Product drying process temperature regulation system, method, and method of preparing a stucco gypsum

Abstract

The application discloses a product drying process temperature adjusting system and method and a preparation method of plastering gypsum, relates to the technical field of temperature regulation, and is characterized in that sensors are arranged at key points of a preparation environment of the plastering gypsum, and a historical / real-time temperature matrix is constructed; historical temperature data are classified based on a particle swarm algorithm, current abnormal temperature is identified, and real-time regulation is performed; a neural network model is trained by using a simulated annealing algorithm, and future temperature trends are accurately predicted; environment temperature is intervened in advance according to a prediction result, the system comprises data acquisition, classification regulation, model training and prediction application modules; when the plastering gypsum is prepared, dynamic temperature control is realized in the raw material mixing, drying and grinding stages; temperature fluctuation is reduced, product drying uniformity, strength qualification rate and production efficiency are improved, and energy consumption and labor cost are reduced.

Description

Technical Field

[0001] This invention belongs to the field of temperature control systems, and more specifically, relates to a temperature control system and method for product drying process, as well as a method for preparing plaster. Background Technology

[0002] In the preparation of plaster plaster, temperature control during the drying stage is one of the key factors affecting product quality and production efficiency. Currently, the widely used drying temperature control methods mainly rely on manually setting temperature thresholds and maintaining them through simple temperature control equipment. These systems typically lack intelligent adjustment capabilities; they can only control the drying process based on preset values ​​and cannot dynamically adjust according to actual temperature changes during the plaster plaster drying process.

[0003] The existing technical solutions have the following drawbacks:

[0004] High energy consumption: Because the temperature cannot be precisely controlled, the system often needs to fluctuate within a large range, resulting in unnecessary energy waste.

[0005] Unstable product quality: Inaccurate temperature control can easily lead to uneven drying of plaster, affecting the strength and setting time of the final product;

[0006] Complex operation: Manually setting and adjusting the temperature requires professional knowledge and experience, which increases the difficulty and cost of operation. Summary of the Invention

[0007] In response to the problems in related technologies, this invention proposes a temperature control system and method for the product drying process, as well as a method for preparing plaster, to overcome the aforementioned technical problems existing in the prior art.

[0008] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0009] This invention relates to a method for temperature control during product drying, comprising the following steps:

[0010] S1. Set up a set of key environmental locations, collect temperature data and collection time data of a large number of key environmental locations, and construct a historical environmental temperature data matrix; set the current environmental temperature collection time point, collect temperature data of each key environmental location point at the current environmental temperature collection time point, and obtain the current environmental temperature data matrix.

[0011] S2. Classify the temperature data in the historical ambient temperature data matrix to obtain a historical ambient temperature category center dataset; classify the temperature data in the current ambient temperature data matrix according to the historical ambient temperature category center dataset to obtain a current ambient temperature data category set; set an abnormal ambient temperature data category set according to the historical ambient temperature category center dataset; compare the current ambient temperature data category set with the abnormal ambient temperature data category set, and decide whether to adjust the current ambient temperature based on the comparison result;

[0012] S3. Construct an initial neural network model. Divide the historical ambient temperature data matrix into a historical ambient temperature data training matrix and a historical ambient temperature data test matrix according to the proportion. Train, test and optimize the initial neural network model to obtain the final neural network model.

[0013] S4. Substitute the current ambient temperature data matrix into the final neural network model to obtain the future ambient temperature data matrix; classify the future ambient temperature data matrix to obtain a future ambient temperature data category set; compare the future ambient temperature data category set with the abnormal ambient temperature data category set; and decide whether to adjust the current ambient temperature based on the comparison result.

[0014] This invention constructs a historical / real-time temperature matrix by deploying sensors at key points in the drying equipment; classifies historical temperatures based on a particle swarm optimization algorithm to identify current abnormal temperatures and adjust them immediately; trains a neural network using a simulated annealing algorithm to improve prediction accuracy; anticipates future temperature trends and proactively adjusts to avoid uneven drying; real-time classification avoids temperature fluctuations, improving the drying uniformity and strength stability of plaster; predictive control reduces ineffective refrigeration / heating, saving labor costs; and proactively interrupts abnormal temperature chains, reducing the incidence of drying defects and improving product qualification rates. It solves the problem of traditional temperature control lag and provides a stable drying environment throughout the entire lifecycle for building materials such as plaster.

[0015] Preferably, step S1 includes the following steps:

[0016] S11. Locate the key environmental locations that best reflect ambient temperature, and place temperature sensors at these locations; define the set of key environmental locations. a i Indicates the first set of key environmental locations There are 16 key environmental locations, where m represents the total number of key environmental locations.

[0017] S12. Collect historical temperature data and corresponding historical environmental temperature acquisition time points from a large number of key environmental locations using temperature sensors. This data includes various normal temperature data and all abnormal temperature data. Construct a historical environmental temperature matrix A, as follows.

[0018] ;

[0019] A im This represents the historical temperature data collected at the m-th key environmental location point at the i-th historical environmental temperature collection time point; n represents the total number of historical environmental temperature collection time points.

[0020] S13. Set the current ambient temperature acquisition period, and divide the ambient temperature acquisition period into time points to obtain the current ambient temperature acquisition time point set. b i This indicates the current ambient temperature collection time point set as the [number]th [time point]. There are 100 ambient temperature data collection time points, where o represents the total number of current ambient temperature data collection time points.

[0021] S14. Using the temperature sensor at the current ambient temperature collection time point set Set of key environmental locations at each data collection time point Temperature data was collected at each key environmental location to obtain the current environmental temperature data matrix B, as shown below.

[0022] ;

[0023] Among them, A im This represents the temperature data collected at the m-th key environmental location point at the i-th current ambient temperature collection time point;

[0024] By placing temperature sensors at key locations in the environment, errors in data collection can be reduced. By collecting historical environmental temperature data, which includes various normal temperature data and all abnormal temperature data, the temperature conditions of the dry environment can be classified, and the classification results can be used to determine whether the current drying temperature is abnormal.

[0025] Preferably, step S2 includes the following steps:

[0026] S21. Classify the temperature data in the historical ambient temperature data matrix A to obtain a central dataset of historical ambient temperature categories. ;where c i represents the center data of the i-th historical ambient temperature data type, and p represents the total number of historical ambient temperature data types;

[0027] S22. Calculate the Euclidean distance between each temperature data point in the current ambient temperature data matrix B and the corresponding temperature category center data point in the ambient temperature category center dataset c, to obtain the current ambient temperature Euclidean distance data matrix C; as follows.

[0028] ;

[0029] Where C im Let represent the Euclidean distance between the temperature data collected at the m-th key environmental location point at the i-th environmental temperature collection time point and the corresponding temperature type center data in the environmental temperature type center dataset c.

[0030] The temperature data type corresponding to the smallest Euclidean distance in each row of the current ambient temperature Euclidean distance data matrix C is taken as the corresponding temperature data type in the current ambient temperature data matrix B, thus obtaining the current ambient temperature data type set. ; where d i Indicates the current environment's first Types of temperature data at each time point;

[0031] S23. Based on the historical environmental temperature type central dataset And historical environmental temperature matrix, define the set of environmental abnormal temperature data categories , where e i q represents the i-th type of abnormal temperature data in the set of abnormal temperature data types, and q represents the total number of abnormal temperature data types.

[0032] The current ambient temperature data type set Data sets of abnormal environmental temperatures The comparison is performed. If a temperature data type belonging to the abnormal temperature data type set exists in the current ambient temperature data type set, the ambient temperature is adjusted; otherwise, no adjustment is needed.

[0033] By classifying the historical ambient temperature data matrix, a central dataset of historical ambient temperature types is obtained, which can more accurately identify the types of dry temperatures and thus more precisely determine whether there are abnormal temperatures in the current dry temperature data. By using the Euclidean distance between each temperature data in the historical ambient temperature data matrix and the central dataset of historical ambient temperature types, the dry temperature data can be monitored in real time and adjusted in a timely manner.

[0034] Preferably, step S21 includes the following steps:

[0035] S211. Randomly select several temperature data points from the historical environmental temperature data in the historical environmental temperature data matrix A as initial cluster centers to obtain the initial historical environmental temperature type center dataset y.

[0036] S212. Construct a particle swarm, setting the size of the particle swarm to l; obtain the particle swarm. ; where r i This represents the i-th particle in the particle swarm.

[0037] Based on the initial historical environmental temperature type center dataset y, the initial positions of each particle in the particle swarm r are randomly generated, resulting in the initial position set g of the particle swarm; the first maximum iteration number is set to h1 and the first current iteration number is set to h2; then the fitness function of the particle swarm is set as follows.

[0038] ;

[0039] In the formula, w represents the sum of the Euclidean distances between the historical ambient temperature data in the historical ambient temperature data matrix A and the selected initial historical ambient temperature type center data;

[0040] S213. Start the iteration. During each iteration, update the position of each particle in the particle swarm according to the fitness function of the particle swarm.

[0041] When h2≥h1, the iteration stops and the historical environmental temperature type center dataset c is output;

[0042] By using the particle swarm optimization algorithm, the historical ambient temperature data matrix can be quickly classified to obtain a central dataset of historical ambient temperature categories. By continuously adjusting the central data of each category to minimize the Euclidean distance of other data to the category center data, the categories of historical ambient temperatures can be identified more accurately.

[0043] Preferably, step S3 includes the following steps:

[0044] S31. Construct an initial neural network model, which includes an input layer, hidden layers, and an output layer. Set the number of nodes in the input layer of the initial neural network model to D, the number of nodes in the hidden layer to E, and the number of nodes in the output layer to F. Set the initial weights of the neural network model to G and the initial threshold to H.

[0045] S32. Set the training data ratio to J and the test data ratio to K. Divide the historical ambient temperature data matrix into a historical ambient temperature data training matrix and a historical ambient temperature data test matrix according to the training data ratio J and the test data ratio K. Set the historical ambient temperature data training matrix to train the label matrix.

[0046] S33. Set a training error threshold, input the historical ambient temperature data training matrix and the historical ambient temperature data training label matrix into the input node of the initial neural network model for training. When the training error of the neural network model is less than the training error threshold, stop training and obtain a trained neural network model.

[0047] S34. Input the historical ambient temperature data test matrix into the trained neural network model for testing. After the test is completed, the final neural network model is obtained.

[0048] By dividing the historical ambient temperature data matrix into a training matrix and a test matrix, we can ensure the model's generalization ability on unknown data, avoid overfitting, and thus improve the accuracy of the prediction model.

[0049] Preferably, step S34 includes the following steps:

[0050] S341. Input the historical ambient temperature data test matrix into the trained neural network model for testing, and obtain the test accuracy k1 after the test is completed.

[0051] S342. Set an accuracy threshold k2; when k1≥k2, use the trained neural network model as the final neural network model; otherwise, adjust and optimize the initial weights G and initial threshold H of the trained neural network model to obtain an optimized neural network model, and use the optimized neural network model as the final neural network model.

[0052] By setting an accuracy threshold, testing can be stopped once a certain accuracy level is achieved on a neural network model, thus avoiding overfitting.

[0053] Preferably, optimizing the initial weights and initial thresholds of the trained neural network model in step S342 to obtain an optimized neural network model includes the following steps:

[0054] S3421. Set the initial temperature parameters, cooling rate, and maximum number of iterations; set the current solution as a combination of initial weights and initial thresholds, denoted as the current solution; set the current optimal solution as the current solution; set the current temperature as the initial temperature parameter;

[0055] S3422. Define the objective function, denoted as the second fitness function;

[0056] Calculate the fitness value of the current solution;

[0057] S3423. Start the iteration operation; in each iteration, perform the following steps: randomly perturb the current solution in its neighborhood to generate a new solution; calculate the fitness value of the new solution; calculate the fitness difference.

[0058] If the fitness difference is less than 0, the new solution is accepted as the current solution; if the fitness difference is better than the optimal fitness difference, the current optimal solution is updated; otherwise, the optimal solution is not updated; the current temperature is updated according to the cooling strategy.

[0059] S3424. When the maximum number of iterations is reached, the iteration operation stops, and the optimal weight and optimal threshold contained in the current optimal solution are output.

[0060] S3425. Substitute the optimal weights and optimal thresholds into the trained neural network model to obtain the final neural network model;

[0061] By using simulated annealing to optimize the initial weights and thresholds of a neural network model, the model parameters can be optimized, thereby improving the accuracy of the neural network model.

[0062] Preferably, step S4 includes the following steps:

[0063] S41. Substitute the current ambient temperature data matrix into the final neural network model to predict the ambient temperature at future times, and obtain the future ambient temperature data matrix.

[0064] S42. Calculate the Euclidean distance between each temperature data in the future ambient temperature data matrix and the corresponding temperature type center data in the ambient temperature type center dataset to obtain the future ambient temperature Euclidean distance data matrix; take the temperature data type corresponding to the smallest Euclidean distance in each row of the future ambient temperature Euclidean distance data matrix as the type of the corresponding temperature data in the future ambient temperature data matrix to obtain the future ambient temperature data type matrix.

[0065] S43, Combine the future ambient temperature data type set with the ambient abnormal temperature data type set. Compare;

[0066] When the set of future environmental temperature data types contains temperature data types that belong to the set of abnormal environmental temperature data types. During this process, the ambient temperature is adjusted, and steps S41, S42, and S43 are repeated until no temperature data category belonging to the abnormal ambient temperature data category set exists in the future ambient temperature data category matrix. If the set of future environmental temperature data types does not contain any temperature data type belonging to the set of abnormal environmental temperature data types. During this time, there is no need to adjust the ambient temperature;

[0067] By using a final neural network model to predict future ambient temperature data, we can understand the development trend of ambient temperature in advance; when anomalies are found in future ambient temperature data, we can adjust the current temperature in time to keep the future ambient temperature at a normal level.

[0068] The product drying process temperature control system includes a data collection module, an ambient temperature classification module, a neural network model training and optimization module, and a neural network model application module.

[0069] The data collection module is used to collect a large amount of historical environmental temperature data and time collection point data, and to collect temperature data at each key environmental location point at the current environmental temperature collection time point.

[0070] The ambient temperature classification module is used to classify the historical ambient temperature data matrix into a historical ambient temperature data type central dataset; and to determine whether the current temperature is abnormal based on the historical ambient temperature data type central dataset; if the current temperature is abnormal, it is adjusted in a timely manner.

[0071] The neural network model training and optimization module is used to train and optimize the neural network model;

[0072] The neural network model application module uses the current ambient temperature of the final neural network model to predict the future ambient temperature and determine whether there is an anomaly in the future ambient temperature; if there is an anomaly in the future ambient temperature, the current temperature is adjusted to the normal level of the future ambient temperature.

[0073] The preparation method of plastering gypsum uses the product drying process temperature control method; it also includes a raw material mixing stage, a drying stage, a grinding stage, and a finished product packaging stage; the raw material mixing stage involves mixing gypsum powder, additives, and water in a certain proportion to obtain gypsum slurry; the drying stage involves feeding the gypsum slurry into the drying equipment and starting the product drying process temperature control system for dynamic temperature control; the grinding stage involves grinding the dried gypsum blocks to obtain plastering gypsum powder.

[0074] The present invention has the following beneficial effects:

[0075] By placing temperature sensors at key locations on the drying equipment, data collection errors are reduced, ensuring data accuracy. By collecting historical ambient temperature data containing various normal and abnormal temperature data, the various temperature conditions of the plaster drying environment can be classified, thereby more accurately determining whether there is an abnormality in the current drying temperature.

[0076] By classifying the historical ambient temperature data matrix, a central dataset of historical ambient temperature types is obtained. Through real-time classification and identification of abnormal temperatures in the dry environment, the system can quickly trigger precise regulation to avoid the impact of temperature fluctuations on the drying quality of plaster. By dividing the historical ambient temperature data matrix into training and testing matrices and optimizing it using simulated annealing algorithm, the accuracy of the neural network model is improved.

[0077] By substituting the current ambient temperature data matrix into the final neural network model, the future ambient temperature data matrix can be obtained, thereby predicting future temperature data. Based on the comparison results between the future ambient temperature data set and the abnormal ambient temperature data set, the drying temperature can be adjusted in advance to avoid uneven drying caused by sudden temperature changes, thus improving the automation and efficiency of the system.

[0078] Real-time monitoring and prediction of drying temperature can reduce human intervention in temperature control, reduce ineffective temperature control operations and energy consumption in the plaster preparation process, and ensure that the drying process is carried out under optimal temperature conditions, thereby improving production efficiency and quality.

[0079] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0080] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, the drawings can be obtained from these drawings without creative effort.

[0081] Figure 1 This is a schematic diagram of the temperature control method for the drying process of the product of the present invention;

[0082] Figure 2 This is a schematic diagram of the process of adjusting the temperature during the drying process of the product of the present invention, in which the type of product is determined by the current ambient temperature and then the decision is made on whether to adjust the current ambient temperature.

[0083] Figure 3 This is a schematic diagram of the process of temperature regulation during the drying process of the product of the present invention, in which a final neural network model is used to predict the future ambient temperature and make a type judgment to determine whether to adjust the ambient temperature in advance.

[0084] Figure 4 This is a schematic diagram of the module for the temperature control method during the drying process of the product of the present invention. Detailed Implementation

[0085] The technical solutions of the embodiments of the invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the invention, and not all embodiments. Based on the embodiments of the invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the invention.

[0086] In the description of this invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inner", etc., which indicate orientation or positional relationship, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the invention.

[0087] Example 1:

[0088] Please see Figure 1 , Figure 2 , Figure 3 This invention discloses a method for temperature control during the product drying process, comprising the following steps:

[0089] S1. Set up a set of key environmental locations, collect temperature data and collection time data of a large number of key environmental locations, and construct a historical environmental temperature data matrix; set the current environmental temperature collection time point, collect temperature data of each key environmental location point at the current environmental temperature collection time point, and obtain the current environmental temperature data matrix.

[0090] S1 includes the following steps:

[0091] S11. Locate the key environmental locations that best reflect ambient temperature, and place temperature sensors at these locations; define the set of key environmental locations. a i Indicates the first set of key environmental locations There are 16 key environmental locations, where m represents the total number of key environmental locations.

[0092] S12. Collect historical temperature data and corresponding historical environmental temperature acquisition time points from a large number of key environmental locations using temperature sensors. This data includes various normal temperature data and all abnormal temperature data. Construct a historical environmental temperature matrix A, as follows.

[0093] ;

[0094] A im This represents the historical temperature data collected at the m-th key environmental location point at the i-th historical environmental temperature collection time point; n represents the total number of historical environmental temperature collection time points.

[0095] S13. Set the current ambient temperature acquisition period, and divide the ambient temperature acquisition period into time points to obtain the current ambient temperature acquisition time point set. b i This indicates the current ambient temperature collection time point set as the [number]th [time point]. There are 100 ambient temperature data collection time points, where o represents the total number of current ambient temperature data collection time points.

[0096] S14. Using the temperature sensor at the current ambient temperature collection time point set Set of key environmental locations at each data collection time point Temperature data was collected at each key environmental location to obtain the current environmental temperature data matrix B, as shown below.

[0097] ;

[0098] Among them, A im This represents the temperature data collected at the m-th key environmental location point at the i-th current ambient temperature collection time point;

[0099] S2. Classify the temperature data in the historical ambient temperature data matrix to obtain a historical ambient temperature category center dataset; classify the temperature data in the current ambient temperature data matrix according to the historical ambient temperature category center dataset to obtain a current ambient temperature data category set; set an abnormal ambient temperature data category set according to the historical ambient temperature category center dataset; compare the current ambient temperature data category set with the abnormal ambient temperature data category set, and decide whether to adjust the current ambient temperature based on the comparison result;

[0100] S2 includes the following steps:

[0101] S21. Classify the temperature data in the historical ambient temperature data matrix A to obtain a central dataset of historical ambient temperature categories. ;where c i represents the center data of the i-th historical ambient temperature data type, and p represents the total number of historical ambient temperature data types;

[0102] S21 includes the following steps:

[0103] S211. Randomly select several temperature data points from the historical environmental temperature data in the historical environmental temperature data matrix A as initial cluster centers to obtain the initial historical environmental temperature type center dataset y.

[0104] S212. Construct a particle swarm, setting the size of the particle swarm to l; obtain the particle swarm. ; where r iThis represents the i-th particle in the particle swarm.

[0105] Based on the initial historical environmental temperature type center dataset y, the initial positions of each particle in the particle swarm r are randomly generated, resulting in the initial position set g of the particle swarm; the first maximum iteration number is set to h1 and the first current iteration number is set to h2; then the fitness function of the particle swarm is set as follows.

[0106] ;

[0107] In the formula, w represents the sum of the Euclidean distances between the historical ambient temperature data in the historical ambient temperature data matrix A and the selected initial historical ambient temperature type center data;

[0108] S213. Start the iteration. During each iteration, update the position of each particle in the particle swarm according to the fitness function of the particle swarm.

[0109] When h2≥h1, the iteration stops and the historical environmental temperature type center dataset c is output;

[0110] S22. Calculate the Euclidean distance between each temperature data point in the current ambient temperature data matrix B and the corresponding temperature category center data point in the ambient temperature category center dataset c, to obtain the current ambient temperature Euclidean distance data matrix C; as follows.

[0111] ;

[0112] Where C im Let represent the Euclidean distance between the temperature data collected at the m-th key environmental location point at the i-th environmental temperature collection time point and the corresponding temperature type center data in the environmental temperature type center dataset c.

[0113] The temperature data type corresponding to the smallest Euclidean distance in each row of the current ambient temperature Euclidean distance data matrix C is taken as the corresponding temperature data type in the current ambient temperature data matrix B, thus obtaining the current ambient temperature data type set. ; where d i Indicates the current environment's first Types of temperature data at each time point;

[0114] S23. Based on the historical environmental temperature type central dataset And historical environmental temperature matrix, define the set of environmental abnormal temperature data categories , where e i q represents the i-th type of abnormal temperature data in the set of abnormal temperature data types, and q represents the total number of abnormal temperature data types.

[0115] The current ambient temperature data type set Data sets of abnormal environmental temperatures The comparison is performed. If a temperature data type belonging to the abnormal temperature data type set exists in the current ambient temperature data type set, the ambient temperature is adjusted; otherwise, no adjustment is needed.

[0116] S3. Construct an initial neural network model. Divide the historical ambient temperature data matrix into a historical ambient temperature data training matrix and a historical ambient temperature data test matrix according to the proportion. Train, test and optimize the initial neural network model to obtain the final neural network model.

[0117] S3 includes the following steps:

[0118] S31. Construct an initial neural network model, which includes an input layer, hidden layers, and an output layer. Set the number of nodes in the input layer of the initial neural network model to D, the number of nodes in the hidden layer to E, and the number of nodes in the output layer to F. Set the initial weights of the neural network model to G and the initial threshold to H.

[0119] S32. Set the training data ratio to J and the test data ratio to K. Divide the historical ambient temperature data matrix into a historical ambient temperature data training matrix and a historical ambient temperature data test matrix according to the training data ratio J and the test data ratio K. Set the historical ambient temperature data training matrix to train the label matrix.

[0120] S33. Set a training error threshold, input the historical ambient temperature data training matrix and the historical ambient temperature data training label matrix into the input node of the initial neural network model for training. When the training error of the neural network model is less than the training error threshold, stop training and obtain a trained neural network model.

[0121] S34. Input the historical ambient temperature data test matrix into the trained neural network model for testing. After the test is completed, the final neural network model is obtained.

[0122] S34 includes the following steps:

[0123] S341. Input the historical ambient temperature data test matrix into the trained neural network model for testing, and obtain the test accuracy k1 after the test is completed.

[0124] S342. Set an accuracy threshold k2; when k1≥k2, use the trained neural network model as the final neural network model; otherwise, adjust and optimize the initial weights G and initial threshold H of the trained neural network model to obtain an optimized neural network model, and use the optimized neural network model as the final neural network model.

[0125] The optimization of the initial weights and initial thresholds of the trained neural network model in step S342 to obtain the optimized neural network model includes the following steps:

[0126] S3421. Set the initial temperature parameter T1, cooling rate α, and maximum number of iterations M; set the current solution as a combination of initial weight G and initial threshold H, denoted as the current solution W; set the current optimal solution as the current solution W; set the current temperature T as the initial temperature parameter T1.

[0127] S3422. Define the objective function, denoted as the second fitness function; the formula for the second fitness function is as follows.

[0128] ;

[0129] Where v1 represents the bias amount;

[0130] Calculate the fitness value f of the current solution;

[0131] S3423. Start the iteration operation; in each iteration, perform the following steps: randomly perturb the current solution W in its neighborhood to generate a new solution W1; calculate the fitness value f1 of the new solution W1; calculate the fitness difference Δf = f - f1;

[0132] If the fitness difference Δf < 0, then accept the new solution W1 as the current solution; if f1 is better than the optimal fitness difference f best If the optimal solution W1 is found, then update the current optimal solution; otherwise, do not update the optimal solution; update the current temperature T according to the cooling strategy. new =αT old ;

[0133] S3424. When the maximum number of iterations M is reached, the iteration operation stops, and the optimal weight R and optimal threshold U contained in the current optimal solution are output.

[0134] S3425. Substitute the optimal weights R and the optimal thresholds U into the trained neural network model to obtain the final neural network model;

[0135] S4. Substitute the current ambient temperature data matrix into the final neural network model to obtain the future ambient temperature data matrix; classify the future ambient temperature data matrix to obtain a future ambient temperature data category set; compare the future ambient temperature data category set with the abnormal ambient temperature data category set; and decide whether to adjust the current ambient temperature based on the comparison result.

[0136] S4 includes the following steps:

[0137] S41. Substitute the current ambient temperature data matrix into the final neural network model to predict the ambient temperature at future times, and obtain the future ambient temperature data matrix.

[0138] S42. Calculate the Euclidean distance between each temperature data in the future ambient temperature data matrix and the corresponding temperature type center data in the ambient temperature type center dataset to obtain the future ambient temperature Euclidean distance data matrix; take the temperature data type corresponding to the smallest Euclidean distance in each row of the future ambient temperature Euclidean distance data matrix as the type of the corresponding temperature data in the future ambient temperature data matrix to obtain the future ambient temperature data type matrix.

[0139] S43, Combine the future ambient temperature data type set with the ambient abnormal temperature data type set. Compare;

[0140] When the set of future environmental temperature data types contains temperature data types that belong to the set of abnormal environmental temperature data types. During this process, the ambient temperature is adjusted, and steps S41, S42, and S43 are repeated until no temperature data category belonging to the abnormal ambient temperature data category set exists in the future ambient temperature data category matrix. If the set of future environmental temperature data types does not contain any temperature data type belonging to the set of abnormal environmental temperature data types. During this time, there is no need to adjust the ambient temperature.

[0141] Example 2:

[0142] Please see Figure 4 The product drying process temperature control system includes a data collection module, an ambient temperature classification module, a neural network model training and optimization module, and a neural network model application module.

[0143] The data collection module is used to collect a large amount of historical environmental temperature data and time collection point data, and to collect temperature data at each key environmental location point at the current environmental temperature collection time point.

[0144] The ambient temperature classification module is used to classify the historical ambient temperature data matrix into a historical ambient temperature data type central dataset; and to determine whether the current temperature is abnormal based on the historical ambient temperature data type central dataset; if the current temperature is abnormal, it is adjusted in a timely manner.

[0145] The neural network model training and optimization module is used to train and optimize the neural network model;

[0146] The neural network model application module uses the current ambient temperature of the final neural network model to predict the future ambient temperature and determine whether there is an anomaly in the future ambient temperature; if there is an anomaly in the future ambient temperature, the current temperature is adjusted to the normal level of the future ambient temperature.

[0147] Example 3:

[0148] The method for preparing plaster plaster uses the temperature control method for the product drying process; it also includes a raw material mixing stage, a drying stage, a grinding stage, and a finished product packaging stage.

[0149] The raw material mixing stage involves adding gypsum powder, additives, and water to a mixer in a specific ratio and stirring at 45±2℃ for 15 minutes to obtain a uniform gypsum slurry.

[0150] The drying stage involves feeding the gypsum slurry into a tunnel drying equipment and connecting it to an automatic temperature optimization and control system. The temperature is controlled by the product drying process temperature regulation system of this invention. Initial stage (0-2 hours): maintain 60±2℃ to promote moisture evaporation; middle stage (2-5 hours): raise the temperature to 75±1℃ to accelerate drying; later stage (5-7 hours): lower the temperature to 55±1℃ to avoid overheating.

[0151] The grinding stage involves feeding the dried gypsum blocks into a Raymond mill and grinding them to a fineness of 120 mesh (testing standard: laser particle size analyzer).

[0152] The finished product packaging stage involves automatic packaging machines to dispense the product into 25kg bags, and sampling tests are conducted (strength test: flexural strength ≥3.5MPa, compressive strength ≥6.0MPa; setting time: initial setting ≥60min, final setting ≤240min).

[0153] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0154] The preferred embodiments of the invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for temperature control during product drying, characterized in that, Includes the following steps: S1. Set up a set of key environmental locations, collect a large amount of temperature data and collection time data of key environmental locations, and construct a historical environmental temperature data matrix. Set the current ambient temperature collection time point, and collect temperature data at each key location in the environment at the current ambient temperature collection time point to obtain the current ambient temperature data matrix; S2. Classify the temperature data in the historical ambient temperature data matrix to obtain a historical ambient temperature category center dataset; classify the temperature data in the current ambient temperature data matrix to obtain a current ambient temperature data category set; set an abnormal ambient temperature data category set based on the historical ambient temperature category center dataset; compare the current ambient temperature data category set with the abnormal ambient temperature data category set, and decide whether to adjust the current ambient temperature based on the comparison result. S3. Construct an initial neural network model. Divide the historical ambient temperature data matrix into a historical ambient temperature data training matrix and a historical ambient temperature data test matrix according to the proportion. Train, test and optimize the initial neural network model to obtain the final neural network model. S4. Substitute the current ambient temperature data matrix into the final neural network model to obtain the future ambient temperature data matrix; The future ambient temperature data matrix is ​​classified to obtain a set of future ambient temperature data categories. The set of future ambient temperature data categories is compared with the set of abnormal ambient temperature data categories, and a decision is made on whether to adjust the current ambient temperature based on the comparison result.

2. The temperature control method for the product drying process according to claim 1, characterized in that, S1 includes the following steps: S11. Locate the key environmental locations that best reflect the ambient temperature, and place temperature sensors at these key locations; set up a set of key environmental locations. S12. Collect historical temperature data of a large number of key environmental locations and corresponding historical environmental temperature collection time points through temperature sensors, and construct a historical environmental temperature matrix. S13. Set the current ambient temperature acquisition period, divide the ambient temperature acquisition period into time points, and obtain the current ambient temperature acquisition time point set. S14. Using the temperature sensor, the temperature data of each key environmental location point in the set of key environmental location points is collected at each collection time point in the current ambient temperature collection time point set, to obtain the current ambient temperature data matrix.

3. The temperature control method for the product drying process according to claim 1, characterized in that, S2 includes the following steps: S21. Classify the temperature data in the historical ambient temperature data matrix to obtain a central dataset of historical ambient temperature categories. S22. Calculate the Euclidean distance between each temperature data in the current ambient temperature data matrix and the corresponding temperature type center data in the ambient temperature type center dataset to obtain the current ambient temperature Euclidean distance data matrix. The temperature data type corresponding to the smallest Euclidean distance in each row of the current ambient temperature Euclidean distance data matrix is ​​taken as the type of temperature data in the current ambient temperature data matrix, thus obtaining the current ambient temperature data type set. S23. Based on the historical ambient temperature type center dataset and the historical ambient temperature matrix, set the ambient abnormal temperature data type set; The current ambient temperature data set is compared with the abnormal ambient temperature data set. If there is a temperature data type in the current ambient temperature data set that belongs to the abnormal ambient temperature data set, the ambient temperature is adjusted; otherwise, no adjustment is needed.

4. The temperature control method for the product drying process according to claim 3, characterized in that, S21 includes the following steps: S211. Randomly select several temperature data points from the historical environmental temperature data in the historical environmental temperature data matrix as initial cluster centers to obtain the initial historical environmental temperature type center dataset. S212. Construct a particle swarm and obtain the historical environmental temperature type center dataset by iteratively updating the position of the particle swarm.

5. The temperature control method for the product drying process according to claim 1, characterized in that, S3 includes the following steps: S31. Construct the initial neural network model and set the relevant parameters of the model; S32. Set the training data ratio to J and the test data ratio to K. Divide the historical ambient temperature data matrix into a historical ambient temperature data training matrix and a historical ambient temperature data test matrix according to the training data ratio J and the test data ratio K. Set the historical ambient temperature data training matrix to train the label matrix. S33. Set a training error threshold, input the historical ambient temperature data training matrix and the historical ambient temperature data training label matrix into the input node of the initial neural network model for training. When the training error of the neural network model is less than the training error threshold, stop training and obtain a trained neural network model. S34. Input the historical ambient temperature data test matrix into the trained neural network model for testing. After the test is completed, the final neural network model is obtained.

6. The temperature control method for the product drying process according to claim 5, characterized in that, S34 includes the following steps: S341. Input the historical ambient temperature data test matrix into the trained neural network model for testing, and obtain the test accuracy k1 after the test is completed. S342. Set an accuracy threshold k2; when k1≥k2, use the trained neural network model as the final neural network model; otherwise, adjust and optimize the initial weights G and initial threshold H of the trained neural network model to obtain an optimized neural network model, and use the optimized neural network model as the final neural network model.

7. The temperature control method for the product drying process according to claim 6, characterized in that, The optimization of the initial weights and initial thresholds of the trained neural network model in step S342 to obtain the optimized neural network model includes the following steps: S3421. Set the initial temperature parameters, cooling rate, and maximum number of iterations; set the current solution as a combination of initial weights and initial thresholds, denoted as the current solution; set the current optimal solution as the current solution; set the current temperature as the initial temperature parameter; S3422. Define the objective function, denoted as the second fitness function; Calculate the fitness value of the current solution; S3423. Start the iteration operation; in each iteration, perform the following steps: randomly perturb the current solution in its neighborhood to generate a new solution; calculate the fitness value of the new solution; calculate the fitness difference. If the fitness difference is less than 0, the new solution is accepted as the current solution; if the fitness difference is better than the optimal fitness difference, the current optimal solution is updated; otherwise, the optimal solution is not updated; the current temperature is updated according to the cooling strategy. S3424. When the maximum number of iterations is reached, the iteration operation stops, and the optimal weight and optimal threshold contained in the current optimal solution are output. S3425. Substitute the optimal weights and optimal thresholds into the trained neural network model to obtain the final neural network model.

8. The method for temperature control during the product drying process according to claim 1, characterized in that, S4 includes the following steps: S41. Substitute the current ambient temperature data matrix into the final neural network model to predict the ambient temperature at future times, and obtain the future ambient temperature data matrix. S42. Calculate the Euclidean distance between each temperature data in the future ambient temperature data matrix and the corresponding temperature type center data in the ambient temperature type center dataset to obtain the future ambient temperature Euclidean distance data matrix; take the temperature data type corresponding to the smallest Euclidean distance in each row of the future ambient temperature Euclidean distance data matrix as the type of the corresponding temperature data in the future ambient temperature data matrix to obtain the future ambient temperature data type matrix. S43. Compare the set of future ambient temperature data types with the set of abnormal ambient temperature data types; When there is a temperature data type in the future ambient temperature data type set that belongs to the set of abnormal ambient temperature data types, the ambient temperature is adjusted, and S41, S42 and S43 are repeated until there is no temperature data type in the future ambient temperature data type matrix that belongs to the set of abnormal ambient temperature data types; if there is no temperature data type in the future ambient temperature data type set that belongs to the set of abnormal ambient temperature data types, then there is no need to adjust the ambient temperature.

9. A temperature control system for the product drying process, characterized in that, The system for implementing the product drying process temperature control method as described in any one of claims 1-8 includes a data collection module, an ambient temperature classification module, a neural network model training and optimization module, and a neural network model application module.

10. A method for preparing plaster, characterized in that, The product drying process temperature control method used as described in any one of claims 1-8; it further includes a raw material mixing stage, a drying stage, a grinding stage, and a finished product packaging stage; the raw material mixing stage involves mixing gypsum powder, additives, and water in a certain proportion to obtain gypsum slurry; the drying stage involves feeding the gypsum slurry into the drying equipment and starting the product drying process temperature control system for dynamic temperature control; The grinding stage involves grinding the dried gypsum blocks to obtain plaster powder.

Images