Method and system for quantitative evaluation of wall sticking facing a spray drying tower

Abstract

The application provides a wall sticking quantitative evaluation method and system for a spray drying tower, relates to the technical field of drying towers, and acquires multidimensional correlation data and a plurality of operating parameter sequences through collection; powder wall sticking analysis is respectively performed by using a powder wall sticking simulation model and a powder wall sticking prediction model, and powder wall sticking simulation distribution and powder wall sticking prediction distribution are output; powdering fluctuation analysis is performed according to the operating parameter sequence, a powdering fluctuation coefficient is output, and a model output weight is configured; the powder wall sticking simulation distribution and the powder wall sticking prediction distribution are fitted according to the model output weight, and a powder wall sticking evaluation result is output. The application solves the technical problems that the prior art lacks working condition fluctuation adaptability due to the dependence on a single physical model or a data model, and the wall sticking evaluation precision is low and the reliability is poor, achieves the technical effects of improving the wall sticking evaluation precision and the reliability, and enhancing the dynamic adaptability to complex working conditions.

Description

Technical Field

[0001] This application relates to the field of drying tower technology, specifically to a method and system for quantitative assessment of wall adhesion in spray drying towers. Background Technology

[0002] Spray drying technology is widely used in the food, pharmaceutical, and chemical industries to rapidly dry liquid materials into powder. During the operation of the spray drying tower, powder particles tend to adhere to the inner wall of the tower, causing a sticking phenomenon. This not only affects product yield and leads to unstable product quality, but also increases cleaning and maintenance costs. Therefore, accurate assessment and prediction of the sticking condition is a key requirement for optimizing the spray drying process.

[0003] Currently, the assessment of wall adhesion in spray drying towers mainly relies on empirical judgment, single-point temperature monitoring, or physical model calculations. For example, physical models use fluid dynamics simulation (CFD) to study the airflow distribution and powder trajectory within the tower, predicting areas at risk of wall adhesion. Empirical models obtain local parameters through manual inspections and sensors (such as temperature and humidity probes), combining historical experience to determine wall adhesion trends. However, traditional physical models rely on idealized assumptions and struggle to fully consider the coupled effects of various factors influencing wall adhesion, such as airflow velocity, temperature gradient, and humidity changes, leading to significant deviations between simulation results and actual operating conditions. Furthermore, empirical models typically use fixed parameters, lacking consideration for fluctuations during production. In actual production, the physical properties of liquid materials and environmental conditions often change dynamically, making it difficult for assessment results to adapt to different operating conditions, thus affecting the stability and reliability of predictions. Summary of the Invention

[0004] This application provides a quantitative assessment method and system for wall adhesion in spray drying towers, which solves the technical problem that existing technologies lack adaptability to operating condition fluctuations due to reliance on a single physical model or data model, resulting in low accuracy and poor reliability in wall adhesion assessment. It achieves the technical effect of improving the accuracy and reliability of wall adhesion assessment and enhancing dynamic adaptability to complex operating conditions.

[0005] In view of the above problems, this application provides a method for quantitative assessment of wall adhesion in spray drying towers. The method includes: collecting multi-dimensional correlation factors in the drying and pulverizing process of liquid materials, and collecting various operating parameters within the spray drying tower at fixed points to obtain multi-dimensional correlation data and several operating parameter sequences; using a powder wall adhesion simulation model and a powder wall adhesion prediction model, performing powder wall adhesion analysis based on the multi-dimensional correlation data and several operating parameter sequences, and outputting a simulated powder wall adhesion distribution and a predicted powder wall adhesion distribution; performing pulverizing volatility analysis based on the several operating parameter sequences, outputting a pulverizing volatility coefficient, and configuring model output weights; and fitting the simulated powder wall adhesion distribution and the predicted powder wall adhesion distribution based on the model output weights to output a powder wall adhesion assessment result.

[0006] On the other hand, this application also provides a quantitative assessment system for wall adhesion in spray drying towers. The system includes: a data acquisition module for acquiring multi-dimensional correlation factors during the drying and pulverizing process of liquid materials, and for acquiring various operating parameters within the spray drying tower at fixed points, thereby obtaining multi-dimensional correlation data and several operating parameter sequences; a simulation analysis module for performing powder wall adhesion analysis based on the multi-dimensional correlation data and several operating parameter sequences using a powder wall adhesion simulation model and a powder wall adhesion prediction model, and outputting a simulated powder wall adhesion distribution and a predicted powder wall adhesion distribution; a fluctuation analysis module for performing pulverization fluctuation analysis based on the several operating parameter sequences, outputting a pulverization fluctuation coefficient, and configuring model output weights; and a fitting evaluation module for fitting the simulated powder wall adhesion distribution and the predicted powder wall adhesion distribution based on the model output weights, and outputting a powder wall adhesion evaluation result.

[0007] One or more technical solutions provided in this application have at least the following beneficial effects:

[0008] By collecting multi-dimensional correlated data and several operating parameter sequences, a complete data foundation is constructed, improving the comprehensiveness and representativeness of the data. Using a powder wall adhesion simulation model and a powder wall adhesion prediction model, powder wall adhesion analysis is performed based on the multi-dimensional correlated data and several operating parameter sequences, outputting the simulated and predicted distributions of powder wall adhesion. This step simulates the powder wall adhesion behavior in the drying tower using a physical model and uses a machine learning model combined with historical data for prediction, achieving a dual evaluation based on mechanism analysis and data-driven approaches, thus improving prediction accuracy. Powdering volatility analysis is performed based on the several operating parameter sequences, considering the volatility in the production process and quantifying the dynamic changes in chemical conditions, outputting a powdering volatility coefficient. The model weights are adjusted according to the volatility state, making the evaluation method more adaptable and enhancing its ability to adapt to complex operating conditions. By adaptively weighting and fusing the analysis results of the physical and data models, accurate evaluation under different operating conditions is achieved, improving the stability, reliability, and generalization ability of the prediction.

[0009] In summary, this application achieves accurate assessment of powder adhesion behavior by comprehensively collecting multi-dimensional operating parameters during the spray drying process and combining physical simulations of microfluidics with machine learning prediction models. Simultaneously, it introduces powder production fluctuation analysis and dynamically adjusts the model output weights to adapt to changes in different production conditions. Finally, by adaptively weighting and fusing physical simulation and data prediction results, the accuracy, reliability, and adaptability to complex operating conditions of the adhesion assessment are improved, thus providing technical support for spray drying process optimization and quality control.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating the quantitative assessment method for wall adhesion in spray drying towers provided in an embodiment of this application.

[0012] Figure 2 This is a schematic diagram of the process for constructing a powder wall adhesion prediction model in the quantitative assessment method for wall adhesion of spray drying towers provided in the embodiments of this application.

[0013] Figure 3 This is a schematic diagram of the structure of a quantitative assessment system for wall adhesion in a spray drying tower provided in an embodiment of this application.

[0014] Figure labeling: Acquisition module 10, Simulation analysis module 20, Fluctuation analysis module 30, Fitting evaluation module 40. Detailed Implementation

[0015] This application provides a quantitative assessment method and system for wall adhesion in spray drying towers, which solves the technical problem of low accuracy and poor reliability of wall adhesion assessment caused by the reliance on a single physical model or data model and the lack of adaptability to operating condition fluctuations in the prior art. It achieves the technical effect of improving the accuracy and reliability of wall adhesion assessment and enhancing the dynamic adaptability to complex operating conditions.

[0016] Example 1, as Figure 1 As shown in the embodiments of this application, a method for quantitatively assessing wall adhesion in spray drying towers is provided, the method comprising:

[0017] Step S1: Collect multi-dimensional correlation factors in the process of drying and pulverizing liquid materials, and collect various operating parameters in the spray drying tower at fixed points to obtain multi-dimensional correlation data and several operating parameter sequences.

[0018] Specifically, liquid materials refer to liquid substances that need to be dried into powder during spray drying, such as milk and protein solutions. Multi-dimensional related factors refer to multiple variables that affect the wall-sticking phenomenon, such as environmental factors (humidity, temperature), material properties (viscosity, solid content), and process parameters (spray pressure, feed rate). Operating parameters include airflow velocity, temperature, humidity, and pressure within the spray drying tower.

[0019] Comprehensive data collection is conducted on the multi-dimensional correlation factors of liquid materials during the drying and powdering process, as well as various operating parameters within the spray drying tower. Specifically, the physicochemical properties of the liquid materials are measured using specialized testing equipment, such as viscometers to measure viscosity, surface tension meters to measure surface tension, and chromatographs to analyze the composition of the liquid materials. Simultaneously, multiple sensors and data acquisition devices are deployed within the spray drying tower to acquire operating parameters, including fixed-point collection of airflow velocity, temperature, humidity, and pressure, forming a sequence of operating parameters. For example, temperature sensors (thermocouples, infrared thermometers) are used to monitor the temperature gradient within the drying tower; humidity sensors are used to detect air moisture content and assess the degree of dryness; wind speed sensors are used to measure airflow velocity and determine drying dynamics; and pressure sensors detect pressure distribution within the tower to optimize spray conditions.

[0020] By collecting multi-dimensional correlation factors and operating parameter sequences, we can comprehensively collect key factors affecting wall adhesion, providing an accurate and rich data foundation for subsequent analysis and improving the accuracy of wall adhesion assessment.

[0021] Step S2: Using the powder adhesion simulation model and the powder adhesion prediction model, perform powder adhesion analysis based on the multi-dimensional correlation data and several operating parameter sequences, and output the simulated distribution and the predicted distribution of powder adhesion.

[0022] Specifically, the powder adhesion simulation model is based on computational fluid dynamics (CFD) and is used to simulate the movement trajectory and adhesion distribution of powder within a spray drying tower. The powder adhesion prediction model is trained using machine learning algorithms to predict the likelihood and extent of powder adhesion based on historical data. Multi-dimensional correlated data and operating parameter sequences are input into the powder adhesion simulation model to simulate the flow field, temperature field, and particle movement trajectory within the tower, obtaining the simulated distribution of powder adhesion. Simultaneously, machine learning models (such as neural networks and support vector machines) are trained on historical data to learn the complex relationships between adhesion and various factors. These are then used to predict the predicted distribution of powder adhesion by inputting current multi-dimensional correlated data and operating parameter sequences.

[0023] By combining two different types of models—physical simulation and data-driven models—powder adhesion behavior can be analyzed from different perspectives. The simulation model explains the adhesion phenomenon from a physical mechanism perspective, while the prediction model predicts the adhesion situation from the statistical laws of data. The two models complement each other, improving the accuracy and comprehensiveness of adhesion analysis.

[0024] Step S3: Perform milling volatility analysis based on the aforementioned series of operating parameters, output the milling volatility coefficient, and configure the model output weights.

[0025] Specifically, volatility analysis is performed on the operating parameter sequence. Statistical analysis methods (such as standard deviation and variance) are used to assess the degree of fluctuation in parameters such as airflow velocity, temperature, and humidity, and the volatility coefficient of each parameter is calculated. This volatility coefficient is an indicator that quantifies the degree of fluctuation in operating parameters, reflecting the degree of influence of the parameter on wall adhesion. The model output weights are configured according to the magnitude of the volatility coefficient; parameters with larger volatility coefficients have a more significant impact on wall adhesion and are therefore given higher weights in the model. Adaptive weighting algorithms (such as weighted average and Bayesian update) are used to dynamically adjust the model weights. For example, when volatility is high, the weight of the data-driven model is increased to adapt to complex operating conditions; when volatility is low, the weight of the physical model is increased to provide more reliable theoretical analysis. For instance, in a milk powder factory, if the feed rate and air humidity of the production line frequently fluctuate, machine learning prediction (data model) may be more adaptable to changes than physical simulation; therefore, the influence weight of the data model is increased.

[0026] By using volatility analysis and weight configuration, the evaluation focus of the model can be dynamically adjusted, enhancing the model's sensitivity to changes in key factors and improving the real-time nature and adaptability of the evaluation results.

[0027] Step S4: Fit the simulated distribution of powder sticking to the wall and the predicted distribution of powder sticking to the wall according to the output weights of the model, and output the evaluation result of powder sticking to the wall.

[0028] Specifically, a weighted fusion algorithm (such as weighted averaging or Bayesian optimization) is used to fuse the physical simulation results (simulated distribution of powder adhesion to the wall) and the machine learning prediction results (predicted distribution of powder adhesion to the wall) to obtain the final powder adhesion assessment result. By fusing the results obtained from the two models, the advantages of both models are combined. The resulting powder adhesion assessment result considers both the results of physical principle simulation and the results of data-driven prediction, making the final assessment result more accurate and reliable, and better reflecting the wall adhesion situation in the drying and powdering process of liquid materials.

[0029] Furthermore, step S1 in the embodiments of this application includes:

[0030] Step S11: Collect multi-dimensional correlation factors during the drying and powdering of liquid materials, and obtain multi-dimensional correlation data. The multi-dimensional correlation data includes at least the characteristics of the liquid material, air environment parameters, predetermined air drying parameters, and predetermined atomization parameters.

[0031] Step S12: During the drying and powdering process of liquid materials, the airflow velocity, gas flow rate, temperature, humidity and pressure in the spray drying tower are collected at fixed points according to a predetermined time interval, and arranged in order of monitoring time to obtain a number of operating parameter sequences. The operating parameter sequence includes K consecutive operating parameters, where K is an integer greater than 20.

[0032] Specifically, sensors, detection instruments, and online monitoring equipment are used to comprehensively collect multi-dimensional correlation data during the drying and powdering of liquid materials. This multi-dimensional correlation data includes at least the characteristics of the liquid materials, air environment parameters, predetermined air drying parameters, and predetermined atomization parameters. Liquid material characteristics refer to the physicochemical properties of the liquid materials, such as the properties of concentrated emulsions (solid content, milk fat content, protein structure, etc.), the type of material, and its proportions. For collecting liquid material characteristics, multiple analytical methods are required. Densitometers, viscometers, and refractometers are used to measure the solid content of the material (e.g., the solid content of milk powder concentrate is typically between 40% and 55%). Fourier transform infrared spectroscopy (FTIR) is used to analyze protein structure and determine the denaturation of whey proteins. A milk fat analyzer is used to detect milk fat content to ensure consistency between different batches of material. Air environment parameters refer to parameters such as temperature, humidity, and flow rate of the air inside the spray drying tower. These parameters directly affect the drying process and powder formation. Temperature and humidity sensors can be used to measure the humidity and temperature inside and outside the spray drying chamber, and a dew point meter can be used to determine the moisture content of the air to ensure the stability of the drying air. Preset air drying parameters refer to the pre-defined air drying conditions during the spray drying process, such as inlet air temperature, outlet air temperature, and airflow rate. Preset atomization parameters refer to the atomizer's operating parameters, such as atomization pressure and atomization speed; these parameters affect droplet size and distribution. Inlet and outlet air temperatures can be measured using thermocouples and infrared thermometers, and nozzle pressure can be monitored using differential pressure sensors to ensure stable atomization. The collection of multi-dimensional correlation factors provides comprehensive data on material, environmental, and process parameters, ensuring accurate input information for the spray drying process and providing data support for subsequent wall adhesion assessment.

[0033] In the process of drying and pulverizing liquid materials, parameters such as airflow velocity, gas flow rate, temperature, humidity, and pressure are collected at fixed points within the spray drying tower at predetermined time intervals. Specifically, a sensor network, including temperature, humidity, and pressure sensors, is installed inside the tower to continuously collect data at K time points at set time intervals (e.g., every minute or every second). The collected data are then arranged chronologically to form an operating parameter sequence. Here, K is an integer greater than 20, representing the number of consecutive operating parameters included in the operating parameter sequence. For example, multiple sensors are installed at different locations within the spray drying tower, collecting data once per minute for 30 minutes to form an operating parameter sequence containing 30 data points. By collecting operating parameters at fixed points and forming a sequence, the dynamic changes of parameters during the drying process can be captured, providing detailed data support for subsequent fluctuation analysis and wall adhesion assessment, thus improving the real-time performance and accuracy of the evaluation.

[0034] Furthermore, the construction process of the powder wall adhesion simulation model includes:

[0035] Step S01: Collect the geometric structure, dimensional information, and inner wall smoothness of the spray drying tower, and construct a three-dimensional geometric model of the spray drying tower using CFD software.

[0036] Step S02: Construct a powder wall adhesion simulation model based on the principles of microfluidics and the three-dimensional geometric model.

[0037] Specifically, information such as the shape and dimensions of the drying tower, including its height, diameter, and the shapes of its inlet and outlet, is obtained through laser scanning or CAD drawings. For example, a typical milk powder spray drying tower is cylindrical, 10–20 meters high, and 5–10 meters in diameter. A surface profilometer is used to measure the smoothness of the tower's inner wall (e.g., Ra value, used to quantify surface roughness). Inner wall smoothness refers to the roughness of the tower's inner wall, affecting powder adhesion behavior; smooth surfaces are less likely to cause powder deposition than rough surfaces. Using CFD preprocessing software (such as ANSYS Fluent, COMSOL Multiphysics, etc.), a three-dimensional geometric model of the spray drying tower is created based on its geometry, dimensions, and inner wall smoothness. This model ensures that key features, such as the locations of the air inlet, outlet, and spray nozzles, are included for subsequent fluid simulation.

[0038] CFD (Computational Fluid Dynamics) software is used to simulate physical phenomena such as fluid flow, heat transfer, and mass transfer. Based on mathematical models (such as the Navier-Stokes equations), it uses numerical calculations to solve for the spatial and temporal distribution of various physical quantities of fluids (such as velocity, pressure, and temperature). Based on the principles of microfluidics and a pre-constructed three-dimensional geometric model, a powder adhesion simulation model is built using CFD software. Specifically, various parameters need to be set in the model, including fluid properties (such as airflow velocity, temperature, and humidity), spray particle information (such as particle size distribution and spray pressure), material characteristics (such as the concentration of milk powder emulsion, milk fat content, and protein structure), and operating conditions (such as inlet air temperature, humidity, and flow rate). At the microscale, van der Waals forces between particles and the tower wall cause particle adhesion. Secondly, the surface of the particles may carry an electric charge, attracting the tower wall surface and generating electrostatic forces that lead to adhesion. Furthermore, the elasticity of the particles and the microscopic roughness of the tower wall surface also affect the adhesion phenomenon. Using the CFD software solver, based on parameters such as the particle size distribution, surface smoothness, and external environment (e.g., humidity, airflow velocity) of the milk powder particles, and grounded in microfluidics principles (e.g., van der Waals forces, electrostatic forces, surface tension), the critical adhesion conditions between the particles and the tower wall are calculated. This allows for the simulation of the particle motion trajectory and adhesion behavior within the tower. Based on the simulation results, the specific areas of particle adhesion are determined, and their location, area, and adhesion strength are analyzed. Simultaneously, the deposition amount of each particle under specific conditions is simulated to further analyze the particle adhesion behavior. Finally, the post-processing function of the CFD software is used to calculate the density and coverage area of ​​the adhered particles, resulting in a quantitative analysis of the adhesion state.

[0039] The powder wall adhesion simulation model constructed through the above steps comprehensively considers the influence of various factors on powder wall adhesion based on the principles of microfluidics. Through detailed parameter settings and numerical simulation calculations, the powder wall adhesion situation in the spray drying tower can be accurately simulated, including the wall adhesion area, wall adhesion particle density, and coverage area, providing a theoretical basis for the analysis and evaluation of powder wall adhesion.

[0040] Furthermore, such as Figure 2 As shown, the construction process of the powder wall adhesion prediction model includes:

[0041] Step S03: Using the similarity tolerance interval, the geometry, size information and inner wall smoothness of the spray drying tower are expanded to generate similarity comparison conditions.

[0042] Step S04: Using the similarity comparison conditions as constraints, retrieve the operation records of similar spray drying towers using big data retrieval, collect sample association datasets and several sample operation parameter sequence sets, and obtain the powder wall adhesion distribution under different sample association data and several sample operation parameter sequences to obtain the sample powder wall adhesion distribution set.

[0043] Step S05: Based on the principle of ensemble learning, a powder adhesion prediction model is constructed using the sample association dataset, several sample running parameter sequence sets, and sample powder adhesion distribution set as training data.

[0044] Specifically, the similarity tolerance interval refers to the permissible range of similarity in terms of geometry, dimensions, and inner wall smoothness, in order to find similar situations. The similarity comparison conditions are derived by expanding the geometry, dimensions, and inner wall smoothness of the spray drying tower, and are used to filter out operating records of similar spray drying towers that meet the requirements from a large dataset. Permissible variations in geometry and dimensions are set, for example: tower height is allowed to vary by ±15%; diameter by ±10%; spray angle by ±5°; and inner wall roughness Ra value by ±0.1µm. Based on these set tolerance intervals, the geometry, dimensions, and inner wall smoothness of the spray drying tower are mathematically expanded to generate the similarity comparison conditions. For example, if the original height of the spray drying tower is h, and the similarity tolerance interval is ±ɑ, then the height range in the similarity comparison conditions is [h(1-ɑ), h(1+ɑ)]. For example, the parameters of a spray drying tower A are: diameter 8 meters, tower height 15 meters, and inner wall Ra value 0.2µm. Based on the set similarity tolerance range, the similarity comparison conditions are as follows: diameter between 7.2 meters and 8.8 meters, tower height between 12.75 meters and 17.25 meters, and Ra value between 0.1µm and 0.3µm.

[0045] Using the generated similarity comparison criteria as constraints, a large number of operation records of spray drying towers are retrieved using a database management system or big data analytics platform (such as Hadoop, Spark, etc.). For spray drying towers that meet the similarity comparison criteria, relevant data are collected, including sample association datasets and several sample operation parameter sequence sets, as well as powder adhesion distribution under different sample association data and several sample operation parameter sequences, to obtain a sample powder adhesion distribution set. Through big data retrieval and sample dataset collection, rich actual operation data can be obtained, providing a realistic and reliable foundation for model training and improving the model's generalization ability and prediction accuracy.

[0046] Based on the principle of ensemble learning, a powder adhesion prediction model is constructed using a sample association dataset, a sample runtime parameter sequence set, and a sample powder adhesion distribution set as training data. Specifically, a suitable ensemble learning algorithm (such as Random Forest or XGBoost) is selected, and the training data is input into the algorithm for training. Taking Random Forest as an example, multiple subsets are first extracted from the sample association dataset and the sample runtime parameter sequence set. Then, a simple decision tree model (weak learner) is built on each subset. Each decision tree model is trained based on the sample powder adhesion distribution set, and the structure of the decision tree (such as node splitting rules and leaf node values) is continuously adjusted during training to minimize the prediction error. Finally, multiple trained decision trees are combined to form a Random Forest model (strong learner), which is the powder adhesion prediction model.

[0047] The powder adhesion prediction model built based on the principle of ensemble learning can leverage the advantages of multiple weak learners to improve the predictive ability of powder adhesion. Compared with a single learning model, it has better accuracy, stability, and generalization ability, and can better adapt to different spray drying tower operating conditions, thus providing a reliable basis for predicting powder adhesion.

[0048] Furthermore, in step S04 of this application embodiment, obtaining the powder adhesion distribution under different sample association data and several sample operating parameter sequences includes:

[0049] Step S04-1: Collect images of the inner wall of the spray drying tower under the first sample association data and several first sample operating parameter sequences, and perform image analysis to determine the first wall adhesion area.

[0050] Step S04-2: After peeling off the powder adhering to the inner wall of the spray drying tower by physical peeling method, weigh it and divide it by the area of ​​the corresponding peeled area to obtain multiple first wall adhesion densities.

[0051] Step S04-3: Construct the first sample powder wall adhesion distribution based on the first wall adhesion area and multiple first wall adhesion densities.

[0052] Specifically, the first sample associated data can be any set of data from multiple sample associated data retrieved from big data retrieval. The first sample associated data corresponds to several operating parameter sequences, i.e., the first sample operating parameter sequence. During sample data collection, an industrial endoscope or high-definition camera is used to acquire images of the inner wall of the spray drying tower under the first sample associated data and several first sample operating parameter sequences, capturing images of the inner wall from different positions and angles within the tower. Then, image processing software (such as ImageJ or OpenCV) is used to analyze the acquired images, identify the wall-adhering areas, and calculate their area. For example, in the milk powder production process, images of the tower's inner wall are captured using an industrial endoscope, and OpenCV software is used to segment the images, distinguishing the wall-adhering and non-wall-adhering areas. For instance, segmentation can be based on color differences (if the wall-adhering powder is a different color from the tower wall) or texture features. Then, the number of pixels in the wall-adhering area is calculated, and the first wall-adhering area is calculated based on known image proportions (such as the actual area corresponding to each pixel). The wall-adhering area information of the spray drying tower's inner wall is directly obtained through image acquisition and analysis. This method is relatively intuitive and can quickly obtain the distribution range of powder adhering to the tower wall, providing important area dimension information for subsequent construction of powder adhering to the wall, which helps to understand the powder adhering situation more comprehensively.

[0053] After the spray drying tower stops operating, the powder adhering to the inner wall of the tower is removed using a physical peeling method, targeting different areas. Specifically, a scraper or a specialized peeling tool can be used to carefully scrape off the powder adhering to the tower wall and collect it in a weighing bottle. Then, the mass of the peeled powder is weighed using a high-precision balance and divided by the area of ​​the corresponding peeled region to obtain several initial wall adhesion densities. For example, if the mass of powder peeled off in a specific peeled area (e.g., 1 square meter) is 50 grams, then the wall adhesion density of that area is 50 grams per square meter.

[0054] The wall adhesion distribution of the first sample powder is constructed based on the first wall adhesion area and multiple first wall adhesion densities. The inner wall of the spray drying tower can be divided into several small regions (based on coordinates or specific partitioning rules). For each small region, the corresponding first wall adhesion area and first wall adhesion density data are integrated. For example, a two-dimensional matrix or data structure can be created, where rows represent different small regions and columns represent information such as wall adhesion area and density, thereby constructing the wall adhesion distribution of the first sample powder. By constructing the wall adhesion distribution of the first sample powder, the wall adhesion situation under specific conditions can be comprehensively and quantitatively described, providing detailed data support for subsequent model training and prediction.

[0055] Furthermore, step S05 includes:

[0056] Step S05-1: Configure Q ensemble learning operators, wherein the ensemble learning operators include at least a feedforward neural network, a random forest, and a gradient boosting tree, and Q is an integer greater than or equal to 3.

[0057] Step S05-2: Use the aforementioned sample association dataset, several sample running parameter sequence sets, and sample powder wall adhesion distribution set as training data, and divide them into Q equal parts, proportionally dividing them into Q training sets and Q test sets.

[0058] Step S05-3: Using the Q training sets and Q test sets, train and test the Q ensemble learning operators respectively to obtain Q powder adhesion prediction branches.

[0059] Step S05-4: Based on the principle of ensemble learning, construct a powder wall adhesion prediction model according to the Q powder wall adhesion prediction branches.

[0060] Specifically, ensemble learning operators refer to the specific algorithms or model structures used to build models in ensemble learning, including at least feedforward neural networks, random forests, and gradient boosting trees. These ensemble learning operators have different characteristics and advantages, and by combining these different operators, a higher-performance powder adhesion prediction model can be constructed. Q ensemble learning operators are selected and configured using machine learning frameworks (such as TensorFlow, PyTorch, and scikit-learn), where Q is an integer greater than or equal to 3, meaning at least three different ensemble learning operators are selected to ensure the accuracy and generalization ability of the powder adhesion prediction model.

[0061] Using data processing software, the sample association dataset, sample running parameter sequence set, and sample powder adhesion distribution set are divided into Q parts. For each dataset, it is divided into training set and test set according to a certain ratio (for example, 80% as training set and 20% as test set), resulting in Q training set and Q test set.

[0062] For each ensemble learning operator, training is performed using the corresponding training set. Taking random forest as an example, the sample association dataset and sample runtime parameter sequence set in the training set are used as input features, and the sample powder adhesion distribution set is used as the target label. Training is performed by adjusting parameters such as the number of decision trees and the depth of the trees in the random forest, enabling the model to learn the relationship between input and output. Then, the trained model is tested using the corresponding test set, and the error metrics (such as mean squared error, accuracy, etc.) between the predicted results and the actual results are calculated. Through this training and testing process, Q ensemble learning operators are trained and tested respectively, resulting in Q powder adhesion prediction branches. Each powder adhesion prediction branch has the ability to predict powder adhesion, and because it is built based on different ensemble learning operators, it has different prediction characteristics and performance, providing multiple components with different prediction properties for building the final powder adhesion prediction model.

[0063] Based on the principle of ensemble learning, a powder adhesion prediction model is constructed using Q powder adhesion prediction branches. First, the performance of each of the Q branches is evaluated on its respective test set. For example, it can be evaluated based on metrics such as the mean squared error between the predicted and actual results, and accuracy. Then, weights are assigned to each branch based on these performance metrics. For example, if a branch has high accuracy and low mean squared error, it can be assigned a higher weight. Let the Q powder adhesion prediction branches be (M1, M2, ..., M...). Q The corresponding weights are (ω1, ω2, ..., ω). Q ), and ω1+ω2+…+ω Q =1. For new input data (such as new sample association data and sample running parameter sequences), each branch M... i Each of these will produce a prediction result ri, then the final prediction result of the powder adhesion prediction model is R=∑ω i r i (i=1,2,…,Q).

[0064] The powder adhesion prediction model, built based on the principle of ensemble learning, combines Q powder adhesion prediction branches by weighted averaging. This integrates the advantages of each branch, improving the accuracy, stability, and generalization ability of the prediction, thereby obtaining a more accurate and reliable powder adhesion prediction distribution.

[0065] Furthermore, in step S3, a milling volatility analysis is performed based on the aforementioned sequence of operating parameters, and a milling volatility coefficient is output, including:

[0066] Step S31: Obtain several operating parameter sequences, including airflow velocity sequence, gas flow rate sequence, temperature sequence, humidity sequence and pressure sequence.

[0067] Step S32: Perform fluctuation analysis on the several operating parameter sequences respectively to obtain several fluctuation coefficients.

[0068] Step S33: Obtain the powder-making fluctuation coefficient by weighting the several fluctuation coefficients, wherein the weight of the operating parameter is positively correlated with the degree of powder sticking to the wall.

[0069] Specifically, the operating parameter sequences include airflow velocity sequences, gas flow rate sequences, temperature sequences, humidity sequences, and pressure sequences, which are continuous data for the corresponding parameters within a certain time interval. Sensors installed inside the spray drying tower collect parameters such as airflow velocity, gas flow rate, temperature, humidity, and pressure at predetermined time intervals to form the corresponding sequence data.

[0070] Fluctuation analysis was performed on several operating parameter sequences to obtain several fluctuation coefficients. Taking the airflow velocity sequence as an example, its fluctuation coefficient was calculated using statistical methods (such as standard deviation, variance, coefficient of variation, etc.). For example, the mean of the airflow velocity sequence could be calculated first, then the sum of squares of the differences between each data point and the mean could be calculated, and then divided by the number of data points to obtain the variance. Finally, the square root of the variance could be taken to obtain the standard deviation σ, which could be used as the fluctuation coefficient of the airflow velocity sequence. Similar methods were used to calculate the fluctuation coefficients of the other operating parameter sequences, resulting in several fluctuation coefficients.

[0071] The powder-making fluctuation coefficient is obtained by weighting several fluctuation coefficients. Specifically, the influence of each operating parameter on powder adhesion to the wall is first determined, thereby determining its weight. Let the fluctuation coefficients of airflow velocity, gas flow rate, temperature, humidity, and pressure be k1, k2, k3, k4, and k5, respectively, with corresponding weights ω1, ω2, ω3, ω4, and ω5. Based on the principle that the weights of operating parameters are positively correlated with the influence of powder adhesion to the wall, these weights are determined through experiments, experience, or data analysis. For example, if experiments show that the influence of temperature on powder adhesion to the wall is twice that of airflow velocity, then in the weight allocation, ω3 = 2ω1. Then, the powder-making fluctuation coefficient K is calculated using the weighted average method, with the formula K = ω1k1 + ω2k2 + ω3k3 + ω4k4 + ω5k5.

[0072] The powder production fluctuation coefficient obtained through the above steps comprehensively considers the fluctuation of various operating parameters and their impact on powder adhesion to the wall. Through weighted calculation, the powder production fluctuation coefficient can more comprehensively and accurately reflect the overall fluctuation of the powder production process, providing an important reference indicator for monitoring and optimizing the process.

[0073] Furthermore, step S32 includes:

[0074] A first sequence of operating parameters is randomly selected, and the standard deviation and mean of the first parameter are calculated; the ratio of the standard deviation and mean of the first parameter is set as the first fluctuation coefficient.

[0075] Specifically, in one implementation, when calculating the fluctuation coefficient, one operating parameter sequence is randomly selected from several sequences as the first operating parameter sequence. For example, the airflow velocity sequence is randomly selected from airflow velocity, gas flow rate, temperature, humidity, and pressure sequences. Then, the standard deviation and mean of this sequence are calculated, and the ratio of the standard deviation to the mean is set as the first fluctuation coefficient. Through fluctuation analysis, the degree of fluctuation of each operating parameter can be quantified, providing a basis for subsequent weight allocation and enhancing the model's sensitivity to changes in key factors.

[0076] Furthermore, step S3 configures the model output weights, including:

[0077] Step S34: Calculate the ratio of the historical average milling fluctuation coefficient to the milling fluctuation coefficient, and set it as the prediction weight adjustment coefficient.

[0078] Step S35: Multiply the prediction weight adjustment coefficient by the predetermined prediction weight to obtain the real-time prediction weight, wherein the predetermined prediction weight is the initial weight of the powder wall adhesion prediction model, and the value is 0.6.

[0079] Step S36: Subtract the real-time prediction weight from 1 to obtain the real-time simulation weight, and use the real-time prediction weight and the real-time simulation weight as the model output weight.

[0080] Specifically, the historical average milling fluctuation coefficient is the average value of the milling fluctuation coefficient over a past period (historical period), reflecting the average fluctuation level of the milling process during historical stages. The historical milling fluctuation coefficient is obtained from historical data storage, and its average value is calculated. Then, the milling fluctuation coefficient K calculated in the current step (obtained in the previous step S33) is obtained. Finally, the ratio of the historical average milling fluctuation coefficient to the current milling fluctuation coefficient is calculated to obtain the prediction weight adjustment coefficient. By calculating the prediction weight adjustment coefficient, a relationship is established between the historical milling fluctuation level and the current milling fluctuation level. This ratio of the historical average milling fluctuation coefficient to the current milling fluctuation coefficient reflects the degree of change in the current milling fluctuation situation relative to the historical situation, providing a basis for subsequent adjustments to the prediction weight.

[0081] The predetermined prediction weight is the initial weight set in advance for the powder adhesion prediction model, with a value of 0.6, indicating the initial importance of the prediction model in the model output. Multiplying the prediction weight adjustment coefficient by the predetermined prediction weight yields the real-time prediction weight, reflecting the impact of current operating parameter fluctuations on the prediction model.

[0082] The real-time simulation weight is obtained by subtracting the real-time prediction weight from 1. This real-time simulation weight is used to adjust the importance of the simulation model in the final evaluation. For example, if the prediction weight adjustment coefficient is 0.8, then the real-time prediction weight is 0.8 × 0.6 = 0.48. The real-time simulation weight = 1 - 0.48 = 0.52. Then, the real-time prediction weight and the real-time simulation weight are used as the model output weights for calculating the final powder adhesion evaluation result.

[0083] The above weight allocation method can reasonably allocate the weights of the prediction and simulation parts in the powder wall adhesion prediction model according to the fluctuation of powder production, so that the model can more flexibly and accurately predict and simulate powder wall adhesion under different powder production fluctuations.

[0084] In summary, the quantitative assessment method for wall adhesion in spray drying towers provided in this application has the following beneficial effects:

[0085] This application's embodiments achieve accurate assessment and prediction of powder adhesion to the walls of a spray drying tower through four core steps: multi-dimensional data acquisition, physical simulation modeling, machine learning predictive modeling, and adaptive weighted fusion. First, in the data acquisition stage, the characteristics of the liquid material, air environment parameters, and predetermined drying and atomization parameters are extracted. During the drying process, key operating parameters such as airflow velocity, temperature, humidity, and pressure are collected at fixed points to form multi-dimensional correlated data and time-series data, providing a comprehensive data foundation for subsequent analysis. Next, in the physical simulation modeling stage, based on the geometric structure, fluid properties, and particle characteristics of the spray drying tower, a three-dimensional simulation model is established using CFD (Computational Fluid Dynamics) software. Combining microfluidics principles, the adhesion force, motion trajectory, and critical adhesion conditions of the particles are simulated, quantifying and predicting the wall adhesion distribution under different conditions. Then, in the machine learning predictive modeling stage, the similarity comparison conditions for spray drying towers are expanded. Operating records of similar equipment are retrieved from big data, and historical wall adhesion distribution data are obtained by combining image analysis and physical peeling experiments. An ensemble learning method (including feedforward neural networks, random forests, and gradient boosting trees) is used to train the predictive model, thereby establishing a data-driven powder wall adhesion prediction model. Then, in the operating condition fluctuation analysis phase, the pulverization volatility is calculated based on the real-time collected operating parameter sequence, and the weights of the prediction model are adjusted using the historical volatility average. This allows for greater reliance on machine learning predictions when volatility is low, while increasing the proportion of physical simulation models when volatility is high, ensuring the stability and reliability of the evaluation results. Finally, through adaptive weighted fusion, the output weights of the physical simulation and machine learning prediction models are dynamically adjusted according to the pulverization volatility coefficient, and the final powder adhesion evaluation result is obtained using weighted calculation.

[0086] Overall, this application's embodiments achieve accurate assessment of powder adhesion behavior by comprehensively collecting multi-dimensional operating parameters during the spray drying process and combining physical simulations of microfluidics with machine learning prediction models. Simultaneously, powder production fluctuation analysis is introduced to dynamically adjust the model's output weights to adapt to changes in different production conditions. Finally, by adaptively weighting and fusing physical simulation and data prediction results, the accuracy, reliability, and adaptability to complex operating conditions of the adhesion assessment are improved, thus providing technical support for spray drying process optimization and quality control.

[0087] Example 2, as Figure 3 As shown, based on the same inventive concept as in Embodiment 1 above, this application provides a quantitative assessment system for wall adhesion in spray drying towers, the system comprising:

[0088] The data acquisition module 10 is used to collect multi-dimensional correlation factors in the process of drying and pulverizing liquid materials, as well as to collect various operating parameters in the spray drying tower at fixed points, and to obtain multi-dimensional correlation data and several operating parameter sequences.

[0089] The simulation analysis module 20 is used to perform powder adhesion analysis based on the multi-dimensional correlation data and several operating parameter sequences using the powder adhesion simulation model and the powder adhesion prediction model, and output the powder adhesion simulation distribution and the powder adhesion prediction distribution.

[0090] The fluctuation analysis module 30 is used to perform milling fluctuation analysis based on the several operating parameter sequences, output milling fluctuation coefficients, and configure model output weights.

[0091] The fitting evaluation module 40 is used to fit the simulated distribution of powder sticking to the wall and the predicted distribution of powder sticking to the wall according to the weights output by the model, and output the evaluation result of powder sticking to the wall.

[0092] Furthermore, in this embodiment of the application, the data acquisition module 10 is also used to perform the following steps:

[0093] Multi-dimensional correlation factors are collected during the drying and pulverization of liquid materials to obtain multi-dimensional correlation data. The multi-dimensional correlation data includes at least the characteristics of the liquid material, air environment parameters, predetermined air drying parameters, and predetermined atomization parameters. During the drying and pulverization process of liquid materials, the airflow velocity, gas flow rate, temperature, humidity, and pressure in the spray drying tower are collected at fixed points according to predetermined time intervals and arranged in chronological order to obtain several operating parameter sequences. The operating parameter sequences include K consecutive operating parameters, where K is an integer greater than 20.

[0094] Furthermore, the system described in this application embodiment also includes a first model building module, which is used to build a powder wall adhesion simulation model, and the execution steps include:

[0095] The geometric structure, dimensions, and inner wall smoothness of the spray drying tower were collected, and a three-dimensional geometric model of the spray drying tower was constructed using CFD software. Based on the principles of microfluidics and the three-dimensional geometric model, a powder adhesion simulation model was constructed.

[0096] Furthermore, the system described in this application embodiment also includes a second model building module, which is used to build a powder adhesion prediction model, and the execution steps include:

[0097] By utilizing a similarity tolerance interval, the geometric structure, dimensional information, and inner wall smoothness of the spray drying tower are expanded to generate similar comparison conditions. Using these similarity comparison conditions as constraints, large-scale data retrieval is performed on the operating records of similar spray drying towers to collect sample association datasets and several sample operating parameter sequence sets. The powder adhesion distribution under different sample association datasets and several sample operating parameter sequences is then obtained, resulting in a sample powder adhesion distribution set. Based on the principle of ensemble learning, the sample association dataset, several sample operating parameter sequence sets, and the sample powder adhesion distribution set are used as training data to construct a powder adhesion prediction model.

[0098] Furthermore, the second model building module is also used to perform the following steps:

[0099] Images of the inner wall of the spray drying tower under the first sample association data and several first sample operating parameter sequences are acquired, and the first wall adhesion area is determined by image analysis. The powder adhering to the inner wall of the spray drying tower is peeled off by physical peeling method, weighed, and divided by the area of ​​the corresponding peeled area to obtain multiple first wall adhesion densities. The first sample powder wall adhesion distribution is constructed based on the first wall adhesion area and multiple first wall adhesion densities.

[0100] Furthermore, the second model building module is also used to perform the following steps:

[0101] Configure Q ensemble learning operators, each including at least a feedforward neural network, a random forest, and a gradient boosting tree, where Q is an integer greater than or equal to 3. Use the aforementioned sample association dataset, several sample running parameter sequence sets, and sample powder adhesion distribution set as training data, and divide them into Q equal parts, proportionally dividing them into Q training sets and Q test sets. Use the Q training sets and Q test sets to train and test the Q ensemble learning operators respectively, obtaining Q powder adhesion prediction branches. Based on the ensemble learning principle, construct a powder adhesion prediction model according to the Q powder adhesion prediction branches.

[0102] Furthermore, in this embodiment of the application, the fluctuation analysis module 30 is also used to perform the following steps:

[0103] Several operating parameter sequences are obtained, including airflow velocity sequence, gas flow rate sequence, temperature sequence, humidity sequence, and pressure sequence; fluctuation analysis is performed on the several operating parameter sequences to obtain several fluctuation coefficients; the powder making fluctuation coefficient is obtained by weighting the several fluctuation coefficients, wherein the weight of the operating parameter is positively correlated with the influence of powder sticking to the wall.

[0104] Furthermore, in this embodiment of the application, the fluctuation analysis module 30 is also used to perform the following steps:

[0105] A first sequence of operating parameters is randomly selected, and the standard deviation and mean of the first parameter are calculated; the ratio of the standard deviation and mean of the first parameter is set as the first fluctuation coefficient.

[0106] Furthermore, in this embodiment of the application, the fluctuation analysis module 30 is also used to perform the following steps:

[0107] The ratio of the historical average of the milling fluctuation coefficient to the milling fluctuation coefficient is calculated and set as the prediction weight adjustment coefficient. The prediction weight adjustment coefficient is multiplied by the predetermined prediction weight to obtain the real-time prediction weight, wherein the predetermined prediction weight is the initial weight of the powder sticking to the wall prediction model and has a value of 0.6. The real-time prediction weight is obtained by subtracting the real-time prediction weight from 1, and the real-time prediction weight and the real-time simulation weight are used as the model output weights.

[0108] Through the foregoing detailed description of the quantitative assessment method for wall adhesion to spray drying towers, those skilled in the art can clearly understand that the quantitative assessment system for wall adhesion to spray drying towers in this embodiment corresponds to the system disclosed in Embodiment 2, and has corresponding functional modules and beneficial effects as it is similar to the method disclosed in Embodiment 1. For relevant details, please refer to the method section.

[0109] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for quantitatively assessing wall adhesion in spray drying towers, characterized in that the method... include: Collect multi-dimensional correlation factors in the process of drying and pulverizing liquid materials, and collect various operating parameters in the spray drying tower at fixed points to obtain multi-dimensional correlation data and several operating parameter sequences; Using the powder adhesion simulation model and the powder adhesion prediction model, powder adhesion analysis is performed based on the multi-dimensional correlation data and several operating parameter sequences, and the simulated distribution and predicted distribution of powder adhesion are output. Based on the aforementioned series of operating parameters, a milling volatility analysis is performed, the milling volatility coefficient is output, and the model output weights are configured. Based on the output weights of the model, the simulated distribution of powder sticking to the wall and the predicted distribution of powder sticking to the wall are fitted to output the powder sticking to the wall evaluation result. Among them, a weighted fusion algorithm is used to fuse the simulated distribution of powder sticking to the wall and the predicted distribution of powder sticking to the wall to obtain the final powder sticking to the wall evaluation result. Obtain the powder adhesion distribution under different sample correlation data and several sample operating parameter sequences, including: Images of the inner wall of the spray drying tower under the first sample correlation data and several first sample operating parameter sequences were acquired, and the first wall adhesion area was determined by image analysis. The powder adhering to the inner wall of the spray drying tower is peeled off by physical peeling method, weighed, and divided by the area of ​​the corresponding peeled area to obtain multiple first wall adhesion densities. A first sample powder wall adhesion distribution is constructed based on the first wall adhesion area and multiple first wall adhesion densities; Configure the model output weights, including: Calculate the ratio of the historical average milling fluctuation coefficient to the milling fluctuation coefficient, and set it as the prediction weight adjustment coefficient; The prediction weight adjustment coefficient is multiplied by the predetermined prediction weight to obtain the real-time prediction weight, wherein the predetermined prediction weight is the initial weight of the powder wall adhesion prediction model, and the value is 0.

6. The real-time simulation weight is obtained by subtracting the real-time prediction weight from 1, and the real-time prediction weight and the real-time simulation weight are used as the model output weights.

2. The method for quantitative assessment of wall adhesion in spray drying towers according to claim 1, characterized in that, Obtain multi-dimensional correlated data and several sequences of runtime parameters, including: Collect multi-dimensional correlation factors during the drying and powdering of liquid materials, and obtain multi-dimensional correlation data. The multi-dimensional correlation data includes at least the characteristics of liquid materials, air environment parameters, predetermined air drying parameters, and predetermined atomization parameters. During the drying and powdering process of liquid materials, the airflow velocity, gas flow rate, temperature, humidity and pressure inside the spray drying tower are collected at fixed points according to a predetermined time interval. The data are then arranged in chronological order of the monitoring time to obtain several operating parameter sequences. Each operating parameter sequence includes K consecutive operating parameters, where K is an integer greater than 20.

3. The method for quantitative assessment of wall adhesion in spray drying towers according to claim 1, characterized in that, The process of constructing the powder wall adhesion simulation model includes: Collect information on the geometric structure, dimensions, and inner wall smoothness of the spray drying tower, and use CFD software to construct a three-dimensional geometric model of the spray drying tower. A powder adhesion simulation model was constructed based on the principles of microfluidics and the aforementioned three-dimensional geometric model.

4. The method for quantitative assessment of wall adhesion in spray drying towers according to claim 3, characterized in that, The process of constructing the powder wall adhesion prediction model includes: By utilizing the similarity tolerance interval, the geometric structure, dimensional information, and inner wall smoothness of the spray drying tower are expanded to generate similarity comparison conditions. The similarity tolerance interval refers to the range of similarity allowed in terms of geometric structure, dimensional information, and inner wall smoothness. With the aforementioned similarity comparison conditions as constraints, big data retrieval is performed on the operation records of similar spray drying towers, sample association datasets and several sample operation parameter sequence sets are collected, and the powder wall adhesion distribution under different sample association data and several sample operation parameter sequences is obtained to obtain the sample powder wall adhesion distribution set. Based on the principle of ensemble learning, a powder adhesion prediction model is constructed using the aforementioned sample association dataset, several sample running parameter sequence sets, and sample powder adhesion distribution set as training data.

5. The method for quantitative assessment of wall adhesion in spray drying towers according to claim 4, characterized in that, Based on the principle of ensemble learning, a powder adhesion prediction model is constructed using the aforementioned sample association dataset, several sample running parameter sequence sets, and sample powder adhesion distribution set as training data, including: Configure Q ensemble learning operators, wherein the ensemble learning operators include at least a feedforward neural network, a random forest, and a gradient boosting tree, where Q is an integer greater than or equal to 3; The sample association dataset, several sample running parameter sequence sets, and sample powder wall adhesion distribution set are used as training data and are divided into Q equal parts, which are then divided into Q training sets and Q test sets according to the proportion. Using the Q training sets and Q test sets, Q ensemble learning operators are trained and tested respectively to obtain Q powder adhesion prediction branches; Based on the principle of ensemble learning, a powder wall adhesion prediction model is constructed according to the Q powder wall adhesion prediction branches.

6. The method for quantitative assessment of wall adhesion in spray drying towers according to claim 1, characterized in that, Based on the aforementioned series of operating parameters, a milling volatility analysis is performed, and a milling volatility coefficient is output, including: Acquire several operating parameter sequences, including airflow velocity sequence, gas flow rate sequence, temperature sequence, humidity sequence, and pressure sequence; Fluctuation analysis was performed on the several operating parameter sequences to obtain several fluctuation coefficients; The powder-making fluctuation coefficient is obtained by weighting the aforementioned fluctuation coefficients, wherein the weight of the operating parameter is positively correlated with the degree of powder adhesion to the wall.

7. The method for quantitative assessment of wall adhesion in spray drying towers according to claim 6, characterized in that, Fluctuation analysis was performed on the aforementioned series of operating parameters, including: A first sequence of operating parameters is randomly selected, and the standard deviation and mean of the first parameter are calculated. The ratio of the standard deviation of the first parameter to the mean of the first parameter is set as the first fluctuation coefficient.

8. A quantitative assessment system for wall adhesion in spray drying towers, characterized in that, The system is used to perform the quantitative assessment method for wall adhesion of a spray drying tower according to any one of claims 1-7, including: The data acquisition module is used to collect multi-dimensional correlation factors in the process of drying and pulverizing liquid materials, as well as to collect various operating parameters in the spray drying tower at fixed points, and to obtain multi-dimensional correlation data and several operating parameter sequences. The simulation analysis module is used to perform powder adhesion analysis based on the multi-dimensional correlation data and several operating parameter sequences using the powder adhesion simulation model and the powder adhesion prediction model, and output the simulated distribution and predicted distribution of powder adhesion. The fluctuation analysis module is used to perform milling fluctuation analysis based on the aforementioned series of operating parameters, output milling fluctuation coefficients, and configure model output weights. The fitting evaluation module is used to fit the simulated distribution of powder adhesion to the wall and the predicted distribution of powder adhesion to the wall based on the weights output by the model, and output the evaluation result of powder adhesion to the wall.

Images