Ceramic production process parameter control method

Abstract

The invention relates to the technical field of ceramic production control. The invention relates to a ceramic production process parameter control method. The method comprises the following steps: S1, collecting historical production records of ceramics, and distributing the historical production records according to finished product records and inferior product records; s2, acquiring real-time image data of a to-be-stamped ceramic green body, acquiring process parameters of the stamping device and the to-be-stamped ceramic green body, and extracting a corresponding historical production record according to the process parameters; s3, combining the historical production record with the real-time image data to carry out stamping matching on the same image; the method realizes high efficiency, intelligence and greenization of ceramic production through a synergistic effect of five dimensions of data intelligence, quality control, efficiency optimization, cost control and technology expansion, and has the advantages that hidden process experience is converted into dominant data assets, and meanwhile, through risk pre-control and a flexible production mechanism, the production efficiency is greatly improved. And the market change coping capability of enterprises is obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of ceramic production control, in particular to a method for controlling ceramic production process parameters. Background Art

[0002] In ceramic production, precise control of process parameters is the core link to ensure product quality and improve production efficiency. Currently, production control is mainly carried out through manual experience or single-dimensional data (such as pressure, temperature and other process parameters). Its core purpose is to ensure the stability of ceramic bodies in key processes such as stamping and sintering, and reduce the rate of defective products.

[0003] At present, traditional processes rely on operators to manually adjust parameters such as stamping pressure and kiln temperature according to the appearance of the green body or equipment operation data. Although basic quality control can be achieved, there are problems such as low decision-making efficiency and poor parameter consistency, making it difficult to meet the production needs of multiple varieties and high precision. At the same time, manual parameter adjustment cannot quickly match the diverse green body characteristics, resulting in long production change time and high trial and error costs. The order delivery cycle is usually as long as several weeks. Secondly, there is a lack of real-time correlation analysis between green body surface defects and process parameters, and the defective product rate is generally as high as 5% to 8%. Therefore, a method for controlling ceramic production process parameters is proposed. Summary of the Invention

[0004] The object of the present invention is to provide a method for controlling process parameters of ceramic production to solve the problems raised in the above background technology.

[0005] To achieve the above object, a method for controlling process parameters of ceramic production is provided, comprising the following steps:

[0006] S1. Collect historical production records of ceramics and classify them into finished product records and defective product records;

[0007] S2. Acquire real-time image data of the ceramic body to be stamped, and simultaneously acquire process parameters of the stamping device and the ceramic body to be stamped, and then extract corresponding historical production records based on the process parameters;

[0008] S3, combining historical production records with real-time image data to perform stamping matching of the same image, and determining the finished product based on the matching result. If the determination result does not show the finished product, proceed to step S4;

[0009] S4, real-time body parameter identification of the green body is performed based on the real-time image data, and at the same time, a risk analysis of the parameters of the inferior green body is performed based on the historical production records. Then, a range comparison is performed between the real-time body parameters and the parameters of the inferior green body, and S5 is performed based on the comparison results;

[0010] S5. Establish a stamping simulation model based on historical production records and the stamping device, and input the blank parameters into the stamping simulation model to perform stamping predictions with different process parameters. Combine the prediction results with the process parameters to generate adjustment parameters and input them into the stamping device for control.

[0011] As a further improvement of the present technical solution, the S1 establishes a data module between the ceramic production devices, and obtains the production data of each ceramic when it passes through the production device through the data module.

[0012] As a further improvement of this technical solution, the steps of S1 are as follows:

[0013] S1.1. Summarize the relevant data of each ceramic as the historical production record of the ceramic;

[0014] S1.2. Obtain the test result of each ceramic. When the test result is a finished product, assign the historical production record corresponding to the ceramic as a finished product record. Conversely, when the test result is a defective product, assign the historical production record corresponding to the ceramic as a defective product record.

[0015] As a further improvement of this technical solution, the steps of S2 are as follows:

[0016] S2.1. Installing image capture devices at multiple angles on the stamping device so that the image capture range of the image capture devices covers the surface of the ceramic body to be stamped, capturing image data of the ceramic body to be stamped by the image capture devices, and then integrating the captured image data as real-time image data of the ceramic body to be stamped;

[0017] S2.2. Obtain process parameters of the stamping device and the ceramic body to be stamped, and then extract the same parameters from the historical production records based on the process parameters to obtain historical production records with the same process parameters in the stamping stage.

[0018] As a further improvement of this technical solution, the steps of S3 are as follows:

[0019] S3.1. Combine the historical production records obtained in S2.2 with the real-time image data to perform stamping matching on the same image;

[0020] S3.2. If the historical production record contains the same historical image data as the real-time image data, and the corresponding historical production record is a finished product record, the stamping device will not be adjusted and the stamping process will proceed normally;

[0021] S3.3. If the historical production record contains the same historical image data as the real-time image data, and the corresponding historical production record is a defective product record, proceed to step S4;

[0022] S3.4. When the historical production records do not contain historical image data identical to the real-time image data, proceed to step S4.

[0023] As a further improvement of the present technical solution, when S3 makes a finished product judgment based on the matching results, when the same historical image data has both finished product records and defective product records, the defective product causes are analyzed for the defective product records. When the defective product causes have no impact on the stamping stage, the record is adjusted to retain only the finished product records. Conversely, when the defective product causes have an impact on the stamping stage, the record is adjusted to retain only the defective product records.

[0024] As a further improvement of this technical solution, the steps of S4 are as follows:

[0025] S4.1. Real-time body parameter recognition of the body based on the real-time image;

[0026] S4.2. Obtain the finished product production standards, then conduct a risk analysis of inferior product body parameters based on the finished product production standards and historical production records, and obtain the inferior product body parameters based on the analysis results;

[0027] S4.3. Compare the real-time green body parameters with the inferior green body parameters. If the real-time green body parameters are not within the inferior green body parameter range, proceed to S5. Otherwise, if the real-time green body parameters are within the inferior green body parameter range, reposition the ceramic green body.

[0028] As a further improvement of this technical solution, the steps of performing the risk analysis of defective green body parameters in S4.2 are as follows:

[0029] FP i ={f li |f li <L i ∨f li >U i};

[0030] Among them, FP i is the set of out-of-limit inferior product data of parameter i, fl i is the specific value of the i-th parameter in the l-th defective product record, L i is the qualified lower limit of the i-th parameter, U i is the qualified upper limit of the i-th parameter, ∨ is a logical operator, which means that if one of the conditions is met, it is established;

[0031]

[0032] in, is the limit-exceeding rate of the i-th parameter, k is the total number of historical defective product records, |FP i ∣ is the number of records of substandard products with the i-th parameter failing to meet the standards;

[0033] when The inferior product interval is [min(FP i ), max(FP i )], otherwise, keep the finished product interval.

[0034] As a further improvement of this technical solution, the steps of S5 are as follows:

[0035] S5.1. Establish a stamping simulation model based on historical production records and the stamping device. Then, input the real-time blank parameters into the stamping simulation model to perform stamping predictions with different process parameters. The predicted process parameters obtained are the same as those of the finished product after stamping.

[0036] S5.2. Count the number of each predicted process parameter and the finished product record, then select the predicted process parameter with the highest number and combine it with the original process parameter set by the stamping device to generate adjustment parameters, and then input the generated adjustment parameters into the stamping device for stamping control.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. In this ceramic production process parameter control method, a cross-modal association of (image-parameter-quality) is achieved by combining real-time image data with process parameters. When the real-time image matches the historical finished product image, the historical successful parameters are directly reused to reduce the trial and error cost. Through over-limit rate analysis and inspection, the risk range of inferior product parameters is delineated. If the real-time parameters reach the risk range, an automatic alarm is issued and the position of the blank is adjusted. The inferior product rate is significantly reduced and the parameter consistency is improved.

[0039] 2. This ceramic production process parameter control method achieves efficient, intelligent and green ceramic production through the synergistic effect of five dimensions: data intelligence, quality control, efficiency optimization, cost control and technology expansion. Its advantage lies in converting implicit process experience into explicit data assets. At the same time, through risk pre-control and flexible production mechanisms, it significantly improves the company's ability to respond to market changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is the overall flow chart of the present invention;

[0041] Figure 2 A flowchart of the present invention for summarizing relevant data of each ceramic;

[0042] Figure 3 A flowchart of the present invention for obtaining historical production records with the same process parameters in the stamping stage;

[0043] Figure 4 This is a flowchart of the normal stamping process of the present invention;

[0044] Figure 5 This is a flowchart of the present invention for obtaining parameters of inferior green bodies based on analysis results;

[0045] Figure 6 This is a flowchart of the present invention for obtaining the same predicted process parameters as the predicted results after stamping of the finished product records. DETAILED DESCRIPTION

[0046] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0047] See also Figures 1-6 As shown, the purpose of this embodiment is to provide a ceramic production process parameter control method, comprising the following steps:

[0048] S1. Collect historical production records of ceramics and classify them into finished product records and defective product records;

[0049] S1 establishes a data module between the ceramic production devices and obtains the production data of each ceramic when it passes through the production device through the data module.

[0050] The OPC UA protocol is used to implement data interaction between the device layer (PLC, sensor) and the edge layer (gateway).

[0051] The steps of S1 are as follows:

[0052] S1.1. Summarize the relevant data of each ceramic as the historical production record of the ceramic;

[0053] Generate a unique ID for each ceramic and associate it with the entire process data. Obtain process parameters (pressure, temperature, etc.), green body physical parameters (thickness, roughness, etc.) and timestamps from each production equipment (stamping machine, kiln, etc.), and splice the data of each link in chronological order to form a single historical record.

[0054] S1.2. Obtain the test results of each ceramic. If the test result is a finished product, assign the corresponding historical production record of the ceramic as a finished product record. Conversely, if the test result is a defective product, assign the corresponding historical production record of the ceramic as a defective product record. The specific steps are as follows:

[0055] Obtaining test results: The ceramics are judged to be qualified through manual inspection or AI visual inspection (such as the YOLO model). The results are uploaded and then collected by the data module. Qualified ceramics are marked as finished products, and unqualified ones are marked as inferior products. The defect type (such as surface cracks and dimensional deviation) is also recorded.

[0056] Historical record classification step: associate the test results with the historical records of the corresponding ceramics. If the test result is a finished product, mark it as a finished product record. If it is a defective product, mark it as a defective product record and add the defect type.

[0057] S2. Acquire real-time image data of the ceramic body to be stamped, and simultaneously acquire process parameters of the stamping device and the ceramic body to be stamped, and then extract corresponding historical production records based on the process parameters;

[0058] The steps of S2 are as follows:

[0059] S2.1. Install image capture devices at multiple angles on the stamping device so that the image capture range of the image capture devices covers the surface of the ceramic body to be stamped. The image capture devices are used to capture image data of the ceramic body to be stamped. The captured image data is then integrated as real-time image data of the ceramic body to be stamped. The specific steps are as follows:

[0060] Determine the shooting angle and number: Use a three-eye vision solution (three cameras on the left, center, and right), with angles of 45°, 0°, and -45° respectively, to ensure that there is no blind spot coverage on the surface of the blank (top and side coverage ≥ 95%). The camera should be installed at a height of 300-500mm from the blank surface, and the focal length should be adjusted to a depth of field that covers the blank thickness range (e.g., 2-10mm).

[0061] Image acquisition execution process: When the blank arrives at the stamping station, the photoelectric sensor sends a trigger signal to the image acquisition. After receiving the signal, the three cameras shoot simultaneously to obtain multi-view images at the same time;

[0062] Image integration: Use the ORB algorithm to detect and describe feature points, generate a set of key points, match cross-view feature points using the Hamming distance, then use the RANSAC algorithm to remove mismatched points, calculate the homography matrix between views, and finally use the fade-in and fade-out weighted averaging to stitch the corrected images.

[0063] S2.2. Obtain process parameters of the stamping device and the ceramic body to be stamped, then extract the same parameters from historical production records based on the process parameters to obtain historical production records with the same process parameters during the stamping stage;

[0064] The equipment parameters are obtained through the PLC interface of the stamping device as the basis for historical record matching, and then records that are completely consistent with the current process parameters are screened out from the historical database.

[0065] S3, combining historical production records with real-time image data to perform stamping matching of the same image, and determining the finished product based on the matching result. If the determination result does not show the finished product, proceed to step S4;

[0066] The steps for S3 are as follows:

[0067] S3.1. Combine the historical production records obtained in S2.2 with the real-time image data to perform stamping matching on the same image. The formula is as follows:

[0068]

[0069] Among them, Sim is the similarity, F real is the feature vector of real-time image data, F hist Feature vectors of images recorded for historical production;

[0070] When Sim<1, it is judged as non-identical images;

[0071] When Sim=1, it is judged as the same image;

[0072] S3.2. If the historical production record contains the same historical image data as the real-time image data, and the corresponding historical production record is a finished product record, the stamping device will not be adjusted and the stamping process will proceed normally;

[0073] S3.3. If the historical production record contains the same historical image data as the real-time image data, and the corresponding historical production record is a defective product record, proceed to step S4;

[0074] S3.4. If the historical production records do not contain the same historical image data as the real-time image data, proceed to step S4;

[0075] When S3 determines the finished product based on the matching results, if the same historical image data has both finished product records and defective product records, the defective product causes are analyzed for the defective product records;

[0076] For example, if the defect of a defective product is glaze bubbles, and analysis shows that it is not related to the stamping pressure (the bubbles are generated during the glazing process), the record will be re-marked as a finished product;

[0077] The defect recorded for defective products is a bottom crack. Since the stamping pressure is 20% higher than the average of the finished products, it is determined to be caused by stamping and the defective product label is retained.

[0078] When the reason for defective products has no impact on the stamping stage, the record is adjusted to retain only the finished product record. Conversely, when the reason for defective products has an impact on the stamping stage, the record is adjusted to retain only the defective product record. The specific formula is as follows:

[0079] G=(C P >0)∧(C F >0);

[0080] Among them, G represents whether it is a mixed record group. When G=1, it represents a mixed record group. When G=0, it represents not a mixed record group. P is the number of finished product records within the group, C F is the number of defective product records in the group, ∧ is a logical operator (satisfy at the same time);

[0081]

[0082] Among them, t is the difference in parameter means between defective products and finished products, α F is the mean value of the defective product record parameter, α P is the mean value of the parameter recorded for the finished product, S is the combined standard deviation, and a critical value (0.05) is set. If t exceeds the critical value, it is considered that the parameter difference is significant and the defect may be related to the stamping stage.

[0083] S4, real-time body parameter identification of the green body is performed based on the real-time image data, and at the same time, a risk analysis of the parameters of the inferior green body is performed based on the historical production records. Then, a range comparison is performed between the real-time body parameters and the parameters of the inferior green body, and S5 is performed based on the comparison results;

[0084] The steps for S4 are as follows:

[0085] S4.1. Real-time body parameter recognition of the body based on the real-time image;

[0086] Using image processing technology, key parameters of the blank, such as thickness, size, surface flatness, etc., are extracted from the real-time image and summarized as real-time blank parameters;

[0087] S4.2. Obtain the finished product production standards, then conduct a risk analysis of inferior product body parameters based on the finished product production standards and historical production records, and obtain the inferior product body parameters based on the analysis results;

[0088] Obtain the various parameter standards required for finished product production from data sources such as production process documents and historical successful production records, including parameter value ranges, tolerances, and other information. Then, combined with the finished product production standards and historical production records, analyze the blank parameter characteristics that cause the product to be defective, identify the parameter patterns associated with defective products, and determine the value range of the defective blank parameters based on the risk analysis results. The formula is as follows:

[0089] FP i ={f li |f li <L i ∨f li >U i};

[0090] Among them, FP i is the set of out-of-limit inferior product data of the i-th parameter, f li is the specific value of the i-th parameter in the l-th defective product record, L i is the qualified lower limit of the i-th parameter, U i is the qualified upper limit of the i-th parameter, ∨ is a logical operator, which means that if one of the conditions is met, it is established;

[0091]

[0092] in, is the limit-exceeding rate of the i-th parameter, k is the total number of historical defective product records, |FP i ∣ is the number of records of substandard products with the i-th parameter failing to meet the standards;

[0093] when The inferior product interval is [min(FP i ), max(FP i )], otherwise, the finished product standard range is retained.

[0094] S4.3. Compare the real-time green body parameters with the inferior green body parameters. If the real-time green body parameters are not within the inferior green body parameter range, proceed to S5. Otherwise, if the real-time green body parameters are within the inferior green body parameter range, reposition the ceramic green body.

[0095] S5. Establish a stamping simulation model based on historical production records and the stamping device, and input the blank parameters into the stamping simulation model to perform stamping predictions with different process parameters. Combine the prediction results with the process parameters to generate adjustment parameters and input them into the stamping device for control.

[0096] The steps for S5 are as follows:

[0097] S5.1. Establish a stamping simulation model based on historical production records and the stamping device. Then, input the real-time blank parameters into the stamping simulation model to perform stamping predictions for different process parameters. The predicted process parameters obtained are the same as those of the finished product after stamping. The formula is as follows:

[0098] z=M(x,y)=σ(W2σ(W1[x;y]+b1)+b2);

[0099] Wherein, z is the predicted finished product quality vector, M is the stamping simulation model function, the input is the blank parameters and stamping parameters, the output is the prediction result, x is the real-time blank parameter vector, y is the stamping device parameter vector, [x; y] is the vector of x and y concatenated by dimension, W1 is the first-layer neural network weight matrix, W2 is the second-layer neural network weight matrix, b1 is the first-layer neural network bias vector, b2 is the second-layer neural network bias vector, and σ is the activation function used to introduce nonlinearity;

[0100] z j =M(x real ,y j );

[0101] Among them, z j is the quality of the finished product predicted using the jth set of stamping parameters yj, x real is the blank parameter vector collected in real time, y j are the parameters of the punching device to be tested in group j;

[0102] S5.2. Count the number of each predicted process parameter and finished product record, then select the predicted process parameter with the highest number and combine it with the original process parameter set for the stamping device to generate an adjustment parameter. Then, input the generated adjustment parameter into the stamping device for stamping control;

[0103] The adjustment parameter is the difference parameter between the original set process parameter and the predicted process parameter. By inputting the adjustment parameter, the original set process parameter is adjusted to be the same as the predicted process parameter.

[0104] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for controlling process parameters of ceramic production, characterized in that: The following steps are involved: S1. Collect historical production records of ceramics and classify them into finished product records and defective product records; S2. Acquire real-time image data of the ceramic body to be stamped, and simultaneously acquire process parameters of the stamping device and the ceramic body to be stamped, and then extract corresponding historical production records based on the process parameters; S3, combining historical production records with real-time image data to perform stamping matching of the same image, and determining the finished product based on the matching result. If the determination result does not show the finished product, proceed to step S4; S4, real-time body parameter identification of the green body is performed based on the real-time image data, and at the same time, a risk analysis of the parameters of the inferior green body is performed based on the historical production records. Then, a range comparison is performed between the real-time body parameters and the parameters of the inferior green body, and S5 is performed based on the comparison results; S5. Establish a stamping simulation model based on historical production records and the stamping device, and input the blank parameters into the stamping simulation model to perform stamping predictions with different process parameters. Combine the prediction results with the process parameters to generate adjustment parameters and input them into the stamping device for control.

2. The ceramic production process parameter control method according to claim 1, characterized in that: Said S1 establishes a data module between the ceramic production devices, and obtains the production data of each ceramic when passing through the production device through the data module.

3. The ceramic production process parameter control method according to claim 1, characterized in that: The steps of S1 are as follows: S1.

1. Summarize the relevant data of each ceramic as the historical production record of the ceramic; S1.

2. Obtain the test result of each ceramic. When the test result is a finished product, assign the historical production record corresponding to the ceramic as a finished product record. Conversely, when the test result is a defective product, assign the historical production record corresponding to the ceramic as a defective product record.

4. The ceramic production process parameter control method according to claim 1, characterized in that: The steps of S2 are as follows: S2.

1. Installing image capture devices at multiple angles on the stamping device so that the image capture range of the image capture devices covers the surface of the ceramic body to be stamped, capturing image data of the ceramic body to be stamped by the image capture devices, and then integrating the captured image data as real-time image data of the ceramic body to be stamped; S2.

2. Obtain process parameters of the stamping device and the ceramic body to be stamped, and then extract the same parameters from the historical production records based on the process parameters to obtain historical production records with the same process parameters in the stamping stage.

5. The ceramic production process parameter control method according to claim 1, characterized in that: The steps of S3 are as follows: S3.

1. Combine the historical production records obtained in S2.2 with the real-time image data to perform stamping matching on the same image; S3.

2. If the historical production record contains the same historical image data as the real-time image data, and the corresponding historical production record is a finished product record, the stamping device will not be adjusted and the stamping process will proceed normally; S3.

3. If the historical production record contains the same historical image data as the real-time image data, and the corresponding historical production record is a defective product record, proceed to step S4; S3.

4. When the historical production records do not contain historical image data identical to the real-time image data, proceed to step S4.

6. The ceramic production process parameter control method according to claim 1, characterized in that: When S3 makes a finished product judgment based on the matching results, if the same historical image data has both finished product records and defective product records, the defective product records are analyzed for their causes. If the defective product causes have no impact on the stamping stage, the records are adjusted to retain only the finished product records. Conversely, if the defective product causes have an impact on the stamping stage, the records are adjusted to retain only the defective product records.

7. The ceramic production process parameter control method according to claim 1, characterized in that: The steps of S4 are as follows: S4.

1. Real-time body parameter recognition of the body based on the real-time image; S4.

2. Obtain the finished product production standards, then conduct a risk analysis of inferior product body parameters based on the finished product production standards and historical production records, and obtain the inferior product body parameters based on the analysis results; S4.

3. Compare the real-time green body parameters with the inferior green body parameters. If the real-time green body parameters are not within the inferior green body parameter range, proceed to S5. Otherwise, if the real-time green body parameters are within the inferior green body parameter range, reposition the ceramic green body.

8. The ceramic production process parameter control method according to claim 7, characterized in that: The steps for performing the risk analysis of defective blank parameters in S4.2 are as follows: FP i ={f li |f li <L i ∨f li >U i }; Among them, FP i is the set of out-of-limit inferior product data of the i-th parameter, f li is the specific value of the i-th parameter in the l-th defective product record, L i is the qualified lower limit of the i-th parameter, U i is the qualified upper limit of the i-th parameter, ∨ is a logical operator, which means that if one of the conditions is met, it is established; in, is the limit-exceeding rate of the i-th parameter, k is the total number of historical defective product records, |FP i ∣ is the number of records of substandard products with the i-th parameter failing to meet the standards; when The inferior product interval is [min(FP i ), max(FP i )], otherwise, keep the finished product interval.

9. The ceramic production process parameter control method according to claim 1, characterized in that: The steps of S5 are as follows: S5.

1. Establish a stamping simulation model based on historical production records and the stamping device. Then, input the real-time blank parameters into the stamping simulation model to perform stamping predictions with different process parameters. The predicted process parameters obtained are the same as those of the finished product after stamping. S5.

2. Count the number of each predicted process parameter and the finished product record, then select the predicted process parameter with the highest number and combine it with the original process parameter set by the stamping device to generate adjustment parameters, and then input the generated adjustment parameters into the stamping device for stamping control.

Images