Cutter wear number-real fusion test method for blisk machining

Abstract

The invention discloses a tool wear number-real fusion test method for blisk machining, and belongs to the field of mechanical engineering, computer science and data analysis. The method comprises the following steps: monitoring an entity, synchronously acquiring vibration signal data in the machining process of the five-axis machining center, and preprocessing the vibration signal data to obtain effective vibration signal data; a number-real fusion prediction model is constructed based on an adaptive random forest algorithm so as to predict effective vibration signal data, a prediction result comprises an acceleration effective value prediction value, and the adaptive random forest algorithm comprises the step of performing hyper-parameter tuning on the random forest algorithm by using Bayesian optimization; a digital and entity operation integrated interface is arranged, and a dynamic threshold monitoring and real-time alarm mechanism is carried out on predicted effective vibration signal data. According to the invention, a more accurate and efficient tool wear test means is provided by combining a digital technology and actual production data, and the production efficiency is improved.

Description

Technical Field

[0001] The present invention belongs to the fields of mechanical engineering, computer science, and data analysis, and specifically relates to a digital-physical fusion test method for tool wear in the machining of blisks. Background Art

[0002] In the rough milling stage of blisk machining on a five-axis machining center in the traditional manufacturing process, relatively large feed rates and cutting depths are usually adopted to achieve rapid material removal. However, this machining method leads to an enlarged contact area between the tool and the workpiece, increased contact stress, and accompanied by relatively large vibrations and impacts. These factors will exacerbate the fatigue and wear of the tool material, resulting in relatively serious tool wear in the rough milling stage. In the rough milling process of blisks on a five-axis machining center, digital-physical fusion testing of the tool wear condition helps to timely adjust the machining parameters, which can extend the tool life, improve the machining accuracy and surface quality. Given that the degree of tool wear is closely related to the intensity of the vibration signal, the vibration signal can be predicted and analyzed to identify and adjust the key machining parameters in advance. This not only ensures the consistency and reliability of the blisk but also realizes cost reduction and efficiency improvement. The intelligent manufacturing system uses digital-physical fusion technology to realize real-time analysis of the vibration data generated when the tool contacts the workpiece, and predicts the trend of tool wear through intelligent algorithms.

[0003] Currently, the commonly used intelligent algorithms include machine learning algorithms. However, when machine learning algorithms are used alone, there are certain difficulties and problems: Support Vector Machines perform poorly in dealing with complex non-linear relationships; the Random Forest regression model has problems such as high computational resource consumption and poor performance for highly linearly correlated features.

[0004] Therefore, based on the above method technologies and the current situation of the rough milling process, there is an urgent need for a suitable intelligent algorithm to accurately test the tool wear condition in the blisk machining process. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides a digital-physical fusion test method for tool wear in the machining of blisks. Based on the difficulties and problems of current machine learning algorithms, it combines the advantages of existing algorithms and reduces the influence of algorithm defects, and establishes an adaptive random forest digital-physical fusion prediction model to achieve accurate test and analysis of the degree of tool wear. By combining the use of a dynamic threshold monitoring mechanism and a real-time alarm system, the adaptive random forest model issues a warning before the vibration signal reaches the critical state, thereby avoiding machining quality problems or equipment damage caused by tool wear.

[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A numerical-physical fusion testing method for tool wear in integral blisk machining, comprising the following steps:

[0008] Through physical monitoring, vibration signal data during the machining process of a five-axis machining center is synchronously collected and preprocessed to obtain effective vibration signal data;

[0009] Based on the adaptive random forest algorithm, a numerical-physical fusion prediction model is constructed to predict the effective vibration signal data. The prediction results include the predicted value of the effective acceleration value. The adaptive random forest algorithm includes using Bayesian optimization to define the hyperparameter space of the random forest, constructing an objective function; using Gaussian process regression to simulate the objective function, using the expected improvement acquisition function to select hyperparameters, training the random forest model based on the selected hyperparameters, feeding back the output result of the random forest model to the Gaussian process, and using Gaussian process regression to simulate the objective function again, iterating until the optimal hyperparameters are found. Based on the optimal hyperparameters, the random forest model is trained again on the training set to obtain the optimal random forest model, that is, the numerical-physical fusion prediction model;

[0010] Set up a digital and physical operation integration interface to implement a dynamic threshold monitoring and real-time alarm mechanism for the predicted effective vibration signal data.

[0011] In a second aspect, the present invention provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned numerical-physical fusion testing method for tool wear in integral blisk machining.

[0012] In a third aspect, the present invention provides a computer-readable storage medium, on which executable instructions are stored, and when the instructions are executed by a processor, the processor can implement the aforementioned numerical-physical fusion testing method for tool wear in integral blisk machining.

[0013] The advantages of the present invention compared with the prior art are as follows:

[0014] Adopt a real-time data acquisition module, and use the cubic spline interpolation method to process the missing values of the vibration signal data set; then perform numerical-physical fusion testing on the processed vibration signal data. Using the adaptive random forest algorithm, find the optimal parameter combination of the random forest through Bayesian optimization, analyze the preprocessed data, and establish an adaptive random forest numerical-physical fusion prediction model for tool wear to ensure the best prediction performance;

[0015] By collecting and analyzing the vibration data of the cutting tool in the machining process of the integral blisk in real time, a digital virtual-real fusion prediction model of tool wear is established. Combining this model with the actual machining environment, it can simulate and predict the wear state of the cutting tool in the virtual environment, so as to provide accurate wear compensation and optimization strategies for actual machining. Such a fusion not only improves the machining accuracy and extends the tool life, but also significantly enhances the machining efficiency and the intelligent level of the overall production process by reducing trial and error and downtime;

[0016] The dynamic threshold monitoring mechanism and the real-time alarm system can timely detect potential machining problems, reduce unexpected downtime, and improve production efficiency. The time-domain and frequency-domain analysis of the vibration signal in the prediction results can realize a comprehensive virtual-real fusion test and evaluation of the tool wear situation. Brief Description of the Drawings

[0017] Figure 1 is a schematic diagram of a virtual-real fusion test method for tool wear in the machining of integral blisks according to the present invention;

[0018] Figure 2 is a flowchart of a virtual-real fusion test method for tool wear in the machining of integral blisks according to the present invention. Detailed Embodiment

[0019] The present invention will be further described in detail below with reference to the accompanying drawings.

[0020] As Figure 1 shown, the present invention relates to a virtual-real fusion test method for tool wear in the machining of integral blisks, and uses the PyCharm integrated development environment (IDE) to write data processing and analysis codes. The tool wear is relatively serious in the rough milling stage of the integral blisk machining process. This method preprocesses the vibration data of the actual cutting tool in the integral blisk machining process collected and analyzed by sensors in real time, and establishes a digital virtual-real fusion prediction model of tool wear. Combining the digital cutting tool of this model with the actual cutting tool in the actual machining environment and performing real-time interaction, it can simulate and predict and analyze the wear state of the cutting tool in the virtual environment, identify and adjust key machining parameters in advance according to the prediction, achieve reducing the amplitude of the vibration signal, or perform early warning control, and effectively reduce tool wear.

[0021] Specifically, as Figure 2 shown, it is a flowchart of a virtual-real fusion test method for tool wear in the machining of integral blisks according to the present invention, and the specific implementation is as follows:

[0022] Entity monitoring and digital signal synchronous acquisition: Through entity monitoring, synchronous acquisition of vibration signal data during the machining process of the five-axis machining center is realized, and the specific implementation is as follows:

[0023] Through research, it is understood that key process parameters such as spindle speed, torque, lubricating oil temperature, and feed speed all have potential impacts on vibration signals. In order to accurately predict vibration signals and optimize the machining process, a quantitative method of sensitivity analysis is used to evaluate the influence degree of the above parameters on vibration signals. After analysis, it is found that the feed speed is the main factor affecting vibration signals, and its change has a significant positive correlation with the intensity of vibration signals. Although the remaining parameters also have certain effects, their influence on vibration signals is relatively small compared to the feed speed. Therefore, the data acquisition module mainly collects data related to vibration signals at different feed speeds.

[0024] The five-axis machining center is equipped with three high-precision sensors distributed on the spindle, beside the spindle motor, and under the machining fixture, which are used to collect vibration signal data during the machining process in real time to realize the digital mapping of the physical machining process. During the machining process, the three sensors respectively collect vibration data at different feed speeds. After being preprocessed, these data are stored in the comma-separated values (CSV) format. The CSV file contains time series data, as well as columns such as 'root mean square acceleration', 'peak acceleration', 'kurtosis coefficient of acceleration', 'root mean square velocity', 'peak velocity', 'root mean square displacement', and 'peak displacement'. Among them, the peak columns can reflect the peaks of the tool vibration signals, representing the maximum vibrations endured by the tool; the root mean square columns reflect the average values of the vibration signals, and the root mean square columns are mainly used for predicting vibration signals; the kurtosis coefficient can reflect potential fault information of the tool, and when mechanical impacts are caused by damage to the working surface, the kurtosis coefficient will increase rapidly. Both the acceleration and velocity data columns can be used to characterize the intensity of the tool vibration signals, and they can be mutually converted through calculation formulas.

[0025] Use the PyCharm integrated development environment to write data processing and analysis code. The code extracts the 'root mean square acceleration' data required by the digital-physical fusion model from the CSV file to characterize the intensity of vibration signals and realizes the fusion analysis of digital and physical signals.

[0026] Combining high-precision sensors with Internet of Things technology can achieve remote monitoring and fault prediction. This is not limited to tool wear but can also be extended to the health management of the entire factory equipment.

[0027] Digital missing value processing: Conduct data preprocessing on the collected data, including digital missing value processing, select appropriate interpolation methods for missing value estimation, and obtain effective vibration signal data. The specific implementation is as follows:

[0028] Select an interpolation method to ensure high-precision digital-physical fusion prediction under various feed speed conditions. Compared with quadratic interpolation, cubic spline interpolation can provide higher precision and smoothness; compared with piecewise Hermite interpolation, it is simpler to handle derivatives; compared with Newton interpolation and Hermite interpolation, it gives more consideration to smoothness and ease of use. Therefore, the present invention uses the cubic spline interpolation method to estimate missing values of vibration signals.

[0029] Use the cubic spline interpolation method to estimate missing values of vibration signals. It fits known data points by constructing a piecewise cubic polynomial function and interpolates between these points. First, determine the known data points and establish interpolation nodes, then construct a system of equations through interpolation conditions, continuity conditions, and boundary conditions, and solve to obtain the coefficients of the cubic polynomial for each sub-interval. After that, construct a complete cubic spline function and use the corresponding polynomial to estimate within the interval where the missing value is located. Finally, verify the accuracy of the interpolation by comparing the interpolation result with the actual value.

[0030] In the adaptive random forest model, the cubic spline interpolation method is used to optimize model parameters and accurately predict input data. For the sub-interval , the cubic spline function can be expressed as:

[0031]

[0032] where are the coefficients of the polynomial, which need to be determined by data points, boundary conditions, and smoothness conditions. u is the independent variable, representing any point within the sub-interval . is the starting point of the sub-interval . represents the difference between u and the starting point of the sub-interval.

[0033] Digital-physical fusion prediction model and adaptive random forest: A digital-physical fusion prediction model is constructed based on the adaptive random forest algorithm to predict effective vibration signal data. The prediction results include the predicted value of the effective value of acceleration. The adaptive random forest algorithm includes using Bayesian optimization to tune the hyperparameters of the random forest algorithm. The specific implementation is as follows:

[0034] After data preprocessing, through feature analysis of different model algorithms, and using them to make trial predictions on existing vibration signal data respectively, and then comparing the predicted values of vibration signals with the true values, the model and algorithm with the best prediction effect can be selected. The specific implementation is as follows:

[0035] First, use machine learning algorithms to predict the vibration signal-related data of the existing feed rate, including but not limited to algorithms such as Support Vector Machine (SVM) and XGBoost.

[0036] SVM is a linear model, and it is not as effective as the adaptive random forest in dealing with complex non-linear relationships; XGBoost is a serial integration, and it takes a long training time to process large-scale data, while the adaptive random forest can automatically perform hyperparameter tuning, process large-scale data sets in a shorter time, and maintain high prediction performance.

[0037] Consider using an adaptive random forest model with missing values filled in by interpolation method to predict the effective vibration signal and fit the vibration signal prediction curve. Through the comparative analysis of the prediction curve of this model and the physical vibration signal, it is found that it is the optimal digital-physical fusion prediction model. Specifically:

[0038] First, use Bayesian optimization to define the hyperparameter space of the random forest and construct the objective function. Then, perform the Bayesian optimization process, use Gaussian process regression to simulate the objective function, use the Expected Improvement acquisition function (EI) to select hyperparameters, train the random forest model with hyperparameters on the training set, feedback the results to the Gaussian process model, iterate the model to improve the effect, and find the optimal values of the hyperparameters. The random forest process uses the optimal hyperparameters determined by Bayesian optimization to train the random forest model on the training set. Then, evaluate the performance of the adaptive random forest model on the test set for model verification. Use the trained adaptive random forest model to predict the tool vibration signal data. Input the processed feature data, and the model will output the prediction result according to the learned rules. The following introduces the specific implementation of each step separately:

[0039] Bayesian optimization can be used for global optimization by constructing a probability model of the objective function and using this model to select the next evaluation point to maximize the expected improvement of the objective function.

[0040] In the adaptive random forest model, the concepts of Bayesian optimization and Gaussian process regression can be integrated into the model training and prediction process to improve the self-adaptability and prediction accuracy of the model.

[0041] In the model training process, use Gaussian process regression to perform probability prediction on the function values of unknown points. The Gaussian process regression model can model the prediction results of each tree in the random forest. The formula of the Gaussian process regression model used in Bayesian optimization is as follows, and its meaning is: the value of the function f at point x follows a Gaussian process distribution with as the mean and as the covariance.

[0042]

[0043] Among them, \(g(x)\) represents the objective function value at the input point \(x\); is the mean function, which is usually assumed to be a constant; is the kernel function, which defines the similarity between any two points \(x\) and \(x'\) in the input space. represents a Gaussian Process, which is a probability process where the joint distribution of any set of points is a multivariate normal distribution.

[0044] In Bayesian optimization, the kernel function used in Gaussian process regression is a function that describes the similarity between any two points \(x\) and \(x'\) in the input space. The kernel function uses the Radial Basis Function (RBF) kernel, which assumes that the similarity decreases exponentially as the distance between points increases. It can be expressed as:

[0045] ,

[0046] where, is the standard deviation of the kernel function, and \(l\) is the length scale parameter.

[0047] When using the Gaussian process regression model to predict a new input point , not only a predicted value is obtained, but also a measure of the uncertainty about this predicted value. The mean and variance of the prediction distribution provide an accurate estimate and confidence interval for the predicted value. These formulas combine the information of the training data and the characteristics of the kernel function to calculate the prediction distribution of the new point. The formulas for the mean and variance of the prediction distribution are as follows:

[0048] ,

[0049] ,

[0050] where, \(X\) represents the set of all input points in the training dataset; is the new input point for which the objective function value is predicted; is the variance of the noise, \(Y\) is the vector of objective values of the training data points, and \(I\) represents the identity matrix of the same dimension as the kernel matrix . and are the vectors composed of the kernel function values between the new input point and all training data points \(X\); is the kernel matrix composed of the kernel function values between the training data points \(X\); is the kernel function value of the new input point with itself, representing its self-similarity.

[0051] In the adaptive random forest, Bayesian optimization is used to guide the search process and select an evaluation point that maximizes the probability that the function value at that point exceeds the current best value. The expected improvement acquisition function (EI) is a commonly used choice, which takes into account the current best observation , aiming to find the point that is expected to bring the greatest improvement:

[0052] ,

[0053] where represents the expected improvement value at point x, is the expectation operator, indicating the expectation of the expression within the parentheses; is the current best objective value, is a very small positive number.

[0054] In the adaptive random forest, the predicted value of each tree for the input sample z is , where b is the index of the tree. Then the predicted output of the random forest can be expressed as:

[0055] ,

[0056] where represents the overall predicted value of the random forest for the target variable t; B is the total number of trees in the random forest.

[0057] The variance estimate of the random forest consists of two parts: the variance of each tree's prediction and the variance between the trees. It can provide information about the prediction uncertainty of the digital-real fusion model:

[0058]

[0059] where represents the variance of the predicted value of the target variable t. is the average of all trees' predictions, represents the variance of each tree's prediction, represents the variance between the trees.

[0060] In the random forest, the importance of each feature can be calculated by averaging the importance levels of that feature across all trees:

[0061]

[0062] where refers to the importance score of feature w. B is the total number of trees in the random forest. is the importance score of feature w in the b-th tree.

[0063] To evaluate the performance of the adaptive random forest model, the error of the model can be estimated by the cross-validation method:

[0064]

[0065] Among them, CV Error refers to the cross-validation error, N is the number of samples, is the true value of the i-th sample, is the predicted value of the i-th sample, and L is the loss function, which is used to measure the difference between the predicted value and the true value of the sample.

[0066] In a specific embodiment, the adaptive random forest algorithm uses Bayesian optimization to tune the hyperparameters of the random forest and obtains the optimal parameters: for the X, Y, and Z axes of the random forest regression, 200, 100, and 100 decision trees are used respectively, the minimum number of samples for splitting is 2, 5, and 2 respectively, and the maximum depth of each tree is limited to 5, 5, and 10 respectively.

[0067] This method combines the traditional machining process with artificial intelligence algorithms. By using the digital-physical fusion model of machine learning to predict tool wear, the machining accuracy and efficiency can be improved.

[0068] Vibration signal prediction: Use the adaptive random forest model to digitally predict the data related to the vibration signal at an unknown feed rate, and draw the curve of the actual vibration signal of the entity and the fitting curve of the vibration signal predicted by the digital model. The resulting curve graph is a real-time dynamic fluctuation curve graph. The graph contains two curves, namely the true value of the effective acceleration and the predicted value of the effective acceleration.

[0069] Based on the digital-physical fusion prediction model, a dynamic threshold monitoring mechanism and a real-time alarm system are additionally added, the curve effects of the vibration signal in the time domain and frequency domain are added, and a digital and physical operation integration interface is designed. The specific implementation is as follows:

[0070] Dynamic threshold monitoring and real-time alarm system: The dynamic threshold monitoring mechanism combines the entity machining data with the prediction results of the digital model. By setting a specific threshold to evaluate the prediction results of the vibration signal, real-time digital-physical fusion testing is achieved. In addition, an alarm counter is set for each axis of the system. If the predicted values of multiple consecutive data points (such as 30) exceed the threshold, the system will automatically trigger the alarm mechanism to remind the staff that under the current parameter settings, the vibration signal will continue to be at a high level, thus exacerbating tool wear. The staff can adjust key parameters such as the feed rate in a timely manner according to the alarm prompt to optimize the vibration signal, reduce tool wear, and ensure the stability of the machining process and the machining quality.

[0071] Tool wear is closely related to the physical properties of materials. By analyzing the vibration signal data, the interaction between the tool and the workpiece material can be inferred, thereby guiding the development of new materials to improve tool performance and durability. By accurately predicting tool wear, material waste and energy consumption can be reduced, meeting the goals of sustainable development. This method helps to reduce the environmental impact of industrial production.

[0072] Digital and physical operation integration interface: In the digital and physical operation integration interface, the user can input specific feed rate values according to the prompts for vibration signal prediction. The result graph integrates the effective acceleration value curves of two known feed rates adjacent to the target feed rate, enabling intuitive comparative analysis to judge the accuracy of the prediction results.

[0073] The above mainly analyzes the vibration signal prediction results in the time domain. To achieve accurate testing of tool wear, a vibration signal prediction curve is plotted in the frequency domain, and characteristic quantities such as the center frequency, frequency variance, and mean square frequency are extracted. Through the analysis method combining the time domain and the frequency domain, a comprehensive digital-physical fusion test and evaluation of tool wear can be realized.

[0074] The whole process reflects the application of data science in Industry 4.0. In the context of Industry 4.0, by collecting and analyzing a large amount of processing data, the production process can be optimized, downtime can be reduced, and production efficiency can be improved. This data-driven method can be extended to the entire production line to achieve comprehensive intelligent manufacturing. In the future, this technology can be further developed into an adaptive control system to adjust processing parameters in real time to extend tool life and improve processing quality.

[0075] In summary, the present invention discloses a digital-physical fusion test method for tool wear in the machining of integral blisks, including: predicting vibration signals using a data acquisition module, an interpolation analysis method, and an adaptive random forest digital-physical fusion prediction model; using a dynamic threshold monitoring mechanism and a real-time alarm system for early warning, adding the curve effects of vibration signals in the time domain and the frequency domain, and designing a digital and physical operation integration interface. This method can achieve digital-physical fusion testing of tool wear in the machining of integral blisks on a five-axis machining center.

[0076] In a second aspect, the present invention provides an electronic device, including: one or more processors; a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned digital-physical fusion test method for tool wear in the machining of integral blisks.

[0077] In a third aspect, the present invention provides a computer-readable storage medium, on which executable instructions are stored, and when the instructions are executed by a processor, the processor can implement the foregoing method for testing the digital-physical fusion of tool wear for the overall blisk machining.

[0078] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A digital-real fusion test method for tool wear in blisk machining, characterized in that: The following steps are involved: Through entity monitoring, the vibration signal data of the five-axis machining center during machining is synchronously collected and pre-processed to obtain effective vibration signal data; A digital-real fusion prediction model is constructed based on the adaptive random forest algorithm to predict effective vibration signal data; Set up an integrated interface of digital and physical operations, and implement dynamic threshold monitoring and real-time alarm mechanism for the predicted effective vibration signal data.

2. The digital-real fusion testing method for tool wear in blisk machining according to claim 1, characterized in that: The synchronous collection of vibration signal data during the machining process of the five-axis machining center is based on high-precision sensors distributed on the main shaft, the motor beside the main shaft, and under the machining fixture. The collected vibration signal data is saved in CSV format; the preprocessing includes using a cubic spline interpolation method to estimate missing values ​​of the vibration signal data.

3. The digital-real fusion testing method for tool wear in blisk machining according to claim 2, characterized in that: The method of using the cubic spline interpolation method to estimate missing values ​​of vibration signal data includes constructing a piecewise cubic polynomial function to fit known vibration signal data points and interpolating between the known vibration signal data points.

4. The digital-real fusion testing method for tool wear in blisk machining according to claim 3, characterized in that: The cubic spline interpolation method specifically includes: Determine the known vibration signal data points and establish interpolation nodes, build a set of equations through interpolation conditions, continuity conditions and boundary conditions, and solve to obtain the cubic polynomial coefficients of each segmented subinterval; Construct a complete cubic polynomial function and use the corresponding cubic polynomial to estimate in the interval where the missing value is located; The accuracy of the interpolation is verified by comparing the interpolation results with the actual values.

5. The digital-real fusion testing method for tool wear in blisk machining according to claim 1, characterized in that: The method of constructing a digital-real fusion prediction model based on an adaptive random forest algorithm to predict effective vibration signal data includes: Evaluate the performance of the digital-real fusion prediction model on the test set and perform model validation; The effective vibration signal data is input into the digital-real fusion model for prediction and the prediction results are output.

6. The digital-real fusion testing method for tool wear in blisk machining according to claim 5, characterized in that: The prediction result includes an acceleration effective value prediction value.

7. The digital-real fusion testing method for tool wear in blisk machining according to claim 5, characterized in that: The adaptive random forest algorithm includes using Bayesian optimization to define the hyperparameter space of the random forest and constructing an objective function; using Gaussian process regression to simulate the objective function, using the expected improvement acquisition function to select hyperparameters, training the random forest model based on the selected hyperparameters, feeding the output results of the random forest model back to the Gaussian process, and again using Gaussian process regression to simulate the objective function, iterating until the optimal hyperparameters are found, and based on the optimal hyperparameters, training the random forest model again on the training set to obtain the optimal random forest model, that is, the digital-real fusion prediction model.

8. The digital-real fusion testing method for tool wear in blisk machining according to claim 1, characterized in that: The dynamic threshold monitoring and real-time alarm mechanism for the predicted effective vibration signal data includes setting a specific threshold to evaluate the predicted results of the effective vibration signal data. If the predicted values ​​of several consecutive data points exceed the threshold, the system automatically triggers the alarm mechanism.

9. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; Among them, when one or more programs are executed by the one or more processors, the one or more processors implement the digital-real fusion testing method for tool wear for integral blade disk processing as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: Executable instructions are stored thereon, which, when executed by a processor, enable the processor to implement a digital-real fusion testing method for tool wear for integral blade disk machining as described in any one of claims 1-8.

Images