Batch online detection method and system for formwork machining

Abstract

The invention discloses a batch online detection method and a batch online detection system for mould base processing, and relates to the technical field of mould base processing detection, and the batch online detection method comprises the following steps: step 1, obtaining data such as the size of a processing part, cutting force, cutter temperature, cooling liquid flow and cutter vibration frequency, and material hardness parameters of a to-be-processed workpiece and used time of a cutter; and 2, calculating the cutter wear degree based on the cutting force and the cutter use time. The method comprises the steps of (1) carrying out data processing, (2) carrying out feature extraction and normalization processing on various types of data, (4) inputting the data and material hardness parameters into a classification model and predicting a processing state, and (5) dynamically adjusting a multi-station processing sequence through a contract network protocol according to a prediction result, a cooling liquid flow and a tool vibration frequency. And 6, the station adjustment information compensates and corrects the machining parameters in real time, and size deviation data are obtained. And step 7, outputting the final machining state of the machining part based on the dimensional deviation data and the machining state prediction result, and realizing efficient detection of mold frame machining.

Description

Technical Field

[0001] The present invention relates to the technical field of mold base processing and detection, and in particular to a batch online detection method and system for mold base processing. Background Art

[0002] In the field of mold base processing, with the continuous increase in the manufacturing industry's demand for high precision, high efficiency, and large-scale production, the importance of processing and detection technology has become increasingly prominent. However, current mold base processing and detection face many challenges and urgently need innovative solutions.

[0003] Traditional detection methods mainly rely on manual use of measuring tools to detect each processing item one by one. In this mode, the operation process is extremely cumbersome. The inspectors need to frequently hold the measuring tools for measurement, consuming a large amount of manpower and time, and the detection efficiency is extremely low. At the same time, the subjective factors introduced by manual operation have a great impact on the detection accuracy. Differences in the techniques and measurement habits of different inspectors will lead to deviations in the measurement results, making it difficult to meet the quality requirements of high-precision mold base processing.

[0004] Summary of the Invention

[0005] To solve the technical problems existing in the prior art, the present invention provides a batch online detection method and system for mold base processing.

[0006] The technical solution adopted by the present invention is as follows:

[0007] The first aspect of the present application provides a batch online detection method for mold base processing, including the following steps:

[0008] Step 1, obtain the dimensional data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, material hardness parameter of the workpiece to be processed, and the used time of the tool;

[0009] Step 2, calculate the tool wear degree data based on the cutting force data and the used time of the tool;

[0010] Step 3, perform feature extraction and normalization processing on the dimensional data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, and tool wear degree data to generate normalized processed data;

[0011] Step 4, use the normalized processed data and the material hardness parameter of the workpiece to be processed as inputs to train a classification model, and output a processing state prediction result; where the processing state prediction result includes qualified, under-processed, and over-processed;

[0012] Step 5, dynamically adjust the multi-station processing sequence based on the processing state prediction result, coolant flow rate data, and tool vibration frequency data through the contract net protocol to generate station adjustment information;

[0013] Step 6: Based on the station adjustment information, compensate and correct the processing parameters of each station in real time, and obtain the dimension deviation data after real-time compensation and correction.

[0014] Step 7: Based on the dimension deviation data after real-time compensation and correction and the processing state prediction result, output the final processing state of the processed part.

[0015] Preferably, the calculation of the tool wear degree data based on the cutting force data and the tool used time includes the following steps:

[0016] Process the cutting force data using the moving average method to calculate the average cutting force data.

[0017] Based on the average cutting force data and the tool used time, calculate the tool wear degree data according to a preset calculation formula.

[0018] Preferably, the feature extraction and normalization processing of the dimension data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, and tool wear degree data to generate the normalized processing data includes the following steps:

[0019] Apply the Canny edge detection algorithm to the dimension data to extract the contour features of the dimension data; use the moving average method for the cutting force data to calculate the trend features of the cutting force data; use the exponential smoothing method for the tool temperature data to eliminate the noise in the tool temperature data and extract the temperature change trend features; perform wavelet decomposition on the coolant flow rate data to extract the characteristic frequency components of the coolant flow rate data; perform fast Fourier transform on the tool vibration frequency data to obtain the frequency domain features of the tool vibration frequency data.

[0020] Perform linear normalization processing on the contour features of the dimension data, the trend features of the cutting force data, the temperature change trend features, the characteristic frequency components of the coolant flow rate data, the frequency domain features of the tool vibration frequency data, and the tool wear degree data respectively to generate the normalized processing data.

[0021] Preferably, based on the processing state prediction result, coolant flow rate data, and tool vibration frequency data, dynamically adjust the multi-station processing sequence through the contract net protocol, and the generation of the station adjustment information includes the following steps:

[0022] Each station sends task tender information to other stations according to the processing state prediction result, coolant flow rate data, and tool vibration frequency data. The tender information includes the processing task requirements of this station and the current processing state, coolant flow rate, and tool vibration frequency data.

[0023] After each work station receives the bidding information from other work stations, it evaluates its acceptance capacity based on its own load rate and the remaining tool life rate, and sends bidding information containing priority evaluation information to the bidding work station. The priority evaluation formula includes the load rate and the remaining tool life rate;

[0024] The bidding work station sorts the bidding work stations according to the bidding information, selects the bidding work station with the highest priority for task allocation and updates its own processing sequence;

[0025] Repeat the process of bidding, tendering and task allocation until the processing sequence adjustment of all work stations is completed, and generate work station adjustment information including the final processing sequence and task allocation of each work station.

[0026] Preferably, based on the work station adjustment information, the processing parameters of each work station are compensated and corrected in real time. The steps for obtaining the size deviation data after real-time compensation and correction are as follows:

[0027] Determine the processing priority and processing sequence of each work station according to the work station adjustment information;

[0028] Monitor the processing status of each work station in real time. If it is predicted to be under-processed and the size deviation exceeds the first preset threshold, adjust the feed speed, cutting depth and spindle speed of this work station;

[0029] After the processing parameters of each work station are compensated and corrected in real time, measure the size of the processed parts of each work station to obtain the size data after real-time compensation and correction;

[0030] Compare the size data after real-time compensation and correction with the target size, and calculate the size deviation data after real-time compensation and correction of each work station.

[0031] Preferably, based on the size deviation data after real-time compensation and correction and the processing status prediction result, the steps for outputting the final processing status of the processed parts are as follows:

[0032] For each processed part, compare its size deviation data after real-time compensation and correction with the second preset threshold;

[0033] According to the comparison result and the processing status prediction result, determine the final processing status according to the preset rules, and output the final processing status of all processed parts.

[0034] Preferably, the determination of the final processing status according to the preset rules includes the following:

[0035] If the size deviation after compensation is less than or equal to the second preset threshold, it is judged as qualified;

[0036] If the size deviation after compensation is greater than the second preset threshold, it is judged as under-processed;

[0037] If it is predicted to be over-processed in step 4, it is judged as over-processed.

[0038] The second aspect of the present application provides a batch online detection system for die set processing, including:

[0039] A data acquisition module, configured to obtain dimensional data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, material hardness parameters of the workpiece to be machined, and the used time of the tool;

[0040] A tool wear calculation module, configured to calculate tool wear degree data based on the cutting force data and the used time of the tool;

[0041] A data processing module, configured to perform feature extraction and normalization processing on the dimensional data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, and tool wear degree data to generate normalized processing data;

[0042] A state prediction module, configured to use the normalized processing data and the material hardness parameters of the workpiece to be machined as inputs, train a classification model, and output a machining state prediction result;

[0043] A station adjustment module, configured to dynamically adjust the multi-station machining sequence through the contract net protocol based on the machining state prediction result, the coolant flow rate data, and the tool vibration frequency data, and generate station adjustment information;

[0044] A parameter compensation module, configured to compensate and correct the machining parameters of each station in real time based on the station adjustment information, and obtain the dimensional deviation data after real-time compensation and correction;

[0045] A result output module, configured to output the final machining state of the machining part based on the dimensional deviation data after real-time compensation and correction and the machining state prediction result.

[0046] The beneficial effects of the present invention are at least one of the following:

[0047] By synchronously collecting data such as the dimensions, cutting force, tool temperature, coolant flow rate, and tool vibration frequency of the machining part, abandoning the traditional manual one-by-one detection mode, the detection cycle is greatly shortened, and the detection efficiency can be improved.

[0048] Considering the material hardness parameters of the workpiece to be machined and the used time of the tool, calculating the tool wear degree through a preset formula, and using a classification model to combine multi-parameter inputs and outputs to obtain the machining state prediction result, it can adapt to complex working conditions such as different material characteristics and tool wear stages, and the accuracy of machining state prediction is improved. Description of the Drawings

[0049] Figure 1 It is a schematic flowchart of the method according to Embodiment 1 of the present invention; Detailed Embodiments

[0050] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0051] Embodiment 1 provides a batch online detection method for die set machining, as Figure 1 shown, including the following steps:

[0052] Step 1, obtain the dimensional data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, material hardness parameter of the workpiece to be machined, and the used time of the tool.

[0053] In the specific implementation process, for Step 1, various types of data can be obtained in the following ways: Utilize a high-precision vision sensor with an accuracy of up to ±0.01 mm, and install it at a suitable position near the processing equipment so that it can clearly capture the contour of the processing part. Through image recognition algorithms, process the collected images to accurately measure the dimensional data of the processing part, such as length, width, hole diameter, etc.

[0054] Install a high-precision force sensor at the connection between the tool and the tool holder, with a measuring range of 0 - 5000 N, which can sense the magnitude and direction of the force generated during the cutting process in real time. The force sensor converts the collected analog signal into a digital signal and transmits it to the data acquisition system for recording and subsequent analysis.

[0055] Use an infrared thermal imager to monitor the tool in real time, with a temperature resolution of up to ±0.5 °C. Align the thermal imager with the working area of the tool to quickly and accurately obtain the temperature distribution and change of the tool during the machining process, thereby obtaining the tool temperature data.

[0056] Install a flow sensor on the coolant delivery pipeline, and accurately measure the coolant flow rate through principles such as electromagnetic induction or turbine rotation. The flow sensor outputs the flow rate data in the form of a standard electrical signal, which is convenient for data acquisition and processing. For example, the change in the coolant flow rate within the range of 0 - 10 L / min can be accurately measured.

[0057] Install a high-frequency vibration sensor on the tool body, with a resolution of up to 0.1 g, which can sensitively detect the vibration of the tool during the machining process. The sensor converts the vibration signal into an electrical signal, and after amplification, filtering and other processing, the tool vibration frequency data is obtained.

[0058] The hardness parameter of the workpiece can be obtained by referring to the technical data of the workpiece material, or a hardness testing device such as a Brinell hardness tester or a Rockwell hardness tester can be used to measure at non-critical parts of the workpiece to obtain accurate material hardness parameters.

[0059] Set a timer in the control system of the processing equipment. Start timing when the tool is first used, and record the cumulative usage time of the tool after each processing task is completed.

[0060] Step 2: Calculate the tool wear degree data based on the cutting force data and the tool usage time.

[0061] In a possible implementation, the calculating the tool wear degree data based on the cutting force data and the tool usage time includes the following steps:

[0062] Process the cutting force data using the moving average method to calculate the average cutting force data;

[0063] Based on the average cutting force data and the tool usage time, calculate the tool wear degree data according to a preset calculation formula.

[0064] Step 3: Perform feature extraction and normalization processing on the dimension data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, and tool wear degree data to generate normalized processing data.

[0065] In a possible implementation, the performing feature extraction and normalization processing on the dimension data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, and tool wear degree data to generate normalized processing data includes the following steps:

[0066] Apply the Canny edge detection algorithm to the dimension data to extract the contour features of the dimension data; use the moving average method for the cutting force data to calculate the trend features of the cutting force data; use the exponential smoothing method for the tool temperature data to eliminate the noise in the tool temperature data and extract the temperature change trend features; perform wavelet decomposition on the coolant flow rate data to extract the characteristic frequency components of the coolant flow rate data; perform a fast Fourier transform on the tool vibration frequency data to obtain the frequency domain features of the tool vibration frequency data;

[0067] Perform linear normalization processing on the contour features of the dimension data, the trend features of the cutting force data, the temperature change trend features, the characteristic frequency components of the coolant flow rate data, the frequency domain features of the tool vibration frequency data, and the tool wear degree data respectively to generate normalized processing data.

[0068] Step 4: Use the normalized processing data and the material hardness parameter of the workpiece to be processed as inputs to train a classification model, and output a processing state prediction result; where the processing state prediction result includes qualified, under-processed, and over-processed.

[0069] In the specific implementation process, the Support Vector Machine (SVM) is selected as the classification model. SVM has good performance in dealing with high-dimensional data and small-sample problems, can effectively avoid overfitting, and can adapt to different data distributions by selecting different kernel functions, showing strong adaptability to complex problems such as the prediction of die set processing status.

[0070] Integrate the normalized data obtained in step 3 with the material hardness parameters of the workpiece to be processed to form a complete input data set. At the same time, prepare the corresponding label data, that is, the known processing status (qualified, under-processed, over-processed). For the convenience of model processing, these labels need to be encoded. For example, "qualified" is marked as class 1, "under-processed" is marked as class 2, and "over-processed" is marked as class 3.

[0071] Divide the integrated data set into a training set and a test set according to a certain proportion. Generally, 70%-80% of the data is used as the training set for model training; the remaining 20%-30% of the data is used as the test set for evaluating the model's performance. When dividing the data set, ensure the randomness and representativeness of the data to avoid bias.

[0072] Input the training set data into the support vector machine model for training. During the training process, the model will automatically adjust its own parameters according to the input data to find an optimal classification hyperplane so that data of different classes can be accurately distinguished. To improve the model's performance, the method of cross-validation can be used to select appropriate hyperparameters, such as the penalty factor and kernel function coefficient, etc.

[0073] Use the test set data to evaluate the trained model. Evaluation metrics can include accuracy, recall rate, F1 value, etc. Accuracy represents the proportion of the number of samples correctly predicted by the model to the total number of samples; recall rate represents the proportion of the number of samples of a certain class correctly predicted by the model to the actual number of samples of that class; the F1 value is the harmonic mean of accuracy and recall rate, comprehensively considering the accuracy and integrity of the model. Through the evaluation metrics, it can be judged whether the model's performance meets the requirements. If not, the hyperparameters of the model can be adjusted or the model can be replaced and retrained.

[0074] Input the normalized data in actual production and the material hardness parameters of the workpiece to be processed into the trained and qualified model, and the model will output the processing status prediction result according to the learned pattern.

[0075] Step 5, based on the middle processing status prediction result, as well as the coolant flow data and tool vibration frequency data, dynamically adjust the multi-station processing sequence through the contract net protocol to generate station adjustment information;

[0076] In a possible implementation, based on the intermediate processing state prediction result, coolant flow data, and tool vibration frequency data, the multi-station processing sequence is dynamically adjusted through the contract net protocol, and generating the station adjustment information includes the following steps:

[0077] Each station sends task bidding information to other stations according to the processing state prediction result, coolant flow data, and tool vibration frequency data. The bidding information includes the processing task requirements of this station and the current processing state, coolant flow, and tool vibration frequency data;

[0078] After each station receives the bidding information from other stations, it evaluates the acceptance ability according to its own load rate and remaining tool life rate, and sends bidding information including priority evaluation information to the bidding station. The priority evaluation formula includes the load rate and the remaining tool life rate;

[0079] The bidding station sorts the bidding stations according to the bidding information, selects the bidding station with the highest priority for task allocation, and updates its own processing sequence;

[0080] Repeat the bidding, bidding, and task allocation processes until the processing sequence adjustment of all stations is completed, and generate station adjustment information including the final processing sequence and task allocation situation of each station.

[0081] Step 6, based on the station adjustment information, compensate and correct the processing parameters of each station in real time, and obtain the size deviation data after real-time compensation and correction.

[0082] In a possible implementation, based on the station adjustment information, compensating and correcting the processing parameters of each station in real time, and obtaining the size deviation data after real-time compensation and correction includes the following steps:

[0083] Determine the processing priority and processing sequence of each station according to the station adjustment information;

[0084] Real-time monitor the processing state of each station. If it is predicted to be under-processed and the size deviation exceeds the first preset threshold, adjust the feed speed, cutting depth, and spindle speed of this station;

[0085] After compensating and correcting the processing parameters of each station in real time, measure the size of the processing part of each station to obtain the size data after real-time compensation and correction;

[0086] Compare the size data after real-time compensation and correction with the target size, and calculate the size deviation data after real-time compensation and correction of each station.

[0087] Step 7, based on the size deviation data after real-time compensation and correction and the processing state prediction result, output the final processing state of the processing part.

[0088] In a possible implementation manner, based on the size deviation data corrected by real-time compensation and the machining state prediction result, outputting the final machining state of the machining part includes the following steps:

[0089] For each machining part, compare the size deviation data corrected by real-time compensation with a second preset threshold;

[0090] According to the comparison result and the machining state prediction result, determine the final machining state according to a preset rule, and output the final machining states of all machining parts.

[0091] In a possible implementation manner, the determining the final machining state according to a preset rule includes the following content:

[0092] If the size deviation after compensation is less than or equal to the second preset threshold, it is determined to be qualified;

[0093] If the size deviation after compensation is greater than the second preset threshold, it is determined to be under-machined;

[0094] If it is predicted to be over-machined in step 4, it is determined to be over-machined.

[0095] In summary, the batch online detection method and system for die set machining of the present invention realizes efficient online detection and real-time control of the die set machining process through multi-sensor data acquisition, data processing algorithms, state prediction models, and dynamic station scheduling strategies.

[0096] Embodiment 2 provides a batch online detection system for die set machining, including:

[0097] A data acquisition module, configured to obtain size data, cutting force data, tool temperature data, coolant flow data, tool vibration frequency data, material hardness parameters of the workpiece to be machined, and the used time of the tool;

[0098] A tool wear calculation module, configured to calculate tool wear degree data based on the cutting force data and the used time of the tool;

[0099] A data processing module, configured to perform feature extraction and normalization processing on the size data, cutting force data, tool temperature data, coolant flow data, tool vibration frequency data, and tool wear degree data to generate normalized processing data;

[0100] A state prediction module, configured to use the normalized processing data and the material hardness parameters of the workpiece to be machined as inputs, train a classification model, and output a machining state prediction result;

[0101] A station adjustment module, configured to dynamically adjust the multi-station machining sequence through the contract net protocol based on the machining state prediction result, the coolant flow data, and the tool vibration frequency data, and generate station adjustment information;

[0102] A parameter compensation module, configured to compensate and correct the processing parameters of each station in real time based on the station adjustment information, and obtain the dimension deviation data after real-time compensation and correction;

[0103] A result output module, configured to output the final processing state of the processed part based on the dimension deviation data after real-time compensation and correction and the processing state prediction result.

[0104] The above embodiments only represent the specific implementation manners of the present invention, and the descriptions thereof are relatively specific and detailed. However, it should not be construed as a limitation to the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. A batch online detection method for mold base processing, characterized in that, It includes the following steps: Step 1: Obtain the dimensional data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, material hardness parameter of the workpiece to be machined, and the used time of the tool; Step 2: Calculate the tool wear degree data based on the cutting force data and the used time of the tool; Step 3: Perform feature extraction and normalization processing on the dimensional data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, and tool wear degree data to generate normalized processing data; Step 4: Use the normalized processing data and the material hardness parameter of the workpiece to be machined as inputs to train a classification model and output the machining state prediction result; where the machining state prediction result includes qualified, under-machined, and over-machined; Step 5: Dynamically adjust the multi-station machining sequence based on the machining state prediction result, coolant flow rate data, and tool vibration frequency data through the contract net protocol to generate station adjustment information; Step 6: Based on the station adjustment information, compensate and correct the machining parameters of each station in real time to obtain the dimension deviation data after real-time compensation and correction; Step 7: Based on the dimension deviation data after real-time compensation and correction and the machining state prediction result, output the final machining state of the machining part.

2. The batch online detection method for die carrier processing according to claim 1, wherein The calculation of the tool wear degree data based on the cutting force data and the used time of the tool includes the following steps: Process the cutting force data using the moving average method to calculate the average cutting force data; Based on the average cutting force data and the used time of the tool, calculate the tool wear degree data according to a preset calculation formula.

3. A batch on-line detection method for die carrier processing according to claim 2, characterized in that, The performance of feature extraction and normalization processing on the dimensional data, cutting force data, tool temperature data, coolant flow rate data, tool vibration frequency data, and tool wear degree data to generate normalized processing data includes the following steps: Apply the Canny edge detection algorithm to the dimensional data to extract the contour features of the dimensional data; use the moving average method for the cutting force data to calculate the trend features of the cutting force data; use the exponential smoothing method for the tool temperature data to eliminate the noise in the tool temperature data and extract the temperature change trend features; perform wavelet decomposition on the coolant flow rate data to extract the characteristic frequency components of the coolant flow rate data; perform fast Fourier transform on the tool vibration frequency data to obtain the frequency domain features of the tool vibration frequency data; Perform linear normalization processing on the contour features of the dimensional data, the trend features of the cutting force data, the temperature change trend features, the characteristic frequency components of the coolant flow rate data, the frequency domain features of the tool vibration frequency data, and the tool wear degree data respectively to generate normalized processing data.

4. A batch online detection method for die set processing according to claim 1, characterized in that, The dynamic adjustment of the multi-station machining sequence based on the machining state prediction result, coolant flow rate data, and tool vibration frequency data through the contract net protocol to generate station adjustment information includes the following steps: Each station sends task tender information to other stations according to the machining state prediction result, coolant flow rate data, and tool vibration frequency data. The tender information includes the machining task requirements of this station and the current machining state, coolant flow rate, and tool vibration frequency data; After each work station receives the bidding information from other work stations, it evaluates its acceptance capacity based on its own load rate and the remaining tool life rate, and sends a bid information containing the priority evaluation information to the bidding work station. The priority evaluation formula includes the load rate and the remaining tool life rate; The bidding work station sorts the bidding work stations according to the bid information, selects the bidding work station with the highest priority for task assignment and updates its own processing sequence; Repeat the bidding, bidding and task assignment processes until the processing sequences of all work stations are adjusted, and generate work station adjustment information including the final processing sequences and task assignment situations of each work station.

5. A batch on-line detection method for die carrier processing according to claim 1, characterized in that, Based on the work station adjustment information, compensating and correcting the processing parameters of each work station in real time, and obtaining the size deviation data after real-time compensation and correction includes the following steps: Determine the processing priorities and processing sequences of each work station according to the work station adjustment information; Monitor the processing status of each work station in real time. If it is predicted to be under-processed and the size deviation exceeds the first preset threshold, adjust the feed speed, cutting depth and spindle speed of this work station; After compensating and correcting the processing parameters of each work station in real time, measure the sizes of the processed parts of each work station to obtain the size data after real-time compensation and correction; Compare the size data after real-time compensation and correction with the target size, and calculate the size deviation data after real-time compensation and correction of each work station.

6. A batch online detection method for die carrier processing according to claim 1, characterized in that Based on the size deviation data after real-time compensation and correction and the processing status prediction result, output the final processing status of the processed parts includes the following steps: For each processed part, compare its size deviation data after real-time compensation and correction with the second preset threshold; According to the comparison result and the processing status prediction result, determine the final processing status according to the preset rules, and output the final processing status of all processed parts.

7. A batch online detection method for die set processing according to claim 6, characterized in that, The determination of the final processing status according to the preset rules includes the following contents: If the size deviation after compensation is less than or equal to the second preset threshold, it is judged as qualified; If the size deviation after compensation is greater than the second preset threshold, it is judged as under-processed; If it is predicted to be over-processed in step 4, it is judged as over-processed.

8. An on-line batch detection system for die set processing, characterized in that, Including: A data acquisition module for obtaining the size data, cutting force data, tool temperature data, coolant flow data, tool vibration frequency data of the processed parts, the material hardness parameters of the workpiece to be processed and the tool usage time; A tool wear calculation module for calculating the tool wear degree data based on the cutting force data and the tool usage time; A data processing module for performing feature extraction and normalization processing on the size data, cutting force data, tool temperature data, coolant flow data, tool vibration frequency data, and tool wear degree data to generate normalized processing data; A status prediction module for using the normalized processing data and the material hardness parameters of the workpiece to be processed as inputs, training a classification model, and outputting a processing status prediction result; A work station adjustment module for dynamically adjusting the multi-work station processing sequence through the contract net protocol based on the processing status prediction result and the coolant flow data and the tool vibration frequency data, and generating work station adjustment information; A parameter compensation module for compensating and correcting the processing parameters of each work station in real time based on the work station adjustment information, and obtaining the size deviation data after real-time compensation and correction; A result output module, configured to output the final machining state of a machined part based on the size deviation data corrected by the real-time compensation and the machining state prediction result.

Images