An on-line control system and method for premix production quality

Abstract

The application discloses an online control system and method for premix production quality, and relates to the technical field of premix production.The system comprises an information acquisition module, which acquires process parameters and environmental parameters of a premix production process; an environmental perception module, which performs spatial discretization processing on a production environment, constructs a spatial distribution model of the production environment, and identifies an environmental sensitive area; a judgment condition generation module, which dynamically adjusts a first quality judgment condition based on the environmental sensitive area, and generates a second quality judgment condition; a quality measurement and evaluation module, which acquires actual production quality index data, combines the second quality judgment condition, evaluates the production quality, and outputs a quality evaluation result; and a control module, which generates a process adjustment strategy for premix production based on the quality evaluation result, and adjusts process parameters. Through fine spatial perception and dynamic quality judgment on the production environment, the application improves the quality control precision and product stability of premix production.

Description

Technical Field

[0001] This application relates to the field of premix production control technology, specifically to an online control system and method for premix production quality. Background Technology

[0002] Premixes are homogeneous mixtures made by mixing one or more trace components with a carrier or diluent in a specified ratio. Their quality indicators, such as mixing uniformity, active ingredient content, and particle size distribution, directly affect the safety and effectiveness of downstream products.

[0003] Currently, quality control in the premix production process mainly employs two methods: offline sampling and testing, and fixed process parameter monitoring. Offline sampling involves taking samples from the production line at fixed intervals and sending them to a laboratory for physicochemical analysis to determine quality indicators. This method suffers from latency; the time from sampling and testing to result feedback can take hours or even days. During this period, substandard products may continue to be produced, leading to batch scrap and increased production costs. Furthermore, sampling only reflects a localized, instantaneous quality status and cannot cover quality fluctuations throughout the entire production process, easily resulting in situations where samples pass but the actual batch fails. Fixed process parameter monitoring indirectly determines quality by monitoring whether core process parameters such as stirring speed, mixing time, and raw material dosage are within preset ranges. While this method provides real-time feedback on parameter status, it only focuses on the process parameters themselves, neglecting the impact of the production environment on the material's mixing state and component stability. Moreover, the parameter thresholds are uniformly fixed, failing to consider the varying sensitivities to parameter fluctuations in different areas of the production equipment. Summary of the Invention

[0004] To address the aforementioned shortcomings, this invention provides an online control system and method for the production quality of premixed agents. This system couples production processes with environmental factors and dynamically adapts to quality judgment standards, solving the problems of low precision in premixed agent production quality control and insufficient product stability in existing technologies. It enables real-time monitoring, evaluation, and dynamic adjustment of quality indicators during the premixed agent production process, thereby improving the stability of product quality.

[0005] In a first aspect, this application discloses an online control system for the production quality of premixes, comprising:

[0006] The information acquisition module is used to acquire process parameters and environmental parameters of the premix production process;

[0007] The environmental perception module is used to construct a spatial distribution model of the production environment based on the process parameters and environmental parameters, and to identify environmentally sensitive areas of the premix production equipment.

[0008] The judgment condition generation module is used to dynamically adjust the local threshold of the first quality judgment condition for premix production based on the environmentally sensitive area, and generate the second quality judgment condition.

[0009] The quality measurement and evaluation module is used to acquire actual production quality index data of the premix, and evaluate the production quality based on the second quality judgment condition, and output the quality evaluation result.

[0010] The control module is used to generate a process adjustment strategy for the production of premix based on the quality assessment results, and to adjust the process parameters based on the process adjustment strategy.

[0011] Optionally, the process parameters include stirring speed, mixing time, material feed rate, and mixing temperature; the environmental parameters include workshop temperature, humidity, and dust concentration.

[0012] Optionally, the construction of the spatial distribution model of the production environment includes:

[0013] The premix production workshop and the area where the production equipment is located are spatially discretized into multiple continuous spatial units, and the types and functional attributes of the equipment components included in the spatial units are associated to form a basic spatial framework.

[0014] By deploying sensing devices in each spatial unit, real-time data of process parameters and environmental parameters are collected synchronously. The correspondence between parameter data and spatial units is established based on timestamps, forming a multidimensional dataset that includes spatial location, time dimension and parameter values.

[0015] Based on the aforementioned multidimensional dataset, an interpolation algorithm is used to process the discrete parameter data to generate a continuous process parameter field and environmental parameter field covering the entire production area.

[0016] By fusing process parameter fields and environmental parameter fields through a multi-field coupling algorithm, a spatial distribution model of the production environment is constructed.

[0017] Optionally, the identification of environmentally sensitive areas of the premix production equipment includes:

[0018] Based on the spatial distribution model, the coupling characteristics of process parameters and environmental parameters of each spatial unit are extracted, and the influence weight of each parameter fluctuation on the quality of the premix is ​​determined.

[0019] A parameter sensitivity assessment model was established, and the parameter fluctuation range of each spatial unit was compared with the historical correlation data of the premix quality index to calculate the comprehensive sensitivity index of each spatial unit.

[0020] Based on the functional attributes of the equipment components, cluster analysis is performed on spatial units whose comprehensive sensitivity index exceeds a preset threshold to form the identification results of environmentally sensitive areas of the premix production equipment, including the boundaries of the environmentally sensitive areas, sensitivity levels, main influencing parameters, and associated equipment components.

[0021] Optionally, the generation of the second quality judgment condition includes:

[0022] The basic threshold system in the first quality judgment condition of premix production is extracted, including the judgment threshold of the quality indicators of premix mixing uniformity, particle size, and component ratio;

[0023] Based on the identification results of environmentally sensitive areas, the basic threshold system is mapped and associated with the sensitivity level, main influencing parameters and related equipment components of the environmentally sensitive areas to determine the quality indicators and corresponding area ranges for local adjustments.

[0024] Differentiated adjustment strategies are adopted for regions with different sensitivity levels. The adjusted local thresholds of each region are integrated to obtain the judgment criteria for quality indicators, forming the second quality judgment condition.

[0025] Optionally, obtaining the actual production quality index data of the premix includes:

[0026] Deploy online detection devices at key process nodes in premix production to cover production stages in environmentally sensitive areas;

[0027] The raw data of quality indicators are collected by the online detection device, preprocessed, and the heterogeneous data output by different online detection devices are converted into standardized data with a unified dimension to obtain the quality indicator data of the actual production of the premix.

[0028] Establish the correspondence between quality indicator data, process parameters and environmental parameters based on timestamps, and construct a comprehensive data set including quality indicators, process status, environmental status and environmentally sensitive area attributes.

[0029] Optionally, the output quality assessment results include:

[0030] The actual production quality index data are classified according to the environmentally sensitive area attributes, and compared with the judgment criteria of the corresponding area in the second quality judgment condition. The deviation of the actual value of each quality index from the corresponding judgment criterion is calculated.

[0031] Based on the identification results of environmentally sensitive areas, and combined with the comprehensive sensitivity index of each spatial unit and the influence weight of parameter fluctuations on the quality of premix, the influence weight of each environmentally sensitive area on the quality of premix is ​​determined.

[0032] By combining the degree of deviation of quality indicators in the corresponding region, the comprehensive quality score of premix production is calculated by weighted summation, and the quality level is determined. The quality assessment results are output, including the degree of deviation of individual quality indicators and the quality level.

[0033] Optionally, determining the influence weight of each environmentally sensitive area on the quality of the premix includes:

[0034] Based on the affiliation between environmentally sensitive areas and spatial units, extract all spatial units for each environmentally sensitive area;

[0035] For each spatial unit, its comprehensive sensitivity index is coupled with the fluctuation impact weight of the main influencing parameters within the corresponding spatial unit to obtain the quality impact value of the corresponding spatial unit.

[0036] The quality impact values ​​of all spatial units within the same environmentally sensitive area are summed to form the initial impact weight of the corresponding environmentally sensitive area. After standardization, the quality impact weight of each environmentally sensitive area is obtained.

[0037] Optionally, adjusting the process parameters based on the process adjustment strategy includes:

[0038] Based on the deviation of individual quality indicators, comprehensive quality score and quality level of the quality assessment results, process adjustment strategies are generated for different scenarios.

[0039] The generated process adjustment strategy is parsed into target values ​​for specific process parameters, associated with the corresponding production execution mechanism, and the actual values ​​of process parameters and environmental parameters in environmentally sensitive areas are collected in real time through the information acquisition module, and the deviation between the actual values ​​and the target values ​​is compared.

[0040] If the deviation is within the allowable range, maintain the current adjustment range; if the deviation exceeds the allowable range, dynamically adjust the adjustment step size based on the quality impact weight of each environmentally sensitive area.

[0041] By monitoring quality indicator data in subsequent production cycles through the quality measurement and evaluation module, the effectiveness of process adjustment strategies in improving quality can be verified.

[0042] Secondly, this application discloses an online method for controlling the production quality of premixes, comprising:

[0043] Obtain process and environmental parameters for the premix production process, construct a spatial distribution model of the production environment, and identify environmentally sensitive areas of the premix production equipment;

[0044] Based on the environmentally sensitive area, the first quality judgment condition for premix production is dynamically adjusted locally to generate a second quality judgment condition.

[0045] Obtain actual production quality index data of the premix, and evaluate the production quality based on the second quality judgment condition, and output the quality evaluation result;

[0046] Based on the quality assessment results, a process adjustment strategy for the production of premixed agents is generated, and the process parameters are adjusted based on the process adjustment strategy.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] This invention constructs a spatial distribution model of the production environment, identifies environmentally sensitive areas, and solves the problem of relying on a single, static global threshold for quality judgment. It can dynamically adjust local thresholds based on the actual process parameters, environmental parameters, and their impact on product quality in different areas, thereby improving the accuracy and adaptability of quality assessment.

[0049] Secondly, through precise location and real-time monitoring of environmentally sensitive areas, the system can make targeted adjustments to process parameters. Based on the quality assessment results, it can associate specific sensitive areas and key influencing parameters to generate differentiated and tiered adjustment strategies, effectively correcting quality deviations and improving the stability of product quality and batch-to-batch consistency. Attached Figure Description

[0050] Figure 1 A schematic diagram of an online control system for the production quality of a premix provided in this application embodiment;

[0051] Figure 2 A flowchart for identifying environmentally sensitive areas of a premix production equipment provided in this application embodiment;

[0052] Figure 3 A flowchart for generating the second quality judgment condition provided in the embodiments of this application;

[0053] Figure 4 A flowchart illustrating an online method for controlling the production quality of a premix, as provided in an embodiment of this application. Detailed Implementation

[0054] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0055] Existing premix production quality control systems mostly rely on fixed process parameter thresholds for quality judgment, without considering the spatial coupling relationship between production environment parameters and process parameters. This results in a delayed response to quality fluctuations in local environmentally sensitive areas and insufficient overall production quality stability.

[0056] To address the aforementioned shortcomings, this application discloses an online control system for the production quality of premixed agents. By constructing a three-dimensional spatial distribution model of the production environment, it identifies environmentally sensitive areas that influence product quality. Based on the actual process parameters, environmental parameters, and their impact on quality in different sensitive areas, it locally and dynamically adjusts the thresholds for quality judgment, generating adaptive quality assessment standards. This accurately and efficiently assesses the current production status and generates targeted process adjustment strategies based on the assessment results. Through deep perception of the production environment and adaptive adjustment of quality control, this application effectively overcomes the shortcomings of existing technologies, such as delayed response and coarse control, improving the stability and batch consistency of premixed agent production quality.

[0057] See Figure 1 This is a schematic diagram of an online control system for the production quality of a premix provided in an embodiment of this application, including:

[0058] The information acquisition module 10 is used to acquire the process parameters and environmental parameters of the premix production process;

[0059] The environmental perception module 20 is used to construct a spatial distribution model of the production environment based on the process parameters and environmental parameters, and to identify environmentally sensitive areas of the premix production equipment.

[0060] The judgment condition generation module 30 is used to dynamically adjust the local threshold of the first quality judgment condition for premix production based on the environmentally sensitive area, and generate the second quality judgment condition.

[0061] The quality measurement and evaluation module 40 is used to acquire actual production quality index data of the premix, and evaluate the production quality based on the second quality judgment condition, and output the quality evaluation result.

[0062] The control module 50 is used to generate a process adjustment strategy for the production of premix based on the quality assessment results, and to adjust the process parameters based on the process adjustment strategy.

[0063] In practical implementation, the information acquisition module 10 continuously and in real time collects various process and environmental parameters during the premix production process, and performs preliminary data integration and preprocessing. Specifically, this module comprehensively captures various data related to the premix production process through interfaces with various sensors and instruments deployed on the production site, as well as existing production control systems, such as programmable logic controllers (PLCs) or distributed control systems (DCS).

[0064] Regarding process parameters, the key focus is on obtaining the stirring speed, which can be acquired through a Hall effect sensor or encoder on the output shaft of the stirring motor, measuring revolutions per minute (RPM). The mixing time can be synchronously recorded by a timer integrated into the PLC control system, starting from the moment the material is fully added to the mixing chamber and automatically stopping when the mixing process ends. The material feed rate can be acquired through weighing sensors deployed at the discharge ports of the main and auxiliary material silos, providing real-time feedback on the single feed rate and cumulative feed rate. The mixing temperature can be continuously measured using temperature sensors deployed inside the mixing container or along the material path.

[0065] Regarding environmental parameters, the temperature and humidity of the workshop are collected by temperature and humidity sensors deployed along the central axis of the production area to capture the environmental conditions of different areas of the workshop; the dust concentration is monitored by a dust detector installed on the top of the workshop, which uses technologies such as laser scattering to detect the real-time concentration of suspended particulate matter in the air.

[0066] Data transmission can employ a combination of industrial Ethernet and wireless LoRa technology. Wired transmission is used to connect fixed devices, such as load cells. The information acquisition module 10 communicates directly with these devices through a network port, periodically polling or subscribing to their output values. Wireless transmission is used in mobile or difficult-to-wire areas, such as workshop temperature and humidity sensors. These sensor nodes package the collected data into LoRa data frames and transmit them wirelessly to a LoRa gateway located in the workshop. This gateway receives signals from multiple LoRa nodes, performs preliminary data decoding, and then sends the aggregated data to the network where the information acquisition module 10 is located via a wired Ethernet connection for unified reception and processing.

[0067] The information acquisition module 10 can also communicate with the PLC or DCS system on the production site through the OPC UA (OLE for Process Control Unified Architecture) server interface to read the process parameters that have been monitored and controlled in the PLC / DCS, such as mixing time and stirring speed recorded by the PLC's internal timer.

[0068] After the data is transmitted to the information acquisition module 10, the module performs data integration and preprocessing. First, all incoming data is marked using a timestamp mechanism, with sensors or gateways adding timestamps when the data is generated. Next, streaming data from different data sources are aggregated to form a unified data stream. Noise reduction and data smoothing are performed through data filtering to remove invalid data that is significantly outside the reasonable physical range, resulting in preprocessed process and environmental parameter values.

[0069] Existing quality control systems struggle to fully and accurately perceive and understand the complex and dynamic spatial distribution characteristics of the production environment, thus failing to accurately identify the sensitivity of equipment or areas to changes in environmental parameters. This can lead to misjudgments or omissions based on globally unified quality judgment conditions.

[0070] Therefore, the environmental perception module 20 uses the multi-dimensional sensor data input by the information acquisition module 10 to construct a spatial distribution model of the production environment and identify the environmentally sensitive areas of the premix production equipment.

[0071] First, the environmental perception module 20 uses the existing three-dimensional geometric model of the workshop as a basis. The three-dimensional geometric model of the workshop can be generated by importing standard CAD files or by directly modeling with professional software based on accurate measurement data. The model includes the geometric outline of the workshop walls, floor and ceiling, as well as the three-dimensional shape, size and preset installation position of all fixed and semi-fixed production equipment, and the three-dimensional installation position of all kinds of sensors deployed in the workshop.

[0072] Based on this, the environmental perception module 20 performs three-dimensional discretization of the physical space of the entire production workshop. This can be achieved using voxelization or mesh generation techniques, dividing the premix production workshop and its equipment area into several non-overlapping, continuous spatial units with fixed geometric boundaries. The size of each spatial unit can be set according to actual production needs and equipment layout. For example, critical mixing areas can be divided into smaller units for monitoring. Each spatial unit is associated with its main production equipment components, such as agitators, feeders, conveyor belts, and sensor locations, as well as their functional attributes, such as the power of the agitator, the accuracy range of the feeder, and the measurement accuracy of the temperature sensor, forming a basic spatial framework.

[0073] Next, after receiving the timestamp, process parameter values, and environmental parameter values ​​provided by the information acquisition module 10, the environmental perception module 20 associates the numerical parameters with specific units or equipment components in the spatial framework based on the preset physical coordinates of the sensors or production equipment within the workshop. For example, temperature data received at a certain point will be linked to the spatial unit where that point is located and the temperature sensor information within that unit. In this way, each data record is not only time and value, but also extended to include spatial location, timestamp, parameter type, and value. By integrating the discrete data collected by all sensors over a period of time with time as the basis, a multidimensional dataset including spatial location, time dimension, and parameter values ​​is generated.

[0074] Due to the limited number of sensors, discrete point-like data is obtained. To understand the parameter distribution across the entire production area, inferences need to be made between discrete points to generate a continuously distributed field. The environmental sensing module 20 employs spatial interpolation algorithms, such as Kriging interpolation, inverse distance weighted interpolation, and radial basis function interpolation. Kriging interpolation can be selected, and a spherical model is used for the variogram. Parameters such as range, nugget value, and sill value are determined through cross-validation. Based on the known spatial location, time, and value of the sensor data points, the environmental sensing module 20 performs smooth interpolation calculations across the entire production space to predict parameter values ​​at non-data acquisition points. Through interpolation calculations, a continuous process parameter field and environmental parameter field covering the entire production area are generated, such as the mixed temperature field T(x, y, z, t) and the dust concentration field M(x, y, z, t), where (x, y, z, t) represents the process or environmental parameter state at each coordinate point (x, y, z) in the production space at time t.

[0075] Taking into account the interaction and coupling effects of process parameters and environmental parameters within different spatial units—for example, when workshop humidity is high, the agglomeration characteristics of dust may be enhanced, further affecting mixing uniformity—this application embodiment constructs a spatial distribution model of the production environment using a multi-field coupling algorithm. Specifically, firstly, based on the generated continuous process parameter field and environmental parameter field, such as the mixing temperature field T(x, y, z, t) and dust concentration field M(x, y, z, t), and the basic spatial framework, the production area is uniformly divided into three-dimensional grids, with each grid node serving as an interpolation target point. After generating a continuous three-dimensional parameter field for each parameter using Kriging interpolation, the correlation degree at each grid node of any two parameter fields is calculated using the Pearson correlation coefficient. Strongly correlated parameter pairs with high absolute values ​​of correlation coefficients are selected, and parameter interaction terms, such as environmental temperature and humidity, are constructed to quantify their synergistic effects. Simultaneously, the spatial gradient of each parameter field is calculated using the first-order difference method. Subsequently, a... A multiple linear regression model, using premix quality indicators as target variables, incorporates parameter field values, interaction terms, and spatial gradient information as predictor variables. The model coefficients are solved using the least squares method based on historical production data. Finally, for each grid node, four types of features are extracted: parameter field values, spatial gradient, correlation coefficient, and interaction term coefficients. These are combined to form a coupled feature vector, which is then normalized to generate a spatial distribution model of the production environment. This spatial distribution model is a digital twin framework of the production environment, using a continuous parameter field as its basic representation. A multi-field coupling algorithm is then used to further analyze the impact of the parameter field on the actual production quality indicators of the premix, thereby identifying the regions most sensitive to parameter changes.

[0076] See Figure 2 A flowchart for identifying environmentally sensitive areas of a premix production equipment provided in this application embodiment includes:

[0077] S101, Based on the spatial distribution model, extract the coupling characteristics of process parameters and environmental parameters of each spatial unit, and determine the influence weight of each parameter fluctuation on the quality of the premix.

[0078] S102, Establish a parameter sensitivity assessment model, compare the parameter fluctuation range of each spatial unit with the historical correlation data of the premix quality index, and calculate the comprehensive sensitivity index of each spatial unit;

[0079] S103, combining the functional attributes of the equipment components, perform cluster analysis on the spatial units whose comprehensive sensitivity index exceeds the preset threshold to form the identification results of the environmentally sensitive areas of the premix production equipment, including the boundary of the environmentally sensitive area, the sensitivity level, the main influencing parameters and the associated equipment components.

[0080] In practice, after the spatial distribution model is constructed, the environmental perception module 20 continuously collects and processes the process parameters and environmental parameters of each spatial unit, analyzes the relationship between parameter fluctuations and historical premix quality indicators, quantifies the degree of influence of each process parameter and environmental parameter on the premix quality indicators when they change in different spatial units, and generates influence weights.

[0081] Specifically, the quantification process first collects historical operational data from multiple batches of premixed agent production, including process parameters, corresponding environmental parameters, and premixed agent quality index data. For each selected premixed agent quality index, a multiple regression model can be constructed, using the quality index as the target variable. The values ​​of process and environmental parameters collected from various spatial units within the same production batch or time period as this quality index, along with their interaction terms and spatial gradient information, are used as predictor variables. The model is fitted using the least squares method or other standard regression algorithms, and the regression coefficients of each predictor variable relative to the target variable are obtained through the fitting process. Based on this, for each parameter and each spatial unit, its fluctuation range in multiple batches of production data is analyzed; for example, the standard deviation is calculated to measure the dispersion of the parameter values.

[0082] Multiply the absolute value of the calculated regression coefficient by its standard deviation to calculate a potential impact value. Then, normalize the calculated potential impact values ​​of all parameters. Specifically, for the i-th parameter, its impact weight is equal to the product of the absolute value of its regression coefficient and its standard deviation, divided by the sum of the products of the absolute values ​​of the regression coefficients and their standard deviations of all parameters.

[0083] Next, the environmental perception module 20 establishes a parameter sensitivity assessment model. By comparing the actual fluctuation range of parameters in each spatial unit with historical correlation data of premix quality indicators, it calculates the comprehensive sensitivity index of each spatial unit.

[0084] Specifically, the parameter sensitivity assessment model can adopt a multilayer perceptron architecture. The network structure includes an input layer, multiple hidden layers, and an output layer. The number of neurons in the input layer matches the dimension of the input features. The input features cover the fluctuation range of core process parameters and environmental parameters, the preset influence weights of each parameter, and the spatial unit equipment component type features after encoding. A Batch Normalization layer is set to standardize the input. The hidden layer consists of two fully connected layers with ReLU activation function. A dropout layer is added to each layer to progressively extract the nonlinear correlation features between parameter fluctuations and premixed agent quality indicators, and to fuse and optimize the dimensions of the features. The output layer has one neuron using the Sigmoid function to output the comprehensive sensitivity index of a single spatial unit.

[0085] Model construction is divided into several stages: data preparation, preprocessing, training, and validation / optimization. In the data preparation stage, multiple batches of data from normal production in recent years are collected, including production data from different seasons, different raw material batches, and different equipment operating conditions. Each batch of data is split into spatial units, with each spatial unit serving as one sample. Each sample includes an input feature set and a label, and is proportionally divided into training, validation, and test sets. In the data preprocessing stage, continuous features, such as parameter fluctuation amplitude, are standardized using Z-Score; discrete features, such as equipment component types, are converted into binary vectors using one-hot encoding. Outliers in the input features are removed using the 3σ rule. In the model training stage, model parameters are initialized using a He normal distribution, with an initial learning rate of 0.001. The Adam optimizer is selected, and the mean squared error loss function is used. In the model validation and optimization stage, mean absolute error (MAE) and coefficient of determination (R²) are used as the core validation metrics.

[0086] The parameter sensitivity assessment model first assesses the actual fluctuations of process and environmental parameters within each spatial unit over a specific operating cycle. The standard deviation of these parameters can be used to quantify their short-term stability. For example, the standard deviations of parameters such as material input, mixing temperature, workshop temperature and humidity, and dust concentration within the unit can be calculated over a period, such as the past hour or a single production batch. Then, these local parameter standard deviations are multiplied by their corresponding influence weights, and the calculation results for all relevant parameters are summed to obtain the comprehensive sensitivity index for the spatial unit. A higher comprehensive sensitivity index indicates that the parameter combination within the spatial unit is more unstable in actual operation, has a greater impact on quality, and is more sensitive to production quality.

[0087] Based on the comprehensive sensitivity index of each spatial unit, cluster analysis is performed on spatial units whose comprehensive sensitivity index exceeds a preset threshold. Specifically, combining the functional attributes of equipment components, the location information of the spatial units is overlaid with the layout diagram of the production equipment to identify the specific equipment area where each spatial unit with a comprehensive sensitivity index exceeding the preset threshold is located, and this area is then associated with a specific production equipment component. The preset threshold is a critical value pre-set based on historical quality data, process standards, and risk control targets for premix production. Specifically, it can be based on historical data statistics, collecting normal production data for a certain period, such as 3 months, calculating the comprehensive sensitivity index of all spatial units, and taking the mean plus the standard deviation multiple corresponding to the statistical significance level as the preset threshold.

[0088] Subsequently, clustering algorithms, such as DBSCAN clustering or K-means clustering, are used to divide spatially adjacent units with high comprehensive sensitivity indices into one or more environmentally sensitive regions. Based on the average or distribution of the comprehensive sensitivity index within each region, these regions are classified into different sensitivity levels, such as high, medium, and low sensitivity. Specifically, the arithmetic mean and dispersion (variance, range) of the comprehensive sensitivity index are first calculated for each clustered region. Then, using a preset threshold as a benchmark, the regions are classified according to the proportion by which the average value exceeds the preset threshold. Regions with average values ​​significantly exceeding the preset threshold by a high proportion and low dispersion are classified as high-sensitivity regions, where parameter fluctuations have a direct and severe impact on quality. Regions with average values ​​exceeding the preset threshold by a moderate proportion and moderate dispersion are classified as medium-sensitivity regions, where parameter fluctuations have a clear but controllable impact on quality. Regions with average values ​​slightly exceeding the preset threshold by a low proportion and high dispersion are classified as low-sensitivity regions, where parameter fluctuations have a relatively small impact on quality.

[0089] Finally, for each environmentally sensitive area, the main influencing parameters and associated equipment components are determined by analyzing the parameters that contribute the most to the comprehensive sensitivity index within that area; that is, the specific production equipment components located within the sensitive area. The main influencing parameters refer to the process or environmental parameters that contribute most significantly to the comprehensive sensitivity index of the environmentally sensitive area and have the most critical impact on the premix quality fluctuation. Specifically, for a single spatial unit, the product of the standard deviation of each parameter and its corresponding influence weight is the single contribution value of that parameter to the comprehensive sensitivity index of the corresponding spatial unit. After clustering to form sensitive areas, the single contribution values ​​of each parameter for all spatial units within the area are summarized to obtain the total regional contribution value of each parameter. Then, the proportion of the total regional contribution value of each parameter to the total comprehensive sensitivity index of the area is calculated. Parameters are sorted from high to low according to their contribution proportion, and those parameters whose cumulative contribution proportion reaches a preset proportion or ranks among the top are selected as the main influencing parameters of the sensitive area. The selected parameter types are consistent with the process and environmental parameters mentioned above.

[0090] The boundaries, sensitivity levels, main influencing parameters, and associated equipment components of the aforementioned environmentally sensitive areas constitute the complete identification results of the environmentally sensitive areas of the premix production equipment.

[0091] After the environmentally sensitive area is identified, the judgment condition generation module 30 dynamically adjusts the local threshold of the first quality judgment condition for premix production to generate the second quality judgment condition.

[0092] See Figure 3 The flowchart for generating the second quality judgment condition provided in the embodiments of this application includes:

[0093] S201, the basic threshold system in the first quality judgment condition of premix production, including the basic judgment threshold of quality indicators such as mixing uniformity, particle size, and component ratio of premix;

[0094] S202, Based on the identification results of environmentally sensitive areas, the basic threshold system is mapped and associated with the sensitivity level, main influencing parameters and related equipment components of the environmentally sensitive areas to determine the quality indicators and corresponding area range for local adjustments;

[0095] S203, adopt a differentiated adjustment strategy for areas with different sensitivity levels, integrate the adjusted local thresholds of each area to obtain the judgment criteria for quality indicators, and form the second quality judgment condition.

[0096] In practical implementation, the first step is to analyze the primary quality judgment criteria of the existing data, extracting basic judgment thresholds corresponding to quality indicators such as mixing uniformity, particle size, and component proportion. These thresholds can be unified judgment standards applicable to the entire production area, determined based on industry standards, production process requirements, and historical qualified batch production data. Specific judgment requirements include mixing uniformity ≥ a certain qualified standard value and component content error ≤ a certain allowable range. Subsequently, the judgment condition generation module 30, based on the environmentally sensitive area identification results, maps the sensitivity level, main influencing parameters, and associated equipment components of each area to the basic judgment thresholds.

[0097] The association mapping process is based on a preset rule base, which follows an IF-THEN logical structure. The rule settings integrate statistical analysis, such as regression models, the quantified weights of parameters' influence on quality indicators, and the inherent knowledge of domain experts based on production experience and the functional attributes of equipment components. Specifically, each rule defines a set of trigger conditions, i.e., the IF part, which matches the currently identified main influencing parameters with the core functions of the associated equipment components. When the IF condition is met, the THEN part of the rule specifies the specific action, clearly indicating the basic judgment threshold of the quality indicator that needs adjustment, which identified sensitive areas(s) it applies to, the direction of adjustment, and the degree of adjustment, such as significant tightening, moderate fine-tuning, or setting a dynamic range that changes in conjunction with real-time parameters. By querying and executing the preset rules, the judgment condition generation module 30 can associate and map the sensitivity level, main influencing parameters, and functional attributes of associated equipment in each area with the basic judgment threshold, thereby determining which quality indicators' basic judgment thresholds need adjustment in which sensitive areas, and the specific direction and degree of adjustment.

[0098] Based on the above mapping and correlation results, the judgment condition generation module 30 adopts differentiated adjustment strategies for regions with different sensitivity levels. Using a unified basic judgment threshold as the adjustment blueprint, it generates specific local judgment thresholds based on the sensitivity characteristics of each region. For example, for regions identified as highly sensitive, the adjustment strategy can tighten the previously relatively broad basic judgment threshold, or, based on real-time monitoring data of the main influencing parameters of the region, transform the basic judgment threshold into a dynamic adjustment range, forming a local judgment threshold for that region. For medium-sensitivity regions, the basic judgment threshold can be slightly tightened, or a smaller dynamic adjustment range can be set to generate a targeted local judgment threshold. For low-sensitivity or non-sensitive regions, the original basic judgment threshold can continue to be used as the quality judgment standard for that region.

[0099] After determining the adjustment strategies for each region, the judgment condition generation module 30 integrates the judgment criteria for all regions. Specifically, it gathers the judgment criteria from different regions to form a quality judgment standard that covers the entire production space but is optimized and refined for key areas. For non-sensitive areas, the basic judgment threshold is used; while for each identified sensitive area, a local judgment threshold adapted to the characteristics of the area, based on the adjustment of the basic judgment threshold, is stored separately, clarifying its applicable spatial range and triggering conditions. Finally, the basic judgment threshold and the local judgment thresholds of each sensitive area are integrated to form the second quality judgment condition.

[0100] After the quality judgment criteria are determined, the quality measurement and evaluation module 40 is responsible for acquiring the actual production quality index data of the premix, and evaluating the production quality based on the second quality judgment condition, and outputting the quality evaluation result.

[0101] In its implementation, the quality measurement and evaluation module 40 first deploys various online detection devices at key process nodes in the premix production flow, such as material mixing, mixing completion, and before finished product delivery. These devices include online laser particle size analyzers, near-infrared component analyzers, and spectrophotometers for uniformity testing, especially covering production stages corresponding to identified environmentally sensitive areas. The online detection devices collect raw data for various quality indicators. The collected data undergoes unified preprocessing, including data denoising, data smoothing, and error value removal. Based on preset dimension conversion rules, heterogeneous data from different sources are uniformly converted into standardized data with the same dimension, thus obtaining the actual quality indicator data for premix production. Then, based on timestamps, the quality indicator data is aligned with corresponding process parameters, environmental parameters, and environmentally sensitive area attributes to construct a comprehensive data set including quality indicators, process parameters, environmental parameters, and environmentally sensitive area attributes.

[0102] The preset dimensional conversion rule converts measured values ​​in different units, such as temperature, weight, time, and particle size, into standard units based on the nature of their physical quantities. For example, the unit of a temperature sensor, whether it is degrees Celsius (°C) or degrees Fahrenheit (°F), is uniformly converted to degrees Celsius (°C) or Kelvin (K). Similarly, weight can be uniformly converted to kilograms (kg), time to seconds (s), and particle size to micrometers (µm), etc.

[0103] Based on the above, the quality measurement and evaluation module 40 evaluates production quality using a comprehensive dataset and in conjunction with the second quality judgment criteria. Specifically, the quality measurement and evaluation module 40 first classifies and manages the quality indicator data obtained from actual production according to the attributes of the environmentally sensitive areas to which they belong. For the quality indicator data in each environmentally sensitive area, the system compares it with the quality indicator judgment criteria for that area in the second quality judgment criteria, and calculates the degree of deviation between the actual value and the judgment criteria.

[0104] Furthermore, the influence weights of each sensitive region on the premix quality are calculated. Specifically, the quality measurement and evaluation module 40 first extracts a list of all spatial units contained in each sensitive region based on the correspondence between sensitive regions and spatial units. Then, for each spatial unit, the quality measurement and evaluation module 40 couples the comprehensive sensitivity index with the fluctuation influence weights of the main influencing parameters within the spatial unit, for example, by using a weighted product or correlation coefficient method, to comprehensively quantify the quality influence value of the spatial unit. Subsequently, for the same sensitive region, the system sums up the quality influence values ​​of all spatial units within it to form the initial influence weight of the region. To ensure that the influence weights of each region can be used for overall weighted evaluation, the quality measurement and evaluation module 40 further standardizes the initial influence weights, with the sum of the weights of each region being 1. The weight allocation results are automatically adjusted according to real-time changes in the spatial environment and parameter status, providing a dynamic and reliable basis for quality assessment and process strategy setting.

[0105] Finally, the quality measurement and evaluation module 40 comprehensively analyzes the degree of deviation of quality indicators in all sensitive areas and their actual impact weight on production quality. Through statistical methods such as weighted summation, it calculates the overall quality score of the premix production batch. Specifically, the calculation of this overall quality score includes: first, for each sensitive area of ​​the premix production batch, calculating the degree of deviation between the actual value of each quality indicator in that sensitive area and the corresponding judgment standard; next, weighting these deviations according to the impact weight of that sensitive area on the overall quality of the premix; and finally, summing the weighted deviation values ​​of all sensitive areas and converting them into a value within the range of 0-100, thus obtaining the overall quality score of the premix production batch.

[0106] Based on the preset grading standards, the system defines the overall quality level, such as excellent, qualified, needs improvement, unqualified, etc., and outputs complete quality assessment results, including the specific deviation of individual quality indicators from their judgment standards, as well as the final determined overall quality level. The pre-defined grading criteria are based on the overall quality score as the core judgment criterion, combined with the deviation of individual quality indicators. Specifically, the criteria are as follows: Excellent grade: Overall quality score is in a high range, and all individual quality indicators corresponding to sensitive areas are without deviation, with no indicator exceeding the local judgment threshold of the second quality judgment condition. The performance of the premix fully meets the highest standards for downstream applications. Qualified grade: Overall quality score is in a medium range, allowing slight deviations in individual quality indicators in some non-core sensitive areas, but the indicators in core sensitive areas must fully meet the threshold requirements, and the basic performance of the premix is ​​unaffected. Need for improvement grade: Overall quality score is in a low range, with at least one individual quality indicator deviation in a core sensitive area, or multiple non-core sensitive area indicators exceeding the qualified range. Some performance of the premix may be affected, requiring process adjustment and optimization. Unqualified grade: Overall quality score is below a low level, with serious deviations in individual quality indicators in core sensitive areas, or multiple indicators exceeding the local judgment threshold of the second quality judgment condition. The performance of the premix does not meet the basic usage requirements.

[0107] The control module 50 is used to generate a process adjustment strategy for the production of premix based on the quality assessment results, and to adjust the process parameters based on the process adjustment strategy.

[0108] Based on the quality assessment results and the weighting of the impact of each sensitive area on the quality of the premix, the control module 50 automatically generates a process adjustment strategy for premix production and implements parameter adjustments. In practice, the control module 50 divides the process into several adjustment scenarios according to the quality assessment results.

[0109] In response to the quality level being excellent, the system generates a process parameter stabilization strategy to maintain the current stable production state and simultaneously records the combination status of relevant process parameters in all sensitive areas as a benchmark reference for subsequent production batches and process optimization.

[0110] A quality grade of "qualified" indicates a slight deviation in the production process, but it remains within acceptable limits. For quality indicators exhibiting this slight deviation, control module 50, considering the corresponding sensitive area and key influencing parameters, generates a fine-tuning plan for process parameters. The adjustment range is based on the historical correlation between parameter fluctuations and quality deviations, providing stable and limited correction to the deviation. For example, a regression model is used to calculate the deviation in the quality indicator caused by a one-unit change in a specific parameter, and the required parameter adjustment is calculated accordingly.

[0111] In response to a quality level indicating "needs improvement," control module 50 prioritizes core sensitive areas with high quality impact weights. Based on the main influencing parameters, historical data, and current comprehensive data set for these core sensitive areas, it designs a tiered parameter adjustment scheme to gradually correct key process parameters causing quality deviations. This tiered parameter adjustment scheme first performs a small, pre-set incremental adjustment to the key process parameter, for example, changing the parameter value by a percentage or a fixed amount. Control module 50 then monitors the actual value of the adjusted parameter and changes in production quality indicators. If the deviation is corrected but not completely eliminated, the system continues with the next small adjustment. This cycle continues until the key process parameters are gradually corrected, and the quality deviation is significantly reduced or eliminated.

[0112] If the quality level is deemed unqualified, a comprehensive adjustment mechanism is triggered. This mechanism reviews real-time data records from all stages of the production process for this batch, analyzes the status of sensitive areas in each stage, identifies the specific sensitive areas and key process parameters that cause the quality problem, and automatically generates combined adjustment strategies, such as re-ratioing the material input, and adjusting the mixing time and temperature simultaneously.

[0113] The control module 50 parses the generated process adjustment strategy into target values ​​for specific process parameters. This parsing process relies on a pre-configured strategy-parameter mapping library integrated with the upper-level control system, such as a PLC / DCS. This library stores rules for converting different strategy types, such as comprehensive adjustment mechanisms and fine-tuning schemes, into specific, executable digital signals or instruction sets. The control module 50 calls the strategy-parameter mapping library in real time, parses the target process parameter values ​​corresponding to the strategy, and generates control commands based on these values, directly issuing them to the actuators of the production equipment. Simultaneously, the information acquisition module 10 continuously monitors the adjusted process parameters and the environmental status of sensitive areas, comparing the deviations between actual and target values. If the adjustment deviation is within the allowable range, the current adjustment magnitude is maintained; if the deviation exceeds the allowable range, the adjustment step size is dynamically adjusted based on the quality influence weight of each area, ensuring the effectiveness and stability of the adjustment process. During the continuous adjustment period, production quality data is monitored and collected by the quality measurement and evaluation module 40. The system automatically verifies the actual effect of the aforementioned process adjustment strategies on improving production quality, and periodically optimizes process parameters and adjustment rules accordingly to achieve continuous and stable improvement and refined management of production quality.

[0114] See Figure 4 The flowchart below illustrates an online method for controlling the production quality of a premix, as provided in this application embodiment. The method includes:

[0115] S301, Obtain process parameters and environmental parameters of the premix production process, construct a spatial distribution model of the production environment, and identify environmentally sensitive areas of the premix production equipment;

[0116] S302, based on the environmentally sensitive area, the first quality judgment condition for premix production is dynamically adjusted locally to generate a second quality judgment condition;

[0117] S303: Obtain actual production quality index data of the premix, and evaluate the production quality based on the second quality judgment condition, and output the quality evaluation result;

[0118] S304, Based on the quality assessment results, a process adjustment strategy for the production of premixed agent is generated, and the process parameters are adjusted based on the process adjustment strategy.

[0119] In its specific implementation, the online control method disclosed in this application first collects process parameters during the premix production process in real time, such as material input, stirring speed, mixing temperature, and mixing time, as well as environmental parameters, such as workshop temperature and humidity, air flow, dust concentration, and spatial location. Based on high-resolution sensors and historical production data, spatial modeling algorithms, such as rasterization, partitioned clustering, and principal component analysis, are used to construct a spatial distribution model of the entire production environment. The model is then used to analyze the distribution characteristics and trends of parameters in each region. Based on this model, statistical methods such as correlation analysis and sensitivity assessment are used to identify various environmentally sensitive areas during the operation of the production equipment, including determining the boundaries of these areas, the main sensitive parameters, and their weights influencing product quality, providing spatial positioning basis for subsequent quality monitoring.

[0120] After identifying environmentally sensitive areas, the thresholds of the first quality judgment criteria are dynamically adjusted based on the identification results. Specifically, the standard quality judgment parameters of the premix, such as mixing uniformity, particle size range, and component fluctuation, are analyzed. Combined with the influence weight of sensitive areas and real-time environmental conditions, the judgment thresholds for key quality indicators corresponding to high-sensitivity areas are appropriately tightened or relaxed, forming a spatially segmented and dynamically changing second quality judgment criterion.

[0121] As production progresses, the system continuously acquires actual production quality indicator data from online quality testing equipment, such as multi-point particle size analyzers and rapid component detection equipment. Simultaneously, the data is mapped and matched with spatial distribution models and secondary quality judgment conditions. Based on dynamic judgment standards for different regions and parameters, the system automatically evaluates the production quality of each batch of premixed agent, comprehensively considering deviations in various quality indicators within sensitive areas, and outputs the overall batch quality level and sub-indicator evaluation results, providing a quantitative basis for quality control during the production process.

[0122] After obtaining the quality assessment results, the system automatically generates process adjustment strategies. Based on equipment operating history, real-time quality data, and quality impact weights, the system intelligently analyzes the spatial areas and parameters that mainly cause quality deviations, and designs targeted process correction schemes, including but not limited to adjusting the feed ratio, optimizing the stirring or mixing process, adjusting the environmental temperature and humidity control, and extending or shortening the mixing time. After the process adjustment command is issued, the production equipment automatically operates according to the new parameters. The system continues to monitor the production process and changes in quality indicators. After verifying the adjustment effect, the system further iterates and optimizes the adjustment strategy to achieve continuous and stable improvement in the quality of premixed agent production and dynamic online closed-loop management.

[0123] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 or all of the technical features therein. 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 this application.

Claims

1. An online control system for the production quality of premixed agents, characterized in that, include: The information acquisition module is used to acquire process parameters and environmental parameters of the premix production process; The environmental perception module is used to construct a spatial distribution model of the production environment based on the process parameters and environmental parameters, and to identify environmentally sensitive areas of the premix production equipment. The judgment condition generation module is used to dynamically adjust the local threshold of the first quality judgment condition for premix production based on the environmentally sensitive area, and generate the second quality judgment condition. The process of dynamically adjusting the local threshold of the first quality judgment condition for premix production to generate a second quality judgment condition includes: The basic threshold system in the first quality judgment condition of premix production is extracted, including the judgment threshold of the quality indicators of premix mixing uniformity, particle size, and component ratio; Based on the identification results of environmentally sensitive areas, the basic threshold system is mapped and associated with the sensitivity level, main influencing parameters and related equipment components of the environmentally sensitive areas to determine the quality indicators and corresponding area ranges for local adjustments. Differentiated adjustment strategies are adopted for areas with different sensitivity levels. The adjusted local thresholds of each area are integrated to obtain the judgment criteria for quality indicators, forming the second quality judgment condition. The quality measurement and evaluation module is used to acquire actual production quality index data of the premix, and evaluate the production quality based on the second quality judgment condition, and output the quality evaluation result. The control module is used to generate a process adjustment strategy for the production of premix based on the quality assessment results, and to adjust the process parameters based on the process adjustment strategy.

2. The online control system for the production quality of premixed agents according to claim 1, characterized in that, The process parameters include stirring speed, mixing time, material feed rate, and mixing temperature; the environmental parameters include workshop temperature, humidity, and dust concentration.

3. The online control system for the production quality of premixed agents according to claim 1, characterized in that, The spatial distribution model for constructing the production environment includes: The premix production workshop and the area where the production equipment is located are spatially discretized into multiple continuous spatial units, and the types and functional attributes of the equipment components included in the spatial units are associated to form a basic spatial framework. By deploying sensing devices in each spatial unit, real-time data of process parameters and environmental parameters are collected synchronously. The correspondence between parameter data and spatial units is established based on timestamps, forming a multidimensional dataset that includes spatial location, time dimension and parameter values. Based on the aforementioned multidimensional dataset, an interpolation algorithm is used to process the discrete parameter data to generate a continuous process parameter field and environmental parameter field covering the entire production area. By fusing process parameter fields and environmental parameter fields through a multi-field coupling algorithm, a spatial distribution model of the production environment is constructed.

4. The online control system for the production quality of premixed agents according to claim 2, characterized in that, The environmentally sensitive areas of the premix production equipment include: Based on the spatial distribution model, the coupling characteristics of process parameters and environmental parameters of each spatial unit are extracted, and the influence weight of each parameter fluctuation on the quality of the premix is ​​determined. A parameter sensitivity assessment model was established, and the parameter fluctuation range of each spatial unit was compared with the historical correlation data of the premix quality index to calculate the comprehensive sensitivity index of each spatial unit. Based on the functional attributes of the equipment components, cluster analysis is performed on spatial units whose comprehensive sensitivity index exceeds a preset threshold to form the identification results of environmentally sensitive areas of the premix production equipment, including the boundaries of the environmentally sensitive areas, sensitivity levels, main influencing parameters, and associated equipment components.

5. The online control system for the production quality of premixed agents according to claim 1, characterized in that, The acquisition of actual production quality index data for the premix includes: Deploy online detection devices at key process nodes in premix production to cover production stages in environmentally sensitive areas; The raw data of quality indicators are collected by the online detection device, preprocessed, and the heterogeneous data output by different online detection devices are converted into standardized data with a unified dimension to obtain the quality indicator data of the actual production of the premix. Establish the correspondence between quality indicator data, process parameters and environmental parameters based on timestamps, and construct a comprehensive data set including quality indicators, process status, environmental status and environmentally sensitive area attributes.

6. The online control system for the production quality of premixed agents according to claim 4, characterized in that, The output quality assessment results include: The actual production quality index data are classified according to the environmentally sensitive area attributes, and compared with the judgment criteria of the corresponding area in the second quality judgment condition. The deviation of the actual value of each quality index from the corresponding judgment criterion is calculated. Based on the identification results of environmentally sensitive areas, and combined with the comprehensive sensitivity index of each spatial unit and the influence weight of parameter fluctuations on the quality of premix, the influence weight of each environmentally sensitive area on the quality of premix is ​​determined. By combining the degree of deviation of quality indicators in the corresponding region, the comprehensive quality score of premix production is calculated by weighted summation, and the quality level is determined. The quality assessment results are output, including the degree of deviation of individual quality indicators and the quality level.

7. The online control system for the production quality of a premixed agent according to claim 6, characterized in that, The determination of the influence weights of each environmentally sensitive area on the quality of the premix includes: Based on the affiliation between environmentally sensitive areas and spatial units, extract all spatial units for each environmentally sensitive area; For each spatial unit, its comprehensive sensitivity index is coupled with the fluctuation impact weight of the main influencing parameters within the corresponding spatial unit to obtain the quality impact value of the corresponding spatial unit. The quality impact values ​​of all spatial units within the same environmentally sensitive area are summed to form the initial impact weight of the corresponding environmentally sensitive area. After standardization, the quality impact weight of each environmentally sensitive area is obtained.

8. The online control system for the production quality of a premix as described in claim 6, characterized in that, Adjusting process parameters based on the aforementioned process adjustment strategy includes: Based on the deviation of individual quality indicators, comprehensive quality score and quality level of the quality assessment results, process adjustment strategies are generated for different scenarios. The generated process adjustment strategy is parsed into target values ​​for specific process parameters, associated with the corresponding production execution mechanism, and the actual values ​​of process parameters and environmental parameters in environmentally sensitive areas are collected in real time through the information acquisition module, and the deviation between the actual values ​​and the target values ​​is compared. If the deviation is within the allowable range, maintain the current adjustment range; if the deviation exceeds the allowable range, dynamically adjust the adjustment step size based on the quality impact weight of each environmentally sensitive area. By monitoring quality indicator data in subsequent production cycles through the quality measurement and evaluation module, the effectiveness of process adjustment strategies in improving quality can be verified.

9. An online control method for the production quality of a premix, implemented based on an online control system for the production quality of a premix according to any one of claims 1-8, characterized in that, include: Obtain process and environmental parameters for the premix production process, construct a spatial distribution model of the production environment, and identify environmentally sensitive areas of the premix production equipment; Based on the environmentally sensitive area, the first quality judgment condition for premix production is dynamically adjusted locally to generate a second quality judgment condition. Obtain actual production quality index data of the premix, and evaluate the production quality based on the second quality judgment condition, and output the quality evaluation result; Based on the quality assessment results, a process adjustment strategy for the production of premixed agents is generated, and the process parameters are adjusted based on the process adjustment strategy.

Images