Milling tool wear detection method based on dual self-attention mechanism

Abstract

The invention belongs to the technical field of mechanical cutting tool wear detection, and particularly discloses a milling tool wear detection method based on a dual self-attention mechanism, which comprises the following steps of sensor arrangement and experimental parameter setting, signal acquisition sample selection, model training, milling tool wear detection and milling tool wear detection. According to the detection method, effective feature extraction and data dimension reduction are carried out on various sensor signals in the tool machining process, higher-precision tool wear prediction is better achieved through the information extraction capacity of the deep learning technology, and the complexity of the model is reduced through more ingenious model optimization design; the requirement of the model on adaptive hardware is reduced, and the real realization of tool wear becomes possible, so that the system reliability is effectively improved, the accident occurrence risk is reduced, and the safe operation of a machine tool is ensured; the method not only overcomes the limitation of a traditional detection method, but also brings a more efficient monitoring means for the milling industry.

Description

Technical Field

[0001] The invention belongs to the technical field of mechanical cutting tool wear detection, and in particular relates to a milling tool wear detection method based on a dual self-attention mechanism. Background Art

[0002] With the rapid development of modern manufacturing, the performance and life requirements of cutting tools are increasing day by day. The monitoring and evaluation of cutting tool wear has become an important part of improving processing quality and efficiency. Existing cutting tool wear detection methods are mainly divided into two categories. One is to perform real-time image analysis of the back face of the cutting tool based on machine vision technology to obtain the wear value of the cutting tool. However, this method is often affected by cutting fluid and chips, and cannot adapt to complex processing environments. Another method is to analyze various sensor signals during the cutting tool processing process through machine learning technology to predict the wear of the cutting tool. This method is divided into two types, one is to use traditional machine learning algorithms, and the other is to use deep learning algorithms.

[0003] Traditional machine learning algorithms are unable to learn complex feature representations and usually rely on feature engineering, which requires manual selection and feature extraction of data. This results in the performance of the model being largely dependent on the quality and relevance of the input features. We also found in our research that as the tool type, cutting conditions, and processing materials change, the correlation between the statistical indicators of the sensor signals extracted by feature engineering during the processing and tool wear will change, that is, feature engineering established by analyzing sensor data obtained under a specific tool, specific processing conditions, and specific processing materials is usually only applicable to this type of working condition. With the development of deep learning, deep learning models that can automatically extract features from data have made it possible to solve this problem. Since the data points collected from a single cutting data of a tool may exceed 200,000, how to perform effective feature extraction and data dimensionality reduction has become a current problem. However, the existing methods of using deep learning models to detect tool wear usually divide the data into N segments, take some statistical data indicators of each segment, such as maximum value, mean value, variance, etc., and then reorganize the data. On the one hand, this method has the limitations of traditional machine learning methods. During the data reconstruction process, a lot of information related to tool wear will be lost, reducing the accuracy of the prediction. On the other hand, these statistical indicators are not necessarily applicable to all tool wear data. The selected statistical indicators may only apply to the current tool. In addition, this method involves a large number of data storage calls, which is not easy to achieve online monitoring.

[0004] Therefore, it is very necessary to develop a new method for detecting tool wear in milling. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the prior art and provide a milling tool wear detection method based on a dual self-attention mechanism, which improves the accuracy and practicability of tool wear prediction through higher-dimensional feature mapping and effective time series modeling, overcomes the limitations of traditional detection methods, and brings more efficient monitoring means to the milling industry, promoting the further development of intelligent manufacturing; reduces the complexity of the model through model optimization design, thereby reducing the model's requirements for compatible hardware, effectively improving system reliability, and reducing the risk of accidents, thereby ensuring the safe operation of the machine tool.

[0006] In order to achieve the above objectives, the technical solution of the present invention is: a milling tool wear detection method based on a dual self-attention mechanism, comprising the following steps:

[0007] (1) Sensor arrangement and setting of experimental parameters: After determining the range of cutting parameters, choose to conduct a single cutting parameter or multiple cutting parameter milling experiment; when conducting a single cutting parameter milling experiment, use bolts to install the dynamometer on the workbench, the workpiece is connected to the dynamometer by bolts, the vibration sensor is installed on the spindle, the acoustic emission sensor is placed near the workpiece, and the total power of the machine tool is collected using a power meter; after the sensor arrangement is completed, cutting is carried out under the same cutting parameters and processing conditions to obtain the data of each sensor during the processing; when conducting a multiple cutting parameter milling experiment, the workpiece is clamped by a vise; the arrangement of the acceleration sensor and the acoustic emission sensor is the same as that of the single cutting parameter milling experiment;

[0008] (2) Selection of signal acquisition samples: Use an electron microscope to measure and record the tool wear after a single machining process, and the wear value is the maximum wear value of the back tool surface; the single machining data obtained is matched with the tool wear value according to the number of machining times; each tool wear value corresponds to multiple tool passes; when extracting vibration signal samples and acoustic emission signals, first select the tool pass process signal corresponding to each tool wear value, and then select the last 4s of the stable sensor signal of each signal in the single tool pass process as the signal acquisition sample, then each tool wear value corresponds to 1 sample, and each sample contains vibration signals in the three directions of XY and Z, cutting force signals in the three directions of XY and Z, and acoustic emission signals;

[0009] (3) Model training: the one-to-one correspondence between the stable sensor signal and the tool wear value in step (2) is input into the tool wear prediction algorithm based on dual attention for training;

[0010] (4) Milling tool wear detection: The milling processing signal to be detected is input into the trained model to predict the wear of the milling tool and give the corresponding evaluation results;

[0011] (5) Milling tool wear detection: The milling processing signal to be detected is input into the trained model to predict the wear of the milling tool and give the corresponding evaluation results.

[0012] In a preferred embodiment of the present invention, in step (1), vibration, acoustic emission and cutting force data of multiple identical machining processes are collected to obtain original input data represented by XS, and single machining data represented by XSn, where n=1, 2, ..., n, and n represents the number of machining times.

[0013] In a preferred embodiment of the present invention, the specific method of training the model in step (3) comprises the following steps:

[0014] ① Data folding: fold the data in the time series dimension based on the number of points collected during one rotation of the tool, and then fold the data in the channel dimension; after folding the data, the number of neurons required for data mapping within the cycle can be effectively reduced, and the complexity of the model can be effectively reduced while performing high-dimensional feature mapping;

[0015] ② Data mapping within a cycle: By building a mapping layer for data within a cycle based on MLP, this mapping layer takes into account the order of the data, comprehensively considers all data points within a cycle, and maps the data in a higher dimension to avoid the impact of errors in various sensor data at different times;

[0016] ③ Channel dimension expansion of data: When performing feature mapping within a cycle, since the design purpose is to perform high-dimensional mapping within the cycle with data points within the cycle, the data is also folded along the channel dimension. When performing feature extraction within the cycle, it is necessary to fuse the features of signals from different channels, so the data needs to be expanded along the channel dimension;

[0017] ④ Feature extraction within the cycle: After the data is expanded along the channel dimension, it is input into the cross-attention module for feature extraction within the cycle, and the connection between different time steps within the cycle is established. In the cross-attention module, the time step dimension of the matrix Q is set to 1, thereby effectively reducing the dimension of the data;

[0018] ⑤ Time series modeling between cycles: The data that has undergone feature extraction within a cycle is input into the self-attention mechanism. Since the data has undergone feature extraction within a cycle, the time step dimension of the data within a tool rotation cycle becomes 1. Then, the connection between different time steps between cycles is established through the self-attention mechanism, thereby performing time series modeling between cycles on the data.

[0019] ⑥Data aggregation: The data after multiple attention mechanisms are input into the data aggregation layer based on MLP. GELU is set as the activation function between each layer to finally get the predicted value;

[0020] ⑦ Calculate the parameters of each layer: The predicted value obtained above and the actual wear value are calculated through the cross entropy loss function, and then the size of each parameter in the above-mentioned layers is calculated through the back propagation algorithm.

[0021] In a preferred embodiment of the present invention, the milling tool wear detection method also includes, in order to effectively extract and reduce the data, after the data passes through the intra-cycle data mapping layer, expanding the data along the channel dimension, and inputting the data into the cross-attention mechanism of the first layer in sequence in the order of the cycle to perform intra-cycle feature extraction, and then performing time series modeling between cycles by calculating the self-attention between cycles.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] 1. The self-attention mechanism used in the present invention can take into account the relationship between various data points when processing data, fully explore its temporal characteristics, and achieve effective fusion of multi-channel data;

[0024] 2. The tool wear detection method of the present invention is based on a dual self-attention mechanism, which improves the accuracy and practicality of tool wear prediction through higher-dimensional feature mapping and effective time series modeling;

[0025] 3. The tool wear detection method based on the dual self-attention mechanism of the present invention can not only overcome the limitations of traditional detection methods, but also bring more efficient monitoring means to the milling industry and promote the further development of intelligent manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of the tool wear detection method of the present invention;

[0027] Figure 2 It is a schematic diagram of the tool wear prediction model of the present invention;

[0028] Figure 3 This is a schematic diagram of the workpiece specifications and processing direction of the present invention;

[0029] Figure 4 This is a diagram showing the verification results of the performance of the model of the present invention. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is described in more detail below with reference to the accompanying drawings and specific embodiments, but the protection scope of the present invention is not limited to these embodiments.

[0031] A milling tool wear detection method based on a dual self-attention mechanism comprises the following steps:

[0032] (1) Sensor arrangement and experimental parameter setting: Determine the cutting parameter range and conduct milling experiments with single or multiple cutting parameters. After the sensor arrangement is completed, perform cutting under the same cutting parameters and processing conditions to obtain the data of each sensor during the processing.

[0033] (2) Signal acquisition sample selection: Use an electron microscope to measure and record the tool wear after a single machining process. The wear value is the maximum wear value of the back tool face. Extract the stable sensor signal of the last 4 seconds of a single machining process as the signal acquisition sample, including vibration signal, cutting force signal and acoustic emission signal. According to the number of machining times, the obtained single machining data is matched with the tool wear value one by one. Each tool wear value corresponds to a single cutting process.

[0034] (3) Model training: The one-to-one correspondence between the stable sensor signal and the tool wear value in step (2) is input into the tool wear prediction algorithm based on dual attention for training;

[0035] (4) Milling tool wear detection: The milling processing signal to be detected is input into the trained model to predict the wear of the milling tool and give the corresponding evaluation results;

[0036] (5) Milling tool wear detection: The milling processing signal to be detected is input into the trained model to predict the wear of the milling tool and give the corresponding evaluation results.

[0037] In the step (1), vibration, acoustic emission and cutting force data of the same machining process are collected for multiple times, and the original input data is represented as XS, and the single machining data is represented as XSn, n=1, 2, ..., n, where n represents the number of machining times.

[0038] When performing a single cutting parameter milling experiment in step (1), a dynamometer is mounted on a workbench using bolts, a workpiece is connected to the dynamometer via bolts, a vibration sensor is mounted on a spindle, and an acoustic emission sensor is placed near the workpiece.

[0039] When the multi-cutting parameter milling experiment is performed in the step (1), the workpiece is clamped by a vise; the arrangement of the acceleration sensor and the acoustic emission sensor is the same as that of the single cutting parameter milling experiment.

[0040] When extracting vibration signal samples and acoustic emission signals in the step (2), firstly, a full-process signal of a single tool pass corresponding to each tool wear value is selected; secondly, the stable signal of the last 4 seconds of the full-process signal of a single tool pass is selected as a sample, then each tool wear value corresponds to one sample, and each sample contains vibration signals in three directions of XY and Z, cutting force signals in three directions of XY and Z, and acoustic emission signals.

[0041] When extracting the power signal samples in the step (2): each tool wear value also corresponds to extracting a power signal of the whole process of tool passing; the last 4s stable signal is extracted from the power signal of the whole process of a single tool passing as a power signal sample, and each sample contains vibration signals in three directions of XY and Z, cutting force signals in three directions of XY and Z, and acoustic emission signals.

[0042] The specific method of training the model in step (3) includes the following steps:

[0043] ① Data folding: fold the data in the time series dimension based on the number of points collected during one rotation of the tool, and then fold the data in the channel dimension; after folding the data, the number of neurons required for data mapping within the cycle can be effectively reduced, and the complexity of the model can be effectively reduced while performing high-dimensional feature mapping;

[0044] ② Data mapping within a cycle: By building a mapping layer for data within a cycle based on MLP, this mapping layer takes into account the order of the data, comprehensively considers all data points within a cycle, and maps the data in a higher dimension to avoid the impact of errors in various sensor data at different times;

[0045] ③ Channel dimension expansion of data: When performing feature mapping within a cycle, since the design purpose is to perform high-dimensional mapping within the cycle with data points within the cycle, the data is also folded along the channel dimension. When performing feature extraction within the cycle, it is necessary to fuse the features of signals from different channels, so the data needs to be expanded along the channel dimension;

[0046] ④ Feature extraction within the cycle: After the data is expanded along the channel dimension, it is input into the cross-attention module for feature extraction within the cycle, and the connection between different time steps within the cycle is established. In the cross-attention module, the time step dimension of the matrix Q is set to 1, thereby effectively reducing the dimension of the data;

[0047] ⑤ Time series modeling between cycles: The data that has undergone feature extraction within a cycle is input into the self-attention mechanism. Since the data has undergone feature extraction within a cycle, the time step dimension of the data within a tool rotation cycle becomes 1. Then, the connection between different time steps between cycles is established through the self-attention mechanism, thereby performing time series modeling between cycles on the data.

[0048] ⑥Data aggregation: The data after multiple attention mechanisms are input into the data aggregation layer based on MLP. GELU is set as the activation function between each layer to finally get the predicted value;

[0049] ⑦ Calculate the parameters of each layer: The predicted value obtained above and the actual wear value are calculated through the MSE loss function, and then the size of each parameter in the above-mentioned layers is calculated through the back propagation algorithm.

[0050] The milling tool wear detection method also includes, in order to effectively extract and reduce the data, after the data passes through the intra-cycle data mapping layer, expanding the data along the channel dimension, and inputting the data into the cross-attention mechanism of the first layer in sequence in the order of the cycle to perform intra-cycle feature extraction, and then performing time series modeling between cycles by calculating the self-attention between cycles.

[0051] Example 1

[0052] The milling tool wear detection method based on the dual self-attention mechanism is used to perform actual detection of tool wear. The specific steps are as follows:

[0053] (1) Sensor arrangement and setting of experimental parameters: The experiment was conducted under a single cutting parameter. The dynamometer was installed on the workbench with bolts. The workpiece was connected to the dynamometer by bolts. The vibration sensor was installed on the spindle and the acoustic emission sensor was placed near the workpiece. After the sensor arrangement was completed, cutting was performed under the same cutting parameters and processing conditions to obtain the data of each sensor during the processing. The cutting condition settings are shown in Table 1. The models of the tool and cutter body are shown in Table 2 and Table 3, respectively.

[0054] The cutting process is as follows:

[0055] ①Workpiece specifications such as Figure 3 As shown, an indexable ball end milling cutter of the above specifications is used to process once along the illustrated processing direction, and the three-dimensional cutting force, vibration signal and acoustic emission signal during the processing are recorded.

[0056] ② After every 5000mm of machining, record the tool flank wear value.

[0057] ③The total processing length of each knife is 60 meters (or stop and mark if the blade breaks in the middle).

[0058] ④ Test three identical knives in the same way under the same processing conditions.

[0059] (2) Selection of signal acquisition samples: Use an electron microscope to measure and record the tool wear after a single machining process, and take the maximum wear value of the back tool face as the wear value. Take the last 4 seconds of machining data obtained from each wear measurement, and correspond the last 4 seconds of machining data measured each time to the tool wear value. Each tool wear value corresponds to 3 axis vibration signals, 3 cutting force signals, and acoustic emission signals. The sampling frequency of the sensor signal is 30 kHz, so a single sample has 120,000 points, with a total of 8 columns of data.

[0060] (3) Model training: The processed tool sensor signal is matched with the tool wear value one by one, and the data set is divided by three-fold cross validation. The tool wear prediction algorithm based on dual attention is input for training. The loss function of the training process takes the MSE function, and the optimizer selects Adam. Among the hyperparameters of the input model, the batch size is set to 16, the number of cycles is set to 600, the single cycle size is 200, the number of heads in the cross attention mechanism (first attention) is 2, the model dimension is 7, the number of heads in the self-attention mechanism (second attention) is set to 256, the drop rate is set to 0.1, the number of heads is set to 4, the feedforward network dimension is set to 512, the number of layers is set to 2, and the activation function is Gelu;

[0061] Specifically comprising the following steps: In an example of applying the present invention, the specific method of training the model in step (3) comprises the following steps:

[0062] ① Data folding: The data is folded in the time series dimension with the number of points collected during one rotation of the tool as the period, and then the data is folded in the channel dimension. The folded data dimension changes from the previous (16, 120000, 7) to (112, 600, 200);

[0063] ② Data mapping within the cycle: A mapping layer for the data within the cycle is established based on MLP. The number of neurons in the mapping layer is 60, and the dropout rate is set to 0.1. After the mapping layer, the data dimension is (112, 600, 200);

[0064] ③ Channel dimension expansion of data: In order to better perform channel information fusion, the data is expanded along the channel dimension, and the data dimension becomes (16, 7, 600, 200);

[0065] ④ Feature extraction within the cycle: After the data is expanded along the channel dimension, it is input into the cross attention module in a cycle cycle for feature extraction within the cycle. The cross attention mechanism (first attention) has a multi-head number of 2, a model dimension of 512, and a discard rate of 0.1. After extraction, the data dimension becomes (16, 7, 600);

[0066] ⑤ Time series modeling between cycles: swap the positions of the second and third dimensions of the data extracted from the features within the cycle, and input them into the self-attention mechanism. The number of multi-heads in the self-attention mechanism (second attention) is set to 256, the drop rate is set to 0.1, the number of multi-heads is set to 4, the feedforward network dimension is set to 512, the number of layers is set to 2, and the activation function is Gelu. After time series modeling, the data dimension is (16, 600, 256);

[0067] ⑥ Data aggregation: The data after multiple attention mechanisms are expanded, and the dimension becomes (16, 153600), and input into the data aggregation layer based on MLP. GELU is set as the activation function between each layer. The number of neurons in each layer is 12000, 600, and 1 respectively, and the predicted value is finally obtained;

[0068] ⑦ Calculate the parameters of each layer: The predicted value obtained above and the actual wear value are calculated through the MSE loss function, and then the size of each parameter in the above-mentioned layers is calculated through the back propagation algorithm.

[0069] (4) Milling tool wear detection: The milling processing signal to be detected is input into the trained model to predict the wear of the milling tool and give the corresponding evaluation results;

[0070] Since the data set was subjected to three-fold cross validation, three different knives were used as test sets to verify the model performance. The final verification results are as follows: Figure 4 As shown, after calculation, the average absolute error between the predicted value and the true value is 12.86μm, and the root mean square error is 18.65μm, which has good performance.

[0071] Table 1 Cutting parameters in Example 1

[0072]

[0073] Table 2 Tool models and specifications in Example 1

[0074] factory Material code / blade model Specification Golden Heron QTD3007V-S QTD3007V-S

[0075] Table 3. Blade model and size in Example 1

[0076] factory Cutter body model Diameter D(mm) Hitachi ABPF30S32W 30

[0077] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some or all of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A milling tool wear detection method based on a dual self-attention mechanism, characterized in that: The following steps are involved: (1) Sensor arrangement and experimental parameter setting: Determine the cutting parameter range and conduct single cutting parameter or multi-cutting parameter milling experiments. After the sensor arrangement is completed, single cutting parameter milling is carried out under the same cutting parameters and processing conditions to obtain the data of each sensor during the processing; multi-cutting parameter milling experiments are carried out under variable cutting parameters. (2) Signal acquisition sample selection: Use an electron microscope to measure and record the tool wear after a single machining process. The wear value is the maximum wear value of the back tool face. Extract the stable sensor signal of the last 4 seconds of a single machining process as the signal acquisition sample, including vibration signal, cutting force signal and acoustic emission signal. According to the number of machining times, the obtained single machining data is matched with the tool wear value one by one. Each tool wear value corresponds to a single cutting process. (3) Model training: The one-to-one correspondence between the stable sensor signal and the tool wear value in step (2) is input into the tool wear prediction algorithm based on dual attention for training; (4) Milling tool wear detection: The milling processing signal to be detected is input into the trained model to predict the wear of the milling tool and give the corresponding evaluation results; (5) Milling tool wear detection: The milling processing signal to be detected is input into the trained model to predict the wear of the milling tool and give the corresponding evaluation results.

2. The milling tool wear detection method based on the dual self-attention mechanism as claimed in claim 1, characterized in that: In the step (1), vibration, acoustic emission and cutting force data of the same machining process are collected for multiple times, and the original input data is represented as XS, and the single machining data is represented as XSn, n=1, 2, ..., n, where n represents the number of machining times.

3. The milling tool wear detection method based on dual self-attention mechanism as claimed in claim 1, characterized in that: When performing a single cutting parameter milling experiment in step (1), a dynamometer is mounted on a workbench using bolts, a workpiece is connected to the dynamometer via bolts, a vibration sensor is mounted on a spindle, and an acoustic emission sensor is placed near the workpiece.

4. The milling tool wear detection method based on dual self-attention mechanism as claimed in claim 1, characterized in that: When the multi-cutting parameter milling experiment is performed in the step (1), the workpiece is clamped by a vise, and the arrangement of the acceleration sensor and the acoustic emission sensor is the same as that of the single cutting parameter milling experiment.

5. The milling tool wear detection method based on dual self-attention mechanism as claimed in claim 1, characterized in that: When extracting vibration signal samples and acoustic emission signals in step (2), firstly, the whole tool path process signal corresponding to each tool wear value is selected, including cutting force, vibration and acoustic emission signals in three directions of XY and Z.

6. The milling tool wear detection method based on dual self-attention mechanism as claimed in claim 1, characterized in that: When extracting the power signal sample in step (2): each tool wear value also corresponds to the stable signal of the last 4 seconds of the collected wear signal.

7. The milling tool wear detection method based on dual self-attention mechanism as claimed in claim 1, characterized in that: The training method of the model in step (3) comprises the following steps: ① Data folding: fold the data in the time series dimension based on the number of points collected during one rotation of the tool, and then fold the data in the channel dimension; after folding the data, the number of neurons required for data mapping within the cycle can be effectively reduced, and the complexity of the model can be effectively reduced while performing high-dimensional feature mapping; ② Data mapping within a cycle: By building a mapping layer for data within a cycle based on MLP, this mapping layer takes into account the order of the data, comprehensively considers all data points within a cycle, and maps the data in a higher dimension to avoid the impact of errors in various sensor data at different times; ③ Channel dimension expansion of data: When performing feature mapping within a cycle, since the design purpose is to perform high-dimensional mapping within the cycle with data points within the cycle, the data is also folded along the channel dimension. When performing feature extraction within the cycle, it is necessary to fuse the features of signals from different channels, so the data needs to be expanded along the channel dimension; ④ Feature extraction within the cycle: After the data is expanded along the channel dimension, it is input into the cross-attention module for feature extraction within the cycle, and the connection between different time steps within the cycle is established. In the cross-attention module, the time step dimension of the matrix Q is set to 1, thereby effectively reducing the dimension of the data; ⑤ Time series modeling between cycles: The data that has undergone feature extraction within a cycle is input into the self-attention mechanism. Since the data has undergone feature extraction within a cycle, the time step dimension of the data within a tool rotation cycle becomes 1. Then, the connection between different time steps between cycles is established through the self-attention mechanism, thereby performing time series modeling between cycles on the data. ⑥Data aggregation: The data after multiple attention mechanisms are input into the data aggregation layer based on MLP. GELU is set as the activation function between each layer to finally get the predicted value; ⑦ Calculate the parameters of each layer: The predicted value obtained above and the actual wear value are calculated through the cross entropy loss function, and then the size of each parameter in the above-mentioned layers is calculated through the back propagation algorithm.

8. The milling tool wear detection method based on dual self-attention mechanism as claimed in claim 1, characterized in that: The milling tool wear detection method also includes expanding the data along the channel dimension after the data passes through the intra-cycle data mapping layer, inputting the data into the cross-attention mechanism of the first layer in sequence according to the cycle order to perform intra-cycle feature extraction, and then performing time series modeling between cycles by calculating the self-attention between cycles.

Images