A method for controlling process parameters of ceramic production

Abstract

The present invention discloses a method for controlling process parameters of ceramic production, including pressing and molding a blank to obtain a ceramic blank, obtaining an initial firing scheme according to the ceramic blank and database cluster analysis, including a reference temperature change curve and a reference holding time of each firing stage; obtaining the actual temperature change curve of the current stage, using wavelet transform to extract the main trend to divide the actual curve and the reference curve into multiple curve segments; based on each curve segment, respectively determining the first change trend of the actual curve and the second change trend of the reference curve, calculating the trend similarity, calculating the data similarity of each temperature data on the actual curve and the reference curve, and determining the target similarity according to the trend similarity and the data similarity; if the target similarity is greater than the first preset threshold, executing the next firing stage according to the initial firing scheme. The present invention achieves the effects of improving firing accuracy, enhancing product quality stability and improving production efficiency through real-time monitoring and parameter adjustment.

Description

Technical Field

[0001] This invention relates to the field of ceramic production technology, and in particular to a method for controlling ceramic production process parameters. Background Technology

[0002] Traditional ceramic firing processes typically rely on the operator's experience and manual adjustments. Before firing, a general firing plan is developed based on past experience, and then the temperature is adjusted and controlled manually during firing, taking into account actual conditions. However, relying on manual temperature adjustments during firing leads to imprecise temperature control, easily influenced by personal experience. Fluctuations in firing temperature result in inconsistent product quality, affecting yield. Furthermore, this manual adjustment is time-consuming and error-prone, reducing production efficiency and cost-effectiveness. Summary of the Invention

[0003] To address at least one of the aforementioned technical problems, this invention provides a method for controlling ceramic production process parameters.

[0004] This invention provides a method for controlling ceramic production process parameters, the method comprising:

[0005] Ceramic blanks are prepared using a ball mill, and the prepared blanks are pressed into shapes to obtain ceramic bodies.

[0006] The characteristic parameters of the ceramic green body and the database generated based on the historical ceramic preparation process are obtained. The characteristic parameters and the database are then subjected to cluster analysis to obtain the initial firing scheme of the ceramic green body. The initial firing scheme includes the reference temperature change curve and reference holding time for each firing stage.

[0007] The actual temperature change curve and the corresponding reference temperature change curve of the ceramic at the current firing stage are obtained as the actual curve and the reference curve. Wavelet transform is used to extract the temperature change trend of the actual curve and the reference curve to divide the actual curve and the reference curve into multiple curve segments.

[0008] Determine the first trend of change for each curve segment in the actual curve and the second trend of change for each curve segment in the reference curve; calculate the trend similarity between the first and second trends, and calculate the data similarity of each temperature data on the actual curve and the reference curve; determine the target similarity based on the trend similarity and data similarity.

[0009] Determine whether the target similarity is greater than the first preset threshold. If the target similarity is greater than the first preset threshold, execute the next firing stage according to the initial firing plan.

[0010] Preferably, the method further includes:

[0011] If the target similarity is less than or equal to the first preset threshold, calculate the deviation between the target similarity and the first preset threshold, and determine whether the deviation is within a preset range;

[0012] If so, the reference temperature change curve and reference holding time for the next firing stage in the initial firing scheme are optimized, and the next firing stage is executed according to the optimized firing scheme.

[0013] If not, stop the current firing process of the ceramic blank and replace it with a new ceramic blank for firing.

[0014] Preferably, before executing the next firing stage according to the optimized firing scheme, the method further includes:

[0015] The image of the ceramic product after the current firing stage is completed is obtained, and the image is denoised, enhanced and converted to grayscale to obtain a grayscale image.

[0016] Feature extraction is performed on grayscale images, including texture features, edge features, and spectral features; the extracted features are then input into a pre-defined RNN model for defect identification, generating identification results.

[0017] The optimized firing scheme is compensated based on the identification results.

[0018] Preferably, the step of compensating the optimized firing scheme based on the identification results includes:

[0019] The defect type is determined based on the identification results, and the defect type includes cracks, bubbles, and uneven color.

[0020] Correlation analysis was performed on the defect type and the reference temperature change curve and reference holding time in the current firing stage of the initial firing scheme to screen out the target defect type with a correlation greater than the second preset threshold.

[0021] The firing temperature and holding time to be adjusted are determined based on the target defect type, and the optimized firing scheme is compensated based on the firing temperature and holding time to be adjusted.

[0022] Preferably, the pressure during the pressing and molding process is:

[0023] P=C(ρ g -ρ0) n

[0024] In the formula, P represents pressure, C represents the degree of response of the ceramic blank to pressure, and ρ g ρ0 represents the target density of the ceramic blank after pressing, ρ0 represents the density of the ceramic blank before pressing, and n represents the relationship between the density of the blank and the pressure, satisfying 2≤n≤5.

[0025] Preferably, the characteristic parameters include the material type of the ceramic green body, its actual density, and the theoretical density of the finished ceramic product.

[0026] Preferably, the method further includes:

[0027] After all firing stages are completed and the ceramic product is obtained, the quality parameters of the current ceramic product are obtained, and the initial firing scheme for the next ceramic production is determined based on the quality parameters.

[0028] The quality parameters of the finished ceramic product include dimensional variation, porosity, and tensile strength.

[0029] Preferably, determining the initial firing scheme for the next ceramic production based on the quality parameters includes:

[0030] A multi-objective optimization function is established for firing temperature, holding time and heating rate, and corresponding constraints are established based on size change, porosity and tensile strength.

[0031] The multi-objective optimization function is solved using the firefly algorithm, and the initial firing scheme for the next ceramic production is determined based on the solution results.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0033] 1) This invention utilizes the characteristic parameters of ceramic green bodies and historical databases for cluster analysis to obtain an initial firing scheme suitable for the current green body. This allows for rapid provision of a reference for the current firing scheme, making it more consistent with the actual characteristics of the green body. Wavelet transform is used to extract the trends of the actual temperature change curve and the reference curve, thereby accurately dividing the curve segments. When comparing the reference curve and the actual curve, the temperature trends of different curve segments can be compared more precisely, allowing for more accurate detection of deviations and matching of corresponding processing strategies. This improves firing accuracy and accelerates process efficiency.

[0034] 2) When comparing the reference curve and the actual curve, this invention compares the similarity of the changing trends of each curve segment and combines this with the data similarity of discrete temperature points within the curve segment. Based on trend similarity and data similarity, a target similarity is determined. This target similarity is then compared with a first preset threshold. If it is greater than the first preset threshold, it indicates that the current actual firing temperature is normal, and the next firing stage can be executed according to the initial firing plan. Conversely, if it is less than or equal to the first preset threshold, it indicates that an abnormality has occurred in the actual firing process. In this case, the deviation between the target similarity and the first preset threshold is determined, and it is judged whether the deviation is within a preset range. If it is, the reference temperature change curve and reference holding time for the next firing stage in the initial firing plan are optimized before executing the next firing stage. If not, it indicates that the current firing abnormality is irreparable, and therefore, there is no need to continue firing; another blank can be used for refiring. This invention, by monitoring the actual temperature change in real time and comparing it with the reference curve, can promptly detect deviations and make adjustments. By considering the similarity of continuous trend changes and the similarity of discrete values, it can more accurately analyze the differences between the actual curve and the reference curve. When the deviation is found to be within a controllable range, it can be remedied by optimizing the parameters in the next stage. Conversely, if the deviation is uncontrollable, there is no need to continue wasting manpower and resources by replacing the blanks with new ones and refiring them. This allows for the prediction of the quality of the finished ceramic products in advance, and timely replacement when firing problems are irreversible, thus avoiding wasting energy and greatly improving the process efficiency of ceramic processing.

[0035] 3) After obtaining the finished ceramic product, this invention acquires the quality parameters of the current finished ceramic product, establishes constraints based on the quality parameters, and establishes constraints based on the temperature, holding time, and heating rate in the firing scheme. Through optimization, the initial firing scheme before the next ceramic firing is determined. Through this self-learning feedback control and adjustment mechanism, the quality of ceramic firing is continuously improved, ultimately greatly increasing the yield of the entire ceramic firing process.

[0036] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the background art, the accompanying drawings used in the embodiments of the present invention or the background art will be described below.

[0038] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the specification, serve to illustrate the technical solutions of this disclosure.

[0039] Figure 1 A schematic flowchart of a method for controlling ceramic production process parameters provided in an embodiment of the present invention;

[0040] Figure 2 A comparative analysis chart of the actual curve and the reference curve of temperature change over time provided for embodiments of the present invention;

[0041] Figure 3 A schematic flowchart of a method for controlling ceramic production process parameters provided in another embodiment of the present invention;

[0042] Figure 4 Provided for embodiments of the present invention Figure 3 A flowchart illustrating the sub-steps of step S80;

[0043] Figure 5 Provided for embodiments of the present invention Figure 4 A flowchart illustrating the sub-steps of step S803. Detailed Implementation

[0044] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] It should be noted that ceramic production mainly includes the following stages:

[0046] 1) Material preparation: Select appropriate raw materials according to the performance requirements of the ceramic, such as clay, quartz, feldspar, etc.; mix the various raw materials evenly according to the formula ratio;

[0047] 2) Blank making: The ceramic raw materials are processed into clay and shaped into blanks through methods such as slip casting, dry pressing, and hand throwing;

[0048] 3) Drying: Remove moisture from the raw clay to prevent cracking during firing;

[0049] 4) Trimming: Trimming the dried green blank to achieve the desired shape and surface smoothness.

[0050] 5) Bisque firing: Gradually increase the temperature from room temperature, maintain the high temperature for a period of time to ensure that the green body is completely sintered, and then slowly cool it to room temperature;

[0051] 6) Glazing: This is done after bisque firing and after the ceramic body has completely cooled. Glaze is applied to the surface of the ceramic body by means of dipping, spraying, or brushing.

[0052] 7) Glazing: The glazed green body is fired at high temperature in a kiln to harden it and form a glassy surface.

[0053] Currently, quality control in various stages of ceramic production relies heavily on manual experience, such as raw material preparation, molding and drying, and temperature control during firing, leading to unsatisfactory yields. To address this, this invention provides a method for controlling ceramic production process parameters. By accurately controlling the temperature at different firing stages, it improves process efficiency and yield, significantly increasing the energy efficiency of ceramic production.

[0054] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for controlling ceramic production process parameters according to an embodiment of the present invention. Figure 1 As shown, the method includes the following steps:

[0055] S10. A ball mill is used to prepare ceramic blanks, and the prepared blanks are pressed into shape to obtain ceramic blanks.

[0056] In this embodiment, the required raw materials, such as alumina powder, binder, and other additives, are first placed into a ball mill. The raw materials are ground using steel balls or other types of grinding media in the ball mill to achieve the desired particle size distribution and uniform mixing. Then, an appropriate amount of water or other solvent is added to form a slurry with a certain degree of fluidity. The prepared slurry is dried to an appropriate moisture content to facilitate subsequent molding. Finally, the dried blank is pressed into the desired shape and size using a pressing device to form a ceramic blank.

[0057] In one embodiment, the magnitude of the pressure during compression molding is:

[0058] P=C(ρ g -ρ0) n

[0059] In the formula, P represents pressure, C represents the degree of response of the ceramic blank to pressure, and ρ g ρ0 represents the target density of the ceramic blank after pressing, ρ0 represents the density of the ceramic blank before pressing, and n represents the relationship between the density of the blank and the pressure, satisfying 2≤n≤5.

[0060] The above formula describes the relationship between pressure and ceramic green body density during the pressing process. C represents the degree of response of the ceramic green body to pressure, a constant related to material properties, while n represents the exponent of the nonlinear relationship between green body change and pressure. Selecting an appropriate pressing pressure P ensures the final product has the expected density, which is crucial for guaranteeing the performance of the ceramic material. Since C and n are parameters related to material properties, experimentally determining these parameters allows for a better understanding of material behavior and optimization of the pressing process. This helps reduce trial-and-error costs and improve production efficiency. Furthermore, accurate assessment of pressure levels can effectively reduce defects such as voids and cracks within the green body, improving its density and uniformity. Precise density control can significantly improve the quality of the final product.

[0061] S20. Obtain the characteristic parameters of the ceramic green body and the database generated based on the historical ceramic preparation process. Perform cluster analysis on the characteristic parameters and the database to obtain the initial firing scheme of the ceramic green body. The initial firing scheme includes the reference temperature change curve and reference holding time for each firing stage.

[0062] First, the firing stage is mainly divided into the following stages:

[0063] Drying: In this stage, the ceramic body is placed at room temperature to remove most of its moisture. This process can occur naturally or be accelerated by heating.

[0064] Bisque firing: After the ceramic body is completely dried, it is placed in a kiln for preliminary firing, usually called bisque firing or the first firing. The temperature is generally between 900 and 1100 degrees Celsius. The purpose of bisque firing is to harden the ceramic body, making it more solid, so that subsequent processing such as glazing can be carried out.

[0065] Glazing: After bisque firing, a layer of glaze is applied to the ceramic body. Glaze is a vitreous material that melts at high temperatures to form a smooth and waterproof surface.

[0066] Glazing and firing: The glazed ceramic body is placed in the kiln for a second firing, usually at a higher temperature, reaching 1000 to 1300 degrees Celsius or higher. During glazing and firing, the glaze melts and bonds with the ceramic surface, forming a glossy surface.

[0067] The firing temperature control profile typically includes the following in its design:

[0068] Preheating stage: from room temperature to low temperature range. The main purpose of this stage is to remove moisture from the billet and prevent cracking. The temperature rise rate should not be too fast, and a holding time can be set to ensure uniform heating.

[0069] Sintering stage: Based on the material properties, the temperature is gradually increased to the firing temperature. This stage is the critical period for the physicochemical changes of ceramic materials. It is necessary to determine the appropriate heating rate and firing temperature according to the sintering characteristics of the material.

[0070] Holding stage: After reaching the predetermined firing temperature, maintain the temperature for a certain period of time to ensure full sintering and enhance the mechanical strength and stability of the ceramic. The holding time is adjusted according to the type and thickness of the ceramic.

[0071] Cooling stage: Slowly cool down to prevent the product from cracking due to stress caused by sudden temperature changes, especially for large or thick-walled products, special attention should be paid to the cooling rate.

[0072] In this embodiment, the characteristic parameters of the ceramic green body are first obtained. Preferably, the characteristic parameters include the material type of the ceramic green body, its actual density, and the theoretical density of the finished ceramic product. The material type determines the basic chemical composition of the green body. The actual density is obtained by measuring the mass and volume of the green body. The theoretical density is the density value when the material is fully densified, and can usually be found in a materials handbook.

[0073] Next, a database based on historical ceramic preparation processes is obtained. This database requires acquiring historical data from literature, experimental records, or other sources. Data preprocessing is then performed, including data cleaning, outlier removal, and standardization of units and formats. Finally, key features related to the ceramic firing process, such as temperature profiles and holding times, are extracted, and the initial body density and the density of the finished product are recorded for different ceramic material types.

[0074] Furthermore, cluster analysis is performed on the characteristic parameters and the database to obtain the initial firing scheme for the ceramic body.

[0075] In this embodiment, any one of the clustering algorithms, such as K-means, hierarchical clustering, or DBSCAN, can be used to cluster based on features such as material type, actual density, and theoretical density. During clustering, the number of clusters is determined first, and then data points from the database are assigned to different clusters. Based on the clustering results, the corresponding production process in the database can be found, thus obtaining the initial firing scheme. Specifically, the initial firing scheme includes reference temperature change curves and reference holding times for each firing stage. The reference temperature curve is defined based on the typical temperature change pattern of each cluster. The reference holding time is determined based on the average holding time of each stage in historical data. The firing process is divided into three stages: heating, holding, and cooling, and the reference temperature and time for each stage are given.

[0076] S30. Obtain the actual temperature change curve and the corresponding reference temperature change curve of the ceramic at the current firing stage, as the actual curve and the reference curve; use wavelet transform to extract the temperature change trend of the actual curve and the reference curve, so as to divide the actual curve and the reference curve into multiple curve segments.

[0077] Specifically, the current firing stage is first determined, for example, the bisque firing stage. Then, the temperature changes during the firing process are monitored in real time using sensors, and the actual temperature data over time is recorded as a curve. Based on the initial firing scheme obtained after the aforementioned cluster analysis, a reference temperature change curve for the bisque firing stage is matched from it. For ease of description, the actual temperature change curve and the corresponding reference temperature change curve are referred to as the actual curve and the reference curve, respectively.

[0078] Furthermore, wavelet transform is used to extract the main trend of the temperature change curve. Wavelet transform can capture the local characteristics of a signal and is very suitable for processing non-stationary signals. Based on the results of the wavelet transform, the actual curve and the reference curve are divided into multiple curve segments. Each segment represents a main trend of temperature change.

[0079] See Figure 2 , Figure 2 A comparative analysis chart of a reference curve and the actual curve is provided. Figure 2 In the diagram, curve ABCD represents the reference curve for temperature change over time, while abcd represents the actual curve for temperature change over time. Wavelet transformation can be used to extract the main features of the two curves to determine their trends, thus dividing both the reference and actual curves into three segments: segments AB and ab show an upward trend; segments BC and bc show a downward trend; and segments CD and cd show an upward trend.

[0080] S40. Determine the first trend of change for each curve segment in the actual curve and the second trend of change for each curve segment in the reference curve; calculate the trend similarity between the first trend and the second trend, and calculate the data similarity of each temperature data on the actual curve and the reference curve; determine the target similarity based on the trend similarity and data similarity.

[0081] like Figure 2 As shown, both the actual curve and the reference curve are divided into three segments. For each segment of the actual curve, its temperature change trend is determined, which is the first trend. For each segment of the reference curve, its temperature change trend is determined, which is the second trend. The trends of the two curve segments can be compared using the correlation coefficient or Euclidean distance to calculate cosine similarity. If the trends of the actual curve and the reference curve significantly mismatch at a certain time period, it may indicate an anomaly. Furthermore, key data points of the two curve segments are compared, such as... Figure 2In segment AB, endpoints A and B, and the midpoint, are selected; in segment ab, endpoints a and b, and the midpoint, are selected. The mean square error (MSE) or mean absolute error (MAE) of these three temperature points is calculated. The data matching similarity is determined based on the calculation results. If the actual curve and the reference curve have a low data matching degree for a certain time period, it may indicate an anomaly. Finally, the target similarity is determined based on trend similarity and data similarity. A weighted average method can be used to determine a comprehensive target similarity. For example, different weights can be assigned to trend similarity and data similarity, and then a weighted average can be calculated as the final target similarity.

[0082] In this embodiment, trend similarity reflects the degree of similarity between the actual firing process and the reference firing scheme in terms of temperature change trends. Data similarity reflects the degree of similarity between the actual temperature data and the reference temperature data. By combining these two factors, considering both continuous temperature changes and discrete key point temperature differences, the difference between the actual firing process and the reference theoretical process can be accurately assessed, thereby determining whether the firing process meets expectations.

[0083] S50. Determine whether the target similarity is greater than the first preset threshold.

[0084] S60. If the target similarity is greater than the first preset threshold, execute the next firing stage according to the initial firing plan.

[0085] The first preset threshold is usually a number close to 1, such as 0.95. If the target similarity is greater than the first preset threshold, it means that the actual temperature change during firing is very close to the temperature change curve of the reference curve. At this time, it means that there is no particular abnormality in the firing process. The firing process of the next firing stage can be executed according to the reference temperature change curve and holding time of the next stage in the initial firing plan.

[0086] Therefore, this embodiment utilizes the characteristic parameters of the ceramic green body and historical databases for cluster analysis to obtain an initial firing scheme suitable for the current green body. This allows for rapid provision of a reference for the current firing scheme, making it more consistent with the actual characteristics of the green body. Wavelet transform is used to extract the trends of the actual temperature change curve and the reference curve, thereby accurately dividing the curve segments. When comparing the reference curve and the actual curve, the temperature trends of different curve segments can be compared more precisely, allowing for more accurate detection of deviations and matching of corresponding processing strategies. This improves firing accuracy and accelerates process efficiency.

[0087] Furthermore, when comparing the reference curve and the actual curve, this invention compares the similarity of the changing trends of each curve segment and also combines the data similarity of discrete temperature points within the curve segment. Based on trend similarity and data similarity, a target similarity is determined. This target similarity is then compared to a first preset threshold. If it exceeds the first preset threshold, it indicates that the current actual firing temperature is normal, and the next firing stage can be executed according to the initial firing plan. By monitoring the actual temperature changes in real time and comparing them with the reference curve, deviations can be detected and adjusted promptly. By considering both the similarity of continuous trend changes and the similarity of discrete values, the differences between the actual curve and the reference curve can be analyzed more accurately. This provides a reliable basis for decision-making in the actual firing process and enhances process efficiency.

[0088] The above embodiments take into account the case where the target similarity is greater than the first preset threshold. In other embodiments, if the target similarity is less than or equal to the first preset threshold, it is necessary to further consider whether to adjust the firing parameters to remedy the defects caused by abnormal temperature changes, or to directly abandon the current firing process to avoid wasting more resources.

[0089] See Figure 3 In one embodiment, the method further includes:

[0090] S70. If the target similarity is less than or equal to the first preset threshold, calculate the deviation between the target similarity and the first preset threshold, and determine whether the deviation is within a preset range.

[0091] S80. If so, optimize the reference temperature change curve and reference holding time for the next firing stage in the initial firing scheme, and execute the next firing stage according to the optimized firing scheme.

[0092] S90. If not, stop the current firing process of the ceramic blank and replace it with a new ceramic blank for firing.

[0093] Understandably, if the temperature anomaly is minor, such as a slight deviation from the set temperature but still within an acceptable range, firing can usually continue. In this case, the temperature anomaly may not significantly affect the quality of the final product. This is because, to compensate for potential defects caused by the current temperature deviation, the reference temperature change curve and reference holding time for the next firing stage will be optimized, and then the next firing stage will be executed according to the optimized firing plan.

[0094] Specifically, during optimization, the first step is to analyze whether the deviation is due to a trend in temperature changes or a problem with the temperature data itself. If the deviation is mainly caused by a mismatch in trends, it can be optimized by adjusting the rate of temperature change. For example, if the actual temperature rises too quickly, the heating rate in the next stage can be appropriately reduced; conversely, the rate of increase can be reduced if the actual temperature rises too quickly. If the deviation is mainly caused by a mismatch in temperature data, it can be optimized by fine-tuning the holding time and temperature value. For example, if the actual temperature is lower than the reference temperature, the holding time in the next stage can be increased; if it is higher than the reference temperature, the holding time can be decreased. Therefore, the specific optimization strategy can be as follows:

[0095] 1) Adjustment of heating rate:

[0096] If the actual curve's heating rate is faster than the reference curve's, the heating rate in the next stage can be slowed down.

[0097] If the actual heating rate is slower than the reference rate, the heating rate in the next stage can be accelerated.

[0098] 2) Adjustment of heat preservation time:

[0099] If the actual temperature curve is lower than expected, the heat preservation time in the next stage can be extended.

[0100] If the actual temperature curve is higher than expected, the heat preservation time in the next stage can be shortened.

[0101] 3) Temperature adjustment:

[0102] If the temperature remains low at a certain stage during the entire firing process, the maximum sintering temperature for the next stage can be appropriately increased.

[0103] If the temperature in a certain stage remains too high, the maximum sintering temperature in the next stage can be appropriately reduced.

[0104] The above optimization process is for cases where the deviation is within the preset range. However, if the temperature changes abnormally severely, such as the temperature being much higher or lower than the set value, or the temperature suddenly dropping or rising, this may have a negative impact on the structure and performance of the ceramic. In this case, to prevent wasting resources, the firing process of the current ceramic body can be stopped and a new ceramic body can be fired.

[0105] In this embodiment, when the deviation range is found to be controllable, it can be remedied by optimizing the parameters of the next stage. Conversely, if the deviation is uncontrollable, there is no need to continue wasting manpower and resources to replace the blank and refire it. In this way, the quality of the finished ceramic product can be predicted in advance, and the blank can be replaced in time when the firing problem is irreversible, thus avoiding wasting energy and greatly improving the process efficiency of ceramic processing.

[0106] See Figure 4In one embodiment, before performing the next firing stage according to the optimized firing scheme, the method further includes:

[0107] S801. Obtain the image of the ceramic product after the current firing stage is completed, and perform noise reduction, image enhancement and grayscale processing on the ceramic product image to obtain a grayscale image.

[0108] Image acquisition: Using a camera or other imaging equipment to capture images of ceramic products after firing.

[0109] Denoising: Random noise in the image is removed by filtering algorithms (such as Gaussian filtering, median filtering, etc.) to improve the accuracy of subsequent processing.

[0110] Image enhancement: Techniques such as contrast enhancement and histogram equalization are used to improve the clarity and contrast of images, facilitating subsequent feature extraction.

[0111] Grayscale conversion: Converting a color image to a grayscale image simplifies image processing while preserving sufficient detail for analysis.

[0112] Therefore, denoising and enhancement processes can significantly improve image quality, making subsequent feature extraction more accurate. Grayscale conversion simplifies the image processing workflow, reduces computational complexity, and helps improve the overall system processing speed.

[0113] S802. Extract features from the grayscale image, including texture features, edge features, and spectral features; input the extracted features into a preset RNN model for defect identification and generate identification results;

[0114] Texture features: Texture information in the image is extracted using methods such as Gray-Level Co-occurrence Matrix (GLCM) and Local Binary Pattern (LBP).

[0115] Edge features: Techniques such as Canny edge detection and Sobel operator are used to detect edges in the image.

[0116] Spectral features: Although grayscale images do not contain color information, spectral features can be indirectly reflected through the distribution of grayscale values, for example, through histogram analysis.

[0117] The extracted features are input into a pre-trained recurrent neural network (RNN) model. The RNN model's memory capability is used to capture the sequential relationships between features for defect identification, thereby determining the defect type and location. Utilizing the RNN model effectively captures the sequential dependencies between image features, improving the accuracy of defect identification.

[0118] S803. Compensate the optimized firing scheme based on the identification results.

[0119] See Figure 5 In one embodiment, the compensation of the optimized firing scheme based on the identification result includes:

[0120] S8031. Determine the defect type based on the identification result, wherein the defect type includes cracks, bubbles, and uneven color;

[0121] S8032. Perform correlation analysis on the defect type and the reference temperature change curve and reference holding time of the current firing stage in the initial firing scheme, and screen out the target defect type with a correlation greater than the second preset threshold.

[0122] S8033. Determine the firing temperature and holding time to be adjusted according to the target defect type, and compensate the optimized firing scheme according to the firing temperature and holding time to be adjusted.

[0123] In this embodiment, the correlation analysis can be performed using the Pearson correlation coefficient. By identifying the defect type and conducting correlation analysis, the relationship between the specific defect type and the firing parameters is clarified, thereby screening out defects that are highly correlated with the reference temperature change curve and the reference holding time, and thus performing targeted parameter optimization and compensation.

[0124] For example, cracks are often related to excessively high cooling rates, which can be addressed by appropriately increasing the holding time or decreasing the cooling rate. Bubbles may be related to insufficient dehumidification, which can be improved by adjusting the heating rate or increasing the holding time. Uneven color may be due to uneven temperature distribution, which can be resolved by optimizing the temperature profile. By identifying defects, the optimized parameters in the above embodiments can be compensated for, further improving the accuracy of temperature change control and holding time.

[0125] In one embodiment, the method further includes:

[0126] After all firing stages are completed and the ceramic product is obtained, the quality parameters of the current ceramic product are obtained, and the initial firing scheme for the next ceramic production is determined based on the quality parameters.

[0127] The quality parameters of the finished ceramic product include dimensional variation, porosity, and tensile strength.

[0128] In this embodiment, in order to learn and analyze the firing control process, the quality parameters of the current ceramic product are also obtained, so as to determine the initial firing scheme for the next ceramic production based on the quality parameters.

[0129] Specifically, the initial firing plan for the next ceramic production is determined based on quality parameters, including:

[0130] A multi-objective optimization function is established for firing temperature, holding time and heating rate, and corresponding constraints are established based on size change, porosity and tensile strength.

[0131] The multi-objective optimization function is solved using the firefly algorithm, and the initial firing scheme for the next ceramic production is determined based on the solution results.

[0132] In this embodiment, the parameter optimization process is as follows:

[0133] 1) Establish a multi-objective optimization function:

[0134] f = f(T, t, v)

[0135] In the formula, T, t, and v represent the firing temperature, holding time, and heating rate, respectively.

[0136] 2) Determine the parametric relationships between dimensional change ΔL, porosity ε, and tensile strength σ:

[0137] ΔL=aT+bt+c

[0138] ε=d-gT

[0139] σ=-pε-qT 2

[0140] Based on the dimensional change ΔL, porosity ε, and tensile strength σ, the following constraint conditions are recommended:

[0141] ΔL min ≤ΔL≤ΔL max

[0142] ε tar -τ≤ε≤ε tar +τ

[0143] σ min ≤σ

[0144] In the formula, a, b, c, d, g, p, and q are all constant terms; ΔL min ΔL max ε represents the minimum and maximum dimensions that are allowed to vary, respectively. tar τ represents the ideal value of porosity, and σ represents the allowable fluctuation error of porosity. min This indicates the minimum tensile strength.

[0145] Furthermore, the firefly algorithm is used to solve the multi-objective function:

[0146] Initialization steps: Set the initial firing temperature T, holding time t, and heating rate v as the initial firing scheme, and initialize the firefly positions, including the firefly population, number of processors, number of assigned tasks, light absorption coefficient, and spatial distance between two fireflies.

[0147] Position update steps: Calculate the firefly brightness using the objective function based on the initialized firefly position; update the firefly position based on the brightness, including controlling ordinary fireflies to move according to an adaptive step size, with the movement direction determined by the firefly with the highest fluorescence brightness, and controlling the optimal firefly to find the current optimal solution and move accordingly.

[0148] Iteration steps: Determine whether the number of iterations has reached the first preset threshold. If not, repeat the position update step until the number of iterations reaches the first preset threshold, and output the global optimal solution.

[0149] In summary, this embodiment, after obtaining the finished ceramic product, acquires its quality parameters and establishes constraints based on these parameters. These constraints are then established using the temperature, holding time, and heating rate from the firing plan. An optimization process is then used to determine the initial firing plan for the next ceramic firing. This self-learning feedback control mechanism continuously improves the quality of the fired ceramic, ultimately significantly increasing the overall yield of the ceramic firing process.

[0150] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0151] Those skilled in the art will readily understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. Those skilled in the art will also readily understand that the various embodiments of the present invention have different focuses, and for the sake of convenience and brevity, the same or similar parts may not be repeated in different embodiments. Therefore, parts not described or not described in detail in one embodiment can be referred to in other embodiments.

Claims

1. A method for controlling process parameters of ceramic production, characterized in that: The method comprises: A ball mill is used to prepare a ceramic blank, and the prepared blank is pressed and formed to obtain a ceramic body; Acquire characteristic parameters of the ceramic body and a database generated according to a historical ceramic preparation process, perform cluster analysis on the characteristic parameters and the database, and obtain an initial firing scheme for the ceramic body, wherein the initial firing scheme includes a reference temperature change curve and a reference holding time at each firing stage; Obtaining an actual temperature change curve of the ceramic in the current firing stage and a corresponding reference temperature change curve as an actual curve and a reference curve; extracting the temperature change trends of the actual curve and the reference curve by wavelet transform, so as to divide the actual curve and the reference curve into a plurality of curve segments; Determine a first change trend of each curve segment in the actual curve and a second change trend of each curve segment in the reference curve; calculate the trend similarity of the first change trend and the second change trend, and calculate the data similarity of each temperature data on the actual curve and the reference curve, and determine the target similarity according to the trend similarity and the data similarity; Determine whether the target similarity is greater than a first preset threshold, and if the target similarity is greater than the first preset threshold, execute the next firing stage according to the initial firing scheme; If the target similarity is less than or equal to a first preset threshold, calculating a deviation between the target similarity and the first preset threshold, and determining whether the deviation is within a preset range; If so, the reference temperature change curve and reference holding time of the next firing stage in the initial firing plan are optimized; Acquire the ceramic product image after the current firing stage is completed, perform denoising, image enhancement and grayscale processing on the ceramic product image to obtain a grayscale image; Extract features from grayscale images, including texture features, edge features, and spectral features; input the extracted features into the preset RNN model for defect recognition and generate recognition results; Determining defect types according to the recognition results, wherein the defect types include cracks, bubbles, and uneven colors; Performing correlation analysis on the defect type and the reference temperature change curve and reference holding time of the current firing stage in the initial firing scheme, and screening out the target defect type with a correlation greater than a second preset threshold value; Determine the firing temperature and holding time to be adjusted according to the target defect type, and compensate the optimized firing scheme according to the firing temperature and holding time to be adjusted; Execute the next firing stage according to the optimized firing plan; If not, the firing process of the current ceramic body is stopped and a new ceramic body is replaced for firing; After all firing stages are completed to obtain a finished ceramic product, quality parameters of the current finished ceramic product are obtained, and an initial firing plan for the next ceramic production is determined according to the quality parameters; The quality parameters of the finished ceramic product include dimensional change, porosity and tensile strength; Establish a multi-objective optimization function for firing temperature, holding time and heating rate, and establish corresponding constraints based on dimensional change, porosity and tensile strength; The multi-objective optimization function is solved according to the firefly algorithm, and the initial firing plan for the next ceramic production is determined according to the solution result.

2. The ceramic production process parameter control method according to claim 1, characterized in that: The pressure during the compression molding is: P=C(ρ g -p0) n In the formula, P represents pressure, C represents the response degree of the ceramic blank to pressure, ρ g represents the target density of the ceramic body after pressing, ρ0 represents the density of the ceramic body before pressing, and n represents the relationship between the density of the body and the pressure, satisfying 2≤n≤5.

3. The ceramic production process parameter control method according to claim 1, characterized in that: The characteristic parameters include the material type and actual density of the ceramic body and the theoretical density of the finished ceramic product.

Images