Cement clinker performance prediction and mineralogical design model, database and applications thereof

Abstract

This invention relates to a cement clinker performance prediction and mineral phase design model, a cement clinker mineral phase database, and their applications. The model is based on a neural network algorithm for clinker performance prediction and mineral phase design. The clinker performance prediction involves inputting known mineral phase parameters of the clinker into the model, which then outputs its prediction results for the known clinker's performance data. The mineral phase design involves inputting the performance data of a target clinker into the model, which then outputs its design results for the target clinker's mineral phase parameters. The technical problem to be solved is how to overcome the technical difficulties of long research cycles and low efficiency in cement clinker research, improve the R&D efficiency of cement materials, and better guide the R&D and manufacturing of cement clinker.

Description

Technical Field

[0001] This invention belongs to the field of cement clinker manufacturing technology, and in particular relates to a cement clinker performance prediction and mineral phase design model, a cement clinker mineral phase database and their applications. Background Technology

[0002] Database technology, as a data-oriented computer technology, initially focused on data storage and management. Traditional materials databases are primarily relational databases based on the relational model, consisting of two-dimensional tables and their interrelationships. These databases can store numerical data on material properties, structures, processes, and literature, mainly providing resource sharing for researchers, such as the electronic materials testing database established by the CALCE Electronic Packaging Research Center. With the development of database technology, online databases and their system design, software development, and management are evolving to higher levels, as exemplified by Asteel Materials Database, MatWeb, and NIMS. As database technology continues to advance, how to efficiently utilize database resources has become a current research hotspot.

[0003] Meanwhile, traditional materials research and development relies heavily on trial and error, with repeated experiments wasting a significant amount of development time. Those skilled in the field are actively working on building data-driven materials database platforms, currently including database management platforms based on high-throughput computing software such as MaterialProject, AFLOW, MedeA, and VNL. However, current materials databases are primarily used for data sharing and lack the ability to perform data mining and predict material properties. Therefore, traditional materials research and development still requires repeated experiments; in particular, the development of databases for the mineral phases and performance characterization parameters of cement clinker is almost nonexistent. Summary of the Invention

[0004] The main objective of this invention is to provide a cement clinker performance prediction and mineral phase design model, a cement clinker mineral phase database, and their applications. The technical problem to be solved is how to establish a cement clinker performance prediction and mineral phase design model and a cement clinker mineral phase database, so as to overcome the technical difficulties of long research cycle and low efficiency of cement clinker, improve the research and development efficiency of cement materials, better guide the research and development and manufacturing of cement clinker, and thus be more suitable for practical use.

[0005] The objective of this invention and the technical problem it solves are achieved through the following technical solution. According to this invention, a cement clinker performance prediction and mineral phase design model is proposed, which uses a neural network algorithm to predict clinker performance and design its mineral phase. The clinker performance prediction involves inputting known mineral phase parameters of the clinker into the cement clinker performance prediction and mineral phase design model, and the model outputs its prediction results for the performance data of the known clinker. The clinker mineral phase design involves inputting the performance data of a target clinker into the cement clinker performance prediction and mineral phase design model, and the model outputs its design results for the mineral phase parameters of the target clinker.

[0006] The objectives of this invention and the technical problems it addresses can be further achieved by the following technical measures.

[0007] Preferably, in the aforementioned cement clinker performance prediction and mineral phase design model, the mineral phase parameters include the mineral phase types, the content of each mineral phase, and the specific surface area of ​​the cement clinker; the performance data includes 1-day flexural strength, 3-day flexural strength, 28-day flexural strength, 1-day compressive strength, 3-day compressive strength, and 28-day compressive strength; the mineral phase parameters input to the cement clinker performance prediction and mineral phase design model include several cement clinker samples, each cement clinker sample including a parameter matrix representing multiple dimensions of the mineral phase; the performance data input to the cement clinker performance prediction and mineral phase design model includes several cement clinker samples, each cement clinker sample including a data matrix representing multiple dimensions of performance.

[0008] Preferably, the aforementioned cement clinker performance prediction and mineral phase design model is trained according to the following steps:

[0009] 1) Obtain the mineral phase parameters of the known clinker or the performance data of the target clinker;

[0010] 2) Input the known clinker mineral phase parameters or the target clinker performance data into the cement clinker performance prediction and mineral phase design model;

[0011] 3) Adjust the number of neural network layers n, the number of output units per layer m, and the number of iterations k. The cement clinker performance prediction and mineral phase design model outputs the prediction results of known clinker performance data or the design results of mineral phase parameters of the target clinker.

[0012] 4) Compare the predicted or designed results with the measured values ​​of the cement clinker, and calculate the mean square error mse and the mean absolute error mae of the predicted or designed results;

[0013] If the mean square error mse or the mean absolute error mae does not meet the preset threshold, then return to step 3);

[0014] If both the mean square error mse and the mean absolute error mae meet the preset threshold, then the cement clinker performance prediction and mineral phase design model are obtained.

[0015] Preferably, in the aforementioned cement clinker performance prediction and mineral phase design model, the number of neural network layers n is a natural number from 1 to 10; the number of output units per layer m is a natural number from 6 to 50; the number of iterations k is a natural number from 1 to 500; and the preset thresholds are mean square error mse < 1 and mean absolute error mae < 1.

[0016] The objective of this invention and the solution to its technical problems are further achieved by the following technical solution. A cement clinker mineral phase database proposed according to this invention includes:

[0017] The software front-end is installed on a smart terminal; the software front-end includes a data upload interface, a data query interface, and a data prediction interface.

[0018] The database backend is hosted on a solid-state server; the database backend connects to the software frontend; the database backend includes:

[0019] The data upload module, through the data upload interface, enables human-computer interaction and is used to upload and manage the basic data information of the cement clinker.

[0020] The data query module enables human-computer interaction through the data query interface, allowing users to query basic data information of the cement clinker using mineral phase parameters or performance data.

[0021] The data prediction module includes the aforementioned cement clinker performance prediction and mineral phase design model; the data prediction module uses the data prediction interface for human-computer interaction to predict the performance data of known clinker or the mineral phase parameters of the design target clinker.

[0022] The objectives of this invention and the technical problems it addresses can be further achieved by the following technical measures.

[0023] Preferably, the aforementioned cement clinker mineral phase database includes mineral phase characterization parameters and performance characterization parameters of cement clinker.

[0024] The mineral phase characterization parameters include the name, abbreviation, chemical formula, molecular formula, structural diagram, sintering temperature, melting point, crystal parameters, carbon emissions, hydration activity, data source, and data upload time of each mineral phase; the crystal parameters are parameters characterizing the mineral phase crystals, including the three sets of edge lengths a, b, and c of the unit cell, the included angles α, β, and γ between the three sets of edges, the number of different elements, the space group number, the Hermann-Mauguin space group symbol, the Hall space group symbol, the reflection residual factor, the ion doping performance, the cohesive energy per unit cell, the formation energy, and / or the density of states;

[0025] The performance characterization parameters include the cement clinker number, clinker system, content of each mineral phase, clinker performance, testing time, and data source; the clinker performance includes alumina content, magnesium oxide content, sulfur trioxide content, free calcium oxide content, loss on ignition, insoluble matter content, alkali content, specific surface area, residue on a 45μm square-hole sieve, initial setting time, final setting time, soundness, 3-day flexural strength, 7-day flexural strength, 28-day flexural strength, 3-day compressive strength, 7-day compressive strength, and / or 28-day compressive strength;

[0026] Querying the basic data information of cement clinker by mineral phase parameters involves inputting the mineral phase name, abbreviation, or 1 to 6 elements contained in the mineral phase of the target cement clinker into the data prediction interface. The cement clinker mineral phase database can output the mineral phase characterization parameters of all cement clinker that meet the input conditions. Querying the basic data information of cement clinker by performance data involves inputting the specific surface area, initial setting time, final setting time, 3-day flexural strength, 28-day flexural strength, 3-day compressive strength, and / or 28-day compressive strength of the target cement clinker into the data prediction interface. The cement clinker mineral phase database can output the performance characterization parameters of all cement clinker that meet the input conditions.

[0027] Preferably, in the aforementioned cement clinker mineral phase database, the software front-end further includes a user management interface; the database back-end further includes a user management module; the user management module performs human-computer interaction through the user management interface for user registration and user permission management; the cement clinker performance prediction and mineral phase design model is dynamically updated based on changes in the basic data information of the cement clinker.

[0028] The objective of this invention and the technical problem it solves are further achieved by the following technical solution. A method for applying the aforementioned cement clinker mineral phase database according to this invention includes the following steps:

[0029] The user inputs basic data information of cement clinker into the cement clinker mineral phase database through the data upload interface; the data upload module determines whether the basic data information is correct; if not, it prompts the user that the data format is incorrect and the specific location of the error; if yes, it uploads the basic data information to the database backend.

[0030] The data query interface inputs the mineral phase parameters or performance data into the cement clinker mineral phase database; the data query interface outputs the basic data information of the cement clinker.

[0031] Select the performance data of the known clinker or the mineral phase parameters of the target clinker through the data prediction interface; input the mineral phase parameters of the known clinker or the performance data of the target clinker according to the prompts of the data prediction interface; the data prediction interface outputs the prediction results of the performance data of the known clinker or the design results of the mineral phase parameters of the target clinker.

[0032] The objective of this invention and the solution to its technical problem are also achieved by the following technical solution. A storage medium according to this invention includes a stored program, which, when the program is running, controls the device containing the storage medium to execute the aforementioned application method.

[0033] The objective of this invention and the technical problem it solves are further achieved by the following technical solution. An electronic device according to this invention includes a storage medium comprising:

[0034] One or more processors, wherein the storage medium is coupled to the processors, and the processors are configured to execute program instructions stored in the storage medium; the program instructions, when executed, perform the aforementioned application method.

[0035] By employing the above technical solutions, the present invention proposes a cement clinker performance prediction and mineral phase design model, a cement clinker mineral phase database, and their applications, which have at least the following advantages:

[0036] This invention proposes a cement clinker performance prediction and mineral phase design model, a cement clinker mineral phase database, and their applications. By aggregating basic cement clinker data into the mineral phase database, and combining experimental, computational, and enterprise data, the basic data of cement materials is enriched, making the data more systematic. Simultaneously, data cleaning is performed through the cement clinker mineral phase database to discover potential patterns among cement material data. Then, a cement clinker performance prediction and mineral phase design model is established based on a neural network algorithm. By inputting known clinker mineral phase parameters into the model, it can automatically predict the performance data of that known clinker. By inputting the performance data of a target clinker into the model, it can automatically design the mineral phase parameters of that target clinker. Furthermore, extensive experimental verification demonstrates that this invention's cement clinker performance prediction and mineral phase design model... The performance data of the known clinker predicted by the design model is accurate and very close to the true value of the physical test results of the known clinker. Cement clinker is prepared according to the mineral phase parameters of the target clinker designed by the cement clinker performance prediction and mineral phase design model of this invention, and then the results are tested. The test results are very close to the performance data of the target clinker. The above results show that the cement clinker performance prediction and mineral phase design model and cement clinker mineral phase database of this invention can effectively guide the research and development and production of cement materials, greatly shorten the experimental time, and save the cost of experimental materials and research and development labor costs, and have very good practical value. Furthermore, the cement clinker performance prediction and mineral phase design model of this invention can be dynamically updated according to the data changes of the cement clinker mineral phase database, so that the prediction results of the cement clinker performance prediction and mineral phase design model are more accurate and reliable.

[0037] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the structure of the cement clinker performance prediction and mineral phase design model of the present invention;

[0039] Figure 2 This is a schematic diagram of the structure of the cement clinker mineral phase database of the present invention. Detailed Implementation

[0040] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features and effects of a cement clinker performance prediction and mineral phase design model, a cement clinker mineral phase database and its application proposed according to the present invention.

[0041] This invention proposes a cement clinker performance prediction and mineral phase design model, as shown in the appendix. Figure 1 As shown, it uses a neural network algorithm for clinker performance prediction and clinker mineral phase design. Clinker performance prediction and clinker mineral phase design are two functions of the cement clinker performance prediction and mineral phase design model, respectively. The required function is selected for each operation, and then the corresponding parameters are input according to the system prompts. The cement clinker performance prediction and mineral phase design model outputs the corresponding results. Specifically, clinker performance prediction involves inputting known clinker mineral phase parameters into the cement clinker performance prediction and mineral phase design model, which then outputs its prediction results for the known clinker performance data. Clinker mineral phase design involves inputting target clinker performance data into the cement clinker performance prediction and mineral phase design model, which then outputs its design results for the target clinker's mineral phase parameters.

[0042] Preferably, the mineral phase parameters include the mineral phase types, the content of each mineral phase, and the specific surface area of ​​the cement clinker. That is, by inputting known mineral phase types, the content of each mineral phase, and the specific surface area of ​​the cement clinker into the cement clinker performance prediction and mineral phase design model, the model can output the prediction results of various properties corresponding to the cement clinker.

[0043] The mineral phase parameters of the cement clinker include, but are not limited to, the following: 3CaO·SiO2, 2CaO·SiO2, 3CaO·Al2O3, 4CaO·Al2O3·Fe2O3, CaO·Al2O3, CaO·2Al2O3, CaO·6Al2O3, 2CaO·Al2O3·SiO2, 12CaO·7Al2O3, 2CaO·Fe2O3, 6CaO·Al2O3·2Fe2O3, 6CaO·2Al2O3·Fe2O3, 4CaO·3Al2O3·CaSO4, 11CaO·7Al2O3·CaF2, 11CaO·7Al2O3·CaCl2, 21CaO·6SiO2·Al2O3·CaCl2, etc.

[0044] Preferably, the performance data includes 1-day flexural strength, 3-day flexural strength, 28-day flexural strength, 1-day compressive strength, 3-day compressive strength, and 28-day compressive strength. That is, by inputting the target performance such as 1-day flexural strength, 3-day flexural strength, 28-day flexural strength, 1-day compressive strength, 3-day compressive strength, and 28-day compressive strength into the cement clinker performance prediction and mineral phase design model, the model can output various possible mineral phase designs for the target cement clinker to guide clinker formulation.

[0045] From the appendix Figure 1As shown, the mineral phase parameters input into the cement clinker performance prediction and mineral phase design model are input in matrix X form, and the data output by the cement clinker performance prediction and mineral phase design model are also output in matrix Y form.

[0046] When the cement clinker performance prediction and mineral phase design model is used for clinker performance prediction, the matrix X is a matrix including several cement clinker samples, each cement clinker sample including a parameter matrix representing multiple dimensions of mineral phase; the matrix Y output by the cement clinker performance prediction and mineral phase design model is a matrix including several cement clinker samples, each cement clinker sample including a parameter prediction result representing multiple dimensions of performance.

[0047] When the cement clinker performance prediction and mineral phase design model is used for clinker mineral phase design, the matrix X is a matrix that includes several cement clinker samples, each of which includes a parameter matrix representing multiple dimensions of performance; the matrix Y output by the cement clinker performance prediction and mineral phase design model is a matrix that includes several cement clinker samples, each of which includes a parameter design result representing multiple dimensions of mineral phase.

[0048] The aforementioned cement clinker performance prediction and mineral phase design model is trained according to the following steps:

[0049] 1) Obtain the mineral phase parameters of the known clinker or the performance data of the target clinker;

[0050] 2) Input the known clinker mineral phase parameters or the target clinker performance data into the cement clinker performance prediction and mineral phase design model;

[0051] 3) Adjust the number of neural network layers n, the number of output units per layer m, and the number of iterations k. The cement clinker performance prediction and mineral phase design model outputs the prediction results of known clinker performance data or the design results of mineral phase parameters of the target clinker.

[0052] 4) Compare the predicted or designed results with the measured values ​​of the cement clinker, and calculate the mean square error mse and the mean absolute error mae of the predicted or designed results;

[0053] If the mean square error mse or the mean absolute error mae does not meet the preset threshold, then return to step 3);

[0054] If both the mean square error mse and the mean absolute error mae meet the preset threshold, then the cement clinker performance prediction and mineral phase design model are obtained.

[0055] During training, the loss function and optimizer are first set, and then the model prediction performance is optimized by continuously adjusting the number of neural network layers n, the number of output units per layer m, and the number of iterations k. After adjusting the above parameters, the cement clinker performance prediction and mineral phase design model runs and outputs the mean square error mse and mean absolute error mae of the cement clinker performance prediction and mineral phase design model at this time. If either the mean square error or the mean absolute error does not meet the preset threshold requirements, the above parameter settings are adjusted again, and the above steps are repeated until the mean square error and the mean absolute error both meet the preset threshold requirements. At this point, the cement clinker performance prediction and mineral phase design model can be considered qualified.

[0056] Preferably, the number of neural network layers n is a natural number from 1 to 10; the number of output units per layer m is a natural number from 6 to 50; the number of iterations k is a natural number from 1 to 500; and the preset threshold is that the mean square error mse and the mean absolute error mae are both less than 1. The cement clinker performance prediction and mineral phase design model trained according to the above preferred parameters has high accuracy and good prediction precision.

[0057] This invention also proposes a cement clinker mineral phase database, consisting of appendices. Figure 2 As shown, it includes a software front-end and a database back-end; the software front-end is installed on a smart terminal, and it only serves as a human-computer dialogue interface to receive data information and display and output data information; the database back-end is hosted on a solid-state server; the database back-end is connected to the software front-end; the software front-end encapsulates the received data and transmits it to the database back-end, and the database back-end encapsulates the calculated data and returns it to the software front-end.

[0058] The software front-end includes a user management interface, a data upload interface, a data query interface, and a data prediction interface. The database back-end includes a user management module, a data upload module, a data query module, and a data prediction module. Each module is implemented through database software.

[0059] The cement clinker mineral phase database of this invention can be hosted on any type of server in the prior art, and its software can be implemented based on any programming language in the prior art. In a specific embodiment of this invention, the database is built on a MySQL 8.0 system and hosted on a solid-state server. The database software is developed using the Python language, and users can access the server through the network to upload data, query data, and predict data.

[0060] The user management module enables human-computer interaction through the user management interface for user account registration. After relevant information is stored in the database backend and successfully reviewed by the administrator to obtain the corresponding user permissions, the user can access the software homepage with their account and password, and then access the data upload interface, data query interface, and data prediction interface through the homepage menu.

[0061] The data upload interface enables human-computer interaction, which is used to upload and manage the basic data information of the cement clinker.

[0062] The basic data information uploaded to the cement clinker mineral phase database preferably includes mineral phase characterization parameters and performance characterization parameters of cement clinker; wherein, the mineral phase characterization parameters include the name, abbreviation, chemical formula, molecular formula, structural diagram, firing temperature, melting point, crystal parameters, carbon emissions, hydration activity, data source, and data upload time of each mineral phase; the crystal parameters are parameters characterizing the mineral phase crystals, including the three sets of edge lengths a, b, and c of the unit cell, the included angles α, β, and γ between the three sets of edges, the number of different elements, space group number, Hermann-Mauguin space group symbol, Hall space group symbol, reflection residual factor, ion doping performance, unit cell cohesive energy, formation energy, and / or density of states; the upload format of the cement clinker mineral phase characterization parameters is a multi-row, two-column table, with each row containing one of the aforementioned mineral phase characterization parameters, the first column being the name of the mineral phase characterization parameter, and the second column being the specific value of that mineral phase characterization parameter, as shown in Table 1:

[0063] Table 1. Upload format for mineral facies characterization parameters

[0064] Mineral phase name Specific mineral phase names of cement clinker Abbreviation The specific abbreviation for cement clinker …… …… Upload time The specific upload time of this data

[0065] The performance characterization parameters include the cement clinker's serial number, clinker system, content of each mineral phase, clinker properties, testing time, and data source. The upload format for the cement clinker's performance characterization parameters is shown in Table 2.

[0066] Table 2 Performance Characterization Parameter Upload Format

[0067]

[0068] Table 2 is only an illustration of the upload format for performance characterization parameters, therefore no specific parameters are filled in. The clinker system mentioned therein includes, but is not limited to, silicate clinker system, aluminate clinker system, sulfoaluminate clinker system, fluoroaluminate clinker system, chloroaluminate clinker system, etc. The mineral phases mentioned include, but are not limited to, 3CaO·SiO2, 2CaO·SiO2, 3CaO·Al2O3, 4CaO·Al2O3·Fe2O3, CaO·Al2O3, CaO·2Al2O3, CaO·6Al2O3, 2CaO·Al2O3·SiO2, 12CaO·7Al2O3, 2CaO·Fe2O3, 6CaO·Al2O3·2Fe2O3, 6CaO·2Al2O3·Fe2O3, 4CaO·3Al2O3·CaSO4, 11CaO·7Al2O3·CaF2, 11CaO·7Al2O3·CaCl2, and 21CaO·6SiO2·Al2O3·CaCl2. The clinker properties refer to various indicators characterizing clinker performance, including but not limited to free calcium oxide content, alumina content, magnesium oxide content, sulfur trioxide content, loss on ignition, insoluble matter content, alkali content, specific surface area, residue on a 45μm square hole sieve, initial setting time, final setting time, stability, 3-day flexural strength, 7-day flexural strength, 28-day flexural strength, 3-day compressive strength, 7-day compressive strength, and 28-day compressive strength.

[0069] Data uploads can be performed on a single data entry or in batches via data files. After a single data entry is uploaded, the database system preprocesses the data to determine if it meets upload requirements and falls within a reasonable data range. If the data is correct, it is uploaded to the database backend; otherwise, it returns a message indicating an incorrect data format and the specific location of the error. After a data file is uploaded, the database system reads the file content and processes the data according to the single data entry processing method described above. If the data is correct, it is uploaded to the database backend; otherwise, it returns a message indicating an incorrect data format and the location of the error.

[0070] The data query interface allows for human-computer interaction, enabling users to query basic data information of cement clinker using mineral phase parameters or performance data, and display the query results.

[0071] Querying basic data information of cement clinker by mineral phase parameters is done through the data prediction interface by inputting the mineral phase name, abbreviation, and / or 1 to 6 elements contained in the mineral phase of the target cement clinker. In the clinker mineral phase query interface, users directly enter the name, abbreviation, or 1 to 6 elements of the mineral phase they wish to search for in the corresponding spaces. All three input boxes do not need to be filled; entering one is sufficient for the query. Entering multiple elements will display results that meet all input requirements. The cement clinker mineral phase database can output the mineral phase characterization parameters of all cement clinker that meet the input conditions.

[0072] Querying the basic data information of cement clinker through performance data is done by inputting the specific surface area, initial setting time, final setting time, 3-day flexural strength, 28-day flexural strength, 3-day compressive strength, and / or 28-day compressive strength of the target cement clinker into the data prediction interface. In the clinker composition performance query interface, the user clicks the keyword selection drop-down menu, selects keywords (specific surface area, initial setting time, final setting time, 3-day flexural strength, 28-day flexural strength, 3-day compressive strength, and 28-day compressive strength, etc.), and then enters the query range in the corresponding position. The results will be displayed according to the query conditions. The cement clinker mineral phase database can output the performance characterization parameters of all cement clinker that meet the input conditions.

[0073] The data prediction module includes the aforementioned cement clinker performance prediction and mineral phase design model. This module uses a user-friendly interface to predict the performance data of known clinker or the mineral phase parameters of the target clinker. First, select the prediction function on the data prediction interface; then input the required parameters according to the software prompts; finally, the prediction results are output.

[0074] As the data in the cement clinker mineral phase database gradually becomes richer, the cement clinker performance prediction and mineral phase design model can be dynamically updated based on changes in the basic data information of the cement clinker, so as to improve the accuracy of its prediction and design results.

[0075] This invention also proposes an application method based on the aforementioned cement clinker mineral phase database, which includes the following steps:

[0076] The user inputs basic data information of cement clinker into the cement clinker mineral phase database through the data upload interface; the data upload module determines whether the basic data information is correct; if not, it prompts the user that the data format is incorrect and the specific location of the error; if yes, it uploads the basic data information to the database backend.

[0077] The data query interface inputs the mineral phase parameters or performance data into the cement clinker mineral phase database; the data query interface outputs the basic data information of the cement clinker.

[0078] Select the performance data of the known clinker or the mineral phase parameters of the target clinker through the data prediction interface; input the mineral phase parameters of the known clinker or the performance data of the target clinker according to the prompts of the data prediction interface; the data prediction interface outputs the prediction result of the performance data of the known clinker or the design result of the mineral phase parameters of the target clinker.

[0079] The above steps do not need to be performed in order; you can choose any one of the functions to execute.

[0080] The present invention also proposes a storage medium including a stored program, which, when the program is running, controls the device where the storage medium is located to execute the aforementioned application method.

[0081] The present invention also proposes an electronic device including a storage medium comprising: one or more processors, the storage medium being coupled to the processors, the processors being configured to execute program instructions stored in the storage medium; the program instructions, when executed, perform the aforementioned application method.

[0082] The present invention will be further described below with reference to specific embodiments, but this should not be construed as a limitation on the scope of protection of the present invention. Some non-essential improvements and adjustments made by those skilled in the art based on the above description of the present invention still fall within the scope of protection of the present invention.

[0083] Unless otherwise specified, all materials and reagents mentioned below are commercially available products well known to those skilled in the art; unless otherwise specified, all methods described are methods known in the art. Unless otherwise defined, the technical or scientific terms used should have the ordinary meaning understood by those skilled in the art to which this invention pertains.

[0084] Example 1:

[0085] This embodiment illustrates the cement clinker performance prediction and mineral phase design model of the present invention using a silicate cement clinker system. The mineral phase composition of the silicate cement clinker is: C3S-C2S-C3A-C4AF. The mineral phase parameters (i.e., the mass fraction of each mineral phase) and performance parameters (1-day, 3-day, and 28-day compressive and flexural strengths) of the cement clinker are all obtained from cement enterprise production data and stored in the database. After cleaning the basic data in the database, the cement clinker performance prediction model is obtained by training according to the training method of the cement clinker performance prediction and mineral phase design model of the present invention.

[0086] The main parameter settings of the cement clinker performance prediction and mineral phase design model in this embodiment are as follows: When the number of neural network layers n is low, the mean square error of the cement clinker performance prediction and mineral phase design model decreases as the number of layers increases; when the number of neural network layers n is too high, the mean square error and mean absolute error do not change much. The number of output units m in each layer is related to the number of neural network layers. Too many output units in a single layer will also cause the model to overfit. For models with fewer input features, the number of output units in a single layer is generally less than 64, so the number of output units in the first layer is preferably 32. Since the final output result of the cement clinker performance prediction and mineral phase design model is clinker performance parameters, in this embodiment, the output clinker performance of the cement clinker performance prediction and mineral phase design model is set to 1-day flexural strength, 3-day flexural strength, 28-day flexural strength, 1-day compressive strength, 3-day compressive strength, and 28-day compressive strength, that is, the number of output units in the nth layer of the neural network is determined to be 6. The number of output units in the intermediate layers is initially set to 32, and is gradually reduced by adjusting the mean square error and mean absolute error of the model. When the number of iterations k is too high, the average absolute error of the cement clinker performance prediction and mineral phase design model will increase. The initial value is set to 500, and the average absolute error is adjusted based on the preferred number of neural network layers n and the number of output units m of each layer.

[0087] After multiple adjustments to the cement clinker performance prediction model based on the above logic, in this embodiment, the preferred neural network layer number n is 5 layers, the number of output units m in each layer is 32, 28, 16, 12, and 6 respectively, and the number of iterations k is 300. At this time, the mean square error mse of the cement clinker performance prediction and mineral phase design model is 0.976, and the mean absolute error mae is 0.711.

[0088] In the cement clinker performance prediction and mineral phase design model, the prediction function is selected as "clinker performance prediction" in the data prediction interface; then, the "known clinker mineral phase parameters" are entered according to the software prompts. Based on the cement clinker performance prediction and mineral phase design model established in this embodiment, three cement clinker samples are selected and XRD quantitative analysis is performed on the cement clinker using the Rietveld method to determine the mineral phase mass fraction of each sample. The obtained parameters are then input into the data prediction module based on this model. In this embodiment, the known clinker mineral phase parameters input include three cement clinker samples. Each cement clinker sample includes four mineral phase types, the content of the four mineral phases, and a matrix of the specific surface area of ​​each cement clinker, totaling nine parameters; as shown in Table 3:

[0089] Table 3. Mineral phase parameters of the clinker known in this embodiment.

[0090] <![CDATA[C3S]]> <![CDATA[C2S]]> <![CDATA[C3A]]> <![CDATA[C4AF]]> Specific surface area 52.41％ 21.38％ 9.20％ 12.34％ <![CDATA[359cm 2 / g]]> 50.33％ 21.08％ 10.79％ 11.92％ <![CDATA[354cm 2 / g]]> 58.11％ 17.22％ 9.65％ 11.86％ <![CDATA[360cm 2 / g]]>

[0091] The matrix of prediction results for the performance data of three cement clinker samples output by the cement clinker performance prediction and mineral phase design model is shown in Table 4 below:

[0092] Table 4. Prediction results of known clinker performance data in this embodiment.

[0093]

[0094] Finally, the three cement clinker samples were used to prepare cement mortar according to GB / T 17671 method, and their flexural and compressive strengths at 1d, 3d, and 28d were tested. The actual test results are shown in Table 5 below:

[0095] Table 5 shows the actual test results of the three cement clinker samples in this embodiment.

[0096]

[0097] Comparing the predicted results of the performance data in Table 4 with the actual test results of the performance data in Table 5, it can be seen that the prediction results of the cement clinker performance prediction and mineral phase design model of the present invention have high accuracy. The comparison results are shown in Tables 6-1, 6-2 and 6-3:

[0098] Table 6-1 Prediction accuracy of Sample 1

[0099]

[0100] Table 6-2 Prediction accuracy of Sample 2

[0101]

[0102] Table 6-3 Prediction accuracy of Sample 3

[0103]

[0104] Example 2

[0105] Following the same model training logic as described in Example 1, after multiple adjustments to the cement clinker performance prediction and mineral phase design model, it is preferred that the neural network layer n in this example has 6 layers, the number of output units m in each layer is 32, 28, 24, 18, 12, and 8 respectively, and the number of iterations k is 300. At this time, the mean square error mse of the cement clinker performance prediction and mineral phase design model is 0.971, and the mean absolute error mae is 0.852.

[0106] In the cement clinker performance prediction and mineral phase design model, select "Clinker Mineral Phase Design" as the prediction function in the data prediction interface; then input the "performance data of the target clinker" according to the software prompts. Based on the cement clinker performance prediction and mineral phase design model established in this embodiment, the performance parameters of the target clinker input in this embodiment require that each cement clinker sample include a matrix of the upper and lower limits of its 1d, 3d, and 28d flexural and compressive strengths, totaling 6 pairs of parameters; as shown in Table 7:

[0107] Table 7 Performance parameter requirements for the target cement clinker in this embodiment

[0108]

[0109] The cement clinker performance prediction and mineral phase design model outputs target clinker mineral phase parameters as upper and lower limits for four mineral phase types and their contents, as shown in Table 8 below:

[0110] Table 8. Predicted results of the mineral phase design of the target clinker in this embodiment.

[0111] <![CDATA[C3S]]> <![CDATA[C2S]]> <![CDATA[C3A]]> <![CDATA[C4AF]]> Remark 49.50％ 21.00％ 9.20％ 12.15％ lower limit 51.50％ 24.00％ 9.45％ 12.40％ upper limit

[0112] Based on the predicted results of the mineral phase design, the content of each component in the cement raw meal was calculated, the cement raw meal ratio and firing process were adjusted, and clinker that meets the requirements in Table 8 was prepared. Three cement clinker samples were selected to prepare cement mortar according to the GB / T17671 method, and their flexural and compressive strengths at 1d, 3d and 28d were tested. The actual test results are shown in Table 9.

[0113] Table 9 shows the test results of three cement clinker samples in this embodiment.

[0114]

[0115] Comparing the results in Table 7 and Table 9, it can be seen that the performance requirements of cement clinker produced based on the clinker mineral phase design results are within the target range and have high accuracy.

[0116] Example 3

[0117] This embodiment illustrates the query function of the cement clinker database, including the query function for clinker mineral phase data and the query function for clinker performance parameters.

[0118] For the clinker mineral facies data query function, you can select "Clinker Mineral Facies Data Query" on the data query interface. Enter the clinker mineral facies name, abbreviation, and / or 1 to 6 elements contained in the mineral facies in the corresponding spaces on the interface. Clicking the query button will automatically display a brief summary of the query results (if no data is entered, a summary of all mineral facies information will be returned in pages). In this embodiment, no data was entered, and the displayed results are shown in Table 10.

[0119] Table 10 Results of clinker mineral phase data query when no query requirements were entered.

[0120]

[0121] Clicking on the mineral name (in Chinese) for each mineral phase will display its complete mineral phase information; clicking on tricalcium silicate will display its complete mineral phase information, as shown in Table 11.

[0122] Table 11 Results of all mineral phase information for tricalcium silicate

[0123]

[0124]

[0125] Some mineral facies information is currently missing; administrators can supplement it through the backend management system.

[0126] To query the performance parameters of cement clinker, select "Clinker Performance Parameter Query" in the data query interface, choose keywords (specific surface area, initial setting time, final setting time, 3-day flexural strength, 28-day flexural strength, 3-day compressive strength, and 28-day compressive strength, etc.), and then enter the query range in the corresponding position. The results will be displayed according to the query conditions.

[0127] In this embodiment, the performance data of cement clinker with a 28-day compressive strength of 55-56 MPa were queried, and the results are shown in Table 12.

[0128] Table 12 Results of data retrieval for cement clinker with a 28-day compressive strength of 55-56.

[0129]

[0130]

[0131] The technical features in the claims and / or specification of this invention can be combined, and the combination is not limited to the combinations obtained through reference in the claims. Technical solutions obtained by combining the technical features in the claims and / or specification are also within the scope of protection of this invention.

[0132] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for constructing a cement clinker performance prediction and mineral phase design model, characterized in that, It uses a neural network algorithm to predict clinker performance and design clinker mineral phases; the clinker performance prediction involves inputting known mineral phase parameters of clinker into the cement clinker performance prediction and mineral phase design model, and the cement clinker performance prediction and mineral phase design model outputs its prediction results of the performance data of the known clinker; The clinker mineral phase design involves inputting the performance data of the target clinker into the cement clinker performance prediction and mineral phase design model, and the cement clinker performance prediction and mineral phase design model outputting its design results for the mineral phase parameters of the target clinker. The training is conducted according to the following steps: 1) Obtain the mineral phase parameters of the known clinker or the performance data of the target clinker; 2) Input the known clinker mineral phase parameters or the target clinker performance data into the cement clinker performance prediction and mineral phase design model; 3) Adjust the number of neural network layers n, the number of output units per layer m, and the number of iterations k. The cement clinker performance prediction and mineral phase design model outputs the prediction results of known clinker performance data or the design results of mineral phase parameters of the target clinker. 4) Compare the predicted or designed results with the measured values ​​of the cement clinker, and calculate the mean square error mse and the mean absolute error mae of the predicted or designed results; If the mean square error mse or the mean absolute error mae does not meet the preset threshold, then return to step 3). If both the mean square error mse and the mean absolute error mae meet the preset threshold, then the cement clinker performance prediction and mineral phase design model are obtained.

2. The method for constructing a cement clinker performance prediction and mineral phase design model according to claim 1, characterized in that, The mineral phase parameters include the mineral phase types, the content of each mineral phase, and the specific surface area of ​​the cement clinker; the performance data include 1-day flexural strength, 3-day flexural strength, 28-day flexural strength, 1-day compressive strength, 3-day compressive strength, and 28-day compressive strength; the mineral phase parameters input to the cement clinker performance prediction and mineral phase design model include several cement clinker samples, each cement clinker sample including a parameter matrix representing multiple dimensions of the mineral phase; the performance data input to the cement clinker performance prediction and mineral phase design model includes several cement clinker samples, each cement clinker sample including a data matrix representing multiple dimensions of performance.

3. The method for constructing a cement clinker performance prediction and mineral phase design model according to claim 1, characterized in that, The number of neural network layers n is a natural number from 1 to 10; the number of output units per layer m is a natural number from 6 to 50; the number of iterations k is a natural number from 1 to 500; and the preset thresholds are mean square error mse < 1 and mean absolute error mae < 1.

4. A cement clinker mineral phase database, characterized in that, It includes: The software front-end is installed on a smart terminal; the software front-end includes a data upload interface, a data query interface, and a data prediction interface. The database backend is hosted on a solid-state server. The database backend is connected to the software frontend; The database backend includes: The data upload module, through the data upload interface, enables human-computer interaction and is used to upload and manage the basic data information of the cement clinker. The data query module enables human-computer interaction through the data query interface, allowing users to query basic data information of the cement clinker using mineral phase parameters or performance data. The data prediction module includes the method for constructing a cement clinker performance prediction and mineral phase design model as described in any one of claims 1 to 3; the data prediction module performs human-computer interaction through the data prediction interface to predict the performance data of known clinker or the mineral phase parameters of the design target clinker.

5. The cement clinker mineral phase database according to claim 4, characterized in that, The basic data information includes mineral phase characterization parameters and performance characterization parameters of cement clinker; The mineral phase characterization parameters include the name, abbreviation, chemical formula, molecular formula, structural diagram, sintering temperature, melting point, crystal parameters, carbon emissions, hydration activity, data source, and data upload time of each mineral phase. The crystal parameters are parameters characterizing mineral phase crystals, including the three sets of edge lengths a, b, and c of the unit cell, the included angles α, β, and γ between the three sets of edges, the number of different elements, the space group number, the Hermann-Mauguin space group symbol, the Hall space group symbol, the reflection residual factor, the ion doping performance, the cohesive energy per unit cell, the formation energy, and / or the density of states. The performance characterization parameters include the cement clinker number, clinker system, content of each mineral phase, clinker performance, testing time, and data source; the clinker performance includes alumina content, magnesium oxide content, sulfur trioxide content, free calcium oxide content, loss on ignition, insoluble matter content, alkali content, specific surface area, residue on a 45μm square-hole sieve, initial setting time, final setting time, soundness, 3-day flexural strength, 7-day flexural strength, 28-day flexural strength, 3-day compressive strength, 7-day compressive strength, and / or 28-day compressive strength; Querying the basic data information of cement clinker by mineral phase parameters involves inputting the mineral phase name, abbreviation, or 1-6 elements contained in the mineral phase of the target cement clinker into the data prediction interface. The cement clinker mineral phase database can output the mineral phase characterization parameters of all cement clinker that meet the input conditions. Querying the basic data information of cement clinker by performance data involves inputting the specific surface area, initial setting time, final setting time, 3-day flexural strength, 28-day flexural strength, 3-day compressive strength, and / or 28-day compressive strength of the target cement clinker into the data prediction interface. The cement clinker mineral phase database can output the performance characterization parameters of all cement clinker that meet the input conditions.

6. The cement clinker mineral phase database according to claim 4, characterized in that, The software front-end also includes a user management interface; the database back-end also includes a user management module; the user management module uses the user management interface for human-computer interaction, and is used for user registration and user permission management; the cement clinker performance prediction and mineral phase design model is dynamically updated based on changes in the basic data information of the cement clinker.

7. A method for applying a cement clinker mineral phase database according to any one of claims 4 to 6, characterized in that, It includes the following steps: The user inputs basic data information of cement clinker into the cement clinker mineral phase database through the data upload interface; the data upload module determines whether the basic data information is correct; if not, it prompts the user that the data format is incorrect and the specific location of the error; if yes, it uploads the basic data information to the database backend. The data query interface inputs the mineral phase parameters or performance data into the cement clinker mineral phase database; the data query interface outputs the basic data information of the cement clinker. Select the performance data of the known clinker or the mineral phase parameters of the target clinker through the data prediction interface; input the mineral phase parameters of the known clinker or the performance data of the target clinker according to the prompts of the data prediction interface; the data prediction interface outputs the prediction results of the performance data of the known clinker or the design results of the mineral phase parameters of the target clinker.

8. A storage medium comprising a stored program, characterized in that, When the program is running, it controls the device containing the storage medium to execute the application method of claim 7.

9. An electronic device, comprising a storage medium, characterized in that, It includes: One or more processors, wherein the storage medium is coupled to the processors, and the processors are configured to execute program instructions stored in the storage medium; wherein the program instructions, when executed, perform the application method of claim 7.

Images