An intelligent material environment adjustment system for a vacuum clay kneading machine

Abstract

The present application discloses an intelligent material environment adjustment system for a vacuum clay kneading machine, which relates to the field of artificial intelligence technology. It solves the technical problems in the prior art that the influence of environmental factors on the mixing of clay materials is not considered, and at the same time, the lack of consideration of adaptively controlling the vacuum clay kneading machine to make the finished clay materials meet the target requirements, resulting in poor finished state of the clay materials and low production efficiency of the vacuum clay kneading machine. By generating material data according to demand data and generating initial equipment parameters therefrom; generating estimated parameters of the to-be-finished product according to the material data, the to-be-finished product data and the environmental data; generating equipment adjustment parameters according to the estimated parameters of the to-be-finished product, considering the influence of environmental factors on the change of the to-be-finished product data in the mixer, predicting in real time the gap between the to-be-finished product data and the finished product data, adding materials in advance or adjusting the equipment parameters, ensuring that the demand data can be obtained on time and with quality at the minimum cost, and improving the accuracy and efficiency of the vacuum clay kneading machine.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence technology, and specifically relates to an intelligent material environment adjustment system for a vacuum pug mill. Background Art

[0002] A vacuum pug mill is a special equipment used in industries such as ceramics and building materials. It is mainly used for degassing, stirring, and homogenizing clay materials to improve the plasticity, density, and forming properties of the clay materials. With the continuous progress of technology, vacuum pug mills are developing towards the direction of intelligence, energy conservation, and multi-functionality, providing more efficient and environmentally friendly solutions for the industry.

[0003] The prior art (a patent application for an invention patent with a publication number of CN113370386A) discloses a clay pugging method for ceramic rollers and its vacuum pug mill. Among them, the clay pugging method for ceramic rollers includes the following steps: mixing ceramic roller clay materials, binders, and water; evacuating and extruding the pugging mixture into strip-shaped clay materials; while pugging, evacuating and extruding to remove the residual gas in the materials, and obtaining ceramic roller clay materials with high density, tiny and uniformly distributed embedded gas pores, effectively improving the structural uniformity, strength, and thermal shock resistance of the prepared ceramic rollers; at the same time, the present invention also discloses a vacuum pug mill, configured with extrusion and vacuum pumping devices, optimizing the extrusion pressure and the aperture of the extrusion head, with a compact structure and reliable performance, which is beneficial to improving the granulation density and uniformity while reducing the complexity of the equipment and improving the production efficiency of ceramic rollers.

[0004] The above solution mixes ceramic roller clay materials, binders, and water and inputs them into a vacuum pug mill for stirring, without considering the influence of environmental factors on the clay material stirring, and at the same time lacks the consideration of adaptively controlling the vacuum pug mill to make the finished clay material meet the target requirements, resulting in poor finished state of the clay material and low production efficiency of the vacuum pug mill; therefore, the intelligent material adjustment system for the vacuum pug mill still needs further improvement. Summary of the Invention

[0005] This application aims to solve at least one of the technical problems existing in the prior art; for this purpose, this application proposes an intelligent material environment adjustment system for a vacuum pug mill, which is used to solve the technical problems that the prior art does not consider the influence of environmental factors on the clay material stirring, and at the same time lacks the consideration of adaptively controlling the vacuum pug mill to make the finished clay material meet the target requirements, resulting in poor finished state of the clay material and low production efficiency of the vacuum pug mill.

[0006] To achieve the above object, the first aspect of this application provides an intelligent material environment adjustment system for a vacuum pug mill, including: a data acquisition module, a data analysis module, a control module, and a database;

[0007] The data acquisition module: obtains environmental data, equipment parameters, semi-finished product data, and demand data through data acquisition devices;

[0008] The data analysis module: generates material data according to demand data; obtains the corresponding material parameter data when the material data enters the mixer; generates initial equipment parameters according to the material data and the material parameter data; obtains the semi-finished product data in the mixer in real time; generates estimated semi-finished product parameters according to the material data, the semi-finished product data, and the environmental data; generates equipment adjustment parameters according to the estimated semi-finished product parameters;

[0009] The control module: transports the corresponding materials to the mixer according to the material data; assigns the equipment adjustment parameters to the corresponding equipment parameters for control.

[0010] Through the above steps, this application considers the influence of environmental factors on the change of semi-finished product data in the mixer, and predicts in real time the gap between the semi-finished product data and the finished product data, adding materials in advance or adjusting the equipment parameters to ensure that the demand data can be obtained on time and with high quality at the lowest cost, improving the accuracy and efficiency of the vacuum clay kneader.

[0011] Further, generating the material data according to the demand data includes:

[0012] Obtaining the demand data; the demand data includes the finished product name, finished product parameters, finished product weight, and demand time;

[0013] Searching for the material data in the material data table according to the finished product name and the finished product weight; the material data table is constructed by experts according to the demand data; the material data includes several material names and their corresponding material weights.

[0014] Further, generating the initial equipment parameters according to the material data and the material parameter data includes:

[0015] Obtaining the material data, the demand data, and the corresponding material parameter data when the material data enters the mixer;

[0016] Inputting the material data, the demand data, and the material parameter data into a parameter estimation model to obtain the initial equipment parameters; the parameter estimation model is constructed through an artificial intelligence model; the initial equipment parameters include operating power, operating time, stirring speed, and equipment vacuum degree.

[0017] Through the above steps, this application automatically generates the required material data according to the demand data, generates the initial equipment parameters according to the material data, the material parameter data, and the demand data, stirs the materials with the initial equipment parameters, and automatically controls the vacuum clay kneader, reducing manual operation and improving the efficiency of the vacuum clay kneader.

[0018] Further, the parameter prediction model is constructed by an artificial intelligence model, including:

[0019] Obtain a number of historical material data, demand data, material parameter data and their corresponding initial equipment parameters;

[0020] Divide a number of historical material data, demand data, material parameter data and their corresponding initial equipment parameters into training data, validation data and test data; perform data preprocessing on the training data, validation data and test data to obtain a training set, a validation set and a test set;

[0021] Select an artificial intelligence model as the basic model;

[0022] Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain a pre-trained model;

[0023] Verify the pre-trained model on the test set, and finally obtain a parameter prediction model that takes material data, demand data and material parameter data as inputs and outputs the initial equipment parameters.

[0024] Further, the generation of the estimated parameters of the to-be-finished product according to the material data, the to-be-finished product data and the environmental data includes:

[0025] Obtain historical to-be-finished product data, historical material data, historical environmental data, historical equipment parameters and recording time in the mixer in real time; the equipment parameters are initial equipment parameters or equipment adjustment parameters; the environmental data includes external environmental data and internal environmental data;

[0026] Integrate a number of historical to-be-finished product data, historical material data, historical environmental data and historical equipment parameters into a number of prediction sequences according to the recording time;

[0027] Input the prediction sequence into the to-be-finished product parameter prediction model for prediction to obtain the to-be-finished product estimated parameters; the to-be-finished product parameter prediction model is constructed by a machine learning model.

[0028] Further, the to-be-finished product parameter prediction model is constructed by a machine learning model, including:

[0029] Obtain a number of historical to-be-finished product data, historical material data, historical environmental data, historical equipment parameters and recording time;

[0030] Perform data preprocessing on a number of historical to-be-finished product data, historical material data, historical environmental data, historical equipment parameters and recording time to obtain the preprocessed historical to-be-finished product data, historical material data, historical environmental data, historical equipment parameters and recording time;

[0031] Integrate the historical semi-finished product data, historical material data, historical environmental data, and historical equipment parameters with the same recording time into a time feature vector;

[0032] Use the principal component analysis method to perform feature reduction on several time feature vectors to obtain the reduced time feature vectors; the reduced time feature vectors are the time feature vectors obtained by the principal component analysis method when the cumulative contribution rate reaches M%; where M is a constant and M ∈ [0, 100];

[0033] Integrate each time feature vector into several prediction sequences in the order of recording time;

[0034] Divide several prediction sequences into a training set, a test set, and a validation set in chronological order;

[0035] Select a machine learning model as the basic model;

[0036] Train the basic model through the training set and adjust the learning rate or other hyperparameters on the validation set to obtain a pre-trained model;

[0037] Verify the pre-trained model on the test set, and finally obtain a semi-finished product parameter prediction model with several prediction sequences as the input,

[0038] and the predicted parameters of the semi-finished product as the output.

[0039] Further, the generation of equipment adjustment parameters according to the predicted parameters of the semi-finished product includes:

[0040] Obtain the predicted parameters of the semi-finished product and the finished product parameters;

[0041] Construct a minimized cost function based on the predicted parameters of the semi-finished product and the finished product parameters;

[0042] Solve the minimized cost function through a deep reinforcement learning model to obtain the optimal solution, and generate an adjustment plan based on this;

[0043] Generate equipment adjustment parameters according to the adjustment plan.

[0044] Further, the construction of the minimized cost function based on the predicted parameters of the semi-finished product and the finished product parameters includes:

[0045] Obtain the predicted parameters of the semi-finished product and the finished product parameters;

[0046] Obtain several parameter differences by taking the difference between the finished product parameters and the predicted parameters of the semi-finished product;

[0047] Calculate the amount of added material according to the parameter differences;

[0048] Generate the estimated change in equipment energy according to the parameter differences;

[0049] Construct the minimized cost function FC through the formula: where TWL represents the amount of added material, DJ represents the unit price corresponding to the added material, DNJ represents the unit energy price, and NYL represents the estimated change in equipment energy.

[0050] Furthermore, calculating the amount of added material according to the parameter difference includes:

[0051] Obtain several parameter differences and data of the to-be-finished product; the data of the to-be-finished product includes the weight of the to-be-finished product DCZ.

[0052] Select the parameter difference CC corresponding to the maximum absolute value of several parameter differences.

[0053] Judge whether the parameter difference is within the corresponding difference range.

[0054] Yes, the amount of added material TWL corresponding to the parameter difference is 0.

[0055] No, when the parameter difference is greater than the maximum value in its corresponding difference range;

[0056] Calculate the amount of added material TWL through the formula:

[0057] When the parameter difference is less than the minimum value in its corresponding difference range;

[0058] Calculate the amount of added material TWL through the formula: where DTWL represents the amount of added material required per unit difference per unit weight of the to-be-finished product.

[0059] Furthermore, generating the estimated change in equipment energy according to the parameter difference includes:

[0060] Obtain several parameter differences, the weight of the to-be-finished product, and the estimated parameters of the to-be-finished product; the estimated parameters of the to-be-finished product include the estimated duration.

[0061] Select the parameter difference corresponding to the maximum absolute value of several parameter differences.

[0062] Judge whether the parameter difference is within the corresponding difference range.

[0063] Yes, the estimated change in equipment energy NYL corresponding to the parameter difference is 0.

[0064] No, input the parameter difference, the weight of the to-be-finished product, and the estimated duration into the energy estimation model to obtain the estimated change in equipment energy; the energy estimation model is constructed through an artificial intelligence model.

[0065] ​This application calculates the corresponding amount of added materials or the estimated change in equipment energy based on the parameter difference between the estimated parameters of the semi-finished product and the corresponding finished product parameters. When adding materials, different parameter differences require adding different materials, providing accurate data support for finding the minimum cost and improving the quality of mud production.

[0066] Further, the energy estimation model is constructed through an artificial intelligence model, including:

[0067] Obtain a number of historical parameter differences, the weight of the semi-finished product, the estimated duration, and their corresponding estimated changes in equipment energy;

[0068] Divide a number of historical parameter differences, the weight of the semi-finished product, the estimated duration, and their corresponding estimated changes in equipment energy into training data, validation data, and test data; perform data preprocessing on the training data, validation data, and test data to obtain a training set, a validation set, and a test set;

[0069] Select an artificial intelligence model as the basic model;

[0070] Train the basic model with the training set and adjust the learning rate and hyperparameters on the validation set to obtain a pre-trained model;

[0071] Verify the pre-trained model on the test set, and finally obtain an energy estimation model with the input parameter difference, the weight of the semi-finished product, and the estimated duration, and the output being the estimated change in equipment energy.

[0072] Compared with the prior art, the beneficial effects of this application are:

[0073] This application generates material data based on demand data; transports the corresponding materials to the mixer according to the material data; obtains the material parameter data corresponding to the material data when entering the mixer; generates the initial equipment parameters according to the material data and the material parameter data; obtains the semi-finished product data in the mixer in real time; generates the estimated parameters of the semi-finished product according to the material data, the semi-finished product data, and the environmental data; generates equipment adjustment parameters according to the estimated parameters of the semi-finished product, assigns the equipment adjustment parameters to the corresponding equipment parameters for control, considers the influence of environmental factors on the change of the semi-finished product data in the mixer, and predicts the gap between the semi-finished product data and the finished product data in real time, adding materials in advance or adjusting the equipment parameters to ensure that the demand data can be obtained on time and with high quality at the lowest cost, improving the accuracy and efficiency of the vacuum clay kneader.

[0074] This application obtains the data of semi-finished products, material data, environmental data, and equipment parameters in the blender in real time according to the time sequence, predicts the estimated parameters of the semi-finished products when the required time arrives, and compares the finished product parameters to detect abnormal conditions of the semi-finished product data, making decisions dynamically in advance, reducing time consumption, and improving the production efficiency and mud quality of the vacuum pug mill.

[0075] This application calculates the gap between the estimated parameters of the semi-finished product and the finished product parameters, and considers adding materials or adjusting the equipment parameters to eliminate the gap. Adding materials or adjusting the equipment will result in different cost consumptions. Therefore, the minimum cost that can eliminate the gap is found between the material cost and the energy cost, so as to achieve the goal of making the parameters of the semi-finished product meet the requirements and reducing resource consumption on the basis of meeting the requirements. Brief Description of the Drawings

[0076] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0077] Figure 1 It is a schematic diagram of the principle of a material environment intelligent regulation system for a vacuum pug mill according to the present application;

[0078] Figure 2 It is a flowchart of a material environment intelligent regulation method for a vacuum pug mill according to the present application;

[0079] Figure 3 It is a flowchart for generating the estimated change amount of equipment energy according to the present application. Detailed Embodiments

[0080] The following will clearly and completely describe the technical solutions of the present application in combination with the embodiments. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0081] Please refer to Figure 1 - Figure 2 , the first aspect embodiment of the present application provides a material environment intelligent regulation system for a vacuum pug mill, including: a data acquisition module, a data analysis module, a control module, and a database;

[0082] Data acquisition module: Obtain environmental data, equipment parameters, semi-finished product data, and demand data through data acquisition devices; the data acquisition devices include various sensors, such as temperature sensors and humidity sensors, etc.;

[0083] Data analysis module: Generate material data according to requirement data. The material data refers to several raw material data that form the requirement data. Obtain the material parameter data corresponding to the material data when it enters the mixer. The material parameter data refers to various parameters of the material, such as material temperature, material humidity, etc. Generate the initial equipment parameters according to the material data and the material parameter data. The initial equipment parameters refer to the initial parameter settings of the mixing equipment. Real-time obtain the data of the semi-finished product in the mixer. The semi-finished product data refers to the data when several raw materials are preparing to form the requirement data in the mixer. Generate the semi-finished product estimated parameters according to the material data, the semi-finished product data and the environmental data. The semi-finished product estimated parameters refer to the parameter data that the semi-finished product data can form at the required time, such as the semi-finished product temperature, the semi-finished product humidity and the semi-finished product viscosity, etc. Generate the equipment adjustment parameters according to the semi-finished product estimated parameters. The equipment adjustment parameters refer to the equipment parameters for adjusting the initial equipment parameters.

[0084] Control module: Transmit the corresponding material to the mixer according to the material data. Assign the equipment adjustment parameters to the corresponding equipment parameters for control.

[0085] In this embodiment, generating the material data according to the requirement data includes:

[0086] Obtain the requirement data. The requirement data includes the finished product name, the finished product parameters, the finished product weight and the required time.

[0087] Search for the material data in the material data table according to the finished product name and the finished product weight. The material data table is constructed by experts according to the requirement data. The material data includes several material names and their corresponding material weights.

[0088] In this embodiment, generating the initial equipment parameters according to the material data and the material parameter data includes:

[0089] Obtain the material data, the requirement data and the material parameter data corresponding to the material data when it enters the mixer.

[0090] Input the material data, the requirement data and the material parameter data into the parameter estimation model to obtain the initial equipment parameters. The parameter estimation model is constructed through an artificial intelligence model. The initial equipment parameters include the operating power, the operating time, the mixing speed and the equipment vacuum degree.

[0091] In this embodiment, constructing the parameter estimation model through an artificial intelligence model includes:

[0092] Obtain several historical material data, requirement data and material parameter data and their corresponding historical initial equipment parameters.

[0093] Divide a number of historical material data, demand data, material parameter data, and their corresponding initial historical equipment parameters into training data, validation data, and test data; perform data preprocessing on the training data, validation data, and test data to obtain a training set, a validation set, and a test set; the ratio among the training set, the test set, and the validation set is 7:2:1;

[0094] Select an artificial intelligence model as the basic model; the artificial intelligence model includes a neural network model, etc.;

[0095] Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain a pre-trained model;

[0096] Verify the pre-trained model on the test set, and finally obtain a parameter prediction model that takes the input material data, demand data, and material parameter data and outputs the initial equipment parameters.

[0097] Generating the estimated parameters of the to-be-finished product according to the material data, the to-be-finished product data, and the environmental data in this embodiment includes:

[0098] Obtain the historical to-be-finished product data, historical material data, historical environmental data, historical equipment parameters, and recording time in the blender in real time; the equipment parameters are the initial equipment parameters or the equipment adjustment parameters; the environmental data includes external environmental data and internal environmental data; when the material is predicted for the first time in the blender, the corresponding equipment parameter is the initial equipment parameter; when the material is predicted for the Nth time in the blender, the corresponding equipment parameter is the equipment adjustment parameter corresponding to the (N - 1)th prediction, where N is an integer and N>1; the external environmental data refers to the environmental data outside the blender, and the internal environmental data refers to the environmental data inside the blender;

[0099] Integrate a number of historical to-be-finished product data, historical material data, historical environmental data, and historical equipment parameters into a number of prediction sequences according to the recording time;

[0100] Input the prediction sequence into the to-be-finished product parameter prediction model for prediction to obtain the to-be-finished product estimated parameters; the to-be-finished product parameter prediction model is constructed by a machine learning model.

[0101] The to-be-finished product parameter prediction model in this embodiment is constructed by a machine learning model, including:

[0102] Obtain a number of historical to-be-finished product data, historical material data, historical environmental data, historical equipment parameters, and recording time;

[0103] Perform data preprocessing on a number of historical to-be-finished product data, historical material data, historical environmental data, historical equipment parameters, and recording time to obtain the preprocessed a number of historical to-be-finished product data, historical material data, historical environmental data, historical equipment parameters, and recording time;

[0104] Integrate the historical semi-finished product data, historical material data, historical environmental data, and historical equipment parameters with the same recording time into a time feature vector;

[0105] Use the principal component analysis method to perform feature reduction on several time feature vectors to obtain the reduced time feature vectors; the reduced time feature vectors are the time feature vectors obtained by the principal component analysis method after the cumulative contribution rate reaches M%; where M is a constant, M ∈ [0, 100], and the specific value is set according to experience. In this embodiment, M is set to 95;

[0106] Integrate each time feature vector into several prediction sequences in the order of recording time;

[0107] Divide several prediction sequences into a training set, a test set, and a validation set in chronological order; the ratio between the training set, the test set, and the validation set is 7:2:1;

[0108] Select a machine learning model as the basic model; the machine learning model includes the LSTM model, etc.;

[0109] Train the basic model with the training set and adjust the learning rate or other hyperparameters on the validation set to obtain a pre-trained model;

[0110] Verify the pre-trained model on the test set, and finally obtain a semi-finished product parameter prediction model with several prediction sequences as the input and the predicted parameters of the semi-finished product as the output.

[0111] In this embodiment, by collecting the semi-finished product data, material data, environmental data, and equipment parameters in the mixer in real time according to the time sequence, predicting the predicted parameters of the semi-finished product at the target time point, and comparing them with the finished product parameters, the abnormal situation of the semi-finished product data can be detected; through early dynamic adjustment and decision-making, the time consumption is effectively reduced, and the production efficiency and mud quality of the vacuum pug mill are significantly improved.

[0112] The generation of equipment adjustment parameters according to the predicted parameters of the semi-finished product in this embodiment includes:

[0113] Obtain the predicted parameters of the semi-finished product and the finished product parameters;

[0114] Construct a minimum cost function according to the predicted parameters of the semi-finished product and the finished product parameters; the minimum cost function is set to minimize the cost required for the raw materials in the mixer to reach the finished product parameters at the required time, including adding materials and adjusting equipment parameters;

[0115] The optimal solution is obtained by solving the minimized cost function through a deep reinforcement learning model, and an adjustment plan is generated based on this; the deep reinforcement learning model is a pre-trained model that can effectively obtain the optimal solution of the minimized cost function and generate an adjustment plan based on this;

[0116] Generate device adjustment parameters according to the adjustment plan.

[0117] In this embodiment, constructing the minimized cost function according to the estimated parameters of the semi-finished product and the finished product parameters includes:

[0118] Obtain the estimated parameters of the semi-finished product and the finished product parameters;

[0119] Obtain a number of parameter differences by taking the difference between the finished product parameters and the estimated parameters of the semi-finished product;

[0120] Calculate the amount of added material according to the parameter differences;

[0121] Generate an estimated change in device energy according to the parameter differences; the estimated change in device energy refers to the estimated amount of energy consumed or saved by adjusting the device parameters. Among them, when adjusting the device parameters, if a certain amount of energy is consumed, then NYL>0; otherwise, NYL<0;

[0122] Through the formula Construct the minimized cost function FC; where TWL represents the amount of added material, DJ represents the unit price corresponding to the amount of added material, DNJ represents the unit energy price, and NYL represents the estimated change in device energy; the amount of added material has an upper limit, that is, 0≤TWL≤TWLmax, and TWLmax represents the maximum value of the amount of added material; when adjusting the semi-finished product parameters by adding materials or adjusting the device parameters, ensure that the adjusted semi-finished product parameters are equal to the finished product parameters; this embodiment is to solve the optimal TWL and NYL, so as to achieve the target requirements with the least cost and reduce cost consumption.

[0123] In this embodiment, by calculating the deviation between the estimated parameters of the semi-finished product and the finished product parameters, it comprehensively considers eliminating the deviation by adding materials or adjusting the device parameters; since adding materials and adjusting the device parameters will generate different cost consumptions, such as material costs and energy costs, the system seeks the minimum cost plan that can eliminate the deviation between the two, so as to adjust the semi-finished product parameters to the target requirements; on the basis of meeting production requirements, it effectively reduces resource consumption and improves the economy and sustainability of production.

[0124] In this embodiment, calculating the amount of added material according to the parameter differences includes:

[0125] Obtain a number of parameter differences and semi-finished product data; the semi-finished product data includes the semi-finished product weight DCZ;

[0126] Select the parameter difference CC corresponding to the maximum absolute value of several parameter differences; that is, sequentially solve the factor with the largest gap between the current semi-finished product data and the finished product data;

[0127] Judge whether the parameter difference is within the corresponding difference range; the difference range is set according to experience;

[0128] Yes, the added material quantity TWL corresponding to the parameter difference = 0; in this embodiment, it is considered that if the difference between the semi-finished product parameter and its corresponding finished product parameter is not large, the obtained finished product parameter will not have a great impact on the subsequent process, so there is no need to add materials for adjustment;

[0129] No, when the parameter difference is greater than the maximum value in its corresponding difference range;

[0130] Through the formula Calculate the added material quantity TWL;

[0131] When the parameter difference is less than the minimum value in its corresponding difference range;

[0132] Through the formula Calculate the added material quantity TWL; where DTWL represents the added material quantity required per unit difference and per unit semi-finished product weight; in this embodiment, it is considered that when the positive and negative of the parameter difference are different, different materials need to be added for adjustment. For example, when the viscosity in the semi-finished product parameter is large, materials to reduce the viscosity need to be added for adjustment, and when the viscosity in the semi-finished product parameter is small, materials to increase the viscosity need to be added for adjustment. These two kinds of materials are not the same kind of materials, so they need to be calculated separately; when the added material quantity required per unit semi-finished product weight is fixed, the larger the parameter difference and the larger the semi-finished product weight, the more material is required, so the added material quantity will increase accordingly.

[0133] In this embodiment, by calculating the difference between the estimated parameters of the semi-finished product and its corresponding finished product parameters, the required material addition amount or the estimated change amount of equipment energy is determined; when adding materials, the system accurately matches the corresponding material types and quantities according to different parameter differences, providing reliable data support for finding the minimum cost solution; not only optimizing the resource utilization efficiency, but also significantly improving the quality stability of mud production.

[0134] Please refer to Figure 3 , the generation of the estimated change amount of equipment energy according to the parameter difference in this embodiment includes:

[0135] Obtain several parameter differences, the semi-finished product weight, and the estimated parameters of the semi-finished product; the estimated parameters of the semi-finished product include the estimated duration; the estimated duration is the duration between the current estimated time and the time when stirring is completed;

[0136] Select the parameter difference corresponding to the maximum absolute value of several parameter differences;

[0137] Determine whether the parameter difference is within the corresponding difference range; the difference range is set according to experience;

[0138] Yes, the estimated change in equipment energy NYL corresponding to the parameter difference = 0; in this embodiment, it is considered that if the parameters of the semi-finished product are not much different from the parameters of its corresponding finished product, the obtained finished product parameters will not have a great impact on the subsequent process. Therefore, there is no need to adjust the equipment parameters and the equipment energy will not change;

[0139] No, input the parameter difference, the weight of the semi-finished product, and the estimated duration into the energy prediction model to obtain the estimated change in equipment energy; the energy prediction model is constructed through an artificial intelligence model.

[0140] The energy prediction model in this embodiment is constructed through an artificial intelligence model, including:

[0141] Obtain a number of historical parameter differences, the weight of the semi-finished product, the estimated duration, and their corresponding estimated changes in equipment energy;

[0142] Divide a number of historical parameter differences, the weight of the semi-finished product, the estimated duration, and their corresponding estimated changes in equipment energy into training data, validation data, and test data; perform data preprocessing on the training data, validation data, and test data to obtain a training set, a validation set, and a test set; the ratio among the training set, the test set, and the validation set is 7:2:1;

[0143] Select an artificial intelligence model as the basic model; the artificial intelligence model includes a BP network model, etc.;

[0144] Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain a pre-trained model;

[0145] Verify the pre-trained model on the test set, and finally obtain an energy prediction model that takes the input parameter difference, the weight of the semi-finished product, and the estimated duration and outputs the estimated change in equipment energy.

[0146] Some data in the above formula are calculated by removing the dimension and taking their numerical values. The formula is obtained by software simulation of a large amount of collected data to get a formula closest to the actual situation; the preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0147] Working principle of this application: By obtaining environmental data, equipment parameters, semi-finished product data and requirement data; generating material data according to the requirement data; transporting corresponding materials to the mixer according to the material data; obtaining the material parameter data corresponding to the material data when entering the mixer; generating initial equipment parameters according to the material data and the material parameter data; obtaining the semi-finished product data in the mixer in real time; generating estimated semi-finished product parameters according to the material data, semi-finished product data and environmental data; generating equipment adjustment parameters according to the estimated semi-finished product parameters, assigning the equipment adjustment parameters to the corresponding equipment parameters for control, considering the influence of environmental factors on the change of semi-finished product data in the mixer, predicting in real time the gap between the semi-finished product data and the finished product data, adding materials in advance or adjusting the equipment parameters to ensure that the requirement data can be obtained on time and with high quality at the lowest cost, improving the accuracy and efficiency of the vacuum clay kneading machine, avoiding the influence of environmental factors on clay mixing not considered in the prior art, and lacking the consideration of adaptively controlling the vacuum clay kneading machine to make the finished clay reach the target requirements, resulting in poor finished state of the clay and low production efficiency of the vacuum clay kneading machine.

[0148] The above embodiments are only used to illustrate the technical solutions of this application rather than to limit them. Although this application has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of this application can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of this application.

Claims

1. A material environment intelligent adjustment system for a vacuum clay machine, characterized in that: include: Data acquisition module, data analysis module, control module and database; The data acquisition module acquires environmental data and equipment parameters, finished product data and demand data through data acquisition equipment; The data analysis module: generates material data according to demand data; obtains material parameter data corresponding to the material data when entering the mixer; generates equipment initial parameters according to the material data and material parameter data; obtains the data of the finished product in the mixer in real time; generates estimated parameters of the finished product according to the material data, the data of the finished product and the environmental data; generates equipment adjustment parameters according to the estimated parameters of the finished product; The control module: transmits the corresponding material to the mixer according to the material data; Assigning device adjustment parameters to control corresponding device parameters; The generating of estimated parameters of the finished product according to the material data, the finished product data and the environmental data includes: Real-time acquisition of historical data of finished products, historical material data, historical environmental data, historical equipment parameters and recording time in the mixer; the equipment parameters are initial equipment parameters or equipment adjustment parameters; the environmental data include external environmental data and internal environmental data; Integrate a number of historical finished product data, historical material data, historical environmental data and historical equipment parameters into a number of forecast sequences according to the recording time; Input the prediction sequence into the finished product parameter estimation model to obtain the estimated parameters of the finished product; The parameter estimation model for the finished product is constructed by a machine learning model, including: Obtain some historical data on finished products, historical material data, historical environmental data, historical equipment parameters and recording time; Performing data preprocessing on a number of historical unfinished product data, historical material data, historical environmental data, historical equipment parameters and recording time to obtain a number of preprocessed historical unfinished product data, historical material data, historical environmental data, historical equipment parameters and recording time; Integrate the historical finished product data, historical material data, historical environmental data, and historical equipment parameters with the same recording time into a time feature vector; Use the principal component analysis method to reduce the features of several time feature vectors to obtain the reduced time feature vector; the reduced time feature vector is the time feature vector obtained by the principal component analysis method after the cumulative contribution rate reaches M%; where M is a constant, M∈[0,100]; Integrate each time feature vector into several prediction sequences according to the order of recording time; Divide several prediction sequences into training set, test set and validation set in chronological order; Select a machine learning model as the base model; Train the basic model with the training set, and adjust the learning rate or other hyperparameters on the validation set to obtain the pre-trained model; By verifying the pre-trained model on the test set, a parameter estimation model for the finished product is finally obtained, which takes several prediction sequences as input and outputs the estimated parameters of the finished product; The generating of equipment adjustment parameters according to the estimated parameters of the finished product includes: Obtain estimated parameters of unfinished products and parameters of finished products; Construct a cost minimization function based on the estimated parameters of the finished product and the parameters of the finished product; The deep reinforcement learning model is used to solve the minimization cost function to obtain the optimal solution, and then an adjustment plan is generated; Generate equipment adjustment parameters according to the adjustment plan; The minimization cost function is constructed according to the estimated parameters of the finished product and the parameters of the finished product, including: Obtain estimated parameters of unfinished products and parameters of finished products; By subtracting the parameters of the finished product from the estimated parameters of the finished product, a number of parameter differences are obtained; Calculate the amount of material to be added based on the parameter difference; Generate estimated equipment energy change based on parameter differences; By formula Construct the minimization cost function FC; among them, TWL represents the amount of added material, DJ represents the unit price corresponding to the amount of added material, DNJ represents the unit energy price, and NYL represents the estimated change in equipment energy.

2. The material environment intelligent adjustment system for a vacuum clay machine according to claim 1, characterized in that: The generating of material data according to demand data includes: Obtaining demand data; the demand data includes finished product name, finished product parameters, finished product weight and demand time; Search material data in the material data table according to the finished product name and finished product weight; the material data includes several material names and their corresponding material weights.

3. The material environment intelligent adjustment system for a vacuum clay machine according to claim 1, characterized in that: The generating of the equipment initial parameters according to the material data and the material parameter data comprises: Obtain material data, demand data and material parameter data corresponding to the material data when entering the mixer; Material data, demand data and material parameter data are input into a parameter estimation model to obtain initial equipment parameters; the parameter estimation model is constructed through an artificial intelligence model; the initial equipment parameters include operating power, operating time, stirring speed and equipment vacuum degree.

4. The material environment intelligent adjustment system for a vacuum clay extruder according to claim 3, characterized in that: The parameter estimation model is constructed by an artificial intelligence model, including: Obtain some historical material data, demand data and material parameter data and their corresponding historical equipment initial parameters; Divide a number of historical material data, demand data and material parameter data and their corresponding historical equipment initial parameters into training data, verification data and test data; perform data preprocessing on the training data, verification data and test data to obtain a training set, a verification set and a test set; Select an AI model as the base model; Train the basic model with the training set, and adjust the learning rate and hyperparameters on the validation set to obtain the pre-trained model; By verifying the pre-trained model on the test set, we finally get the input material data, demand data and material parameter data, and output the parameter estimation model of the equipment initial parameters.

5. The material environment intelligent adjustment system for a vacuum clay machine according to claim 1, characterized in that: The method of calculating the amount of added material according to the parameter difference includes: Acquire a number of parameter differences and data of finished products; the data of finished products include the weight DCZ of the finished products; Select a parameter difference CC corresponding to the maximum absolute value of several parameter differences; Determine whether the parameter difference is within the corresponding difference range; Yes, the amount of added material corresponding to the parameter difference TWL=0; No, when the parameter difference is greater than the maximum value in its corresponding difference range; By formula Calculate the amount of added material TWL; When the parameter difference is less than the minimum value in its corresponding difference range; By formula Calculate the added material amount TWL; where DTWL is expressed as the amount of added material required for the unit difference unit weight of the finished product.

6. The material environment intelligent adjustment system for a vacuum clay machine according to claim 1, characterized in that: The generating of the estimated change amount of equipment energy according to the parameter difference includes: Obtaining several parameter differences, the weight of the finished product and the estimated parameters of the finished product; the estimated parameters of the finished product include the estimated duration; Select the parameter difference corresponding to the maximum absolute value of several parameter differences; Determine whether the parameter difference is within the corresponding difference range; Yes, the estimated change in equipment energy corresponding to the parameter difference NYL=0; No, the parameter difference, the weight of the finished product and the estimated duration are input into the energy estimation model to obtain the estimated change in equipment energy; the energy estimation model is constructed through an artificial intelligence model.

Images