Multi-technology fused silver powder production quality control method and system, and storage medium

Abstract

The application relates to the technical field of production quality control, and discloses a silver powder production quality control method and system fusing multiple technologies and a storage medium. The method comprises the following steps: collecting silver powder production data through a distributed sensor to obtain a quality characteristic correlation matrix, adjusting a gamma distribution hyperparameter to obtain an adaptive parameter by using a self-adaptive Bayesian algorithm based on a defect rate parameter, generating a device cooperative control instruction set by optimizing a control strategy through a decision fusion algorithm, performing linkage control on a ball mill, an atomization drying tower and a grading screening device to obtain an execution result, evaluating the execution result to obtain a performance index, and correcting the control instruction set according to the performance index. The application solves technical problems such as performance degradation of a traditional Bayesian control under a large defect rate parameter condition in a silver powder production process, insufficient control precision caused by a lack of cooperative decision fusion of multiple quality characteristics, and an influence on overall quality control effect caused by a lack of cooperative linkage control among production devices.

Description

Technical Field

[0001] This application relates to the field of production quality control technology, and in particular to a method, system and storage medium for quality control of silver powder production that integrates multiple technologies. Background Technology

[0002] Existing quality control technologies for silver powder production primarily employ single statistical process control methods, such as traditional frequency theory control charts and basic Bayesian control methods. These methods typically rely on fixed control parameters and pre-defined statistical distribution assumptions, judging the production status by monitoring single quality characteristics such as particle size distribution and conductivity during silver powder production. When dealing with complex quality variations in silver powder production, traditional control methods mainly employ independent equipment control strategies, lacking effective coordination between various production equipment such as ball mills, atomizing drying towers, and grading and screening equipment.

[0003] However, when the defect rate parameter is large during silver powder production, traditional Bayesian control charts suffer from a sharp degradation in control performance due to improper selection of hyperparameters for the gamma prior distribution, failing to accurately identify abnormal quality states. Secondly, single-technology control methods cannot effectively handle the complex relationships between multiple quality characteristics in silver powder production, resulting in a lack of comprehensiveness and accuracy in control decisions. Furthermore, the independent control modes among various production equipment lack coordination; when the parameters of one piece of equipment are adjusted, other equipment cannot respond synchronously, affecting the overall quality control effect.

[0004] Based on the above analysis, the challenges become clearer: how to achieve intelligent quality control integrating multiple technologies in the silver powder production process, particularly how to solve the adaptive optimization problem of Bayesian control parameters under conditions of high defect rates, while simultaneously constructing a decision-making fusion mechanism for the collaboration of multiple quality characteristics, and further realizing coordinated and linked control among production equipment. These interrelated technical problems require overcoming the limitations of traditional single-technology control and establishing a comprehensive quality management system that integrates multiple advanced control technologies. Summary of the Invention

[0005] This application provides a method, system, and storage medium for quality control in silver powder production that integrates multiple technologies. It solves the technical problems such as the degradation of traditional Bayesian control performance under high defect rate parameters in the silver powder production process, insufficient control accuracy due to the lack of collaborative decision-making integration of multiple quality characteristics, and the impact of the lack of collaborative linkage control between production equipment on the overall quality control effect.

[0006] Firstly, this application provides a method for quality control in silver powder production that integrates multiple technologies. The method includes: real-time data acquisition and processing of the silver powder production line via a distributed sensor network to obtain a silver powder quality characteristic correlation matrix; dynamically adjusting the gamma prior distribution hyperparameters based on the defect rate parameter in the silver powder quality characteristic correlation matrix using an adaptive Bayesian algorithm to obtain adaptive shape parameters and adaptive rate parameters; optimizing the silver powder production control strategy based on the adaptive shape parameters and adaptive rate parameters using a decision fusion algorithm to obtain a collaborative control instruction set for silver powder production equipment; performing coordinated control of the ball mill, atomizing drying tower, and grading and screening equipment based on the collaborative control instruction set for silver powder production equipment to obtain silver powder quality control execution results; evaluating the effectiveness of the silver powder quality control execution results to obtain a performance index set; and modifying the collaborative control instruction set for silver powder production equipment based on the performance index set.

[0007] Optionally, the silver powder quality characteristic correlation matrix includes particle size defect rate parameters, electrical conductivity parameters, and morphological characteristic data. The step of obtaining the silver powder quality characteristic correlation matrix through real-time data acquisition and processing of the silver powder production line via a distributed sensor network includes:

[0008] The particle size distribution of silver powder is detected and processed in real time by a laser particle size analyzer at the discharge port of the ball mill to obtain a particle size distribution dataset. Based on the particle size distribution dataset, defect statistics are calculated to obtain the particle size defect rate parameter.

[0009] The conductivity of silver powder is continuously monitored by a conductivity sensor at the feed end of the atomizing drying tower to obtain time-series data of conductivity performance. The time-series data of conductivity performance is then filtered and noise-reduced to obtain conductivity performance parameters.

[0010] The silver powder morphology image is acquired and processed by the machine vision detection unit at the back end of the grading and screening equipment to obtain silver powder morphology image data. The silver powder morphology image data is then input into an image feature extraction algorithm for morphology analysis and processing to obtain morphology feature data.

[0011] The particle size defect rate parameter, the electrical conductivity parameter, and the morphological feature data are input into a correlation analysis algorithm for coupling relationship calculation to obtain the correlation matrix of silver powder quality characteristics.

[0012] Optionally, the step of dynamically adjusting the hyperparameters of the gamma prior distribution using an adaptive Bayesian algorithm based on the defect rate parameter in the silver powder quality characteristic correlation matrix to obtain adaptive shape parameters and adaptive rate parameters includes:

[0013] Based on the defect rate parameter in the silver powder quality characteristic correlation matrix, a threshold judgment process is performed to obtain the defect rate parameter status identifier. When the defect rate parameter is greater than a preset threshold, the defect rate parameter status identifier is a high-risk state.

[0014] Based on the defect rate parameter status identifier and the basic shape parameter, the shape parameter is calculated using a logarithmic mapping algorithm to obtain the adaptive shape parameter, wherein the adaptive shape parameter is equal to the sum of the products of the basic shape parameter, the adjustment coefficient, and the logarithmic value of the defect rate parameter;

[0015] The adaptive shape parameter and the expected value of the defect rate parameter are input into the rate parameter calculation formula to solve for the rate parameter, thereby obtaining the adaptive rate parameter, wherein the adaptive rate parameter is equal to the ratio of the adaptive shape parameter to the product of the expected value of the defect rate parameter and the adjustment coefficient;

[0016] Based on the adaptive shape parameter and the adaptive rate parameter, gamma distribution parameter verification processing is performed to obtain parameter validity identifier. When the parameter validity identifier is invalid, the shape parameter calculation processing and rate parameter solution are re-executed.

[0017] Optionally, the gamma distribution parameter verification process based on the adaptive shape parameter and the adaptive rate parameter includes:

[0018] The adaptive shape parameter is subjected to a numerical range test to obtain a shape parameter validity mark. When the adaptive shape parameter is less than zero or greater than a preset upper limit value for the shape parameter, the shape parameter validity mark is invalid.

[0019] The adaptive rate parameter is subjected to a numerical range test to obtain a rate parameter validity mark. When the adaptive rate parameter is less than zero or greater than a preset rate parameter upper limit value, the rate parameter validity mark is invalid.

[0020] Based on the adaptive shape parameter and the adaptive rate parameter, the convergence of the gamma distribution is tested to obtain the convergence judgment result. The convergence is judged by calculating whether the ratio of the mean to the variance of the gamma distribution is within the preset convergence interval.

[0021] The parameter validity identifier is obtained by performing a logical AND operation on the shape parameter validity marker, the rate parameter validity marker, and the distribution convergence judgment result.

[0022] Optionally, the optimization of the silver powder production control strategy based on adaptive shape parameters and adaptive rate parameters using a decision fusion algorithm to obtain a collaborative control instruction set for the silver powder production equipment includes:

[0023] The adaptive shape parameter and the adaptive rate parameter are input into the Bayesian control decision module for probability distribution calculation to obtain the silver powder quality control probability distribution function, wherein the silver powder quality control probability distribution function is constructed based on the gamma distribution.

[0024] Based on the conductivity parameters and morphological feature data in the silver powder quality characteristic correlation matrix, multi-dimensional feature fusion processing is performed to obtain a comprehensive evaluation index of silver powder quality.

[0025] The silver powder quality control probability distribution function and the silver powder quality comprehensive evaluation index are input into a three-layer fusion decision architecture for hierarchical decision processing to obtain a fusion decision weight vector. The three-layer fusion decision architecture includes data layer fusion, feature layer fusion and decision layer fusion.

[0026] Based on the fusion decision weight vector, the ball mill speed control parameters, atomizing drying tower temperature control parameters, and grading and screening equipment frequency control parameters are synergistically optimized to obtain the equipment control parameter optimization vector.

[0027] Based on the optimized vector of the equipment control parameters, the instruction encapsulation process is performed to obtain a collaborative control instruction set for silver powder production equipment. The collaborative control instruction set for silver powder production equipment includes ball mill control instructions, atomizing drying tower control instructions, and grading and screening equipment control instructions.

[0028] Optionally, the step of performing coordinated control of the ball mill, atomizing drying tower, and grading and screening equipment according to the coordinated control instruction set of the silver powder production equipment to obtain the silver powder quality control execution result includes:

[0029] The ball mill speed is adjusted and controlled according to the ball mill control instructions in the collaborative control instruction set of the silver powder production equipment to obtain ball mill operating status data, wherein the ball mill operating status data includes actual speed value, current consumption value and vibration amplitude value;

[0030] The feed temperature and discharge temperature of the atomizing drying tower are synchronously adjusted according to the control instructions of the atomizing drying tower in the collaborative control instruction set of the silver powder production equipment to obtain the operating status data of the atomizing drying tower. The operating status data of the atomizing drying tower includes the measured value of the feed temperature, the measured value of the discharge temperature, and the humidity control value.

[0031] According to the control instructions for the grading and screening equipment in the collaborative control instruction set of the silver powder production equipment, the grading and screening frequency and screen spacing of the grading and screening equipment are coordinated and adjusted to obtain the operating status data of the grading and screening equipment. The operating status data of the grading and screening equipment includes the measured value of the grading frequency, the measured value of the screen spacing and the grading efficiency value.

[0032] The operating status data of the ball mill, the operating status data of the atomizing drying tower, and the operating status data of the grading and screening equipment are integrated to obtain the silver powder quality control execution result, which includes equipment linkage execution records and silver powder product quality test data.

[0033] Optionally, the effect evaluation of the silver powder quality control execution results is performed to obtain a set of performance indicators, and the collaborative control instruction set of the silver powder production equipment is modified according to the set of performance indicators. The set of performance indicators includes: average run length, equipment misalignment rate, and quality prediction accuracy, including:

[0034] Based on the quality control process data in the silver powder quality control execution results, the number of consecutive normal operation times is statistically processed to obtain the value of each run length. The sum of the values ​​of each run length is then divided by the total number of runs to obtain the average run length.

[0035] Based on the equipment control action records in the silver powder quality control execution results, erroneous control action identification processing is performed to obtain the number of erroneous control adjustments. The number of erroneous control adjustments is then divided by the total number of equipment control adjustments to obtain the equipment misadjustment rate.

[0036] The actual silver powder quality data and the expected silver powder quality data in the silver powder quality control execution results are compared and judged item by item to obtain the number of correct quality predictions. The number of correct quality predictions is then divided by the total number of quality predictions to obtain the quality prediction accuracy.

[0037] The performance deviation value is obtained by calculating the difference between the average run length, the equipment misadjustment rate and the quality prediction accuracy and their respective preset standard values.

[0038] Based on the performance deviation value, the control parameters in the collaborative control instruction set of the silver powder production equipment are numerically adjusted to obtain the corrected collaborative control instruction set of the silver powder production equipment.

[0039] Secondly, this application provides a multi-technology integrated silver powder production quality control system, which includes:

[0040] The data acquisition module is used to acquire and process real-time data from the silver powder production line through a distributed sensor network to obtain a correlation matrix of silver powder quality characteristics.

[0041] The adjustment module is used to dynamically adjust the hyperparameters of the gamma prior distribution based on the defect rate parameter in the silver powder quality characteristic correlation matrix using an adaptive Bayesian algorithm, so as to obtain the adaptive shape parameter and the adaptive rate parameter.

[0042] The processing module is used to optimize the silver powder production control strategy based on the adaptive shape parameters and adaptive rate parameters through a decision fusion algorithm, and obtain a set of collaborative control instructions for silver powder production equipment.

[0043] The control module is used to perform linkage control on the ball mill, atomizing drying tower and grading and screening equipment according to the collaborative control instruction set of the silver powder production equipment, so as to obtain the silver powder quality control execution result;

[0044] The correction module is used to evaluate the effect of the silver powder quality control execution results, obtain a set of performance indicators, and correct the collaborative control instruction set of the silver powder production equipment according to the set of performance indicators.

[0045] Thirdly, a computer-readable storage medium is provided, wherein instructions are stored therein, which, when executed on a computer, cause the computer to perform the aforementioned silver powder production quality control method integrating multiple technologies.

[0046] The technical solution provided in this application obtains a silver powder quality characteristic correlation matrix by real-time data acquisition and processing of the silver powder production line through a distributed sensor network. This achieves unified fusion and structured expression of multi-source heterogeneous sensor data, effectively solving the problem of incomplete information acquisition by traditional single sensors and providing a complete and accurate data foundation for subsequent intelligent control decisions. Based on the defect rate parameter in the silver powder quality characteristic correlation matrix, an adaptive Bayesian algorithm is used to dynamically adjust the hyperparameters of the gamma prior distribution to obtain adaptive shape parameters and adaptive rate parameters. The core contribution of the adaptive Bayesian algorithm lies in its ability to automatically adjust the statistical parameters of the gamma distribution according to the real-time defect rate level. In particular, the logarithmic mapping transformation and parameter verification mechanism ensure the stability and accuracy of the control system under high defect rate parameters. Compared with the traditional fixed-parameter Bayesian method, the adaptive algorithm of this invention significantly improves the control accuracy and effectively avoids the control performance degradation problem caused by parameter mismatch.

[0047] The silver powder production control strategy is optimized using a decision fusion algorithm to obtain a collaborative control instruction set for the silver powder production equipment. This algorithm employs a three-layer fusion architecture, hierarchically integrating probability distribution functions, comprehensive quality evaluation indicators, and expert rules. Its contribution lies in its ability to comprehensively consider the coupling relationships of multi-dimensional quality characteristics such as silver powder particle size, conductivity, and morphology. By dynamically allocating weights and using fuzzy logic reasoning, it generates the optimal control strategy, overcoming the limitations of single-technology decision-making and achieving intelligent control through multi-technology collaborative optimization. Based on this collaborative control instruction set, the ball mill, atomizing drying tower, and grading and screening equipment are linked for quality control of the silver powder. The technical contribution of this collaborative control mechanism lies in achieving time synchronization and parameter coordination among multiple devices. When the parameter of one device is adjusted, other devices can respond synchronously and make compensatory adjustments, avoiding the mutual interference problems between devices in traditional independent control modes and significantly improving the stability and response speed of the overall control system. The performance evaluation results of silver powder quality control are used to obtain a set of performance indicators. Based on the performance indicator set, the collaborative control instruction set of silver powder production equipment is modified. The performance evaluation mechanism comprehensively evaluates the control performance through multi-dimensional indicators such as average run length, equipment misadjustment rate and quality prediction accuracy. Based on the performance deviation, the control parameters are automatically optimized and adjusted, forming a closed-loop self-learning intelligent control system that continuously improves the control effect and adapts to changes in production conditions. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a schematic diagram of one embodiment of the silver powder production quality control method integrating multiple technologies in this application.

[0050] Figure 2 This is a schematic diagram of the process for dynamically adjusting the hyperparameters of the gamma prior distribution using an adaptive Bayesian algorithm in an embodiment of this application.

[0051] Figure 3 This is a schematic diagram of one embodiment of the silver powder production quality control system that integrates multiple technologies in this application. Detailed Implementation

[0052] This application provides a method, system, and storage medium for quality control in silver powder production that integrates multiple technologies. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0053] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the silver powder production quality control method integrating multiple technologies in this application includes:

[0054] Step S101: Real-time data acquisition and processing of the silver powder production line is performed through a distributed sensor network to obtain the correlation matrix of silver powder quality characteristics;

[0055] Step S102: Based on the defect rate parameter in the silver powder quality characteristic correlation matrix, the hyperparameters of the gamma prior distribution are dynamically adjusted using an adaptive Bayesian algorithm to obtain the adaptive shape parameter and the adaptive rate parameter.

[0056] Step S103: Optimize the silver powder production control strategy using a decision fusion algorithm based on the adaptive shape parameters and adaptive rate parameters to obtain a collaborative control instruction set for the silver powder production equipment;

[0057] Step S104: Perform linkage control on the ball mill, atomizing drying tower and grading and screening equipment according to the collaborative control instruction set of silver powder production equipment to obtain the silver powder quality control execution results;

[0058] Step S105: Evaluate the effect of the silver powder quality control execution results, obtain a set of performance indicators, and modify the set of collaborative control instructions for silver powder production equipment according to the set of performance indicators.

[0059] It is understood that the executing entity of this application can be a silver powder production quality control system integrating multiple technologies, or it can be a terminal or a server; the specific implementation is not limited here. This application's embodiment uses a server as an example for illustration.

[0060] Specifically, a distributed sensor network is used to collect data from the silver powder production line. A laser particle size analyzer detects the silver powder particle size distribution at the ball mill outlet. By measuring the proportion of silver powder particles in different size ranges, the ratio of particles exceeding the standard particle size range (e.g., 0.5-5 micrometers) to the total number of particles is calculated to obtain the particle size defect rate parameter. A conductivity sensor continuously measures the conductivity of the silver powder suspension at the feed end of the atomizing drying tower. After recording the time-series data, a low-pass filter is used to remove high-frequency noise and extract stable conductivity parameters. A machine vision inspection unit acquires images of the silver powder morphology, identifies the outline of silver powder particles using an edge detection algorithm, and calculates morphological feature parameters such as roundness and aspect ratio. These parameters are input into a correlation analysis algorithm to calculate the correlation coefficient matrix between various quality characteristics, forming a silver powder quality characteristic correlation matrix. An adaptive Bayesian algorithm dynamically adjusts the hyperparameters of the gamma prior distribution based on the defect rate parameter. When the detected particle size defect rate parameter exceeds a preset threshold of 0.8, the algorithm sets the defect rate parameter status to a high-risk state. Shape parameter calculation is performed using a logarithmic mapping algorithm. The basic shape parameter is multiplied by the logarithmic values ​​of the adjustment coefficient and the defect rate parameter, and then summed to obtain the adaptive shape parameter. Rate parameter calculation involves dividing the adaptive shape parameter by the product of the expected value of the defect rate parameter and the adjustment coefficient to obtain the adaptive rate parameter. The parameter verification process checks whether the adaptive shape parameter and rate parameter are within their valid ranges. Convergence is assessed by calculating the ratio of the mean to variance of the gamma distribution. If a parameter is invalid, the adjustment coefficient is halved and recalculated.

[0061] The decision fusion algorithm calculates the probability distribution of adaptive shape and rate parameters to construct a probability distribution function for silver powder quality control based on gamma distribution. Conductivity parameters and morphological feature data are fused using a weighted average method to calculate a comprehensive evaluation index for silver powder quality. The three-layer fusion decision architecture performs weighted fusion of multi-source sensor data at the data layer, extracts key quality feature components at the feature layer, and generates a fusion decision weight vector at the decision layer through fuzzy logic reasoning. Ball mill speed control parameters, atomizing dryer temperature control parameters, and grading and screening equipment frequency control parameters are collaboratively optimized based on the weight vector to form an optimized equipment control parameter vector, which is then encapsulated into a collaborative control instruction set. Equipment linkage control is executed according to this collaborative control instruction set. The ball mill adjusts its actual speed according to the speed control command and records current consumption and vibration amplitude. The atomizing dryer synchronously adjusts its feed and discharge temperatures, recording measured temperature and humidity control values. The grading and screening equipment coordinates the adjustment of screening frequency and screen spacing, recording corresponding operating parameters. The equipment operating status data is summarized and integrated to form the silver powder quality control execution result, including complete equipment linkage execution records and silver powder product quality testing data.

[0062] The performance evaluation calculates the average run length by counting the number of consecutive normal operations during the quality control process, summing the run lengths, and then dividing by the total number of runs. The equipment misalignment rate is calculated by identifying the number of erroneous operations in the control action records and dividing this number by the total number of equipment controls to obtain the misalignment ratio. The quality prediction accuracy is calculated by comparing actual silver powder quality data with expected quality data, and the ratio of correct predictions to the total number of predictions. The performance deviation is calculated by subtracting the average run length, equipment misalignment rate, and quality prediction accuracy from their respective standard values. Based on the deviation, the control parameter values ​​in the collaborative control instruction set are adjusted to achieve instruction correction and parameter optimization.

[0063] In one specific embodiment, the process of performing step S101 may specifically include the following steps:

[0064] The particle size distribution of silver powder is detected and processed in real time by a laser particle size analyzer at the discharge port of the ball mill to obtain a particle size distribution dataset. Based on the particle size distribution dataset, defect statistics are calculated to obtain the particle size defect rate parameter.

[0065] The conductivity of silver powder is continuously monitored by a conductivity sensor at the feed end of the atomizing drying tower to obtain time-series data of conductivity performance. The time-series data of conductivity performance is then filtered and noise-reduced to obtain conductivity performance parameters.

[0066] The machine vision inspection unit at the back end of the grading and screening equipment collects and processes the silver powder morphology image to obtain silver powder morphology image data. The silver powder morphology image data is then input into an image feature extraction algorithm for morphology analysis and processing to obtain morphology feature data.

[0067] The particle size defect rate parameter, electrical conductivity parameter and morphological feature data are input into the correlation analysis algorithm for coupling relationship calculation to obtain the correlation matrix of silver powder quality characteristics.

[0068] Specifically, the laser particle size analyzer detects the particle size distribution of silver powder using the principle of laser scattering. When a laser beam irradiates a silver powder sample, it generates scattered light. Silver powder particles of different sizes produce scattered light signals with different angles and intensities. A photodetector collects the scattered light signals and converts them into electrical signals. The data processing unit calculates the particle number distribution for each particle size range based on Mie scattering theory, forming a particle size distribution dataset. Defect statistics calculations calculate the ratio of the number of particles exceeding the standard particle size range in the particle size distribution dataset to the total number of particles. The standard particle size range is set according to the application requirements of the silver powder. Photovoltaic-grade silver powder typically requires a particle size in the range of 0.5-5 micrometers. Particles exceeding this range are considered defective particles. The number of defective particles divided by the total number of particles yields the particle size defect rate parameter. The conductivity sensor measures the conductivity of the silver powder suspension based on electrochemical principles. An AC voltage is applied between the sensor electrodes, and the current intensity passing through the silver powder suspension is measured. The conductivity value is calculated according to Ohm's law. During continuous monitoring, conductivity data is collected once per second to form time-series data on conductivity performance. The filtering and noise reduction process uses a digital filtering algorithm to remove high-frequency noise and random interference from the time-series data. The low-pass filter is set to a cutoff frequency of one-tenth of the sampling frequency to filter out noise components above this frequency and retain the low-frequency signal that reflects the true changes in the conductivity of silver powder. The stable value after filtering is used as the conductivity parameter.

[0069] The machine vision inspection unit acquires images of the silver powder morphology using a CCD camera. The image sensor converts the optical signals into digital image signals, forming silver powder morphology image data. The image feature extraction algorithm first preprocesses the image, including grayscale conversion, noise removal, and contrast enhancement. Then, an edge detection algorithm is used to identify the contour boundaries of the silver powder particles, calculating the geometric parameters of each particle, including area, perimeter, major axis, and minor axis length. Based on these geometric parameters, morphology feature parameters are calculated. Circularity is equal to four times pi multiplied by the area divided by the square of the perimeter, aspect ratio is equal to the major axis length divided by the minor axis length, and surface roughness is calculated through the curvature change of the contour lines. These morphology feature parameters constitute the morphology feature data.

[0070] The correlation analysis algorithm uses particle size defect rate parameters, electrical conductivity parameters, and morphological characteristics as input variables to calculate the Pearson correlation coefficient between each parameter. The correlation coefficient is equal to the product of the covariance of the two variables divided by their respective standard deviations. The covariance reflects the strength of the linear relationship between the two variables, while the standard deviation measures the dispersion of a single variable. The coupling relationship calculation process constructs a multiple linear regression model to analyze the mutual influence between various quality characteristics. The regression coefficient represents the degree of influence of one quality characteristic on another. All correlation coefficients and regression coefficients form a silver powder quality characteristic correlation matrix, and the matrix elements reflect the quantitative relationship between different quality characteristics.

[0071] In one specific embodiment, the process of performing step S102 may specifically include the following steps:

[0072] Threshold judgment is performed based on the defect rate parameter in the correlation matrix of silver powder quality characteristics to obtain the defect rate parameter status identifier. When the defect rate parameter is greater than the preset threshold, the defect rate parameter status identifier is a high-risk state.

[0073] Based on the defect rate parameter status indicator and the basic shape parameter, the shape parameter is calculated using a logarithmic mapping algorithm to obtain the adaptive shape parameter, where the adaptive shape parameter is equal to the sum of the products of the basic shape parameter, the adjustment coefficient, and the logarithmic value of the defect rate parameter;

[0074] The adaptive shape parameter and the expected value of the defect rate parameter are input into the rate parameter calculation formula to solve for the rate parameter, and the adaptive rate parameter is obtained. The adaptive rate parameter is equal to the ratio of the adaptive shape parameter to the product of the expected value of the defect rate parameter and the adjustment coefficient.

[0075] The gamma distribution parameters are verified based on adaptive shape parameters and adaptive rate parameters to obtain parameter validity indicators. When the parameter validity indicator is invalid, the shape parameter calculation and rate parameter solution are re-executed.

[0076] Specifically, such as Figure 2The diagram illustrates the process of dynamically adjusting the hyperparameters of the gamma prior distribution using an adaptive Bayesian algorithm in this embodiment. The threshold judgment process compares the defect rate parameter in the silver powder quality characteristic correlation matrix with a preset threshold. The preset threshold is determined based on silver powder product quality standards and historical production data statistical analysis. The preset threshold for photovoltaic-grade silver powder is typically set to 0.1, and for conductive coating-grade silver powder, it is set to 0.15. When the defect rate parameter value is greater than the preset threshold, the judgment logic sets the defect rate parameter status to a high-risk state, indicating that there is a quality anomaly in the current silver powder production process and the control strategy needs adjustment. When the defect rate parameter is less than or equal to the preset threshold, the status is set to a normal state, indicating that the production process is stable. The logarithmic mapping algorithm is a nonlinear numerical transformation method that mathematically transforms the defect rate parameter using the natural logarithm function. Logarithmic transformation can compress the dynamic range of the data and enhance its linearity. The shape parameter calculation process first reads the basic shape parameter values, which are initial values ​​preset based on the characteristics of the silver powder production process and gamma distribution. Then, the natural logarithm of the defect rate parameter is calculated. The adjustment coefficient is a proportional factor controlling the adjustment range of the shape parameters, determined according to the response characteristics and control precision requirements of the silver powder production line. The adaptive shape parameter calculation process uses the basic shape parameter as a reference value. The adjustment coefficient is multiplied by the logarithm of the defect rate parameter to obtain the adjustment amount. The reference value and the adjustment amount are added to obtain the final adaptive shape parameter. This calculation method ensures that the shape parameters can be dynamically adjusted according to the actual defect rate level.

[0077] The rate parameter calculation formula is designed based on the mathematical characteristics of the gamma distribution. The rate parameter and shape parameter together determine the shape of the probability density function of the gamma distribution. The expected value of the defect rate parameter is the statistical mean of the defect rate calculated based on historical production data, reflecting the average quality level of the silver powder production process. The adjustment coefficient is a dimensionless correction factor used to balance the influence of the shape parameter and the expected value of the defect rate on the rate parameter. The rate parameter solution process uses the adaptive shape parameter as the numerator and the product of the expected value of the defect rate parameter and the adjustment coefficient as the denominator, and obtains the adaptive rate parameter through division. This calculation method ensures that the numerical matching relationship between the rate parameter and the shape parameter conforms to the statistical requirements of the gamma distribution. The gamma distribution parameter verification process includes two aspects: numerical range testing and distribution characteristic testing. The numerical range testing ensures that the shape parameter and rate parameter are mathematically valid. The shape parameter must be greater than zero and less than a preset upper limit value, and the rate parameter also needs to meet the positive number condition and upper limit constraint. The distribution characteristic test determines the rationality of parameter settings by calculating the mean and variance of the gamma distribution. The mean of the gamma distribution equals the shape parameter divided by the rate parameter, and the variance equals the shape parameter divided by the square of the rate parameter. When the ratio of the mean to the variance is within the preset convergence interval, the distribution convergence is considered good. The parameter validity indicator uses a logical AND operation to comprehensively judge the validity of the shape parameter, the validity of the rate parameter, and the distribution convergence. Only when all three conditions are met is the parameter validity indicator in a valid state; otherwise, it is marked as invalid, triggering parameter recalculation.

[0078] In one specific embodiment, the process of performing gamma distribution parameter verification based on the adaptive shape parameter and the adaptive rate parameter may specifically include the following steps:

[0079] The adaptive shape parameters are subjected to a numerical range test to obtain a shape parameter validity mark. When the adaptive shape parameter is less than zero or greater than the preset upper limit value of the shape parameter, the shape parameter validity mark is invalid.

[0080] The adaptive rate parameter is subjected to a numerical range test to obtain a rate parameter validity mark. When the adaptive rate parameter is less than zero or greater than the preset rate parameter upper limit, the rate parameter validity mark is invalid.

[0081] The convergence test of the gamma distribution is performed based on the adaptive shape parameter and the adaptive rate parameter to obtain the convergence judgment result. The convergence is judged by calculating whether the ratio of the mean to the variance of the gamma distribution is within the preset convergence interval.

[0082] The parameter validity identifier is obtained by performing a logical AND operation on the shape parameter validity marker, the rate parameter validity marker, and the distribution convergence judgment result.

[0083] Specifically, the adaptive shape parameter is compared with a zero value. When the adaptive shape parameter is less than zero, it indicates a mathematical inconsistency because the shape parameter of the gamma distribution must be positive. In this case, the shape parameter validity flag is set to invalid. The verification process also compares the adaptive shape parameter with a preset upper limit value. This upper limit value is determined based on the physical constraints and control precision requirements of the silver powder production process. In photovoltaic-grade silver powder production, the upper limit is typically set to 10, and for conductive coating-grade silver powder, it is set to 15. When the adaptive shape parameter exceeds the upper limit value, it indicates that the parameter adjustment has deviated excessively from the normal range, and the shape parameter validity flag is also set to invalid. Only when the adaptive shape parameter is both greater than zero and less than the preset upper limit value is the shape parameter validity flag set to valid. This double-boundary check ensures that the shape parameter is feasible both statistically and in engineering practice. The numerical range verification of the adaptive rate parameter uses the same boundary judgment logic. The rate parameter also needs to meet the mathematical constraints of the gamma distribution, i.e., it must be positive. The verification algorithm compares the adaptive rate parameter with a zero value. When the rate parameter is less than or equal to zero, it violates the basic requirements of the gamma distribution, and the rate parameter validity flag is set to invalid. The preset upper limit of the rate parameter is determined based on the response capability of the silver powder production equipment and the numerical stability of the control algorithm. The upper limit is set to 50 for the ball mill control, 30 for the atomizing drying tower control, and 40 for the grading and screening equipment control. When the adaptive rate parameter exceeds the corresponding upper limit, it indicates that the control response is too aggressive and does not conform to the equipment characteristics; the rate parameter validity flag is then set to invalid. Dual verification conditions ensure that the rate parameter remains within the valid value range, satisfying both mathematical requirements and engineering constraints.

[0084] The convergence test of the gamma distribution assesses the rationality of the distribution parameters through statistical characteristic calculations. The convergence test is based on numerical analysis using the formulas for calculating the mean and variance of the gamma distribution. The mean of the gamma distribution is calculated by dividing the adaptive shape parameter by the adaptive rate parameter, and the variance is calculated by dividing the adaptive shape parameter by the square of the adaptive rate parameter. The ratio of the mean to the variance equals the value of the adaptive rate parameter; this ratio reflects the central tendency and dispersion of the distribution. The preset convergence interval is determined based on the accuracy requirements of silver powder quality control and the statistical characteristics of historical data. The lower limit of the convergence interval is typically set to 5 to indicate that the distribution should not be too dispersed, and the upper limit is set to 100 to indicate that the distribution should not be too concentrated. When the ratio of the mean to the variance falls within this interval, it indicates that the gamma distribution has reasonable statistical characteristics, and the convergence judgment result is set to convergence; otherwise, it is set to divergence, indicating that the parameter combination is inappropriate, leading to abnormal distribution characteristics.

[0085] The logical AND operation combines three independent validity judgments using Boolean logic. The AND operation outputs a true result only when all input conditions are simultaneously true. The operation takes the shape parameter validity flag, the rate parameter validity flag, and the distribution convergence judgment result as three Boolean variables as input. The AND operation outputs a true result only when the shape parameter validity flag is valid, the rate parameter validity flag is valid, and the distribution convergence judgment result is convergent; in this case, the parameter validity flag is set to valid. If any input condition is false, the AND operation outputs false, the parameter validity flag is set to invalid, and the parameter recalculation process is triggered. This strict logical judgment ensures that only parameter combinations that fully satisfy all constraints pass verification.

[0086] For example, in the production of silver powder, the calculated adaptive shape parameter is 3.2. First, it is compared with zero to confirm a positive value. Then, it is compared with the preset upper limit of 10 to confirm it is less than the upper limit. Therefore, the shape parameter validity flag is set to valid. The calculated adaptive rate parameter is 25.6. It is compared with zero to confirm a positive value. It is compared with the preset upper limit of 50 to confirm it is less than the upper limit. Therefore, the rate parameter validity flag is set to valid. The gamma distribution mean is calculated as 3.2 divided by 25.6, which equals 0.125. The variance is calculated as 3.2 divided by the square of 25.6, which equals 0.0049. The ratio of mean to variance is 25.6, which is within the preset convergence interval of 5 to 100. The distribution convergence judgment result is convergence. The logical AND operation combines the three valid states. Since all conditions are met, the parameter validity flag is set to valid. The verification passes, and the process proceeds to the next step of control strategy optimization. If any verification step fails, the entire verification process fails. The adjustment coefficient is halved, and the calculation and verification process of shape parameters and rate parameters is re-executed until an effective combination of parameters that satisfies all constraints is obtained.

[0087] In one specific embodiment, the process of executing step S103 may specifically include the following steps:

[0088] The adaptive shape parameters and adaptive rate parameters are input into the Bayesian control decision module for probability distribution calculation to obtain the silver powder quality control probability distribution function, which is constructed based on the gamma distribution.

[0089] Based on the conductivity parameters and morphological feature data in the correlation matrix of silver powder quality characteristics, multi-dimensional feature fusion processing is performed to obtain a comprehensive evaluation index of silver powder quality.

[0090] The probability distribution function of silver powder quality control and the comprehensive evaluation index of silver powder quality are input into the three-layer fusion decision architecture for hierarchical decision processing to obtain the fusion decision weight vector. The three-layer fusion decision architecture includes data layer fusion, feature layer fusion and decision layer fusion.

[0091] Based on the fusion decision weight vector, the ball mill speed control parameters, atomizing drying tower temperature control parameters, and grading and screening equipment frequency control parameters are synergistically optimized to obtain the equipment control parameter optimization vector.

[0092] Based on the optimization vector of equipment control parameters, the instruction encapsulation process is performed to obtain the collaborative control instruction set for silver powder production equipment. The collaborative control instruction set for silver powder production equipment includes ball mill control instructions, atomizing drying tower control instructions, and grading and screening equipment control instructions.

[0093] Specifically, the Bayesian control decision module converts the adaptive shape parameter and adaptive rate parameter into a mathematical expression of the gamma distribution through probability density function calculation. The probability distribution calculation process constructs a silver powder quality control probability distribution function based on the standard form of the gamma distribution. The probability density function calculation process uses the adaptive rate parameter as a scale parameter and the adaptive shape parameter as a shape parameter, generating a complete probability distribution expression through a combination of gamma and exponential functions. The silver powder quality control probability distribution function describes the probability of silver powder defect rates occurring within different numerical ranges. The peak position of the distribution function reflects the most likely defect rate level, and the width of the distribution reflects the degree of variation in the defect rate. This probabilistic expression provides a statistical basis for subsequent fusion decision-making. The multidimensional feature fusion process numerically integrates the conductivity parameters and morphological feature data in the silver powder quality characteristic correlation matrix. The fusion algorithm uses a weighted average method to unify quality features from different dimensions to the same evaluation scale. The conductivity parameters are converted into dimensionless values ​​through standardization. The standardization process subtracts the historical mean from the original conductivity value and divides by the standard deviation to obtain the standardized conductivity score. The morphological feature data includes multiple parameters such as roundness, aspect ratio, and surface roughness. Each parameter is also standardized and converted into a standard score. Then, weight coefficients are assigned according to the importance of each morphological feature to the quality of silver powder. The calculation of the comprehensive evaluation index of silver powder quality involves multiplying the standardized conductivity score by the conductivity weight coefficient, multiplying each standardized morphological feature score by its corresponding weight coefficient, and summing all weighted scores to obtain the final comprehensive evaluation index value.

[0094] The three-layer fusion decision architecture achieves the orderly integration of multi-source information through hierarchical data processing. The data layer fusion initially integrates the probability distribution function of silver powder quality control and the comprehensive evaluation index of silver powder quality as basic input data. The data layer fusion employs a probabilistic weighting method, determining weight allocation based on the probability density values ​​of the probability distribution function at different defect rate levels, with higher weights for high probability density regions and lower weights for low probability density regions. The feature layer fusion performs feature extraction and dimensionality reduction on the data layer fusion results, identifying the main feature components affecting silver powder quality control through principal component analysis, retaining principal components with an explained variability exceeding 85% as key features. The decision layer fusion, based on fuzzy logic reasoning, converts the principal components of the feature layer into specific control decisions. The fuzzy reasoning rules are formulated based on the experience of silver powder production process experts and historical data statistical patterns. The reasoning process generates a fusion decision weight vector, with each element corresponding to the control weight of the ball mill, atomizing drying tower, and grading and screening equipment. Collaborative optimization calculations adjust the control parameters of each device based on the fusion decision weight vector. For example, optimizing the ball mill speed control parameter involves multiplying the current speed setpoint by the corresponding weight coefficient to obtain the adjusted speed parameter. The optimization of temperature control parameters for the atomizing drying tower involves multiplying the feed temperature and discharge temperature by a temperature weighting coefficient, and coordinating adjustments based on the coupling relationship of temperature control. The optimization of frequency control parameters for the grading and screening equipment involves multiplying the screening frequency by a frequency weighting coefficient, while simultaneously adjusting the screen spacing parameter according to particle size distribution requirements. The equipment control parameter optimization vector classifies and organizes the optimized control parameters of all equipment according to equipment type and parameter type, forming a structured parameter set.

[0095] The instruction encapsulation process optimizes the vector of equipment control parameters into an executable control instruction format. Ball mill control instructions include speed setpoints, start / stop control signals, and safety protection parameters, and the instruction format conforms to the communication protocol requirements of the ball mill controller. Atomizing drying tower control instructions include control parameters such as feed temperature setpoints, discharge temperature setpoints, fan speed, and heating power; the instruction encoding uses an industry-standard data frame format. Grading and screening equipment control instructions include control information such as screening frequency, amplitude adjustment, and screen position; instruction transmission achieves real-time communication via the industrial Ethernet protocol. The collaborative control instruction set for silver powder production equipment arranges the three types of equipment instructions according to their execution sequence, ensuring the coordination and synchronization of control actions across all equipment.

[0096] In one specific embodiment, the process of executing step S104 may specifically include the following steps:

[0097] The ball mill speed is adjusted and controlled according to the ball mill control instructions in the collaborative control instruction set of silver powder production equipment to obtain ball mill operating status data, which includes actual speed value, current consumption value and vibration amplitude value.

[0098] Based on the control instructions of the atomizing drying tower in the collaborative control instruction set of silver powder production equipment, the feed temperature and discharge temperature of the atomizing drying tower are synchronously adjusted to obtain the operating status data of the atomizing drying tower. The operating status data of the atomizing drying tower includes the measured value of the feed temperature, the measured value of the discharge temperature, and the humidity control value.

[0099] According to the control instructions of the grading and screening equipment in the collaborative control instruction set of silver powder production equipment, the grading and screening frequency and screen spacing of the grading and screening equipment are coordinated and adjusted to obtain the operating status data of the grading and screening equipment. The operating status data of the grading and screening equipment includes the measured value of the grading frequency, the measured value of the screen spacing and the grading efficiency value.

[0100] The operating status data of the ball mill, the atomizing drying tower, and the grading and screening equipment are integrated to obtain the silver powder quality control execution results. The silver powder quality control execution results include equipment linkage execution records and silver powder product quality test data.

[0101] Specifically, the ball mill speed regulation and control process receives control commands and executes speed adjustments via a frequency converter. The frequency converter adjusts the motor power supply frequency according to the speed setpoint in the control command, and the motor speed is directly proportional to the power supply frequency to achieve precise speed control. The actual speed value is obtained by measuring the encoder installed on the ball mill spindle. The encoder outputs a fixed number of pulse signals per revolution, and the controller counts the number of pulses per unit time to calculate the actual speed value. The current consumption value is detected by a current transformer to detect the three-phase current of the motor. The transformer converts the large current signal into a standard current signal and inputs it into the data acquisition module. The acquisition module calculates the effective value of the three-phase current as the current consumption parameter. The vibration amplitude value is monitored by an accelerometer to monitor the vibration intensity of the ball mill body. The acceleration signal output by the sensor is converted into vibration displacement amplitude through integration. The vibration amplitude reflects the operating stability and mechanical state of the ball mill.

[0102] The synchronous regulation of the atomizing drying tower involves the coordinated control of the feed temperature and the discharge temperature. Feed temperature control tracks the setpoint by adjusting the heater power. The temperature controller uses a PID algorithm to calculate the heater power output, performing proportional, integral, and derivative operations based on the deviation between the setpoint and the measured value. Discharge temperature control regulates the temperature by adjusting the cooling fan speed and cooling water flow rate. Cooling control and heating control form a closed-loop temperature regulation system. The measured feed and discharge temperatures are obtained by thermocouple temperature sensors. The millivolt-level voltage signals generated by the thermocouples are amplified and linearized by a signal conditioning circuit and converted into temperature values. Humidity control measures the relative humidity inside the drying tower using a humidity sensor. Based on the principle of capacitance or resistance change, the humidity sensor outputs an electrical signal proportional to the humidity, reflecting the moisture removal effect during the silver powder drying process.

[0103] The grading and screening equipment coordinates and adjusts two key parameters simultaneously: screening frequency and screen spacing. Screening frequency control is achieved through variable frequency speed regulation of the vibrating motor. The excitation force generated by the vibrating motor drives the screens to vibrate periodically. The screening frequency directly affects the screening efficiency and separation accuracy of silver powder particles. Screen spacing adjustment is achieved through a mechanical transmission mechanism that adjusts the gap between the upper and lower screens. After receiving control commands, the spacing adjustment mechanism drives a lead screw or hydraulic cylinder to change the screen position. The measured screening frequency is obtained through a vibration frequency sensor, which detects the frequency signal of the screen vibration and converts it into a digital frequency value. The measured screen spacing is obtained by measuring the actual distance between the screens using a displacement sensor. A laser displacement sensor emits a laser beam to measure the positional change on the screen surface, calculating the precise spacing value. The screening efficiency is calculated by comparing the particle size distribution of silver powder before and after screening. The screening efficiency is equal to the ratio of the mass of silver powder within the target particle size range to the total mass before screening.

[0104] Data aggregation and integration processes synchronize the operational status data of the three devices according to timestamps, ensuring the consistency and correlation of data across all devices. The ball mill's operational status data, including actual rotational speed, current consumption, and vibration amplitude, is recorded along with corresponding timestamps. Similarly, the atomizing dryer's operational status data, including measured feed temperature, measured discharge temperature, and humidity control values, is also time-stamped. The grading and screening equipment's operational status data, including measured screening frequency, measured screen spacing, and screening efficiency, is also recorded with corresponding timestamps. Equipment linkage execution records track the temporal relationship of the entire control execution process by recording the sending, execution, and completion times of control commands for each device. Silver powder product quality inspection data is obtained through online quality inspection equipment, including key quality indicators such as particle size distribution, conductivity, and morphological characteristics of the final product. This quality inspection data is correlated with the equipment operational data to form a complete dataset of silver powder quality control execution results.

[0105] In one specific embodiment, the process of executing step S105 may specifically include the following steps:

[0106] Based on the quality control process data in the silver powder quality control execution results, the number of consecutive normal operation times is statistically processed to obtain the value of each run length. The average run length is obtained by summing the values ​​of each run length and dividing by the total number of runs.

[0107] Based on the equipment control action records in the silver powder quality control execution results, erroneous control actions are identified and processed to obtain the number of erroneous control actions. The number of erroneous control actions is then divided by the total number of equipment control actions to obtain the equipment misadjustment rate.

[0108] The actual silver powder quality data and the expected silver powder quality data in the silver powder quality control execution results are compared and judged item by item to obtain the number of correct quality predictions. The number of correct quality predictions is then divided by the total number of quality predictions to obtain the quality prediction accuracy.

[0109] The performance deviation value is obtained by calculating the difference between the average run length, equipment misadjustment rate and quality prediction accuracy and their respective preset standard values.

[0110] The control parameters in the collaborative control instruction set for silver powder production equipment are numerically adjusted based on the performance deviation values ​​to obtain the corrected collaborative control instruction set for silver powder production equipment.

[0111] Specifically, the continuous normal operation count statistical processing identifies the operating cycles that continuously meet quality standards during the quality control process by analyzing the quality control process data in the silver powder quality control execution results. The statistical algorithm checks the quality status of each control cycle item by item, starting from the first record of the quality control process data. When all silver powder quality indicators of a control cycle meet the preset standards, it is recorded as normal operation; when quality indicators exceed the standards, it is recorded as abnormal operation, and the current run count ends. The run length refers to the number of consecutive normal operation cycles between the end of one abnormality and the beginning of the next. Whenever a quality abnormality is detected, the current consecutive normal operation count is recorded as a run length value, and then the counting of the next run restarts. The cumulative summation of all recorded run length values ​​is performed by adding the sum and dividing the result by the total number of runs to obtain the average run length. The average run length reflects the average stable operating capability of the silver powder production process.

[0112] Error control action identification and processing is based on the equipment control action records in the silver powder quality control execution results to identify abnormal control. The identification algorithm judges the correctness of the control action by comparing the expected execution result of the control command with the actual execution result. The equipment control action record contains information such as the sending time of each control command, target parameters, actual achieved parameters, and execution status. The judgment criteria for error control actions include control command execution failure, deviation of actual parameters from target parameters exceeding the allowable range, and control actions causing equipment abnormalities. The number of misadjusted control actions is counted by traversing all equipment control action records and checking whether each control action meets the error judgment criteria. The cumulative count of control actions that meet the criteria is obtained as the number of misadjusted control actions. The total number of equipment control actions is the total number of equipment control action records. The result of the division of the number of misadjusted control actions and the total number of equipment control actions is the equipment misadjustment rate. The misadjustment rate reflects the reliability and accuracy level of equipment control.

[0113] The item-by-item comparison and judgment process involves a one-to-one numerical comparison between the actual silver powder quality data and the expected silver powder quality data in the silver powder quality control execution results. The comparison process includes data matching across multiple quality dimensions, such as particle size distribution, conductivity, and morphology. Actual silver powder quality data is derived from product inspection after the production process, while expected silver powder quality data is calculated based on production process parameters and a quality prediction model. The accuracy of the prediction is determined by calculating the relative error between the actual and expected data. A prediction is considered correct when the relative error is less than a preset threshold, and incorrect when it exceeds the threshold. The number of correct predictions is calculated by accumulating all correct predictions. The total number of quality predictions equals the total number of quality data items compared. The accuracy rate is obtained by dividing the number of correct predictions by the total number of quality predictions. The accuracy rate reflects the reliability and predictive ability of the quality prediction model.

[0114] The difference calculation process involves subtracting the average run length, equipment misalignment rate, and quality prediction accuracy from their respective preset standard values. These preset standard values ​​are determined based on the quality requirements of silver powder production and historical statistical data. The preset standard value for average run length is typically set to more than 100 consecutive normal operating cycles, the preset standard value for equipment misalignment rate is set to less than 5%, and the preset standard value for quality prediction accuracy is set to more than 90%. The difference between the actual calculated value and the standard value reflects the degree of deviation between the current control performance and the expected target. The performance deviation value is calculated by weighting the three differences according to their importance. The weight for average run length deviation is set to 0.4, the weight for equipment misalignment rate deviation is set to 0.3, and the weight for quality prediction accuracy deviation is set to 0.3. The weighted sum is the comprehensive performance deviation value.

[0115] Numerical adjustment processing corrects key control parameters in the collaborative control instruction set of the silver powder production equipment based on the positive or negative direction and magnitude of the performance deviation. The adjustment strategy uses a proportional adjustment method to convert performance deviations into parameter correction values. When the average run length deviation is negative, it indicates insufficient system stability and requires reducing control sensitivity. In this case, the response time constant of the ball mill speed control is increased, the proportional gain of the atomizing drying tower temperature control is decreased, and the adjustment amplitude of the grading and screening equipment frequency control is reduced. When the equipment misadjustment rate deviation is positive, it indicates insufficient control accuracy and requires optimization of control parameters. Adjustments include recalibrating sensor measurement accuracy, correcting the dead zone range of the control algorithm, and updating equipment response characteristic parameters. When the quality prediction accuracy deviation is negative, it indicates that the prediction model needs improvement. Adjustment measures include increasing sample data for the quality characteristic correlation matrix, retraining the weight parameters of the prediction algorithm, and optimizing the weight allocation of multi-technology fusion decision-making. The corrected collaborative control instruction set of the silver powder production equipment includes all adjusted control parameter values.

[0116] The above describes the silver powder production quality control method integrating multiple technologies in the embodiments of this application. The following describes the silver powder production quality control system integrating multiple technologies in the embodiments of this application. Please refer to [link / reference]. Figure 3 One embodiment of the silver powder production quality control system integrating multiple technologies in this application includes:

[0117] The data acquisition module is used to acquire and process real-time data from the silver powder production line through a distributed sensor network to obtain a correlation matrix of silver powder quality characteristics.

[0118] The adjustment module is used to dynamically adjust the hyperparameters of the gamma prior distribution based on the defect rate parameter in the silver powder quality characteristic correlation matrix using an adaptive Bayesian algorithm, so as to obtain the adaptive shape parameter and the adaptive rate parameter.

[0119] The processing module is used to optimize the silver powder production control strategy based on the adaptive shape parameters and adaptive rate parameters through a decision fusion algorithm, and obtain a set of collaborative control instructions for silver powder production equipment.

[0120] The control module is used to perform linkage control on the ball mill, atomizing drying tower and grading and screening equipment according to the collaborative control instruction set of the silver powder production equipment, so as to obtain the silver powder quality control execution result;

[0121] The correction module is used to evaluate the effect of the silver powder quality control execution results, obtain a set of performance indicators, and correct the collaborative control instruction set of the silver powder production equipment according to the set of performance indicators.

[0122] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the silver powder production quality control method integrating multiple technologies.

[0123] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for multi-technology fusion quality control of silver powder production, characterized by, The method comprises: Real-time data acquisition and processing of the silver powder production line through a distributed sensor network to obtain a silver powder quality characteristic correlation matrix; According to the defect rate parameter in the silver powder quality characteristic correlation matrix, the prior distribution hyperparameter of gamma is dynamically adjusted through an adaptive Bayesian algorithm to obtain adaptive shape parameters and adaptive rate parameters, including: threshold judgment processing based on the defect rate parameter in the silver powder quality characteristic correlation matrix to obtain a defect rate parameter state identifier, wherein when the defect rate parameter is greater than a preset threshold, the defect rate parameter state identifier is in a high-risk state; shape parameter calculation processing according to the defect rate parameter state identifier and the basic shape parameter through a logarithmic mapping algorithm to obtain adaptive shape parameters, wherein the adaptive shape parameters are equal to the sum of the product of the basic shape parameters and the adjustment coefficient and the logarithmic value of the defect rate parameter; the adaptive shape parameters and the expected value of the defect rate parameter are input into the rate parameter calculation formula to solve the adaptive rate parameters, wherein the adaptive rate parameters are equal to the ratio of the adaptive shape parameters to the product of the expected value of the defect rate parameter and the adjustment coefficient; gamma distribution parameter verification processing based on the adaptive shape parameters and the adaptive rate parameters to obtain a parameter validity identifier, and when the parameter validity identifier is invalid, the shape parameter calculation processing and the rate parameter solving are re-executed; According to the adaptive shape parameters and the adaptive rate parameters, the silver powder production control strategy is optimized through a decision fusion algorithm to obtain a silver powder production equipment collaborative control instruction set; According to the silver powder production equipment collaborative control instruction set, the ball mill, the atomization drying tower and the classification screening equipment are linked and controlled to obtain a silver powder quality control execution result; Effect evaluation is performed on the silver powder quality control execution result to obtain a performance index set, and the silver powder production equipment collaborative control instruction set is corrected according to the performance index set.

2. The multi-technology fused silver powder production quality control method according to claim 1, characterized by, The silver powder quality characteristic correlation matrix includes particle size defect rate parameters, conductivity performance parameters and morphology feature data, the real-time data acquisition and processing of the silver powder production line through a distributed sensor network to obtain a silver powder quality characteristic correlation matrix, including: Real-time detection and processing of the particle size distribution of silver powder through a laser particle size analyzer at the discharge port of the ball mill to obtain a particle size distribution data set, and defect statistical calculation processing based on the particle size distribution data set to obtain particle size defect rate parameters; Continuous monitoring and processing of the conductivity performance of silver powder through a conductivity sensor at the feed end of the atomization drying tower to obtain conductivity performance time series data, and filtering and noise reduction processing of the conductivity performance time series data to obtain conductivity performance parameters; Silver powder morphology image data is obtained by collecting and processing silver powder morphology images through a machine vision detection unit at the back end of the classification screening equipment, and the silver powder morphology image data is input into an image feature extraction algorithm for morphology analysis processing to obtain morphology feature data; The particle size defect rate parameter, the conductive performance parameter and the morphology characteristic data are input into a correlation analysis algorithm for coupling relationship calculation processing, so as to obtain a silver powder quality characteristic correlation matrix.

3. The multi-technology fused silver powder production quality control method according to claim 1, characterized by, The gamma distribution parameter verification processing based on the adaptive shape parameter and the adaptive rate parameter comprises: The adaptive shape parameter is subjected to numerical range inspection processing, so as to obtain a shape parameter validity mark, wherein when the adaptive shape parameter is less than zero or greater than a preset upper limit value of the shape parameter, the shape parameter validity mark is invalid; The adaptive rate parameter is subjected to numerical range inspection processing, so as to obtain a rate parameter validity mark, wherein when the adaptive rate parameter is less than zero or greater than a preset upper limit value of the rate parameter, the rate parameter validity mark is invalid; The gamma distribution convergence inspection processing based on the adaptive shape parameter and the adaptive rate parameter comprises: the mean and variance ratio of the gamma distribution is calculated to determine whether the convergence is within a preset convergence interval, so as to obtain a distribution convergence judgment result; Logical and operation processing is performed on the shape parameter validity mark, the rate parameter validity mark and the distribution convergence judgment result, so as to obtain a parameter validity identifier.

4. The multi-technology fused silver powder production quality control method according to claim 1, characterized by, The adaptive shape parameter and the adaptive rate parameter are input into a Bayesian control decision module for probability distribution calculation processing, so as to obtain a silver powder quality control probability distribution function, wherein the silver powder quality control probability distribution function is constructed based on a gamma distribution. The conductive performance parameter and the morphology characteristic data in the silver powder quality characteristic correlation matrix are subjected to multi-dimensional feature fusion processing, so as to obtain a silver powder quality comprehensive evaluation index. The silver powder quality control probability distribution function and the silver powder quality comprehensive evaluation index are input into a three-layer fusion decision architecture for hierarchical decision processing, so as to obtain a fusion decision weight vector, wherein the three-layer fusion decision architecture comprises data layer fusion, feature layer fusion and decision layer fusion. The ball mill rotation speed control parameter, the atomization drying tower temperature control parameter and the classification screening equipment frequency control parameter are subjected to collaborative optimization calculation processing according to the fusion decision weight vector, so as to obtain a device control parameter optimization vector. The device control parameter optimization vector is subjected to instruction packaging processing, so as to obtain a silver powder production equipment collaborative control instruction set, wherein the silver powder production equipment collaborative control instruction set comprises a ball mill control instruction, an atomization drying tower control instruction and a classification screening equipment control instruction. The ball mill, the atomization drying tower and the classification screening equipment are subjected to linkage control according to the silver powder production equipment collaborative control instruction set, so as to obtain a silver powder quality control execution result, comprising:

5. The multi-technology fused silver powder production quality control method according to claim 1, characterized by, The ball mill rotation speed is adjusted and controlled according to the ball mill control instruction in the silver powder production equipment collaborative control instruction set, so as to obtain ball mill operation state data, wherein the ball mill operation state data comprises an actual rotation speed value, a current consumption value and a vibration amplitude value. ​ According to the atomization drying tower control instruction in the silver powder production equipment cooperative control instruction set, the feeding temperature and the discharging temperature of the atomization drying tower are synchronously adjusted and processed, and the atomization drying tower running state data is obtained, wherein the atomization drying tower running state data includes the feeding temperature measured value, the discharging temperature measured value and the humidity control value; According to the classification screening equipment control instruction in the silver powder production equipment cooperative control instruction set, the screening frequency and the screen mesh spacing of the classification screening equipment are coordinately adjusted and processed, and the classification screening equipment running state data is obtained, wherein the classification screening equipment running state data includes the screening frequency measured value, the screen mesh spacing measured value and the screening efficiency value; The ball mill running state data, the atomization drying tower running state data and the classification screening equipment running state data are integrated, and the silver powder quality control execution result is obtained, wherein the silver powder quality control execution result includes the equipment linkage execution record and the silver powder product quality detection data.

6. The multi-technology-fused silver powder production quality control method according to claim 1, characterized by, The silver powder quality control execution result is evaluated, the performance index set is obtained, and the silver powder production equipment cooperative control instruction set is corrected according to the performance index set, wherein the performance index set includes: the average run length, the equipment misadjustment rate and the quality prediction accuracy rate, including: Based on the quality control process data in the silver powder quality control execution result, the number of continuous normal operation times is statistically processed, the respective run length values are obtained, and the average run length is obtained by performing accumulation summation processing on the respective run length values and then dividing by the total number of runs; According to the error control action identification processing of the equipment control action record in the silver powder quality control execution result, the misadjustment control times are obtained, and the equipment misadjustment rate is obtained by performing division operation processing on the misadjustment control times and the total control times of the equipment; The actual silver powder quality data and the expected silver powder quality data in the silver powder quality control execution result are compared and judged item by item, the quality prediction correct judgment times are obtained, and the quality prediction accuracy rate is obtained by performing division operation processing on the quality prediction correct judgment times and the total number of quality predictions; According to the average run length, the equipment misadjustment rate and the quality prediction accuracy rate, the performance deviation values are obtained by performing difference calculation processing on the respective preset standard values; Based on the performance deviation values, the control parameters in the silver powder production equipment cooperative control instruction set are adjusted in value, and the corrected silver powder production equipment cooperative control instruction set is obtained.

7. A multi-technology fused silver powder production quality control system, characterized by, The fusion multi-technology silver powder production quality control method as claimed in any one of claims 1 to 6, the fusion multi-technology silver powder production quality control system comprises: A collection module is configured to collect real-time data of a silver powder production line through a distributed sensor network, and obtain a silver powder quality characteristic correlation matrix; An adjustment module is configured to dynamically adjust hyperparameters of a gamma prior distribution through an adaptive Bayesian algorithm according to a defect rate parameter in the silver powder quality characteristic correlation matrix, and obtain adaptive shape parameters and adaptive rate parameters; An adjustment module is configured to dynamically adjust hyperparameters of a gamma prior distribution through an adaptive Bayesian algorithm according to a defect rate parameter in the silver powder quality characteristic correlation matrix, and obtain adaptive shape parameters and adaptive rate parameters; The processing module is configured to perform optimization processing on the silver powder production control strategy by a decision fusion algorithm according to the adaptive shape parameter and the adaptive rate parameter, and obtain a set of silver powder production equipment collaborative control instructions. The control module is configured to perform linkage control on the ball mill, the atomization drying tower and the grading screening equipment according to the set of silver powder production equipment collaborative control instructions, and obtain a silver powder quality control execution result. The correction module is configured to perform effect evaluation on the silver powder quality control execution result, obtain a set of performance indicators, and perform instruction correction on the set of silver powder production equipment collaborative control instructions according to the set of performance indicators.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that The computer program, when executed by a processor, causes the processor to perform the silver powder production quality control method of fusing multiple technologies according to any one of claims 1 to 6.

Images