Technology

[0002] In machine tool processing, tool wear is receiving increasing attention. Tool wear directly affects the surface quality and machining accuracy of products, while also reducing production efficiency. Therefore, monitoring tool wear is of significant economic importance. There are two main methods for tool wear monitoring: direct monitoring and indirect monitoring. Direct monitoring is easily affected by the machining environment and cannot be performed online, greatly limiting its application. Indirect monitoring has little or no impact on the cutting process and is widely used for online tool monitoring. However, the signal acquisition process of indirect monitoring is highly susceptible to external interference, leading to signal quality degradation. Effectively eliminating noise interference is crucial for this technology. Wavelet transform, as an evolution of Fourier transform, offers superior time-frequency domain analysis capabilities and is highly effective in removing noise from frequency bands different from the target signal.

[0003] The Whale Optimization Algorithm (WOA) is a swarm intelligence optimization algorithm that originated from the hunting behavior of humpback whales, large predators. Compared with most optimization algorithms, WOA is easier to implement and has more outstanding optimization effects. It has been applied in various optimization fields, such as optimal parameter selection, mechanical fault diagnosis, and neural network weight optimization.

[0004] Chinese patent CN112147226B discloses a method for determining the optimal number of decomposition layers for wavelet denoising based on laser ultrasonic signals. The method includes the following steps: First, an excitation laser is used to excite ultrasonic waves on the surface of a workpiece, and a receiving laser is used to receive the ultrasonic signal; the ultrasonic signal is then subjected to time-domain averaging to obtain the time-domain averaged signal; second, the signal is decomposed into j-level wavelet layers using the dmey wavelet basis function, and the wavelet entropy at the j-level decomposition is calculated. If the wavelet entropy at the j+1-level decomposition is smaller than that at the j-level decomposition, then the j-level is selected as the optimal number of decomposition layers. This invention provides a method for determining the optimal number of decomposition layers for wavelet denoising based on laser ultrasonic signals. However, this invention only optimizes the number of decomposition layers, while the dmey wavelet basis function is directly given. This means that the method does not provide selection criteria for wavelet basis functions based on specific application scenarios and signal characteristics, and cannot guarantee that the selected dmey wavelet basis function is the optimal one, potentially leading to inappropriate selection.

[0005] Chinese patent application CN110009593A discloses an image adaptive enhancement method based on the whale optimization algorithm. This method uses the whale optimization algorithm to search for the optimal parameter combination of a non-complete Beta function to enhance the image. However, the whale optimization algorithm suffers from several problems, including low optimization accuracy, susceptibility to local optima, and slow convergence speed. These problems primarily stem from the whale optimization algorithm's use of random search to explore the search space, which can easily lead to the algorithm getting trapped in local optima. Summary of the Invention

[0006] Different wavelet basis functions are suitable for different types of signals. Only by selecting appropriate wavelet basis functions and decomposition levels can signal features be captured more accurately during wavelet denoising, thereby achieving better denoising results. In practical applications, the optimization process often involves a large amount of computation and search. However, selecting wavelet basis functions and decomposition levels as the main optimization objects can reduce the search space, lower computational complexity, and improve the efficiency of the algorithm.

[0007] In view of this, the purpose of this invention is to provide a tool signal noise reduction method based on IWOA-optimized wavelet parameters, which selects appropriate combinations of wavelet basis functions and decomposition layers through the IWOA optimization algorithm to improve the accuracy and reliability of tool signal processing.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A tool signal denoising method based on IWOA-optimized wavelet parameters includes the following steps:

[0010] Step 1: Preprocess the raw tool signal; initialize the parameters of the IWOA optimization algorithm;

[0011] Step 2: Use the Circle chaotic mapping strategy to generate a whale population, so that individual whales are randomly generated and relatively uniformly distributed in the search space, and map the position of individual whales to a two-dimensional array P containing wavelet basis functions and decomposition levels;

[0012] Step 3: Calculate the fitness of each individual whale, and record the optimal fitness and individual position of the current whale population as the historical optimal fitness and historical optimal individual position;

[0013] Step 4: Select an elite individual with the best global position and a worst global position from the current whale population. Use a combination of pinhole imaging back-learning mechanism and best-worst back-learning strategy to update the positions of the elite and worst individuals. Calculate the first fitness of the elite individual after the update and the second fitness of the worst individual after the update. Compare the first fitness, second fitness, and historical best fitness: if the first fitness or second fitness is better than the historical best fitness, update the historical best fitness and replace the historical best individual position with the updated elite or worst individual position; otherwise, keep the historical best fitness and historical best individual position unchanged.

[0014] Step 5: Determine if the current iteration count has reached the set maximum iteration count: if yes, proceed to Step 9; if no, proceed to Step 6.

[0015] Step 6: Update the chaotic adaptive inertia weight vector and the parameters of the IWOA optimization algorithm;

[0016] Step 7: Update the position of the whale population, calculate the fitness of each individual whale, and obtain the optimal fitness of the current whale population; compare the optimal fitness of the current whale population with the historical optimal fitness: if the optimal fitness of the current whale population is better than the historical optimal fitness, then update the historical optimal fitness and replace the historical optimal individual position with the optimal individual position of the current whale population; otherwise, the historical optimal fitness and the historical optimal individual position remain unchanged.

[0017] Step 8: Repeat step 4;

[0018] Step 9: Use the wavelet basis function and decomposition level obtained by mapping the historical best individual position as the optimal parameters for wavelet analysis, and use wavelet analysis to denoise the tool signal.

[0019] Furthermore, in step one, the method for preprocessing the original tool signal is as follows: the acquired tool cutting force signal is processed to generate an identifiable file, and the invalid value truncation method is used to truncate the data at the beginning and end by 5%.

[0020] Furthermore, in step two, the principle behind the Circle chaotic mapping strategy for generating the whale population is as follows:

[0021]

[0022] Where, x i This represents the number of the i-th chaotic sequence; mod(a,b) represents the modulo operation of a with respect to b.

[0023] Furthermore, in step four, the method for updating the position of elite individuals is as follows:

[0024] X * best =(a j +b j ) / 2+(a j +b j ) / 2n-X best / n.

[0025] Among them, X * best Indicates the updated location of the elite individual; X best This represents the position of the elite individual before the update; assuming that in a certain space, there is a flame p with height h, the projection of which onto the x-axis is the global optimal position X. best The upper and lower limits of the coordinate axes, a j b j This represents the upper and lower limits of the j-th dimension of the current solution. A pinhole screen with a small hole is placed at the base point o. The flame can pass through the small hole and obtain a height of h on the receiving screen. * The inverted p * At this point, the optimal position X obtained through pinhole imaging is obtained on the x-axis. * best ; n is the ratio of the height before and after projection, and

[0026] Furthermore, in step four, the method for updating the position of the worst individual is as follows:

[0027] X worst (t+1)=a j +rand·(b j -X worst (t))

[0028] Among them, X worst (t+1) represents the position of the worst individual after the update; X worst (t) represents the worst individual position before the update; rand is a random number in [0,1].

[0029] Furthermore, in step six, the method for updating the chaotic adaptive inertia weight vector is as follows:

[0030]

[0031] in, ω Represents the chaotic adaptive inertia weights; ω s ω represents the initial value of the inertia weight. e The inertia weight represents the maximum number of iterations; t is the current iteration number; T max This indicates the maximum number of iterations.

[0032] Furthermore, in step six, the parameter update method for the IWOA optimization algorithm is as follows:

[0033] A = 2a × rand - a

[0034] C = 2 × rand

[0035] a=(2-2t / T max )

[0036] Where a, A, and C are parameters of the IWOA optimization algorithm, and a is the convergence factor, C is the generated random number; t is the current iteration number; T max This indicates the maximum number of iterations.

[0037] Furthermore, in step seven, the method for updating the whale population's location is as follows:

[0038] X(t+1)=ω·X * (t)-A×D,|A|<1,p<0.5

[0039] X(t+1)=ω·X rand -A×D rand ,|A|…1,p<0.5

[0040] X(t+1)=ω·X * (t)+D'×e bl ×cos(2πl),p…0.5

[0041] Where X(t+1) is the updated position of the individual whale; X(t) represents the current position of the individual whale; X * (t) represents the current global optimal position; X rand ω represents the position vector of a randomly selected whale in the population; ω = [ω1, ω2] is the chaotic adaptive inertia weight vector; ω1 and ω2 both represent the values ​​of the generated adaptive weights; l represents a random number between [-1, 1]; b is a constant constrained to a logarithmic spiral form; A·D represents the encirclement step size; D represents the distance between the whale and the individual; D rand denoted as , where represents the distance between the whale and the current random individual; D' represents the distance between the whale and the current global best individual; p represents a random number generated in the range [0,1].

[0042] Furthermore, the fitness function is:

[0043]

[0044] Where F represents the fitness of an individual whale; Q is the weighting factor; R SNR E represents the signal-to-noise ratio (SNR) value. RMSE This represents the result value of the root mean square error; and:

[0045]

[0046]

[0047] Where x(i) is the cutting force signal in the x-axis direction; x'(i) is the denoised cutting force signal in the x-axis direction.

[0048] The beneficial effects of this invention are as follows:

[0049] This invention presents a tool signal denoising method based on IWOA-optimized wavelet parameters. By optimizing wavelet parameters using the IWOA algorithm, it better adapts to different types of tool signals compared to traditional fixed parameter settings. Specifically, this invention addresses the problems of insufficient population diversity, susceptibility to local optima, and insufficient exploratory ability inherent in traditional WOA algorithms. First, it replaces the randomly generated initial population in the original algorithm with a population generated by Circle mapping, achieving population diversity. Second, it combines pinhole imaging back-learning with optimal / worst-optimal back-learning strategies to improve the algorithm's ability to escape local optima. Finally, it proposes a chaotic adaptive inertial weight to enhance the algorithm's exploratory ability. In summary, this invention's tool signal denoising method based on IWOA-optimized wavelet parameters achieves a better combination of wavelet parameters for tool signals by improving the whale optimization algorithm, which has significant practical application value for improving the accuracy and reliability of tool signal processing.

[0050] The present invention also has the following advantages:

[0051] 1) The IWOA algorithm has stronger global search characteristics, which enables it to quickly and accurately obtain the optimal wavelet basis function and decomposition level in the solution space, thereby significantly improving the denoising effect;

[0052] 2) By introducing Circle mapping, the search range of whale pods in space can be expanded, the diversity of pod locations can be increased, and the defect of the algorithm being prone to getting trapped in local optima can be improved to some extent.

[0053] 3) By comprehensively considering the signal-to-noise ratio and root mean square error, and using the linear weighted summation method to calculate the comprehensive index F as the objective function, the quality of the denoising effect can be evaluated more comprehensively, ensuring the accuracy and reliability of the final result.

[0054] 4) The use of inertial weights based on chaos adaptation can adjust the size of the inertial weights according to the iteration process of the algorithm. Larger weights are used for global exploration in the early stage and smaller weights are used for local optimization in the later stage, which effectively balances the needs of global search and local search and further enhances the optimization effect.

[0055] 5) Finding two feasible solutions and finding their inverse solutions enhances the algorithm's ability to escape local extrema. Attached Figure Description

[0056] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration:

[0057] Figure 1 This is a flowchart of an embodiment of the tool signal noise reduction method based on IWOA optimized wavelet parameters according to the present invention;

[0058] Figure 2 This is a comparison image of the original tool signal before and after preprocessing;

[0059] Figure 3 This is a comparison chart showing the tool signal before and after noise reduction processing using optimized wavelet parameters. Detailed Implementation

[0060] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0061] like Figure 1 As shown, this embodiment of the tool signal denoising method based on IWOA-optimized wavelet parameters includes the following steps:

[0062] Step 1: Preprocess the raw tool signal; initialize the parameters of the IWOA optimization algorithm.

[0063] Specifically, in this embodiment, a three-axis force gauge is used to collect cutting force signals in the X, Y, and Z directions. The original tool signal in this embodiment is only the cutting force signal in the x-axis direction. The method for preprocessing the original tool signal in this embodiment is as follows: the collected tool cutting force signal is processed to generate a recognizable file, and the data at the beginning and end of the line is truncated by the invalid value truncation method.

[0064] In this embodiment, the parameters of the IWOA optimization algorithm include: whale population size N, parameter dimension dim, and maximum number of iterations T. max The search space has an upper bound of ub and a lower bound of lb.

[0065] Step 2: Use the Circle chaotic mapping strategy to generate a whale population, so that individual whales are randomly generated and relatively uniformly distributed in the search space, and the position of individual whales is mapped to a two-dimensional array P containing wavelet basis functions and decomposition levels.

[0066] Specifically, this embodiment uses Circle chaotic mapping instead of random generation for initializing the whale population during WOA population initialization. This expands the search range of the whale pod in space, increases the diversity of the pod's location, and to some extent improves the algorithm's tendency to get trapped in local optima. The principle of the Circle chaotic mapping strategy for generating the whale population is as follows:

[0067]

[0068] Where, x i This represents the number of the i-th chaotic sequence; mod(a,b) represents the modulo operation of a with respect to b.

[0069] In this embodiment, a two-dimensional array P[i,j] is created to store combinations of different wavelet basis functions and decomposition levels. This combination will serve as the search space during the optimization process. Each two-dimensional array P[i,j] represents the parameters of a corresponding set of wavelet basis functions and decomposition levels. The indices i and j are identifiers of the wavelet basis functions and decomposition levels. Specifically, the ranges of common wavelet basis functions and decomposition levels are shown in Table 1.

[0070] Table 1 Wavelet Transform Parameter Range

[0071]

[0072] Step 3: Calculate the fitness of each individual whale and record the best fitness and the best position of the current whale population as the historical best fitness and the historical best position.

[0073] Specifically, this embodiment employs a Circle chaotic mapping strategy to initialize N whale individuals with relatively uniform positional distribution in the search space. Then, it substitutes the values ​​of N sets of wavelet basis functions and the decomposition level corresponding to the N whale individuals into a wavelet thresholding denoising method based on soft thresholding for noise reduction, obtaining the denoised cutting force signal in the x-axis direction. The SNR and RMSE are then calculated based on the cutting force signals in the x-axis direction before and after noise reduction. Finally, the fitness function is defined by comprehensively considering the values ​​of SNR and RMSE to calculate F. In this embodiment, the fitness function is:

[0074]

[0075] Where F represents the fitness of an individual whale; Q is a weighting factor used to adjust the relative importance of signal-to-noise ratio and root mean square error in the overall index; R SNR E represents the signal-to-noise ratio (SNR) value. RMSE This represents the result value of the root mean square error; and:

[0076]

[0077]

[0078] Where x(i) is the cutting force signal in the x-axis direction; x'(i) is the denoised cutting force signal in the x-axis direction.

[0079] Step 4: Select an elite individual with the best global position and a worst global position from the current whale population. Use a combination of pinhole imaging back-learning mechanism and best-worst back-learning strategy to update the positions of the elite and worst individuals. Calculate the first fitness F of the elite individual after the position update. best The second fitness F after updating the position of the worst individual worst .

[0080] Based on the principle of pinhole imaging, we can conclude that:

[0081]

[0082] In this context, it is assumed that in a certain space, the projection of a flame p of height h onto the x-axis is the globally optimal position X. best The upper and lower limits of the coordinate axes, a j b j This represents the upper and lower limits of the j-th dimension of the current solution. A pinhole screen with a small hole is placed at the base point o. The flame can pass through the small hole and obtain a height of h on the receiving screen. * The inverted p * At this point, the optimal position X obtained through pinhole imaging is obtained on the x-axis. * best (Reverse point).

[0083] make X is obtained through transformation. * best The expression, i.e., the position update method for elite individuals, is:

[0084] X * best =(a j +b j ) / 2+(a j +b j ) / 2n-X best / n

[0085] Among them, X * best Indicates the updated location of the elite individual; X best This represents the position of the elite individual before the update; assuming that in a certain space, there is a flame p with height h, the projection of which onto the x-axis is the global optimal position X. best The upper and lower limits of the coordinate axes, a j b jThis represents the upper and lower limits of the j-th dimension of the current solution. A pinhole screen with a small hole is placed at the base point o. The flame can pass through the small hole and obtain a height of h on the receiving screen. * The inverted p * At this point, the optimal position X obtained through pinhole imaging is obtained on the x-axis. * best ; n This represents the ratio of the height before and after projection.

[0086] When n=1, we can obtain:

[0087] X * best =(a j +b j )-X best

[0088] A random back-learning strategy is used for the worst-performing individual at the global worst position. Specifically, in each iteration of the algorithm, a position selection is performed. The global optimal solution is updated by comparing the fitness value after back-learning with the previous fitness value. The method for updating the position of the worst-performing individual is as follows:

[0089] X worst (t+1)=a j +rand·(b j -X worst (t))

[0090] Among them, X worst (t+1) represents the position of the worst individual after the update; X worst (t) represents the worst individual position before the update; rand is a random number in [0,1]; a j and b j It is dynamically changing and has a more accurate search range compared to other reverse learning strategies, which can improve the optimization accuracy of the algorithm.

[0091] Compare first fitness F best Second fitness F worst and historical best fitness If the first fitness F best Or second fitness F worst Better than the best fitness in history If the historical best fitness is updated, the historical best individual position is replaced with the updated elite individual position or the worst individual position; otherwise, the historical best individual position remains unchanged.

[0092] Step 5: Determine if the current iteration number t has reached the set maximum iteration number T. max If yes, proceed to step nine; otherwise, proceed to step six.

[0093] Step 6: Update the chaotic adaptive inertia weight vector and the parameters of the IWOA optimization algorithm.

[0094] By employing chaotic adaptive inertia weights, larger weights should be used in the early stages of algorithm iteration to allow the algorithm to quickly reach the vicinity of the target value with larger steps; while smaller weights should be used in the later stages of iteration to allow the algorithm to move with smaller steps, thus enabling precise searching in the vicinity of the target solution and better local optimization. Specifically, the update method for the chaotic adaptive inertia weight vector is as follows:

[0095]

[0096] in, ω Represents the chaotic adaptive inertia weights; ω s ω represents the initial value of the inertia weight. e The inertia weight represents the maximum number of iterations; t is the current iteration number; T max This indicates the maximum number of iterations.

[0097] Each update of the IWOA optimization algorithm parameters involves the current iteration number t and the maximum iteration number T. max The calculation updates a and A based on these parameters, while simultaneously generating a random number C to adjust the position update. Specifically, the parameter update method of the IWOA optimization algorithm is as follows:

[0098] A = 2a × rand - a

[0099] C = 2 × rand

[0100] a=(2-2t / T max )

[0101] Where a, A, and C are parameters of the IWOA optimization algorithm, and: a is the convergence factor, which decreases linearly from 2 to 0 as the number of iterations increases; C is the generated random number; t is the current iteration number; T max This indicates the maximum number of iterations.

[0102] Step 7: Update the position of the whale population, calculate the fitness of each individual whale, and obtain the optimal fitness of the current whale population; compare the optimal fitness of the current whale population with the historical optimal fitness: if the optimal fitness of the current whale population is better than the historical optimal fitness, then update the historical optimal fitness and replace the historical optimal individual position with the optimal individual position of the current whale population; otherwise, the historical optimal fitness and the historical optimal individual position remain unchanged.

[0103] Specifically, in this embodiment, the method for updating the location of the whale population is as follows:

[0104] X(t+1)=ω·X *(t)-A×D,|A|<1,p<0.5

[0105] X(t+1)=ω·X rand -A×D rand ,|A|…1,p<0.5

[0106] X(t+1)=ω·X * (t)+D'×e bl ×cos(2πl),p…0.5

[0107] Where X(t+1) is the updated position of the individual whale; X(t) represents the current position of the individual whale; X * (t) represents the current global optimal position; X rand The vector represents the position of a randomly selected whale in the population; ω = [ω1, ω2] is the chaotic adaptive inertia weight vector; ω1 represents ...; ω2 represents ...; l represents a random number between [-1, 1]; b is a constant constrained to a logarithmic spiral form; A·D represents the encirclement step size; D represents ...; D rand D' represents the distance between the whale and the current global best individual; p represents a random number generated in the range [0,1].

[0108] Step 8: Repeat step 4.

[0109] Step 9: Use the wavelet basis function and decomposition level obtained by mapping the historical best individual position as the optimal parameters for wavelet analysis, and use wavelet analysis to denoise the tool signal.

[0110] Specifically, the wavelet basis functions and decomposition levels obtained by using the traditional Whale Optimization Algorithm (WOA) and the improved optimization algorithm (IWOA) of this embodiment are shown in Table 2 and Table 3, respectively.

[0111] Table 2 shows the optimal solution obtained based on the traditional WOA algorithm.

[0112]

[0113] Table 3 shows the optimal solution obtained based on the IWOA algorithm.

[0114]

[0115] Specifically, the optimal wavelet basis obtained using the improved IWOA optimization algorithm in this embodiment is coif5, with a decomposition level of 4, consistent with the conclusions drawn from the tool signal denoising method based on IWOA-optimized wavelet parameters. Meanwhile, it was observed that the traditional whale algorithm WOA optimization algorithm gets trapped in local optima when searching for the optimal wavelet basis. This embodiment introduces IWOA as an optimization tool, which can effectively search the wavelet parameter space and avoid getting trapped in local optima, thus achieving more accurate optimization. Compared to the traditional WOA in searching for optimal wavelet parameters, the method in this embodiment exhibits better performance and stronger global search capability. In summary, this embodiment has significant advantages in tool signal denoising.

[0116] The above-described embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A tool signal noise reduction method based on IWOA-optimized wavelet parameters, characterized in that: Includes the following steps: Step 1: Preprocess the raw tool signal; initialize the parameters of the IWOA optimization algorithm; Step 2: A whale population is generated using a Circle chaotic mapping strategy, allowing individual whales to be randomly generated and relatively uniformly distributed within the search space. The positions of individual whales are mapped to a two-dimensional array containing wavelet basis functions and the number of decomposition levels. P ; Step 3: Calculate the fitness of each individual whale, and record the optimal fitness and individual position of the current whale population as the historical optimal fitness and historical optimal individual position; Step 4: Select an elite individual with the best global position and a worst global position from the current whale population. Use a combination of pinhole imaging back-learning mechanism and best-worst back-learning strategy to update the positions of the elite and worst individuals. Calculate the first fitness of the elite individual after the update and the second fitness of the worst individual after the update. Compare the first fitness, second fitness, and historical best fitness: if the first fitness or second fitness is better than the historical best fitness, update the historical best fitness and replace the historical best individual position with the updated elite or worst individual position; otherwise, keep the historical best fitness and historical best individual position unchanged. Step 5: Determine if the current iteration count has reached the set maximum iteration count: if yes, proceed to Step 9; if no, proceed to Step 6. Step 6: Update the chaotic adaptive inertia weight vector and the parameters of the IWOA optimization algorithm; Step 7: Update the position of the whale population, calculate the fitness of each individual whale, and obtain the optimal fitness of the current whale population; compare the optimal fitness of the current whale population with the historical optimal fitness: if the optimal fitness of the current whale population is better than the historical optimal fitness, then update the historical optimal fitness and replace the historical optimal individual position with the optimal individual position of the current whale population; otherwise, the historical optimal fitness and the historical optimal individual position remain unchanged. Step 8: Repeat steps 4 through 7. Step 9: Use the wavelet basis function and decomposition level obtained by mapping the historical best individual position as the optimal parameters for wavelet analysis, and use wavelet analysis to denoise the tool signal. In step six, the method for updating the chaotic adaptive inertia weight vector is as follows: in, This represents the chaotic adaptive inertia weight; This represents the initial value of the inertia weight; The inertia weight represents the maximum number of iterations; This represents the current iteration number; Indicates the maximum number of iterations; In step seven, the method for updating the whale population location is as follows: in, This is the updated location of the individual whale; Indicates the current location of the individual whale; This indicates the current globally optimal position; This represents the position vector of a whale randomly selected from the population; This is the chaotic adaptive inertia weight vector; and All represent the values ​​of the generated adaptive weights; express Random numbers between; It is a constant that defines the logarithmic spiral form; Indicates the enclosing step size; Indicates the distance between the whale and the individual; This represents the distance between the whale and the current random individual; This represents the distance between the whale and the current global best individual; Indicates in Random numbers generated within a specified range.

2. The tool signal denoising method based on IWOA-optimized wavelet parameters according to claim 1, characterized in that: In step one, the method for preprocessing the original tool signal is as follows: the acquired tool cutting force signal is processed to generate an identifiable file, and the invalid value truncation method is used to truncate the data at the beginning and end by 5%.

3. The tool signal denoising method based on IWOA-optimized wavelet parameters according to claim 1, characterized in that: In step two, the principle behind the Circle chaotic mapping strategy for generating a whale population is as follows: in, Indicates the first Number of chaotic sequences; express right The remainder operation.

4. The tool signal denoising method based on IWOA-optimized wavelet parameters according to claim 1, characterized in that: In step four, the method for updating the position of elite individuals is as follows: in, Indicates the updated location of elite individuals; This represents the position of an elite individual before the update; assuming there is a space with a height of... flames exist The projection on the axis is the global optimal position. Upper and lower limits of coordinate axes , Indicates the current solution. The upper and lower limits of the dimension, at the base point A pinhole screen with a small hole is placed on top. The flame can be reflected on the receiving screen at a height of [missing information]. inverted image At this time x The axis obtains the optimal position generated by pinhole imaging. ; This is the ratio of the height before and after projection, and .

5. The tool signal denoising method based on IWOA optimized wavelet parameters according to claim 1, characterized in that: In step four, the method for updating the position of the worst individual is as follows: in, Indicates the position of the worst individual after the update; Indicates the position of the worst individual before the update; for Random numbers on the array.

6. The tool signal denoising method based on IWOA-optimized wavelet parameters according to claim 1, characterized in that: In step six, the parameter update method for the IWOA optimization algorithm is as follows: in, 、 and All of these are parameters of the IWOA optimization algorithm, and The convergence factor is For generated random numbers; This represents the current iteration number; This indicates the maximum number of iterations.

7. The tool signal noise reduction method based on IWOA optimized wavelet parameters according to claim 1, characterized in that: The fitness function is: in, Indicates the fitness level of an individual whale; As a weighting factor; This represents the signal-to-noise ratio (SNR) value. This represents the result value of the root mean square error; and: in, for Cutting force signal in the axial direction; For noise reduction Cutting force signal in the axial direction.

Images