A preparation method and terminal device of a green cement-based material based on crystal form regulation

Abstract

This application is applicable to the field of material preparation technology, and provides a method for preparing a green cement-based material based on crystal form control and a terminal device. The method includes: training a crystal prediction model to obtain a trained crystal prediction model, where the crystal prediction model includes a category prediction model and an image prediction model. When training the crystal prediction model, the parameter influence degree corresponding to each sample output by the category prediction model is used to train the image prediction model, so that the image prediction model refers to the influence of the parameters when outputting crystal images, making the generated crystal images more accurate. Inputting new raw material ratio parameters into the trained crystal prediction model to obtain the crystal form category and crystal image corresponding to each raw material ratio parameter, and finally determining the optimal raw material ratio according to the crystal form category and crystal image. When making the cement-based material, calcium carbonate crystals with better performance are generated according to the optimal raw material ratio, and the generated calcium carbonate crystals are used to generate the cement-based material.

Description

Technical Field

[0001] This application belongs to the technical field of material preparation, and particularly relates to a method for preparing a green cement-based material based on crystal form regulation and a terminal device. Background Art

[0002] Cement-based materials are widely used in infrastructure such as buildings, roads, and bridges. As an admixture for cement-based materials, calcium carbonate has many remarkable advantages and has brought an innovative impact to the field of civil engineering. Currently, calcium carbonate mainly exists in three crystal forms: calcite, aragonite, and vaterite. Due to the differences in crystal structures and physical and chemical properties, different crystal forms of calcium carbonate show different effects as admixtures for cement-based materials.

[0003] Currently, the production methods of calcium carbonate mainly include chemical synthesis and carbonization methods. These artificial synthesis methods have high energy consumption, long production cycles, and high requirements for equipment, and cannot meet the production needs. Therefore, how to quickly prepare calcium carbonate with better performance to obtain better cement-based materials is a problem that needs to be solved currently. Summary of the Invention

[0004] The embodiments of this application provide a method for preparing a green cement-based material based on crystal form regulation and a terminal device, which can solve the problem of how to prepare calcium carbonate with better performance to improve the quality of cement-based materials.

[0005] In a first aspect, the embodiments of this application provide a method for preparing a green cement-based material based on crystal form regulation, including:

[0006] Inputting training samples into a crystal prediction model to be trained to train the crystal prediction model, and obtaining a trained crystal prediction model; wherein, the training samples are composed of multiple groups of first raw material ratio parameters for generating calcium carbonate crystals, and each group of first raw material ratio parameters is a sample; the crystal prediction model includes a category prediction model and an image prediction model, and the parameter influence degree corresponding to each sample output in the category prediction model is used to train the image prediction model, and the parameter influence degree represents the contribution degree of each parameter in the input sample to the result;

[0007] Inputting multiple groups of second raw material ratio parameters into the trained crystal prediction model to obtain the crystal form category and crystal image of the calcium carbonate crystals generated by each group of the second raw material ratio parameters;

[0008] Based on the crystal form category and the crystal image, determining the optimal raw material ratio parameters of the calcium carbonate crystals;

[0009] Preparing calcium carbonate crystals according to the optimal raw material ratio parameters, and using the prepared calcium carbonate crystals to prepare cement-based materials.

[0010] In one implementable manner, the first raw material ratio parameter includes calcium ion concentration, crystal form regulator concentration, and crystal form regulator type;

[0011] Training the crystal prediction model by inputting the training samples into the crystal prediction model to be trained to obtain the trained crystal prediction model includes:

[0012] Obtaining the j-th sample from the training samples, where j is a positive integer from 1 to N - 1, and N is the total number of samples in the training samples;

[0013] Inputting the j-th sample into the category prediction model to train the category prediction model to obtain the parameter influence degree corresponding to the j-th sample, and sending the parameter influence degree corresponding to the j-th sample to the image prediction model;

[0014] Inputting the j-th sample into the image prediction model, and training the image prediction model by using the parameter influence degree corresponding to the j-th sample and the j-th sample;

[0015] Inputting the (j + 1)-th sample in the training samples into the crystal prediction model to train the crystal prediction model until the maximum number of training times is reached to obtain the trained crystal prediction model.

[0016] In one implementable manner, determining the optimal raw material ratio parameter of the calcium carbonate crystal based on the crystal form category and the crystal image includes:

[0017] Calculating the similarity between the predicted crystal image and a preset standard image, where the standard image corresponding to the crystal image is determined according to the crystal form category of the crystal image;

[0018] Calculating the peak signal-to-noise ratio of the crystal image;

[0019] Calculating the distribution difference between the crystal image and a preset standard image;

[0020] Calculating the morphological matching degree between the crystal image and a preset standard image;

[0021] Screening out the optimal raw material ratio parameter corresponding to each crystal form category according to the similarity, the peak signal-to-noise ratio, the distribution difference, and the morphological matching degree.

[0022] In one implementable manner, screening out the optimal raw material ratio parameter corresponding to each crystal form category according to the similarity, the peak signal-to-noise ratio, the distribution difference, and the morphological matching degree includes:

[0023] Calculate the total score of the crystal image according to the similarity, the peak signal-to-noise ratio, the distribution difference, and the morphological matching degree;

[0024] According to the total score of each crystal image, screen out the optimal raw material ratio parameters corresponding to each crystal form category.

[0025] In one implementable manner, generating training samples based on multiple calcium carbonate crystals includes:

[0026] Obtain initial images of multiple calcium carbonate crystals, where each calcium carbonate crystal corresponds to one initial image, and different calcium carbonate crystals are generated according to different first raw material ratio parameters;

[0027] Perform clustering processing on multiple initial images to obtain the crystal form category of the calcium carbonate crystal in each initial image;

[0028] According to the crystal form category of each calcium carbonate crystal, label each group of the first raw material ratio parameters and each initial image to obtain the first raw material ratio parameters with labeled crystal form categories and the initial images with labeled crystal form categories. The first raw material ratio parameters and initial images with labeled crystal form categories form training samples.

[0029] In one implementable manner, the above method may further include: obtaining solid waste, where calcium ions exist in the solid waste;

[0030] Generate multiple groups of initial extraction conditions for extracting calcium ions from the solid waste to obtain an initial population. The initial extraction conditions include temperature, extraction time, extractant, and solid-liquid ratio;

[0031] For each group of the initial extraction conditions, based on the initial extraction conditions, extract the calcium ions from the solid waste to obtain the performance parameters of the calcium ion extraction process under the initial extraction conditions;

[0032] Based on the performance parameters, calculate the fitness value of each group of the initial extraction conditions;

[0033] Based on the fitness value of each group of the initial extraction conditions, screen genetic individuals from the initial population, and construct a new population according to the genetic individuals to obtain the i-th candidate population, where 1 ≤ i ≤ M and M is the maximum number of iterations;

[0034] Perform iterative loops until the preset maximum number of iterations is reached. Screen genetic individuals from the M-th candidate population to obtain the optimal extraction conditions. The optimal extraction conditions are used to extract the calcium ions from the solid waste, and the calcium ions are used to prepare calcium carbonate crystals when preparing cement-based materials. The extraction conditions corresponding to the genetic individuals screened from the M-th candidate population are the optimal extraction conditions.

[0035] In one implementable manner, the performance parameters include an extraction rate, energy consumption, and waste liquid volume. Based on the performance parameters, calculating the fitness value of each group of the initial extraction conditions includes:

[0036] For each group of the initial extraction conditions, determine the weight of each performance parameter according to the extraction rate corresponding to the initial extraction conditions, and preset the weights of the performance parameters corresponding to different extraction rates;

[0037] Based on the weight of each performance parameter and each performance parameter, calculate the fitness of each group of the initial extraction conditions.

[0038] In a second aspect, an embodiment of the present application provides a device for preparing a green cement-based material based on crystal form regulation, including:

[0039] A model training module for inputting training samples into a crystal prediction model to be trained to train the crystal prediction model, and obtaining a trained crystal prediction model; wherein, the training samples are composed of multiple groups of first raw material ratio parameters for generating calcium carbonate crystals, and each group of first raw material ratio parameters is a sample; the crystal prediction model includes a category prediction model and an image prediction model, and the parameter influence degree corresponding to each sample output in the crystal prediction model is used to train the image prediction model, and the parameter influence degree characterizes the contribution degree of each parameter in the input sample to the result;

[0040] A prediction module for inputting multiple groups of second raw material ratio parameters into the trained crystal prediction model to obtain the crystal form category and crystal image of the calcium carbonate crystals generated by each group of the second raw material ratio parameters;

[0041] A parameter determination module for determining the optimal raw material ratio parameters of the calcium carbonate crystals based on the crystal form category and the crystal image, wherein the optimal raw material ratio parameters are used to prepare the calcium carbonate crystals required for preparing the cement-based material when preparing the cement-based material.

[0042] In a third aspect, an embodiment of the present application provides a terminal device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for preparing a green cement-based material based on crystal form regulation according to any one of the first aspects is implemented.

[0043] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the method for preparing a green cement-based material based on crystal form regulation according to any one of the first aspects is implemented.

[0044] In a fifth aspect, an embodiment of the present application provides a computer program product. When the computer program product runs on a terminal device, the terminal device is caused to execute the method for preparing a green cement-based material based on crystal form regulation described in any one of the first aspects above.

[0045] The beneficial effects of the embodiment of the first aspect of the present application compared with the prior art are as follows: The crystal prediction model is trained using training samples to obtain a trained crystal prediction model. The crystal prediction model includes a class prediction model and an image prediction model. When training the crystal prediction model, the parameter influence degree corresponding to each sample output by the class prediction model is used to train the image prediction model. The image prediction model refers to the influence of the parameters when outputting the crystal image, making the generated crystal image more accurate. After obtaining the trained crystal prediction model, new raw material ratio parameters are input into the trained crystal prediction model to obtain the crystal form class and crystal image corresponding to each raw material ratio parameter. Finally, the optimal raw material ratio is determined based on the crystal form class and crystal image. Through the trained crystal prediction model of the present application, the crystal form class and crystal image of calcium carbonate crystals generated by the raw material ratio parameters can be quickly predicted, eliminating the process of real experiments, making the crystal form class and crystal image of calcium carbonate crystals generated by the determined raw material ratio parameters faster; the optimal raw material ratio of calcium carbonate crystals is finally determined based on a large number of crystal images. The performance of calcium carbonate generated according to the optimal raw material ratio is better, and the calcium carbonate with better performance is used to prepare the cement-based material, improving the quality of the cement-based material. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a schematic flowchart of a method for preparing a green cement-based material based on crystal form regulation provided by an embodiment of the present application;

[0047] Figure 2 is a schematic flowchart of a method for training a crystal prediction model provided by an embodiment of the present application;

[0048] Figure 3 is a schematic flowchart of a method for generating training samples provided by an embodiment of the present application;

[0049] Figure 4 is a schematic flowchart of a method for extracting solid waste calcium source provided by an embodiment of the present application;

[0050] Figure 5 is a schematic structural diagram of a device for preparing a green cement-based material based on crystal form regulation provided by an embodiment of the present application;

[0051] Figure 6 is a schematic structural diagram of a terminal device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0053] Cement-based materials are widely used in infrastructure such as buildings, roads, and bridges, but their production process is accompanied by high energy consumption and large carbon emissions. Calcium carbonate, as an admixture of cement-based materials, has many significant advantages and has brought revolutionary impacts to the field of civil engineering.

[0054] In nature, calcium carbonate mainly exists in three crystal forms: calcite, aragonite and vaterite. Different crystal forms of calcium carbonate as admixtures for cement-based materials show different effects due to their differences in crystal structure and physical and chemical properties. Among them, calcite calcium carbonate is the most common crystal form with a stable crystal structure. In cement-based materials, it can effectively improve the working performance of cement-based materials. For example, the particle morphology of calcite calcium carbonate is regular and the surface is relatively smooth. It plays a "ball effect" in cement paste, greatly improving the fluidity of the paste and facilitating the pouring and construction of concrete. The crystal structure of aragonite calcium carbonate is relatively special, with a high surface area and surface energy. In improving the working performance of cement-based materials, aragonite calcium carbonate can interact more closely with cement particles and water with its high specific surface area, further optimize water retention and reduce water seepage to a certain extent. Vaterite calcium carbonate is a metastable phase with a unique spherical structure. In terms of the working performance of cement-based materials, the spherical structure of vaterite calcium carbonate makes it more prominent in improving the fluidity in cement paste, which can effectively improve the construction and workability of concrete.

[0055] At present, the production pathways of calcium carbonate mainly come from chemical synthesis and carbonization methods. These artificial synthesis methods have high energy consumption, long production cycles, and high equipment requirements. With the deepening of the concept of green building and sustainable development, the application prospects of green and low-carbon calcium carbonate preparation technology are extremely broad. Biomineralization technology can induce the crystallization of calcium carbonate minerals through organisms such as urease or bacteria. Its production method has the advantages of green environmental protection. How to use biomineralization technology to generate calcium carbonate more efficiently is a problem that needs to be solved at present.

[0056] The calcium sources in biomineralization technology are mostly derived from chemical products, such as calcium chloride, calcium acetate, etc. In addition, since calcium carbonate crystals have a variety of isomorphic variants, the morphology and size of mineral crystals are difficult to accurately control during the mineralization process, which will affect the mechanical properties and durability of cement-based materials. Therefore, how to extract calcium ions in a more environmentally friendly way is a problem that needs to be solved at present.

[0057] Based on this, the present application proposes a method for preparing a crystal form-regulated green cement-based material. First, the present application uses training samples to train a crystal prediction model to obtain a trained crystal prediction model. A large number of raw material ratio parameters are input into the trained crystal prediction model, and the crystal prediction model can predict the crystal form category and crystal image of calcium carbonate that can be generated according to the raw material ratio parameters. Finally, the optimal crystal image is selected from the crystal images to obtain the optimal crystal image corresponding to each crystal form category. When preparing the cement-based material, calcium carbonate crystals can be generated according to the optimal raw material ratio parameters, and then the cement-based material is prepared using the calcium carbonate crystals.

[0058] By searching for the optimal raw material ratio parameters of calcium carbonate, the present application can prepare calcium carbonate with better performance, thereby making the performance of the prepared cement-based material better.

[0059] Regarding the calcium ions used for preparing calcium carbonate crystals, the present application extracts calcium ions from solid waste to achieve the purpose of environmental protection. Specifically, through a screening method, the optimal extraction conditions for extracting calcium ions from solid waste are obtained, so as to efficiently extract calcium ions from solid waste using the optimal extraction conditions.

[0060] Finally, the calcium ions extracted from solid waste are fused with other raw materials according to the optimal raw material ratio parameters to obtain calcium carbonate crystals with better performance. The calcium carbonate crystals are then mixed with other substances to prepare a cement-based material.

[0061] The following combines Figure 1 to detail the method for preparing a crystal form-regulated green cement-based material according to the embodiments of the present application.

[0062] Figure 1 shows a schematic flowchart of the method for preparing a crystal form-regulated green cement-based material provided by the present application. Referring to Figure 1 , the detailed description of the optimized method is as follows:

[0063] S101, input training samples into the crystal prediction model to be trained to train the crystal prediction model, and obtain a trained crystal prediction model; wherein, the training samples are composed of multiple groups of first raw material ratio parameters for generating calcium carbonate crystals, and each group of first raw material ratio parameters is a sample; the crystal prediction model includes a category prediction model and an image prediction model, and the parameter influence degree corresponding to each sample output by the category prediction model is used to train the image prediction model, and the parameter influence degree characterizes the contribution degree of each parameter in the input sample to the result.

[0064] In this embodiment, the crystal form prediction model is used to predict the crystal form category of calcium carbonate crystals generated using the input raw material ratio parameters. The crystal form categories may include calcite, aragonite, and vaterite. The image prediction model is used to predict the crystal image of calcium carbonate crystals generated using the input raw material ratio parameters.

[0065] In this embodiment, the first raw material ratio parameter is a randomly selected parameter.

[0066] A set of raw material ratio parameters may include crystal form regulator concentration, crystal form regulator type, urease source, urease concentration, calcium source type, calcium ion concentration, and urea concentration, etc. Other raw materials may be crystal form regulators, urease, etc.

[0067] As an example, the crystal form regulator concentration may be 0.05M, 0.1M, or 0.2M, etc. The crystal form regulator type may include magnesium chloride, magnesium sulfate, magnesium hydroxide, etc. The urease source may be: drying and grinding pumpkin seeds, watermelon seeds, or soybeans, stirring and mixing the ground powder with deionized water for 1 hour, and taking the supernatant as the urease solution. The urease concentration may be 20g / L. The calcium source type may be calcium chloride, etc. The calcium ion concentration may be 0.1 mol / L or 0.2mol / L, etc. The urea concentration may be 0.1mol / L, etc. Take pictures and classify the generated calcium carbonate crystals, and label the images of the calcium carbonate crystals and the first raw material ratio parameters corresponding to the calcium carbonate crystals according to the category of the calcium carbonate crystals to obtain training samples. The training samples include the first raw material ratio parameters with labeled crystal form categories and the images of calcium carbonate crystals with labeled crystal form categories.

[0068] After inputting the first raw material ratio parameter into the crystal form prediction model, the crystal form prediction model outputs the crystal form category of the calcium carbonate crystals corresponding to the first raw material ratio parameter, and the parameter influence degree of each parameter in the first raw material ratio parameter. For example, the parameter influence degree of the crystal form regulator concentration in the first raw material ratio parameter is 10%, the parameter influence degree of the crystal form regulator is 5%, the parameter influence degree of the urease source is 7%, etc. The parameter influence degrees corresponding to different first raw material ratio parameters may be different.

[0069] The image prediction model will output the corresponding image according to the parameter influence degree. For example, the size of calcium carbonate crystals generated by a low concentration of crystal form regulator may be small, and as the concentration of the crystal form regulator increases, the size of the generated calcium carbonate crystals increases. When training the image prediction model, adding the parameter influence degree generated by the crystal form prediction model can make the trained image prediction model more accurate.

[0070] S102, Input multiple groups of second raw material ratio parameters into the trained crystal prediction model to obtain the crystal form category and crystal image of calcium carbonate crystals generated by each group of the second raw material ratio parameters.

[0071] In this embodiment, the number of the second raw material ratio is much larger than that of the first raw material ratio parameter, so that a small sample can be used to train the crystal prediction model, and then a large sample and the trained crystal prediction model can be used to find the optimal solution.

[0072] S103. Determine the optimal raw material ratio parameter of the calcium carbonate crystal based on the crystal form category and the crystal image.

[0073] In one way, classify the crystal images to obtain all the crystal images under the same crystal form category, and then input all the images under the same crystal form category into the neural network model to obtain the optimal crystal image under this crystal form category. The second raw material ratio parameter corresponding to the optimal crystal image is the optimal raw material ratio parameter corresponding to this crystal form category.

[0074] In another way, determine the optimal crystal image by analyzing the crystal morphology, size and distribution of the calcium carbonate crystal in the crystal image, and then obtain the optimal raw material ratio parameter.

[0075] S104. Prepare calcium carbonate crystals according to the optimal raw material ratio parameter, and use the prepared calcium carbonate crystals to prepare cement-based materials.

[0076] In actual use, when it is necessary to prepare A-type calcium carbonate crystals, the raw materials are proportioned into a biological solution according to the optimal raw material ratio parameter corresponding to the A-type calcium carbonate crystals, and then a biomineralization reaction is carried out to obtain a precipitate of calcium carbonate crystals. The calcium carbonate crystals are mixed with other materials to obtain cement-based materials.

[0077] In this application, first use the training samples to train the crystal prediction model to obtain the trained crystal prediction model. The crystal prediction model includes a category prediction model and an image prediction model. When training the crystal prediction model, the parameter influence degree corresponding to each sample output by the category prediction model is used to train the image prediction model. The image prediction model refers to the influence of the parameters when outputting crystal images, so that the generated crystal images are more accurate. After obtaining the trained crystal prediction model, input the new raw material ratio parameters into the trained crystal prediction model to obtain the crystal form category and crystal image corresponding to each raw material ratio parameter. Finally, determine the optimal raw material ratio according to the crystal form category and crystal image. Through the trained crystal prediction model of this application, the crystal form category and crystal image of the calcium carbonate crystal generated by the raw material ratio parameter can be quickly predicted, saving the process of real experiments, and making the crystal form category and crystal image of the calcium carbonate crystal generated by the determined raw material ratio parameter faster; finally determine the optimal raw material ratio of the calcium carbonate crystal according to a large number of crystal images. The performance of the calcium carbonate generated according to the optimal raw material ratio is better. Use the calcium carbonate with better performance to prepare cement-based materials to improve the quality of cement-based materials.

[0078] As Figure 2 shown, in a possible implementation, the implementation process of step S101 may include:

[0079] S1011, obtain the j-th sample from the training samples, where j is a positive integer from 1 to N - 1, and N is the total number of samples in the training samples.

[0080] In this embodiment, the j-th sample includes the first raw material ratio parameter for annotating the crystal form category and the initial image of the calcium carbonate crystal for annotating the crystal form category. The initial image in the j-th sample is an image of the calcium carbonate crystal generated using the first raw material ratio parameter in the j-th sample.

[0081] S1012, input the j-th sample into the category prediction model to train the category prediction model, obtain the parameter influence degree corresponding to the j-th sample, and send the parameter influence degree corresponding to the j-th sample to the image prediction model.

[0082] In this embodiment, the category prediction model can be constructed based on a random forest classifier. Specifically, the hyperparameters of the classifier (such as the number of trees, maximum depth) can be adjusted using grid search to train the model. Adjusting the hyperparameters of the classifier using grid search can help find a relatively optimal set of hyperparameters, enabling the model to obtain better performance on the given dataset, reducing the overfitting or underfitting phenomenon of the model, and improving the generalization ability of the model.

[0083] After inputting the j-th sample into the category prediction model, the category prediction model outputs the predicted crystal form category of the calcium carbonate crystal that can be generated using the j-th sample and the parameter influence degree of each parameter in the j-th sample. Specifically, the SHAP (SHapley Additive exPlanations) value is used to characterize the parameter influence degree, and the larger the SHAP value, the greater the influence degree of the parameter on the crystal form category.

[0084] After obtaining the predicted crystal form category of the calcium carbonate crystal, compare the predicted crystal form category of the calcium carbonate crystal with the true crystal form category generated by the j-th sample to generate a comparison result, and update the parameters in the category prediction model according to the comparison result to obtain the category prediction model with updated parameters. The category prediction model with updated parameters is used for the next training.

[0085] S1013, input the j-th sample into the image prediction model, and use the parameter influence degree corresponding to the j-th sample and the j-th sample to train the image prediction model.

[0086] In this embodiment, the parameter weight in the image prediction model is set by using the parameter influence degree corresponding to the j-th sample, and then the crystal image of calcium carbonate crystals is predicted and generated according to the j-th sample. The parameters in the image prediction model are updated according to the predicted crystal image, and the image prediction model after updating the parameters is obtained. The image prediction model after updating the parameters is used for the next training.

[0087] In this embodiment, the image prediction model can be constructed based on an adversarial network, and the adversarial network includes a generator and a discriminator.

[0088] The input of the generator is the first raw material ratio parameter, and the output is the predicted crystal image. As an example, the crystal image can be an image of 256×256, and the crystals in the image can be needle-shaped crystals, spherical crystals, rhombic crystals, etc. Specifically, a transposed convolution network can be used in the generator to gradually generate the image. For example, if the initial input is an embedding vector of 1×1×N, the vector size is enlarged by each transposed convolution layer to generate a higher-resolution image. Then, the generated image is subjected to normalization processing and activation function processing to make the generated image more stable. In the generator, low-resolution feature maps and high-resolution feature maps can also be fused to obtain feature maps that can better represent detailed features, and using the feature maps that better represent detailed features to generate the crystal image can make the obtained crystal image more accurate. The loss function of the generator can include adversarial loss and reconstruction loss. The adversarial loss is used to train the generator and the discriminator. The goal of the generator is to minimize the probability that the generated image is judged to be fake by deceiving the discriminator. The reconstruction loss is used to constrain the pixel-level similarity between the generated image and the real image to ensure the same morphology.

[0089] The input of the discriminator is the generated image of the generator and the real crystal image in the j-th sample; the output of the discriminator is the probability used to represent that the generated image is a real crystal image, and the output probability of the discriminator is between 0 and 1. A convolutional network is used in the discriminator to extract image features; an activation function is added to capture complex features, for example, the LeakyReLU activation function; the probability is output through a fully connected layer and a Sigmoid activation function. The loss function of the discriminator can be binary cross-entropy loss.

[0090] When training the adversarial network, if the classification accuracy of the discriminator is close to a preset value (for example, 50%), and / or the structural similarity index of the generated image reaches above the preset value (for example, above 0.85), it indicates that the adversarial network reaches a convergence state and the training of the adversarial network ends.

[0091] During actual training, the images in the training samples can also be rotated, flipped, cropped, etc. to increase the diversity of the samples, so that the trained image prediction model can adapt to various situations, and at the same time, the accuracy of the image prediction model can also be increased.

[0092] S1014. Input the (j + 1)-th sample in the training samples into the crystal prediction model to train the crystal prediction model until the maximum number of training times is reached, and obtain the trained crystal prediction model.

[0093] In this embodiment, after training the crystal prediction model with one sample, continue to train the crystal prediction model with another sample, and iterate in this way until the maximum number of training times is reached, then the trained crystal prediction model can be obtained.

[0094] After obtaining the trained crystal prediction model, use the trained crystal prediction model to quickly predict the crystal form category and crystal image of calcium carbonate crystals that can be obtained from the input raw material ratio parameters, reducing the cost of trial and error in experiments.

[0095] In one embodiment, after the training of the crystal prediction model is completed, obtain validation samples, use the validation samples to evaluate the stability of the crystal prediction model, output validation metrics, and then modify the parameters of the trained crystal prediction model to obtain the final trained crystal prediction model. The validation metrics can include accuracy, recall rate, and F1 score, etc. Among them, the accuracy rate represents the proportion of correct predictions by the crystal prediction model; the recall rate represents the proportion of correct predictions for a certain crystal form category; the F1 score is a balanced metric that comprehensively considers precision and recall rate.

[0096] After the training of the crystal prediction model is completed, the cross-validation method can also be used to verify the trained crystal prediction model to determine the stability of the crystal prediction model and ensure that the performance of the model is consistent under different data partitions.

[0097] After the training of the crystal prediction model is completed, analyze the samples with prediction errors, modify the samples for high-error samples, and then use the modified samples to continue training the crystal prediction model to obtain the trained crystal prediction model.

[0098] In this application, when training the crystal prediction model, each time the image prediction model is trained, the parameter influence degree of the parameters in the samples output by the category prediction model will be used. Guided by the parameter influence degree, generate the image of calcium carbonate crystals corresponding to the samples, making the obtained image of calcium carbonate crystals more accurate.

[0099] After obtaining the trained crystal prediction model, in order to find the optimal raw material ratio parameters, the crystal prediction model can be used to predict the generation results of a large number of parameters, and then obtain the optimal raw material ratio parameters from the large number of parameters.

[0100] Specifically, the implementation method of the above step S103 includes:

[0101] S1031. Calculate the similarity between the predicted crystal image and a preset standard image, where the standard image corresponding to the crystal image is determined according to the crystal form category of the crystal image.

[0102] In this embodiment, standard images corresponding to different crystal form categories are pre-stored.

[0103] After obtaining a crystal image predicted by a crystal prediction model, according to the crystal form category of the predicted crystal image, find the standard image corresponding to this crystal form category.

[0104] In one way, the mean square error is used to determine the similarity between the predicted crystal image and the preset standard image.

[0105] In another way, the structural similarity index is used to characterize the similarity between the predicted crystal image and the preset standard image. The structural similarity index is within 0 - 1, and the closer the structural similarity index is to 1, the more similar the structures are.

[0106] Specifically, through , where is the structural similarity index; is the average value of the intensity values of all pixels in the predicted crystal image; is the average value of the intensity values of all pixels in the standard image; is the variance of the intensity values of the pixels in the predicted crystal image; is the variance of the intensity values of the pixels in the standard image; is the covariance between the predicted crystal image and the standard image; is a preset parameter; is a preset parameter.

[0107] S1032. Calculate the peak signal-to-noise ratio of the crystal image.

[0108] In this embodiment, the peak signal-to-noise ratio is used to measure the image clarity, and the higher the peak signal-to-noise ratio, the better the image quality.

[0109] Specifically, the peak signal-to-noise ratio is calculated according to . Where PSNR is the peak signal-to-noise ratio; is the maximum pixel value; MSE is the mean square error value.

[0110] S1033. Calculate the distribution difference between the crystal image and the preset standard image.

[0111] In this embodiment, the smaller the distribution difference between the crystal image and the preset standard image, the closer the feature distribution of the generated image is to the feature distribution of the real image.

[0112] The distribution difference can be based on Calculated. Among them, FID is the distribution difference between the predicted crystal image and the preset standard image; is the mean vector of the standard image in the feature space; is the mean vector of the predicted crystal image in the feature space; is the covariance matrix of the standard image in the feature space; is the covariance matrix of the predicted crystal image in the feature space; is the trace of the matrix, that is, the sum of the diagonal elements of the matrix. Characterizes the square of the Euclidean distance between the mean vector of the standard image and the mean vector of the predicted crystal image.

[0113] The calculation method of the mean vector of the standard image in the feature space is as follows: Extract the D-dimensional feature vectors of the standard image to obtain Y feature vectors; D is the dimension of the feature space. Calculate the mean of the Y feature vectors to obtain the mean vector. The calculation method of the mean vector of the predicted crystal image in the feature space is the same as that of the mean vector of the standard image, and will not be elaborated here.

[0114] S1034. Calculate the morphological matching degree between the crystal image and the preset standard image.

[0115] In this embodiment, calculate the intersection over union (IoU) between the predicted crystal image and the standard image to obtain the morphological matching degree. Specifically, calculate the intersection area and the union area between the predicted crystal image and the standard image, and then according to calculate the IoU; IoU is the intersection over union. IoU is between 0 and 1. The closer the IoU is to 1, the higher the overlap degree of the two images and the better the morphological matching degree.

[0116] S1035. According to the similarity, the peak signal-to-noise ratio, the distribution difference, and the morphological matching degree, screen out the optimal raw material ratio parameters corresponding to each crystal form category.

[0117] In this embodiment, according to the similarity, the signal-to-noise ratio, the distribution difference, and the morphological matching degree, calculate the total score of the crystal image; according to the total score of each crystal image, screen out the optimal raw material ratio parameters corresponding to each crystal form category.

[0118] In one way, the calculation method of the total score of the crystal image is:

[0119] Using , calculate the total score. Among them, L1 is the weight of similarity, SSIM is the similarity, L2 is the weight of peak signal-to-noise ratio, PSNR is the peak signal-to-noise ratio; FID is the distribution difference, L3 is the weight of the distribution difference, L4 is the weight of the morphological matching degree, and P is the morphological matching degree. L1, L2, L3, and L4 can be set as needed. For example, L1 can be set to 0.5, L2 can be set to 0.2, L3 can be set to 0.2, and L4 can be set to 0.1.

[0120] In another way, the calculation method of the total score of the crystal image is:

[0121] Use to calculate the total score. Characterize the joint contribution of similarity and peak signal-to-noise ratio; 1 + FID represents the penalty term of the distribution difference. The greater the difference, the larger the denominator and the lower the overall score. and are preset constants. If similarity and peak signal-to-noise ratio are more important for the total score, can be set; if the morphological matching degree is more important for the total score, then .

[0122] In another way, the calculation method of the total score of the crystal image is:

[0123] ; among them, and are preset constants. is an exponential function, representing the exponential operation with the natural constant e (approximately equal to 2.71828) as the base.

[0124] In this embodiment, after obtaining the total score of each crystal image, all crystal images are classified according to the crystal form type to obtain all crystal images corresponding to each crystal form category.

[0125] For each crystal form category, screen the crystal image with the highest total score, and the raw material ratio parameter corresponding to the crystal image with the highest total score is the optimal raw material ratio parameter of this crystal form category.

[0126] In another embodiment, after obtaining the similarity, signal-to-noise ratio, distribution difference, and morphological matching degree corresponding to each crystal image, crystal images that meet preset conditions are searched for. The preset conditions include: the similarity is greater than a first preset value (for example, 0.85), the peak signal-to-noise ratio is greater than a second preset value (for example, 30 dB), the distribution difference is less than a third preset value (for example, 10), and the morphological matching degree is greater than a fourth preset value. Calculate the total score of each crystal image that meets the preset conditions; according to the total scores of each crystal image that meets the preset conditions, obtain the optimal raw material ratio parameters corresponding to each crystal form category. Screening the crystal images can reduce the amount of data calculation and improve efficiency.

[0127] As Figure 3 shown, in a possible implementation manner, the method for generating training samples in step S101 may include:

[0128] S201, obtain initial images of multiple calcium carbonate crystals, where each calcium carbonate crystal corresponds to one initial image, and different calcium carbonate crystals are generated according to different first raw material ratio parameters.

[0129] In this embodiment, the raw materials configured according to a set of first raw material ratio parameters are put into a reactor to generate calcium carbonate crystals. Take pictures of each generated calcium carbonate crystal to obtain the initial image of the calcium carbonate crystal generated this time. For example, use a scanning electron microscope to take the initial image of the calcium carbonate crystal. Repeat this cycle to obtain multiple initial images for different first raw material ratio parameters.

[0130] After obtaining the initial images, the initial images can be preprocessed to obtain the processed initial images for subsequent clustering of the images. The preprocessing may include image size normalization and denoising processing, etc. The denoising processing can use an adaptive Gaussian filtering algorithm to remove the noise in the initial images.

[0131] S202, perform clustering processing on the multiple initial images to obtain the crystal form categories of the calcium carbonate crystals in each initial image.

[0132] In this embodiment, feature extraction is performed on each initial image to obtain the feature vector of each initial image. For example, input the initial image into a trained convolutional neural network for feature extraction to obtain the feature vector of the initial image.

[0133] Based on the feature vectors of each initial image, clustering processing is performed to obtain the crystal form categories of each initial image. When performing clustering processing on the initial images, the Euclidean distance can be used to calculate the similarity between the feature vectors, and the category to which the initial image belongs is determined through the similarity between the feature vectors.

[0134] S203. According to the crystal form category of each calcium carbonate crystal, label each group of the first raw material ratio parameters and each initial image to obtain the first raw material ratio parameters and initial images labeled with crystal form categories. The first raw material ratio parameters labeled with crystal form categories and the initial images labeled with crystal form categories form training samples.

[0135] In this embodiment, each sample in the training samples includes the first raw material ratio parameters labeled with crystal form categories and the corresponding initial images labeled with crystal form categories.

[0136] In this application, by performing clustering processing on the initial images of the prepared calcium carbonate crystals, the crystal form category of the calcium carbonate crystals in each initial image can be obtained. Then, by automatically labeling the initial images and the first raw material ratio parameters, accurate training samples can be obtained, and at the same time, the time and cost of manual labeling can be saved.

[0137] After introducing the production method of calcium carbonate, the method for extracting calcium ions when generating calcium carbonate crystals will be introduced below.

[0138] Specifically, as Figure 4 shown, the method for extracting calcium ions may include:

[0139] S301. Obtain solid waste, wherein calcium ions exist in the solid waste.

[0140] In this embodiment, the solid waste may be phosphogypsum, power plant desulfurized gypsum, papermaking lime mud, recycled fine powder, etc.

[0141] S302. Generate multiple groups of initial extraction conditions for extracting calcium ions from the solid waste to obtain an initial population. The initial extraction conditions include temperature, extraction time, extractant, and solid-liquid ratio.

[0142] In this embodiment, the ranges of temperature, extraction time, types of extractants, and solid-liquid ratio are preset. Randomly select temperature values, extraction times, extractants, and solid-liquid ratios from the range of temperature (20 - 80 °C), the range of extraction time (1 - 12 hours), the types of extractants, and the range of solid-liquid ratio (1:1 - 10:1) to generate initial extraction conditions.

[0143] The extractant may be dilute hydrochloric acid (0.1 - 1 M), pure water, ammonium chloride, etc.

[0144] S303. For each group of the initial extraction conditions, based on the initial extraction conditions, extract the calcium ions from the solid waste to obtain the performance parameters of the calcium ion extraction process under the initial extraction conditions.

[0145] In this embodiment, the performance parameters may include extraction rate, energy consumption, and waste liquid volume.

[0146] S304. Calculate the fitness value of each group of the initial extraction conditions based on the performance parameters.

[0147] In this embodiment, by calculate the fitness value, where F(X) is the fitness value, E is the extraction rate, C is the energy consumption, and W is the waste liquid volume; the maximum extraction rate among all extraction rates, is the minimum energy consumption among all energy consumptions, is the minimum waste liquid volume among all waste liquid volumes. R1, R2, and R3 are all weight values.

[0148] In another way, , is a preset parameter.

[0149] In another way, ; where and are preset parameters.

[0150] In one embodiment, R1, R2, and R3 are weight values preset according to needs in advance. R1 can be 0.6, R2 can be 0.3, and R3 can be 0.1.

[0151] In another embodiment, R1, R2, and R3 can also be dynamically adjusted according to the extraction rate so that the calculated fitness is more accurate.

[0152] Specifically, for each group of the initial extraction conditions, determine the weight of each performance parameter according to the extraction rate corresponding to the initial extraction condition, and preset the weights of the performance parameters corresponding to different extraction rates; calculate the fitness of each group of the initial extraction conditions based on the weight of each performance parameter and each performance parameter.

[0153] As an example, if the extraction rate is less than 85%, then R1 = 0.8, R2 = 0.1, R3 = 0.1. If the extraction rate is greater than or equal to 85%, then R1 = 0.5, R2 = 0.3, R3 = 0.2.

[0154] S305. Screen genetic individuals from the initial population based on the fitness value of each group of the initial extraction conditions, construct a new population according to the genetic individuals, and obtain the i-th candidate population, where 1 ≤ i ≤ M and M is the maximum number of iterations.

[0155] In this embodiment, sort the fitness values from high to low, and select the initial extraction conditions corresponding to a preset number of fitness values as genetic individuals from high to low. Perform crossover and mutation on the genetic individuals to obtain a new generation population, denoted as the candidate population.

[0156] S306, perform cyclic iteration until the preset maximum number of iterations is reached, and screen genetic individuals from the Mth candidate population to obtain the optimal extraction conditions.

[0157] In this embodiment, the optimal extraction conditions are used to extract the calcium ions from the solid waste, and the calcium ions are used to prepare calcium carbonate crystals when preparing cement-based materials.

[0158] In this application, extracting calcium ions from solid waste can achieve the purpose of environmental protection. According to the initial extraction conditions, the optimal extraction conditions of calcium ions are obtained by means of cyclic iteration, which is convenient for extracting high-quality calcium ions when preparing cement-based materials.

[0159] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution is prior or posterior. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of this application.

[0160] Corresponding to the method for preparing a green cement-based material based on crystal form regulation described in the above embodiments, Figure 5 The structural block diagram of the device for preparing a green cement-based material based on crystal form regulation provided by the embodiments of this application is shown. For the convenience of description, only the parts related to the embodiments of this application are shown.

[0161] Refer to Figure 5 , the device 400 may include: a model training module 410, a prediction module 420, and a parameter determination module 430.

[0162] Among them, the model training module 410 is used to input training samples into the crystal prediction model to be trained to train the crystal prediction model, and obtain the trained crystal prediction model; wherein, the training samples are composed of multiple groups of first raw material ratio parameters for generating calcium carbonate crystals, and each group of first raw material ratio parameters is a sample; the crystal prediction model includes a category prediction model and an image prediction model, and the parameter influence degree corresponding to each sample output by the category prediction model is used to train the image prediction model, and the parameter influence degree characterizes the contribution degree of each parameter in the input sample to the result;

[0163] The prediction module 420 is used to input multiple groups of second raw material ratio parameters into the trained crystal prediction model to obtain the crystal form category and crystal image of the calcium carbonate crystals generated by each group of the second raw material ratio parameters;

[0164] The parameter determination module 430 is used to determine the optimal raw material ratio parameters of the calcium carbonate crystals based on the crystal form category and the crystal image, wherein the optimal raw material ratio parameters are used to prepare the calcium carbonate crystals required for the cement-based materials when preparing the cement-based materials.

[0165] In a possible implementation, the model training module 410 may specifically be configured to:

[0166] Obtain the j-th sample from the training samples, where j is a positive integer from 1 to N - 1, and N is the total number of samples in the training samples;

[0167] Input the j-th sample into the category prediction model to train the category prediction model, obtain the parameter influence degree corresponding to the j-th sample, and send the parameter influence degree corresponding to the j-th sample to the image prediction model;

[0168] Input the j-th sample into the image prediction model, and use the parameter influence degree corresponding to the j-th sample and the j-th sample to train the image prediction model;

[0169] Input the (j + 1)-th sample in the training samples into the crystal prediction model to train the crystal prediction model until the maximum number of training times is reached, and obtain the trained crystal prediction model.

[0170] In a possible implementation, the parameter determination module 430 may specifically be configured to:

[0171] Calculate the similarity between the predicted crystal image and a preset standard image, where the standard image corresponding to the crystal image is determined according to the crystal form category of the crystal image;

[0172] Calculate the peak signal-to-noise ratio of the crystal image;

[0173] Calculate the distribution difference between the crystal image and a preset standard image;

[0174] Calculate the morphological matching degree between the crystal image and the preset standard image;

[0175] According to the similarity, the peak signal-to-noise ratio, the distribution difference, and the morphological matching degree, screen out the optimal raw material ratio parameters corresponding to each crystal form category.

[0176] In a possible implementation, the parameter determination module 430 may specifically be configured to:

[0177] According to the similarity, the peak signal-to-noise ratio, the distribution difference, and the morphological matching degree, calculate the total score of the crystal image;

[0178] According to the total score of each crystal image, screen out the optimal raw material ratio parameters corresponding to each crystal form category.

[0179] In a possible implementation, the apparatus 400 further includes:

[0180] An image acquisition module, configured to acquire initial images of multiple calcium carbonate crystals, where each calcium carbonate crystal corresponds to one initial image, and different calcium carbonate crystals are generated according to different first raw material ratio parameters;

[0181] A clustering module, configured to perform clustering processing on multiple initial images to obtain the crystal form categories of the calcium carbonate crystals in each initial image;

[0182] A sample generation module, configured to perform tagging processing on each set of first raw material ratio parameters and each initial image according to the crystal form category of each calcium carbonate crystal, to obtain first raw material ratio parameters with labeled crystal form categories and initial images with labeled crystal form categories, and the first raw material ratio parameters with labeled crystal form categories and the initial images form training samples.

[0183] In a possible implementation manner, the above device 400 may further include:

[0184] A waste acquisition module, configured to acquire solid waste, where calcium ions exist in the solid waste;

[0185] A data generation module, configured to generate multiple sets of initial extraction conditions for extracting calcium ions from the solid waste to obtain an initial population, where the initial extraction conditions include temperature, extraction time, extractant, and solid-liquid ratio;

[0186] A parameter determination module, configured to, for each set of initial extraction conditions, extract the calcium ions from the solid waste based on the initial extraction conditions to obtain performance parameters of the calcium ion extraction process under the initial extraction conditions;

[0187] A fitness determination module, configured to calculate the fitness value of each set of initial extraction conditions based on the performance parameters;

[0188] A population update module, configured to, based on the fitness value of each set of initial extraction conditions, screen genetic individuals from the initial population, construct a new population according to the genetic individuals to obtain the i-th candidate population, 1 ≤ i ≤ M, where M is the maximum number of iterations;

[0189] A result output module, configured to perform iterative loop until the preset maximum number of iterations is reached, screen genetic individuals from the M-th candidate population to obtain the optimal extraction conditions, where the optimal extraction conditions are used to extract the calcium ions from the solid waste, and the calcium ions are used to prepare calcium carbonate crystals when preparing cement-based materials, and the extraction conditions corresponding to the genetic individuals screened from the M-th candidate population are the optimal extraction conditions.

[0190] In a possible implementation manner, the fitness determination module may specifically be configured to:

[0191] For each set of the initial extraction conditions, determine the weight of each performance parameter according to the extraction rate corresponding to the initial extraction conditions, and preset the weights of the performance parameters corresponding to different extraction rates.

[0192] Calculate the fitness of each set of the initial extraction conditions based on the weight of each performance parameter and each performance parameter.

[0193] It should be noted that for the information interaction, execution process, etc. between the above-mentioned device / units, since they are based on the same concept as the method embodiments of the present application, their specific functions and the technical effects brought thereby can be specifically referred to the method embodiment part, and will not be elaborated here.

[0194] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the above-mentioned division of each functional unit and module is used as an example for illustration. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiments can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction and do not limit the protection scope of the present application. The specific working processes of the units and modules in the above system can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated here.

[0195] The embodiments of the present application also provide a terminal device. Refer to Figure 6 , the terminal device 500 may include: at least one processor 510, a memory 520, and a computer program stored in the memory 520 and executable on the at least one processor 510. When the processor 510 executes the computer program, it implements the steps in any of the above method embodiments, such as Figure 1 the steps S101 to S104 in the illustrated embodiment. Or, when the processor 510 executes the computer program, it implements the functions of each module / unit in the above device embodiments, such as Figure 5 the functions of the model training module 410 to the parameter determination module 430 shown.

[0196] Exemplarily, a computer program may be divided into one or more modules / units. One or more modules / units are stored in the memory 520 and executed by the processor 510 to implement this application. The one or more modules / units may be a series of computer program segments capable of performing specific functions, and these program segments are used to describe the execution process of the computer program in the terminal device 500.

[0197] Those skilled in the art can understand that Figure 6 merely examples of terminal devices, which do not constitute a limitation on terminal devices, may include more or fewer components than those shown in the figure, or combine certain components, or different components, such as input / output devices, network access devices, buses, etc.

[0198] The processor 510 may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0199] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience in representation, the buses in the drawings of this application are not limited to only one bus or one type of bus.

[0200] The method for preparing a green cement-based material based on crystal form regulation provided by the embodiments of this application can be applied to terminal devices such as computers, tablet computers, laptop computers, netbooks, personal digital assistants (PDAs), etc. The embodiments of this application do not impose any restrictions on the specific types of terminal devices.

[0201] In the above embodiments, the descriptions of the respective embodiments have their own emphases. For parts not detailed or recorded in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0202] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional technician can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0203] In the embodiments provided in this application, it should be understood that the disclosed terminal devices, apparatuses, and methods can be implemented in other ways. For example, the terminal device embodiments described above are merely illustrative. For example, the division of the modules or units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0204] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0205] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0206] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above-mentioned method embodiments of this application, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by one or more processors, the steps of the above-mentioned method embodiments can be implemented.

[0207] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above-described embodiment methods of this application, it can also be completed by a computer program instructing relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by one or more processors, the steps of the above-described various method embodiments can be implemented.

[0208] Similarly, as a computer program product, when the computer program product runs on a terminal device, it enables the terminal device to implement the steps in the above-described various method embodiments when executed.

[0209] Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device that can carry the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0210] The above-described embodiments are only used to illustrate the technical solutions of this application, rather than to limit it; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A preparation method of a green cement-based material based on crystal form regulation, characterized in that, Including: Inputting training samples into a crystal prediction model to be trained to train the crystal prediction model, and obtaining a trained crystal prediction model; wherein, the training samples are composed of multiple groups of first raw material ratio parameters for generating calcium carbonate crystals, and each group of first raw material ratio parameters is a sample; the crystal prediction model includes a category prediction model and an image prediction model, and the parameter influence degree corresponding to each sample output in the category prediction model is used to train the image prediction model, and the parameter influence degree characterizes the contribution degree of each parameter in the input sample to the result; Inputting multiple groups of second raw material ratio parameters into the trained crystal prediction model to obtain the crystal form category and crystal image of calcium carbonate crystals generated by each group of the second raw material ratio parameters; Based on the crystal form category and the crystal image, determining the optimal raw material ratio parameters of the calcium carbonate crystals; Preparing calcium carbonate crystals according to the optimal raw material ratio parameters, and using the prepared calcium carbonate crystals to prepare a cement-based material; The determining the optimal raw material ratio parameters of the calcium carbonate crystals based on the crystal form category and the crystal image includes: Calculating the similarity between the predicted crystal image and a preset standard image, and the standard image corresponding to the crystal image is determined according to the crystal form category of the crystal image; Calculating the peak signal-to-noise ratio of the crystal image; Calculating the distribution difference between the crystal image and a preset standard image; Calculating the morphological matching degree between the crystal image and a preset standard image; According to the similarity, the peak signal-to-noise ratio, the distribution difference and the morphological matching degree, screening out the optimal raw material ratio parameters corresponding to each crystal form category.

2. The preparation method of the green cement-based material based on crystal form regulation according to claim 1, wherein, The first raw material ratio parameters include calcium ion concentration, crystal form regulator concentration and crystal form regulator type; The inputting training samples into a crystal prediction model to be trained to train the crystal prediction model and obtaining a trained crystal prediction model includes: Obtaining the j-th sample from the training samples, where j is a positive integer from 1 to N-1, and N is the total number of samples in the training samples; Inputting the j-th sample into the category prediction model to train the category prediction model, obtaining the parameter influence degree corresponding to the j-th sample, and sending the parameter influence degree corresponding to the j-th sample to the image prediction model; Inputting the j-th sample into the image prediction model, and using the parameter influence degree corresponding to the j-th sample and the j-th sample to train the image prediction model; Inputting the (j + 1)-th sample in the training samples into the crystal prediction model to train the crystal prediction model until the maximum number of training times is reached, and obtaining a trained crystal prediction model.

3. The preparation method of the green cement-based material based on crystal form regulation according to claim 1, wherein, The screening out the optimal raw material ratio parameters corresponding to each crystal form category according to the similarity, the peak signal-to-noise ratio, the distribution difference and the morphological matching degree includes: Calculating the total score of the crystal image according to the similarity, the peak signal-to-noise ratio, the distribution difference and the morphological matching degree; According to the total score of each crystal image, screening out the optimal raw material ratio parameters corresponding to each crystal form category.

4. The preparation method of the green cement-based material based on crystal form regulation according to any one of claims 1 to 3, characterized in that, The method further includes: Obtaining initial images of a plurality of calcium carbonate crystals, wherein each of the calcium carbonate crystals corresponds to an initial image, and different calcium carbonate crystals are generated according to different first raw material ratio parameters; Performing clustering processing on the plurality of initial images to obtain the crystal form categories of the calcium carbonate crystals in each of the initial images; According to the crystal form category of each calcium carbonate crystal, performing tagging processing on each group of the first raw material ratio parameters and each initial image to obtain the first raw material ratio parameters with the crystal form category marked and the initial images with the crystal form category marked. The first raw material ratio parameters and the initial images with the crystal form category marked form training samples.

5. The preparation method of the green cement-based material based on crystal form regulation according to claim 1, characterized in that The method further includes: Obtaining solid waste, wherein calcium ions exist in the solid waste; Generating multiple groups of initial extraction conditions for extracting calcium ions from the solid waste to obtain an initial population, where the initial extraction conditions include temperature, extraction time, extractant, and solid-liquid ratio; For each group of the initial extraction conditions, based on the initial extraction conditions, extracting the calcium ions from the solid waste to obtain the performance parameters of the calcium ion extraction process under the initial extraction conditions; Based on the performance parameters, calculating the fitness value of each group of the initial extraction conditions; Based on the fitness value of each group of the initial extraction conditions, screening genetic individuals from the initial population, and constructing a new population according to the genetic individuals to obtain the i-th candidate population, where 1 ≤ i ≤ M and M is the maximum number of iterations; Performing iterative loop until the preset maximum number of iterations is reached, screening genetic individuals from the M-th candidate population to obtain the optimal extraction conditions, where the optimal extraction conditions are used to extract the calcium ions from the solid waste, and the calcium ions are used to prepare calcium carbonate crystals when preparing a cement-based material. The extraction conditions corresponding to the genetic individuals screened from the M-th candidate population are the optimal extraction conditions.

6. The preparation method of the green cement-based material based on crystal form regulation according to claim 5, characterized in that The performance parameters include extraction rate, energy consumption, and waste liquid volume.

7. The preparation method of the green cement-based material based on crystal form regulation according to claim 6, wherein, The calculating the fitness value of each group of the initial extraction conditions based on the performance parameters includes: For each group of the initial extraction conditions, determining the weight of each performance parameter according to the extraction rate corresponding to the initial extraction conditions, and presetting the weights of the performance parameters corresponding to different extraction rates; Based on the weight of each performance parameter and each performance parameter, calculating the fitness of each group of the initial extraction conditions.

8. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for preparing a green cement-based material based on crystal form regulation according to any one of claims 1 to 7.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for preparing a green cement-based material based on crystal form regulation according to any one of claims 1 to 7.

Images