Online detection method and system for dispersion degree of high-toughness silicon carbide ceramic slurry

Abstract

This invention relates to the field of data processing, specifically to an online detection method and system for the dispersion of high-toughness silicon carbide ceramic slurry. The method includes: acquiring surface images, nuclear magnetic resonance T2 relaxation imaging, magnetic field strength and temperature data, and slurry solid content information; evaluating the bubble presentation intensity of the ceramic slurry and assessing the magnetotemperature stability factor; calculating the initial detection emphasis coefficient of the slurry sample; extracting the T2 relaxation signal from the T2 relaxation imaging and performing curve fitting, calculating the overall average deviation, and correcting the initial detection emphasis coefficient to obtain the detection emphasis coefficient; selecting high-confidence slurry samples from the slurry sample based on the detection emphasis coefficient; calculating the slurry degradation performance; calculating the dispersion excellence score; and generating the final dispersion detection index. This invention can improve the quality controllability and product consistency of the ceramic slurry preparation process.

Description

Technical Field

[0001] This invention relates to the field of data processing, and specifically to an online detection method and system for the dispersion of high-toughness silicon carbide ceramic slurry. Background Technology

[0002] The dispersion degree of ceramic slurry is a key indicator for measuring the uniformity and stability of solid powder dispersion in a liquid medium. It directly reflects the distribution state and agglomeration degree of powder particles. For high-toughness silicon carbide ceramic slurries, achieving accurate and online detection of its dispersion degree is a core technical step to ensure the stability of the slurry preparation process, the consistency of the final product performance, and to improve production efficiency.

[0003] Currently, low-field nuclear magnetic resonance (NMR) technology is commonly used in the industry for this detection. The technical principle is that water molecules in the slurry exist in two states: free water and bound water, with bound water adsorbed on the surface of silicon carbide particles. NMR, by measuring the hydrogen proton relaxation time (T² relaxation time) of the solvent, can effectively distinguish the proportion of water molecules in these two states. Theoretically, the shorter the T² relaxation time, the higher the proportion of bound water, indicating better powder particle dispersion and better slurry dispersion.

[0004] However, this traditional method faces significant challenges in practical online detection scenarios, severely limiting detection accuracy and efficiency: the imaging quality of low-field NMR is highly dependent on the stability of the detection environment. Fluctuations in the magnetic field strength or temperature drift of the equipment, as well as the introduction of air bubbles into the slurry due to insufficient mixing, all introduce noise and artifacts, resulting in severe distortion of the acquired T2 relaxation signal, thus failing to accurately reflect the true moisture state. Summary of the Invention

[0005] This invention provides an online detection method and system for the dispersion of high-toughness silicon carbide ceramic slurry to solve existing problems.

[0006] The present invention provides an online detection method for the dispersion of high-toughness silicon carbide ceramic slurry, which adopts the following technical solution:

[0007] One embodiment of the present invention provides an online detection method for the dispersion of high-toughness silicon carbide ceramic slurry, the method comprising the following steps:

[0008] Collect surface images of at least two slurry samples, nuclear magnetic resonance T2 relaxation imaging, magnetic field strength and temperature data of the equipment during nuclear magnetic resonance detection, and information on the solid content of the slurry;

[0009] For each slurry sample, the bubble intensity of the ceramic slurry was evaluated based on the grayscale surface image, and the magnetotemperature stability factor of the slurry sample during NMR detection was evaluated based on magnetic field strength and temperature data.

[0010] The initial detection emphasis coefficient of the slurry sample is calculated by combining the bubble intensity and the magnetic temperature stability factor.

[0011] T2 relaxation signal is extracted from T2 relaxation imaging and curve fitting is performed. The overall average deviation between the signal and the fitted curve is calculated to obtain the signal distortion parameter. The initial detection emphasis coefficient is then corrected using the signal distortion parameter to obtain the detection emphasis coefficient.

[0012] Based on the detection emphasis coefficient, high-confidence slurry samples are selected from the slurry samples;

[0013] For each high-confidence slurry sample, the slurry degradation performance of the sample is calculated based on the difference between the T2 relaxation time in the T2 relaxation imaging and the historical baseline time, as well as the ratio of the slurry solid content information of the sample to the historical highest solid content.

[0014] The slurry degradation performance of all high-confidence slurry samples is aggregated, and a dispersion excellence score, which characterizes the overall dispersion level of all slurry samples, is calculated by combining their average value and dispersion.

[0015] The excellent dispersibility score is compared with the scoring benchmark of similar historical slurries to generate the final dispersibility test index.

[0016] Optionally, the bubble representation intensity of the ceramic slurry is evaluated based on the grayscale surface image, specifically including:

[0017] The surface image is converted to a grayscale image, and the surface area of ​​the slurry is extracted;

[0018] Calculate the average grayscale value of all pixels in the slurry surface area;

[0019] Calculate the absolute value of the difference between the gray value and the mean gray value of each pixel in the slurry surface region;

[0020] The average of all calculated absolute values ​​is determined as the bubble intensity.

[0021] Optionally, the magnetotemperature stability factor of the slurry sample during NMR detection is evaluated based on magnetic field strength and temperature data, specifically including:

[0022] Construct data sequences of magnetic field strength variation and temperature variation in chronological order;

[0023] Calculate the standard deviation of the magnetic field strength variation data series;

[0024] Perform linear fitting on the temperature change data sequence and obtain the slope of the fitted line;

[0025] Taking the reciprocal of the standard deviation yields the magnetic field evaluation parameters;

[0026] The temperature evaluation parameters are obtained by taking the absolute value and reciprocal of the slope of the fitted line.

[0027] The magnetic field assessment parameters and temperature assessment parameters are multiplied together and then normalized to obtain the magnetic temperature stability factor.

[0028] Optionally, the initial detection emphasis coefficient of the slurry sample is calculated by combining the bubble presentation intensity and the magnetic temperature stability factor, specifically including:

[0029] The inverse of the bubble intensity is multiplied by the magnetic temperature stability factor, and the result is normalized to obtain the initial detection emphasis coefficient.

[0030] Optionally, the overall average deviation between the signal and the fitted curve is calculated, specifically including:

[0031] Determine the effective signal range of the T2 relaxation signal;

[0032] Within the effective signal range, for each time point, the absolute value of the difference between the measured signal strength value of the T2 relaxation signal at that time point and the theoretical signal strength value of the fitted curve at the same time point is calculated to obtain the signal offset at that time point.

[0033] Calculate the arithmetic mean of the signal offset at all time points;

[0034] The calculated arithmetic mean is used as the overall average deviation.

[0035] Optionally, the initial detection emphasis coefficient is corrected using signal distortion parameters to obtain the detection emphasis coefficient, specifically including:

[0036] Calculate the correction factor based on signal distortion parameters;

[0037] Multiply the correction factor and the initial detection emphasis coefficient, and determine the result of the multiplication as the detection emphasis coefficient;

[0038] The correction factor is obtained in the following way:

[0039] Calculate the hyperbolic tangent function value of the reciprocal of the signal distortion parameter, and then add one to this function value.

[0040] Optionally, based on the difference between the T2 relaxation time in T2 relaxation imaging and the historical baseline time, and the ratio between the slurry solids content information of this sample and the historical highest solids content, the slurry degradation performance of this sample is calculated, specifically including:

[0041] Calculate the difference between the T2 relaxation time and the historical baseline time;

[0042] The initial degradation factor is calculated based on the difference.

[0043] Calculate the ratio of the slurry solids content information to the historical highest solids content;

[0044] The product of the initial degradation factor and the ratio is determined as the slurry degradation performance.

[0045] The initial degradation factor is the negative power of the difference from the natural base.

[0046] Optionally, the slurry degradation performance of all high-confidence slurry samples is aggregated, and a dispersion excellence score, characterizing the overall dispersion level of all slurry samples, is calculated by combining their average value and dispersion. Specifically, this includes:

[0047] Calculate the arithmetic mean of the slurry degradation performance for all high-confidence slurry samples;

[0048] Calculate the standard deviation of the slurry degradation performance for all high-confidence slurry samples;

[0049] Multiply the reciprocal of the arithmetic mean by the reciprocal of the standard deviation, and normalize the result to obtain the dispersion score.

[0050] Optionally, the excellent dispersibility score is compared with the scoring benchmark of similar historical slurries to generate the final dispersibility test index, which specifically includes:

[0051] Calculate the difference between the excellent dispersion score and the historical benchmark score for similar slurries;

[0052] The difference is normalized to obtain the final dispersion detection index.

[0053] This invention proposes an online detection system for the dispersion of high-toughness silicon carbide ceramic slurry, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the online detection method for the dispersion of high-toughness silicon carbide ceramic slurry as described above.

[0054] The beneficial effects of the technical solution of the present invention are:

[0055] In this embodiment of the invention, an invalid data affected by environmental noise and sample defects is effectively eliminated through a dual-weighted screening mechanism based on operating conditions and signals, thereby improving the signal-to-noise ratio of the NMR signal analysis. Furthermore, a multi-sample spatial distribution aggregation evaluation mechanism overcomes the deficiency of single-point sampling in terms of overall representativeness. Thus, higher precision and reliability of online detection of ceramic slurry dispersion are achieved under complex operating conditions. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 A flowchart illustrating an online detection method for the dispersion of high-toughness silicon carbide ceramic slurry according to an embodiment of the present invention;

[0058] Figure 2 This is a structural diagram of an online detection system for the dispersion of high-toughness silicon carbide ceramic slurry, provided in one embodiment of the present invention. Detailed Implementation

[0059] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an online detection method for the dispersion of high-toughness silicon carbide ceramic slurry proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0061] The following description, in conjunction with the accompanying drawings, details the specific scheme of the online detection method for the dispersion of high-toughness silicon carbide ceramic slurry provided by the present invention.

[0062] This invention provides an online detection method and system for the dispersion of high-toughness silicon carbide ceramic slurry. Please refer to [link / reference]. Figure 1 The diagram illustrates a flowchart of an online detection method for the dispersion of high-toughness silicon carbide ceramic slurry according to an embodiment of the present invention. The method includes the following steps:

[0063] S101. Collect surface images of at least two slurry samples, nuclear magnetic resonance T2 relaxation imaging, magnetic field strength and temperature data of the equipment during nuclear magnetic resonance detection, and slurry solid content information.

[0064] For example, high-toughness silicon carbide ceramic slurry is a semi-solid suspension system prepared by mixing silicon carbide ceramic raw material powder, liquid medium (such as water or organic solvent), and necessary dispersing agents. In this system, the uniformity of ceramic powder dispersion is a key precursor indicator that determines the success or failure of subsequent molding and sintering processes and the performance of the final product, and is usually quantitatively characterized by dispersion.

[0065] The dispersion of a slurry directly reflects the distribution and agglomeration degree of powder particles in the liquid phase. Ideal dispersion means that there is no severe agglomeration between powder particles and that the spatial distribution is uniform; conversely, poor dispersion indicates the presence of agglomerates or uneven distribution. Poor dispersion will directly lead to a series of process and product defects, such as:

[0066] Molding stage: Uneven rheological properties of the slurry affect the consistency of the grouting or coating thickness.

[0067] Sintering stage: Local density differences may cause cracking, deformation or residual stress.

[0068] Final product: The non-uniform microstructure leads to discrete mechanical properties (such as toughness and strength), resulting in decreased product reliability.

[0069] Currently, the industry generally uses low-field nuclear magnetic resonance technology for online detection of ceramic slurry dispersion, which is usually achieved with the help of a nuclear magnetic resonance particle surface characteristic analyzer.

[0070] The underlying principle is that the solvent water molecules in the slurry system are not completely homogeneous, but coexist in two states: free water and bound water. Bound water's molecular movement is significantly restricted due to physical or chemical adsorption on the surface of silicon carbide particles within the slurry. This difference in state manifests in nuclear magnetic resonance phenomena as different transverse relaxation times (T2 relaxation times)—the T2 relaxation time of bound water is much shorter than that of free water.

[0071] Therefore, by obtaining the T2 relaxation NMR spectrum or imaging of the slurry, the relative ratio of free water to bound water can be determined. Theoretically, a higher proportion of bound water indicates a tighter and more effective bond between solvent molecules and ceramic particles, indirectly reflecting a better dispersion state and lower degree of agglomeration of the powder particles. Based on this, the T2 relaxation time (or the proportion of bound water it reflects) is used as a key indicator for evaluating the dispersion of ceramic slurries.

[0072] However, the accuracy of the aforementioned traditional detection methods is significantly limited in practical industrial scenarios, with bottlenecks mainly in two aspects: first, the quality of NMR signals is easily affected by the detection environment and defects in the sample itself; second, single-point sampling analysis is difficult to characterize the overall spatial homogeneity of the slurry system. These factors lead to large fluctuations and insufficient reliability in the results of directly assessing dispersion based on T2 relaxation time.

[0073] To achieve a refined online evaluation of the dispersion of high-toughness silicon carbide ceramic slurries, it is first necessary to construct a multi-source heterogeneous dataset containing slurry apparent state, internal NMR information, detection conditions, and composition information. The specific implementation process of this step is as follows:

[0074] Multiple sample preparation: To overcome the limitations of single-point sampling and assess the spatial homogeneity of the slurry, N representative independent slurry samples (N≥2, usually set according to the total amount of slurry and homogeneity requirements) are sequentially obtained from the same batch of ceramic slurry to be tested using a repeated-replacement random sampling method. This method aims to ensure that the samples are comparable and cover the state of different locations within the slurry.

[0075] Multi-source data synchronous acquisition: Using an integrated data acquisition system, the following key data are collected synchronously or sequentially from each slurry sample:

[0076] Slurry surface image data: The surface of each slurry sample was captured by RGB imaging under standard lighting conditions using a high-resolution industrial camera to capture its macroscopic appearance and provide visual basis for subsequent evaluation of mixing uniformity and bubble interference.

[0077] Raw NMR data: Each slurry sample is placed in the detection area of ​​a low-field NMR analyzer (such as an NMR particle surface property analyzer). The equipment's NMR operation monitoring module reads and records the time-series data of the magnetic field strength and temperature of the sample throughout the entire detection cycle in real time, quantifying the stability of the detection environment. Simultaneously, the equipment's NMR imaging module acquires and generates T2 relaxation-weighted NMR images or relaxation spectra reflecting the molecular relaxation characteristics within the sample.

[0078] Slurry composition data: The slurry composition recording module connected to the production control system automatically reads or manually inputs the slurry solid content information corresponding to each sample. This information is a key parameter for correcting the baseline effect of solid content differences on T2 relaxation time.

[0079] Data preprocessing and integration: The raw data collected above undergoes preliminary cleaning and standardization preprocessing, such as unifying image format and resolution, aligning timestamps of each sensor, and removing obvious abnormal collection points. Subsequently, the preprocessed multi-dimensional data (images, magnetic fields, temperature, T2 imaging, solid content) are correlated and encapsulated on a sample-by-sample basis to form structured data records.

[0080] Data Upload and Storage: Upload the processed structured dataset to the database or designated storage area of ​​the central data acquisition and analysis system to provide complete and reliable input for subsequent intelligent analysis processes.

[0081] S102. For each slurry sample, evaluate the bubble intensity of the ceramic slurry based on the grayscale surface image, and evaluate the magnetotemperature stability factor of the slurry sample during NMR detection based on magnetic field strength and temperature data.

[0082] In this embodiment, the bubble intensity of the ceramic slurry is evaluated based on the grayscale surface image, specifically including:

[0083] The surface image is converted to a grayscale image, and the surface area of ​​the slurry is extracted;

[0084] Calculate the average grayscale value of all pixels in the slurry surface area;

[0085] Calculate the absolute value of the difference between the gray value and the mean gray value of each pixel in the slurry surface region;

[0086] The average of all calculated absolute values ​​is determined as the bubble intensity.

[0087] The magnetotemperature stability factor of slurry samples during NMR detection was evaluated based on magnetic field strength and temperature data, specifically including:

[0088] Construct data sequences of magnetic field strength variation and temperature variation in chronological order;

[0089] Calculate the standard deviation of the magnetic field strength variation data series;

[0090] Perform linear fitting on the temperature change data sequence and obtain the slope of the fitted line;

[0091] Taking the reciprocal of the standard deviation yields the magnetic field evaluation parameters;

[0092] The temperature evaluation parameters are obtained by taking the absolute value and reciprocal of the slope of the fitted line.

[0093] The magnetic field assessment parameters and temperature assessment parameters are multiplied together and then normalized to obtain the magnetic temperature stability factor.

[0094] For example, this step aims to address the problem of NMR signals being easily distorted by sample defects and detection environment interference in traditional methods. By quantitatively evaluating these two key influencing factors, an initial confidence weight (i.e., detection emphasis coefficient) is assigned to each sample as the basis for subsequent data screening.

[0095] Low-field nuclear magnetic resonance (NMR) technology indirectly assesses dispersion by detecting the hydrogen proton relaxation behavior of the solvent (mainly water) in a slurry. Its measurement accuracy is extremely sensitive to the following factors:

[0096] If the process control is not good during slurry mixing or transfer, air bubbles may be introduced. The magnetic susceptibility of air bubbles differs greatly from that of the slurry. This not only occupies the effective detection volume and causes a decrease in NMR signal intensity, but also seriously interferes with the accurate measurement of relaxation time due to local magnetic field distortion, directly reducing the image signal-to-noise ratio and the reliability of quantitative analysis.

[0097] The uniformity and stability of the magnetic field in a low-field NMR spectrometer, as well as the temperature rise control during equipment operation, are the physical basis for ensuring the repeatability and accuracy of relaxation time measurements. Magnetic field fluctuations or temperature drift directly cause changes in the resonance frequency and relaxation rate of hydrogen protons, introducing systematic errors and reducing the comparability of data collected at different times or under different conditions.

[0098] Therefore, in order to improve the robustness of the final dispersion assessment, each slurry sample must be pre-assessed from two dimensions: "degree of bubble interference" and "stability of testing conditions".

[0099] For the assessment of bubble interference, the internal homogeneity and potential bubble content of the slurry sample are indirectly reflected by analyzing its static surface appearance. The basic principle is that a well-mixed slurry with few internal defects usually has a smooth and uniform static surface; while a poorly mixed slurry or one with more bubbles will show obvious undulations, wrinkles, or local differences in reflectivity on its surface.

[0100] The specific implementation steps are as follows:

[0101] After sample preparation and before NMR detection, static or quasi-static surface images of the liquid surface of each slurry sample are captured using a fixed high-resolution industrial camera or high-speed camera unit.

[0102] The acquired RGB images were converted to grayscale images to simplify analysis. Subsequently, image segmentation algorithms (such as threshold-based segmentation or lightweight semantic segmentation models) were applied to accurately extract the set of pixels belonging to the "slurry surface region" in the image, excluding interference from container edges, background, and other interfering parts.

[0103] Suppose that the extracted slurry surface area has a total of M valid pixels, where the grayscale value of the i-th pixel is... First, calculate the average gray value of all pixels in the region. :

[0104] ;

[0105] Then, calculate the absolute deviation between the grayscale value of each pixel and the average grayscale value. The bubble rendering intensity is defined as the arithmetic mean of the absolute deviations of all pixels, and the calculation formula is as follows:

[0106] ;

[0107] In the formula, the average gray value This represents the overall brightness level of the slurry surface area. The bubble intensity is calculated as the average absolute deviation of all pixel grayscale values ​​relative to this average. For an ideal slurry sample that is well-mixed and has a smooth surface, the reflective properties across its surface should be consistent, reflected in the grayscale image as the grayscale values ​​of each pixel. They are close to each other, and close to the average. The deviation is very small, so the calculated bubbles appear to have low intensity. Conversely, if the mixing is insufficient or there are many bubbles inside, the surface of the slurry will show local bumps, ripples, or specular reflections, leading to increased differences in brightness between different areas in the image, i.e., pixel grayscale values. Increased dispersion, compared to the average value The absolute deviation increases significantly, thus increasing the intensity of the bubbles.

[0108] Therefore, the bubble intensity can serve as an objective indicator for quantifying the surface uniformity of the slurry and indirectly assessing the degree of internal bubble interference and mixing quality. The lower the bubble intensity, the more uniform the sample appears, the less likely it is to be affected by defects such as bubbles, and the higher its initial confidence level should be in subsequent NMR analysis.

[0109] The stability of the magnetic field of the equipment is a key prerequisite for ensuring the accuracy of nuclear magnetic resonance (NMR) detection. During long-term use, the magnetic system (including permanent magnets or electromagnets) may experience field strength drift or decreased uniformity due to aging or interference from external environmental magnetic fields. This instability of the magnetic field will directly cause the resonance frequency of hydrogen protons in the slurry to shift, thereby interfering with the accurate measurement of T2 relaxation time and ultimately affecting the reliability of the dispersion assessment results.

[0110] Meanwhile, the heat generated during equipment operation leads to a temperature rise, and this temperature rise effect affects the tested slurry sample through heat conduction. Temperature changes alter the thermal motion rate of solvent molecules within the slurry, for example, intensifying Brownian motion. This causes a systematic deviation in the measured T2 relaxation time (usually manifested as a shortened relaxation time), resulting in distorted dispersion assessments. Therefore, it is essential to simultaneously monitor and quantitatively evaluate the magnetic field stability and temperature changes during NMR detection.

[0111] Therefore, for the entire NMR detection period of the current slurry sample, with time as the x-axis, the data sequences of magnetic field intensity changes output by the equipment and the measured temperature changes within that period were recorded, forming magnetic field change datasets and temperature change datasets. The standard deviation of the magnetic field intensity sequence was calculated. The least squares method was used to fit a straight line to the temperature sequence to obtain the slope K of its changing trend.

[0112] Magnetic temperature stability factor The calculation formula can be:

[0113] ;

[0114] in, The function representing the normalization process. Magnetic temperature preference factor. The value directly reflects the quality of the NMR detection environment. If the magnetic field fluctuates during the detection process (…), the quality of the NMR detection environment will be affected. Value) and rate of temperature change ( If all values ​​remain at a low level, then the calculated values ​​will be... A higher value indicates more ideal and stable detection conditions. Under these conditions, the T2 relaxation MRI signal obtained from the sample is least affected by environmental interference, and the dispersion assessment results derived from it have higher credibility and reliability.

[0115] Optionally, 0.01 in the formula is a minimum value to keep the denominator non-zero. In the calculation of the magnetic temperature stability factor, the denominator can be kept non-zero to make the calculation meaningful. It can be set to a minimum value of other values ​​according to actual needs, which will not be elaborated on here.

[0116] S103. Combine the bubble intensity and magnetic temperature stability factor to calculate the initial detection emphasis coefficient of the slurry sample.

[0117] In this embodiment, the inverse of the bubble intensity is multiplied by the magnetic temperature stability factor, and the result of the multiplication is normalized to obtain the initial detection emphasis coefficient.

[0118] For example, for the ceramic slurry sample to be analyzed, the lower the value of the bubble intensity f, the more uniformly the sample is stirred and the less interference from bubbles; at the same time, its magnetic temperature preference factor... A higher value indicates more stable environmental conditions during NMR detection. A lower bubble intensity f is associated with a lower magnetic temperature dominance factor. The higher the value of the NMR data, the better the quality of the data and the higher the reliability of the information. Therefore, it should be given greater weight in the subsequent dispersion analysis.

[0119] Based on the above principles, the initial detection emphasis coefficient for online dispersity detection of slurry samples is currently determined. It can be calculated in the following ways:

[0120] ;

[0121] Optionally, if the calculated bubble intensity f is zero, it indicates that the sample surface is completely uniform and there are no signs of bubbles or uneven mixing. In this case, the initial detection emphasis coefficient F can be set to 1, which means that the sample has the highest initial confidence weight.

[0122] S104. Extract the T2 relaxation signal from the T2 relaxation imaging and perform curve fitting. Calculate the overall average deviation between the signal and the fitted curve to obtain the signal distortion parameter. Then, use the signal distortion parameter to correct the initial detection emphasis coefficient to obtain the detection emphasis coefficient.

[0123] In this embodiment, calculating the overall average deviation between the signal and the fitted curve specifically includes:

[0124] Determine the effective signal range of the T2 relaxation signal;

[0125] Within the effective signal range, for each time point, the absolute value of the difference between the measured signal strength value of the T2 relaxation signal at that time point and the theoretical signal strength value of the fitted curve at the same time point is calculated to obtain the signal offset at that time point.

[0126] Calculate the arithmetic mean of the signal offset at all time points;

[0127] The calculated arithmetic mean is used as the overall average deviation.

[0128] The initial detection emphasis coefficient is corrected using signal distortion parameters to obtain the detection emphasis coefficient, specifically including:

[0129] Calculate the correction factor based on signal distortion parameters;

[0130] Multiply the correction factor and the initial detection emphasis coefficient, and determine the result of the multiplication as the detection emphasis coefficient;

[0131] The correction factor is obtained in the following way:

[0132] Calculate the hyperbolic tangent function value of the reciprocal of the signal distortion parameter, and then add one to this function value.

[0133] For example, this step, based on the evaluation of the sample appearance and the detection environment, further supplements the analysis by combining the quality of the T2 relaxation imaging signal itself. When the slurry contains air bubbles or the magnetic temperature conditions are unstable during detection, it will directly lead to a decrease in the signal-to-noise ratio of the T2 relaxation signal, causing the actual acquired signal waveform to deviate from the ideal shape of smooth attenuation and exhibit irregular fluctuations.

[0134] The specific implementation method is as follows: T2 relaxation imaging of the ceramic slurry sample to be analyzed is acquired, the T2 relaxation signal is extracted, and a curve fitting algorithm is used to generate the corresponding fitting curve. The effective signal range of the T2 relaxation signal is defined, and the absolute value of the difference between the original signal intensity and the theoretical intensity corresponding to the fitting curve is calculated point by point within this range. Then, the arithmetic mean of the absolute values ​​corresponding to all positions is calculated, and this is used as the signal distortion parameter G of the sample, i.e., the overall average deviation.

[0135] A higher signal distortion parameter indicates a greater average deviation between the T2 relaxation signal and the fitted curve at corresponding points, reflecting a higher likelihood of noise interference or distortion and poorer data quality. Therefore, the initial detection emphasis coefficient of the sample needs to be corrected based on the signal distortion parameter to reduce the weight of low-quality data in the final analysis.

[0136] Therefore, the initial detection emphasis coefficient is corrected to obtain the detection emphasis coefficient. The calculation formula is as follows:

[0137] ;

[0138] in, This represents the hyperbolic tangent function.

[0139] S105. Based on the detection emphasis coefficient, select high-confidence slurry samples from the slurry samples.

[0140] For example, multiple slurry samples are obtained by sampling from the ceramic slurry to be tested, and the detection emphasis coefficient corresponding to each sample is calculated according to the aforementioned steps. The higher the value of this coefficient, the less the MRI imaging of the sample is affected by environmental interference and its own defects, and the higher the data reliability; conversely, it indicates that the sample data may have a large error.

[0141] Based on this, based on the detection emphasis coefficient High-confidence slurry samples for final analysis can be selected from all samples using any of the following methods or combinations thereof:

[0142] Proportional threshold screening method: All slurry samples are selected according to their detection emphasis coefficient. Sort the samples from highest to lowest and select the top 30% (e.g., 30%) to form a high-confidence slurry sample set.

[0143] Absolute quantity screening method: All slurry samples are screened according to their detection emphasis coefficient. Sort the samples from highest to lowest and select the top-ranked samples (e.g., 5) to form a high-confidence slurry sample set.

[0144] Numerical threshold screening method: Set a detection emphasis coefficient. Numerical threshold (e.g.) >0.7), all Samples with values ​​greater than the threshold are selected to form a high-confidence slurry sample set.

[0145] S106. For each high-confidence slurry sample, calculate the slurry degradation performance based on the difference between the T2 relaxation time in the T2 relaxation imaging and the historical baseline time, as well as the ratio between the slurry solids content information of the sample and the historical highest solids content.

[0146] In this embodiment, based on the difference between the T2 relaxation time in T2 relaxation imaging and the historical baseline time, and the ratio between the slurry solid content information of the sample and the historical highest solid content, the slurry degradation performance of the sample is calculated, specifically including:

[0147] Calculate the difference between the T2 relaxation time and the historical baseline time;

[0148] The initial degradation factor is calculated based on the difference.

[0149] Calculate the ratio of the slurry solids content information to the historical highest solids content;

[0150] The product of the initial degradation factor and the ratio is determined as the slurry degradation performance.

[0151] The initial degradation factor is the negative power of the difference from the natural base.

[0152] For example, when the slurry dispersion is good, the surface of silicon carbide particles can effectively adsorb water molecules, resulting in a significant increase in the proportion of bound water in the slurry. Since the transverse relaxation rate of bound water is much faster than that of free water, its corresponding T2 relaxation time is shorter. Therefore, the ceramic slurry with higher dispersion has a shorter overall T2 relaxation time, which manifests as a faster decay rate in the nuclear magnetic resonance signal.

[0153] Therefore, a high-confidence slurry sample was selected, and the T2 relaxation time of the signal was extracted from its T2 relaxation imaging. Simultaneously, the average T2 relaxation time of similar qualified slurry samples in history was obtained. Based on this, the slurry degradation factor of the current sample is calculated. The calculation formula can be:

[0154] ;

[0155] In this computational model, the T2 relaxation time of the current high-confidence slurry sample Compared with historical benchmarks The difference It has a clear physical meaning.

[0156] like If the relaxation time of the current sample is longer than the historical normal level, then the difference is... The value is negative. This directly indicates that a higher proportion of water molecules in the slurry are in a relatively free state, and the binding effect between water molecules and the surface of silicon carbide particles is weak. Since good dispersibility is manifested by the effective adsorption of water molecules on the particle surface, this situation initially suggests that the dispersibility of this sample may be poor.

[0157] Conversely, if If the relaxation time is shorter than the historical benchmark, the difference is positive, which initially suggests that the dispersion may be relatively good.

[0158] As a solid-liquid mixture, the solid content R of ceramic slurry determines the total volume concentration of solid particles. When the solid content is high, the average distance between particles decreases, and the interactions between particles and between particles and solvent are enhanced. This change in the physical environment itself accelerates the transverse relaxation of hydrogen protons, resulting in a shorter T2 relaxation time.

[0159] Therefore, a short T2 relaxation time alone cannot distinguish whether it stems from good dispersibility (highly dispersed particles, maximized surface area) or is simply a physical effect caused by high solids content. Without this distinction, the relaxation shortening caused by high solids content might be misjudged as excellent dispersibility, thus introducing significant evaluation errors.

[0160] To correct for the baseline effect of solids content on relaxation time, it is necessary to obtain the solids content of the current high-confidence slurry sample. And compared with the maximum allowable solids content in the production of similar slurries in history. Compare the solid content ratios. Compared with the aforementioned slurry deterioration factors Combined, the slurry degradation performance was calculated. :

[0161] ;

[0162] solid content ratio This serves as a correction factor in the calculation. The larger the value, the closer the solids content of the current slurry is to or reaches the highest level allowed in history, and the stronger the T2 relaxation shortening effect contributed by the high solids content itself.

[0163] In this case, if the uncorrected slurry degradation factor is used directly... The evaluation (based solely on the relaxation time difference) may overestimate the actual dispersion performance of the slurry due to the "interference" of high solids content (i.e., a sample with a short relaxation time simply because of high solids content may be misjudged as having good dispersion).

[0164] Therefore, As a multiplier, it acts on .when When the solid content is >1 (i.e., the solid content is higher than historically common levels), it will be magnified proportionally. The value is thus used in the final evaluation index of slurry deterioration performance. In this way, the overly optimistic misjudgments that may be caused by interference from solid content are offset, making the evaluation results more focused on reflecting the true dispersion state.

[0165] S107. Aggregate the slurry degradation performance of all high-confidence slurry samples, and calculate the dispersion excellence score, which characterizes the overall dispersion level of all slurry samples, by combining their average value and dispersion.

[0166] In this embodiment, the slurry degradation performance of all high-confidence slurry samples is aggregated, and a dispersion excellence score, which characterizes the overall dispersion level of all slurry samples, is calculated by combining their average value and dispersion. Specifically, this includes:

[0167] Calculate the arithmetic mean of the slurry degradation performance for all high-confidence slurry samples;

[0168] Calculate the standard deviation of the slurry degradation performance for all high-confidence slurry samples;

[0169] Multiply the reciprocal of the arithmetic mean by the reciprocal of the standard deviation, and normalize the result to obtain the dispersion score.

[0170] For example, good dispersibility requires not only that the particles and water molecules be tightly bound (i.e., "closely spaced"), but also that this tight binding is evenly distributed throughout the slurry system. The aforementioned slurry degradation performance IV primarily assesses the bonding state between particles and the medium at a single sampling point, but excellent performance at a single or a few points does not represent the whole.

[0171] Therefore, if multiple high-confidence samples taken from the same batch of slurry show significant differences in their calculated slurry degradation performance index (IV) values ​​(i.e., high dispersion), it indicates that the dispersion state in different regions of the slurry is uneven, with aggregation or fluctuations in distribution. This lack of overall uniformity is also an important manifestation of poor dispersion.

[0172] To comprehensively evaluate the overall dispersion level of the slurry, both the average binding state and spatial distribution uniformity need to be examined. Based on this, the following polymerization analysis was performed on all selected high-confidence slurry samples:

[0173] Calculate the slurry degradation performance of all high-confidence slurry samples. The arithmetic mean, denoted as This value reflects the overall average degree of dispersion degradation of the slurry; The smaller the value, the better the average binding state. Calculate the standard deviation of the slurry degradation performance IV for all high-confidence slurry samples, denoted as . This value reflects the degree of difference in dispersion among the sampling points; The smaller the value, the better the spatial uniformity of the slurry.

[0174] Combining the above two dimensions, the dispersibility excellence score of the currently tested ceramic slurry is calculated. :

[0175] ;

[0176] In the formula, the excellent dispersion score The composition logic is based on the following principles:

[0177] Average state dimension : The smaller the value, the lower the average slurry degradation performance of all high-confidence samples, meaning that the overall particle-media bonding state of the slurry is generally better.

[0178] Uniformity dimension : The smaller the value, the higher the degree of slurry degradation in each high-confidence sample. The closer the values ​​are to each other, the smaller the difference in the dispersion state of the slurry at different spatial locations, and the more uniform the spatial distribution.

[0179] middle The smaller the value, the larger its reciprocal. The greater the positive contribution. middle The smaller the value, the larger its reciprocal. The greater the positive contribution.

[0180] normalization function Used to map the product to a standardized scoring range (such as 0-1 or 0-100) for easy comparison and threshold judgment.

[0181] Therefore, the slurry satisfies "good average bonding state" if and only if it simultaneously meets the following conditions. Small) and "uniform spatial distribution" When both conditions are met (small), the calculated dispersion score is excellent. Only then can a higher value be achieved, thus comprehensively judging the current ceramic slurry's dispersion performance as excellent. Any deficiency in any aspect (poor average condition or poor uniformity) will result in a lower score. reduce.

[0182] S108. Compare the excellent dispersibility score with the historical scoring benchmarks of similar slurries to generate the final dispersibility test index.

[0183] In this embodiment, the excellent dispersibility score is compared with the scoring benchmark of similar historical slurries to generate the final dispersibility detection index, which specifically includes:

[0184] Calculate the difference between the excellent dispersion score and the historical benchmark score for similar slurries;

[0185] The difference is normalized to obtain the final dispersion detection index.

[0186] For example, an excellent dispersibility score is obtained for the current ceramic slurry to be tested. Then, it needs to be calibrated within the context of historical production data to generate the final detection indicators that are directly comparable and have guiding significance.

[0187] Therefore, samples of high-toughness silicon carbide ceramic slurries of the same type that passed offline verification were retrieved from the historical database. Dispersion excellence scores calculated using the same method (low-field NMR online detection) were obtained for these samples in the past, and the arithmetic mean of these historical scores was calculated and recorded as the historical baseline score. .

[0188] The current slurry score Compared to this historical benchmark score Compare them and calculate their relative differences:

[0189] ;

[0190] Subsequently, regarding this difference Normalization is performed. The normalization function can be set according to the actual data distribution and process requirements, such as using Z-score standardization or mapping to a specific range (such as the 0-1 range). The final normalized value is defined as the dispersion detection index for high-toughness silicon carbide ceramic slurry currently being tested online.

[0191] This indicator quantifies the degree of deviation of the current slurry dispersion from the historical acceptable level. A positive and larger indicator value indicates that the dispersion is better than the historical normal level; an indicator value close to zero indicates that it is within the normal fluctuation range; a negative and smaller indicator value indicates that the dispersion may not meet the standard and process intervention is required.

[0192] After obtaining the dispersion detection index, it needs to be converted into specific production guidance based on preset process quality thresholds. In this embodiment, two key thresholds are set: 0.55 and 0.85, dividing the detection index into three intervals, each corresponding to different slurry states and process adjustment strategies.

[0193] Excellent range (index > 0.85): If the real-time dispersion index of the current ceramic slurry is higher than 0.85, its dispersion performance is considered excellent and meets the process requirements. At this time, no additional intervention is required, and the slurry can flow into the subsequent processes.

[0194] Optimization Range (0.55 ≤ Index ≤ 0.85): If the measured index falls within this range, the slurry dispersion is not optimal and there is room for optimization. In this case, the system should automatically or prompt the operator to add a predetermined amount of dispersant to the slurry and perform moderate stirring to improve the powder dispersion.

[0195] Unacceptable Range (Index < 0.55): If the real-time measured index is below 0.55, the slurry dispersion is deemed unacceptable. This indicates that supplementing with the standard dose of dispersant may be insufficient to resolve the issue. In this case, more intensive measures are required, such as adding a significantly larger quantity of dispersant and / or extending the ball milling time to forcefully break down any potential agglomerates and promote uniform dispersion. Resampling and testing are necessary after treatment until the index falls within the acceptable range.

[0196] Optionally, the key thresholds here can be set and modified according to actual needs. There are no specific numerical restrictions here. The two key thresholds mentioned above are only the actual thresholds in the preferred embodiment.

[0197] In summary, this invention employs a dual-weighted correction mechanism to automatically screen high-confidence data from multiple samples by integrating slurry appearance imaging, NMR monitoring, and T2 relaxation signal multi-dimensional quality assessment. Furthermore, through degradation analysis correcting for the influence of solid content and statistical evaluation of the uniformity of polymerization spatial distribution, high-precision and robust online detection of the dispersion of high-toughness silicon carbide ceramic slurries is achieved. This method effectively overcomes the inherent limitations of traditional single-sample NMR detection, which is susceptible to environmental interference and cannot assess overall uniformity. Through preset quantitative indicators and grading thresholds, a closed-loop control system from real-time diagnosis to process adjustment is formed, significantly improving the quality controllability and product consistency of the ceramic slurry preparation process.

[0198] This invention also proposes an online detection system for the dispersion of high-toughness silicon carbide ceramic slurry. Please refer to [link / reference]. Figure 2 The diagram shows a structural diagram of an online detection system for the dispersion of high-toughness silicon carbide ceramic slurry provided in an embodiment of the present invention. The system includes: a data acquisition module 101, a data processing module 102, and a slurry detection module 103.

[0199] The data acquisition module 101 is used to acquire surface images of at least two slurry samples, nuclear magnetic resonance T2 relaxation imaging, magnetic field strength and temperature data of the equipment during nuclear magnetic detection, and solid content information of the slurry.

[0200] The data processing module 102 is used to evaluate the bubble presentation intensity of the ceramic slurry based on the grayscale surface image for each slurry sample, and to evaluate the magnetotemperature stability factor of the slurry sample during NMR detection based on magnetic field strength and temperature data.

[0201] The initial detection emphasis coefficient of the slurry sample is calculated by combining the bubble intensity and the magnetic temperature stability factor.

[0202] T2 relaxation signal is extracted from T2 relaxation imaging and curve fitting is performed. The overall average deviation between the signal and the fitted curve is calculated to obtain the signal distortion parameter. The initial detection emphasis coefficient is then corrected using the signal distortion parameter to obtain the detection emphasis coefficient.

[0203] Based on the detection emphasis coefficient, high-confidence slurry samples are selected from the slurry samples;

[0204] For each high-confidence slurry sample, the slurry degradation performance of the sample is calculated based on the difference between the T2 relaxation time in the T2 relaxation imaging and the historical baseline time, as well as the ratio of the slurry solid content information of the sample to the historical highest solid content.

[0205] The slurry degradation performance of all high-confidence slurry samples is aggregated, and a dispersion excellence score, which characterizes the overall dispersion level of all slurry samples, is calculated by combining their average value and dispersion.

[0206] The slurry testing module 103 is used to compare the excellent dispersibility score with the historical scoring benchmark of similar slurries to generate the final dispersibility test index.

[0207] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the online detection system for the dispersion of high-toughness silicon carbide ceramic slurry and the online detection method for the dispersion of high-toughness silicon carbide ceramic slurry provided in the above embodiments belong to the same concept. The specific implementation process is detailed in the method embodiments and will not be repeated here.

[0208] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0209] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0210] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for on-line detection of dispersion of high-toughness silicon carbide ceramic slurry, characterized in that, include: Collect surface images of at least two slurry samples, nuclear magnetic resonance T2 relaxation imaging, magnetic field strength and temperature data of the equipment during nuclear magnetic resonance detection, and information on the solid content of the slurry; For each slurry sample, the bubble intensity of the ceramic slurry was evaluated based on the grayscale surface image, specifically including: The surface image is converted to a grayscale image, and the surface area of ​​the slurry is extracted; Calculate the average grayscale value of all pixels in the slurry surface area; Calculate the absolute value of the difference between the gray value and the mean gray value of each pixel in the slurry surface region; The average of all calculated absolute values ​​is determined as the bubble intensity. The magnetic temperature stability factor of the slurry sample during NMR detection was evaluated based on magnetic field strength and temperature data, specifically including: Construct data sequences of magnetic field strength variation and temperature variation in chronological order; Calculate the standard deviation of the magnetic field strength variation data series; Perform linear fitting on the temperature change data sequence and obtain the slope of the fitted line; Taking the reciprocal of the standard deviation yields the magnetic field evaluation parameters; The temperature evaluation parameters are obtained by taking the absolute value and reciprocal of the slope of the fitted line. The magnetic field assessment parameters and temperature assessment parameters are multiplied and then normalized to obtain the magnetic temperature stability factor. Combining bubble intensity and magnetic temperature stability factor, the initial detection emphasis coefficient of the slurry sample is calculated, specifically including: The inverse of the bubble intensity is multiplied by the magnetic temperature stability factor, and the result is normalized to obtain the initial detection emphasis coefficient. T2 relaxation signal is extracted from T2 relaxation imaging and curve fitting is performed. The overall average deviation between the signal and the fitted curve is calculated to obtain the signal distortion parameter. The initial detection emphasis coefficient is then corrected using the signal distortion parameter to obtain the detection emphasis coefficient. Based on the detection emphasis coefficient, high-confidence slurry samples are selected from the slurry samples; For each high-confidence slurry sample, the slurry degradation performance of the sample is calculated based on the difference between the T2 relaxation time in the T2 relaxation imaging and the historical baseline time, as well as the ratio of the slurry solid content information of the sample to the historical highest solid content. The slurry degradation performance of all high-confidence slurry samples is aggregated, and a dispersion excellence score, which characterizes the overall dispersion level of all slurry samples, is calculated by combining their average value and dispersion. The excellent dispersibility score is compared with the scoring benchmark of similar historical slurries to generate the final dispersibility test index.

2. The method according to claim 1, wherein the method is characterized by, The calculation of the overall average deviation between the signal and the fitted curve specifically includes: Determine the effective signal range of the T2 relaxation signal; Within the effective signal range, for each time point, the absolute value of the difference between the measured signal strength value of the T2 relaxation signal at that time point and the theoretical signal strength value of the fitted curve at the same time point is calculated to obtain the signal offset at that time point. Calculate the arithmetic mean of the signal offset at all time points; The calculated arithmetic mean is used as the overall average deviation.

3. The method according to claim 1, wherein the method is characterized by, The step of correcting the initial detection emphasis coefficient using signal distortion parameters to obtain the detection emphasis coefficient specifically includes: Calculate the correction factor based on signal distortion parameters; Multiply the correction factor and the initial detection emphasis coefficient, and determine the result of the multiplication as the detection emphasis coefficient; The correction factor is obtained in the following way: Calculate the hyperbolic tangent function value of the reciprocal of the signal distortion parameter, and then add one to this function value.

4. The online detection method of high-toughness silicon carbide ceramic slurry dispersity according to claim 1, characterized in that, The method of calculating the slurry degradation performance of a sample based on the difference between the T2 relaxation time in T2 relaxation imaging and the historical baseline time, and the ratio of the slurry solid content information of the sample to the historical highest solid content, specifically includes: Calculate the difference between the T2 relaxation time and the historical baseline time; The initial degradation factor is calculated based on the difference. Calculate the ratio of the slurry solids content information to the historical highest solids content; The product of the initial degradation factor and the ratio is determined as the slurry degradation performance. The initial degradation factor is the negative power of the difference from the natural base.

5. The method for online detection of dispersion of high-toughness silicon carbide ceramic slurry according to claim 1, characterized in that, The degradation performance of all high-confidence slurry samples was aggregated, and a dispersion excellence score, characterizing the overall dispersion level of all slurry samples, was calculated by combining their average value and dispersion. Specifically, this score included: Calculate the arithmetic mean of the slurry degradation performance for all high-confidence slurry samples; Calculate the standard deviation of the slurry degradation performance for all high-confidence slurry samples; Multiply the reciprocal of the arithmetic mean by the reciprocal of the standard deviation, and normalize the result to obtain the dispersion score.

6. The method for online detection of dispersion of high-toughness silicon carbide ceramic slurry according to claim 1, characterized in that, The process of comparing the excellent dispersibility score with the historical scoring benchmarks of similar slurries to generate the final dispersibility detection index specifically includes: Calculate the difference between the excellent dispersion score and the historical benchmark score for similar slurries; The difference is normalized to obtain the final dispersion detection index.

7. A high-toughness silicon carbide ceramic slurry dispersity on-line detection system comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the online detection method for the dispersion of high-toughness silicon carbide ceramic slurry as described in any one of claims 1-6.

Images