Artificial intelligence generation method and system for ferroelectric material component-temperature phase diagram

Abstract

The invention relates to an artificial intelligence generation method and system for a ferroelectric material component-temperature phase diagram, and the method comprises the steps: obtaining a chemical formula of a ferroelectric material with a doping variable, and setting the chemical formula in a character string form; setting a doping proportion, a temperature change range and a step length; chemical elements and doping proportions in the chemical formula of the ferroelectric material with the doping variable in the character string form are processed, and a chemical formula numerical vector list is obtained; forming a temperature list according to the set temperature change range and step length, and performing normalization processing to obtain a normalized temperature list; based on the chemical formula numerical vector list and the normalized temperature list, merging, inputting into a pre-trained crystal phase transformation artificial intelligence model, and outputting a crystal structure type; based on the type of the crystal structure, drawing a component-temperature phase diagram of the chemical formula of the ferroelectric material with the doping variable under different doping proportions and different temperature conditions; compared with the prior art, the method has the advantages of automatically generating the component-temperature phase diagram and the like.

Description

Technical Field

[0001] The present invention relates to the field of new material technology, and in particular to an artificial intelligence generation method and system for a ferroelectric material component-temperature phase diagram. Background Art

[0002] With the rapid development of cutting-edge fields such as new generation information technology, new energy, and integrated circuits, the demand for electronic functional materials has increased dramatically. Ferroelectric materials, as basic electronic functional materials, are widely used in piezoelectric transducers, infrared sensors, energy conversion, and other fields. Traditional typical ferroelectric materials such as barium titanate and lead titanate can no longer meet the rich requirements of material performance in cutting-edge fields. Therefore, finding a feasible solution that can accurately and quickly develop new high-performance ferroelectric materials is a key technology that needs to be solved urgently. Researchers have found that in the component-temperature phase diagram of ferroelectric materials, ferroelectric materials at the intersection of multiple crystal structures, that is, at the phase boundary, have a variety of excellent electrical properties. Taking calcium and zirconium co-doped barium titanate materials as an example, at room temperature, the doping ratio at the cubic, tetragonal, and trigonal phase interfaces generally improves its piezoelectric coefficient, dielectric constant, and other electrical properties by more than one order of magnitude compared to pure barium titanate materials. Designing doped ferroelectric materials based on the phase interface of multiple crystal structures has become a guiding principle for the rapid development of ferroelectric materials in this field. Therefore, by accurately and quickly constructing the component-temperature phase diagram of ferroelectric materials, it is possible to directly determine the interface between multiple phases of ferroelectric materials and directly synthesize ferroelectric materials with high electrical properties. However, in the traditional method of constructing component-temperature phase diagrams, it is common to design experiments, synthesize a large number of materials with different doping ratios, characterize the crystal structure and phase transition temperature of this series of materials one by one, and then draw the component-temperature phase diagram of the doped ferroelectric material, and find the multi-phase interface based on it. For ferroelectric materials with a wide variety of doping ratios, traditional methods cannot achieve rapid phase diagram construction and phase interface search.

[0003] Patent application CN115410654A discloses a machine learning prediction method for the ferroelectric-paraelectric phase transition temperature of a molecular ferroelectric, which collects the chemical formula of the molecular ferroelectric and the experimental value of the ferroelectric-paraelectric phase transition temperature from the literature; uses the literature, first principle calculation, and topological structure to calculate and obtain characteristic parameters for describing sample characteristics; quickly reduces the dimension of the feature, and then embeds the learner in combination with the subset search strategy to obtain the subset of the best feature; takes the ferroelectric-paraelectric phase transition temperature as the target variable and the best feature as the independent variable, and uses the support vector machine regression algorithm to establish a quantitative prediction model for the ferroelectric-paraelectric phase transition temperature of the molecular ferroelectric; collects new molecular ferroelectric samples, obtains the characteristic parameters of the new samples, and predicts the ferroelectric-paraelectric phase transition temperature of the new molecular ferroelectric samples according to the quantitative prediction model. The input data of this method requires information such as first principle calculation and topological structure, and also needs to obtain the characteristic parameters of the sample characteristics. In addition, this method does not involve the prediction of the crystal structure type, nor does it involve the generation of the component-temperature phase diagram of the ferroelectric material, nor does it point out the automatic generation of the component-temperature phase diagram of the ferroelectric material. Summary of the invention

[0004] The purpose of the present invention is to provide an artificial intelligence generation method and system for a ferroelectric material component-temperature phase diagram that realizes automatic generation of the component-temperature phase diagram.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] An artificial intelligence generation method for a ferroelectric material component-temperature phase diagram comprises the following steps:

[0007] Get the chemical formula of the ferroelectric material with doping variables and set it in a string form, wherein the chemical formula of the ferroelectric material with doping variables includes chemical elements and corresponding doping ratios, and the doping variables are included in the doping ratio;

[0008] Based on the chemical formula of the ferroelectric material with doping variables in string form, set the range and step size of the doping ratio and temperature;

[0009] According to the set variation range and step size of the doping ratio, the chemical elements and doping ratios in the chemical formula of the ferroelectric material with doping variables in the string form are processed to obtain a list of chemical formula numerical vectors;

[0010] According to the set temperature variation range and step length, a temperature list is formed and normalized to obtain a normalized temperature list;

[0011] Based on the chemical formula numerical vector list and the normalized temperature list, they are merged and input into a pre-trained crystal phase transition artificial intelligence model as input data to output the crystal structure type;

[0012] Based on the crystal structure type, a component-temperature phase diagram of the chemical formula of the ferroelectric material with doping variables under different doping ratios and different temperature conditions is drawn.

[0013] Furthermore, the chemical formula of the ferroelectric material with doping variables in the string form uses symbols to distinguish chemical elements and doping ratios, specifically:

[0014] Only the chemical element is placed in the symbol and the doping ratio is placed on the side of the corresponding chemical element, or both the chemical element and the doping ratio are placed in different symbols, where the doping variable is included in the doping ratio.

[0015] Furthermore, the step of obtaining the chemical formula numerical vector list includes:

[0016] Based on the chemical formula of the ferroelectric material with doping variables in the string form, chemical formulas with different doping ratios are generated in batches according to the set range and step size of the doping ratio to form a list of chemical formulas with different doping ratios, wherein the number of chemical formulas generated in batches is:

[0017]

[0018] Where N1 is the number of chemical formulas generated in batches, R1 and R2 are the upper and lower limits of the doping ratio, respectively, and ΔR is the step size;

[0019] Based on the chemical formula list of different doping ratios, the doping ratio is normalized using a logarithmic normalization method to obtain a logarithmic normalized doping ratio:

[0020] The chemical elements and logarithmically normalized doping ratios of the chemical formulas in the chemical formula list are embedded using the periodic table to obtain an initial list of chemical formula numerical vectors, where the embedded expression is:

[0021]

[0022] In the formula, v[j] is the numerical vector of the jth initialization formula, m′ i is the logarithm-normalized doping ratio of the ith chemical element, E j is the jth chemical formula containing n elements, ε i is a list of elements in the periodic table;

[0023] Based on the initial chemical formula numerical vector list, a principal component analysis algorithm is used to perform dimensionality reduction to obtain a final chemical formula numerical vector list, wherein the dimensionality reduction is expressed as:

[0024] z p =v j U k

[0025] In the formula, z p is the numerical vector of the chemical formula after dimension reduction, v j is the jth initialization formula numerical vector, U k is a low-dimensional matrix.

[0026] Furthermore, the step of obtaining a logarithmically normalized doping ratio comprises:

[0027] Based on the chemical formula list of different doping ratios, the doping ratio is normalized, and the normalized expression is:

[0028]

[0029] In the formula, m * i is the normalized doping ratio coefficient of the ith chemical element, m i is the doping ratio of the i-th chemical element;

[0030] The normalized doping ratio is logarithmically transformed to obtain a logarithmically normalized doping ratio, and the expression of the logarithmically normalized doping ratio is:

[0031] m′ i =log(1+m * i )

[0032] In the formula, m′ i is the logarithm-normalized doping ratio of the ith chemical element.

[0033] Furthermore, the step of obtaining a normalized temperature list includes:

[0034] According to the set temperature range and step size, each temperature point is generated in batches to form a temperature list, where each temperature point represents the total number of calculations of each chemical formula numerical vector in the chemical formula numerical vector list. The calculation expression is:

[0035]

[0036] Where N2 is the total number of calculations of each chemical formula numerical vector at different temperature points, T1 and T2 are the upper and lower limits of the temperature, respectively, and ΔT is the step size;

[0037] Based on the temperature list, the maximum and minimum normalization algorithm is used for normalization processing to obtain a normalized temperature list, wherein the expression of the maximum and minimum normalization algorithm is:

[0038]

[0039] In the formula, T′ iis the ith normalized temperature point, T i is the i-th original temperature, T min , T max are the minimum and maximum temperature respectively.

[0040] Furthermore, the crystal phase transition artificial intelligence model includes an input layer, a hidden layer and an output layer. The input layer is used to input input data obtained by merging the chemical formula numerical vector list and the normalized temperature list, the hidden layer is used to extract deep-level implicit features of the input data, and the output layer is used to perform classification prediction based on the extracted deep-level implicit features, and output the predicted crystal structure type.

[0041] Furthermore, the hidden layer is provided with multiple layers, each hidden layer is added with a batch normalization operation and a discarding operation, and the execution steps of the hidden layer include:

[0042] In the first hidden layer, the input data is processed to obtain the initial output vector h of the first hidden layer. 1 :

[0043] h 1 =ReLU(W (1) Q+b (1) )

[0044] Where W (1) is the weight matrix of the first hidden layer, Q is the input vector, b (1) is the bias vector of the first hidden layer;

[0045] The initial output vector h of the first layer 1 Perform batch normalization to obtain the initial output vector after batch normalization

[0046]

[0047] Where γ, β and ∈ are training parameters, μ batch and are the mean and variance of the current batch;

[0048] Based on the initial output vector after batch normalization The discard operation is performed to obtain the final output vector of the first hidden layer, and it is used as the input of the next hidden layer. The above steps are repeated until the processing of all hidden layers is completed.

[0049] Furthermore, the final output vector of the first hidden layer is expressed as:

[0050]

[0051] In the formula, is the final output vector of the first hidden layer, and m is the mask.

[0052] Furthermore, the step of outputting the crystal structure type comprises:

[0053] Based on the chemical formula numerical vector list and the normalized temperature list, the total number of executions of the pre-trained crystal phase transition artificial intelligence model is calculated, and the calculation expression of the total number of executions is:

[0054] N=N1×N2

[0055] Where, N is the total number of executions, N1 is the number of chemical formula numerical vectors in the chemical formula numerical vector list, that is, the number of chemical formulas with different doping ratios, and N2 is the number of temperature points in the normalized temperature list, that is, the total number of calculations of each chemical formula numerical vector at different temperature points;

[0056] The chemical formula numerical vectors in the chemical formula numerical vector list and the normalized temperatures in the normalized temperature list are merged as input data for one execution, and the pre-trained crystal phase transition artificial intelligence model is used for processing, and the crystal structure type of this time is output. This step is repeated N-1 times to obtain the crystal structure type of all chemical formula numerical vectors at each temperature point.

[0057] The present invention also provides an artificial intelligence generation system for a ferroelectric material component-temperature phase diagram, comprising:

[0058] Pre-trained artificial intelligence module: used to obtain the crystal phase transition data set of electrical materials, train the multi-layer artificial neural network model, perform hyperparameter optimization and cross-validation during the training process, and obtain a pre-trained crystal phase transition artificial intelligence model;

[0059] Phase diagram automatic generation module: used to read the chemical formula of ferroelectric materials with doping variables, the range of doping ratio and temperature, the step size, and perform standardization processing, cyclically call the pre-trained crystal phase transition artificial intelligence model for processing, output the crystal structure type, and draw the component-temperature phase diagram of the chemical formula of the ferroelectric material with doping variables under different doping ratios and different temperature conditions according to the crystal structure type, wherein the chemical formula of the ferroelectric material with doping variables includes chemical elements and corresponding doping ratios, and the doping variables are included in the doping ratio.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] (1) The present invention sets the doping ratio and temperature of the chemical formula of the ferroelectric material with doping variables to generate chemical formulas with different doping ratios, and uses a crystal phase transition artificial intelligence model to predict the crystal structure type by combining with different temperature points, so as to draw a component-temperature phase diagram under different doping ratios and different temperature conditions. The present invention only needs to input the chemical formula of the ferroelectric material, and does not require any other additional information such as material properties, structure, atoms, etc., which not only realizes the automatic generation of the component-temperature phase diagram of the ferroelectric material, but also reduces the difficulty of generation.

[0062] (2) By using a pre-trained crystal phase transition artificial intelligence model, there is no need to perform unnecessary optimization processes such as parameter adjustment on the model when predicting the crystal structure of ferroelectric materials, thereby improving execution efficiency.

[0063] (3) Compared with the traditional experimental method for constructing a component-temperature phase diagram, the automatic generation method using artificial intelligence in the present invention greatly shortens the time consumed in generating a complete component-temperature phase diagram and reduces R&D costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is a flow chart of the artificial intelligence generation method of the present invention;

[0065] Figure 2 A schematic diagram of an artificial intelligence generation system of the present invention;

[0066] Figure 3 A schematic diagram of the data processing flow of the chemical formula of the ferroelectric material of the present invention;

[0067] Figure 4 The accuracy of the pre-trained crystal phase transition artificial intelligence model in the present invention in the training set and the validation set;

[0068] Figure 5 The calcium zirconium doped barium titanate ferroelectric material (1-x)Ba(Zr 0.2 Ti 0.8 )O3-xBa 0.7 Ca 0.3 Composition-temperature phase diagram of TiO3;

[0069] Figure 6 The tin-doped barium titanate ferroelectric material BaSn generated in the present invention x Ti 1-x Composition-temperature phase diagram of O3;

[0070] Figure 7 The lead zirconate titanate ferroelectric material PbZr generated in the present invention 1-x Ti x Composition-temperature phase diagram of O3. DETAILED DESCRIPTION

[0071] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0072] Example 1

[0073] This embodiment provides an artificial intelligence method for generating a ferroelectric material composition-temperature phase diagram, such as Figure 1 As shown, the method comprises the following steps:

[0074] (1) Natural language processing methods are used to obtain the crystal phase transition data of ferroelectric materials from existing literature, and the collected data are used to construct a ferroelectric material crystal phase transition dataset.

[0075] The ferroelectric material crystal phase transition dataset is a ferroelectric material crystal phase transition information obtained in batches from an accessible scientific research literature library using natural language processing methods or manual acquisition methods. The dataset contains data such as chemical formula, temperature, crystal structure, document identification code, journal name, publication year, and their source information; the crystal phase transition dataset is a linked list data structure, and can also be other data structures such as arrays, trees, and graphs;

[0076] (2) Build a multi-layer artificial neural network intelligent model, train it with the crystal structure transformation data of ferroelectric materials, and obtain the crystal phase transformation artificial intelligence model through cross-validation.

[0077] The Artificial Neural Network module contains a pre-trained multi-layer artificial neural network model trained on a dataset of ferroelectric material crystal phase transitions.

[0078] The multi-layer artificial neural network model consists of one or more input layers, fully connected layers, and output layers. The data received by the input layer is the processed data of the original data. The proportion of chemical elements in the original chemical formula is normalized. For a chemical formula, E d (i=1,2,3…) represents a chemical formula containing n elements, with m d (i=1,2,3…) represents the doping ratio of each element, and the normalized doping ratio coefficient m is calculated based on the following formula: * i (i=1,2,3…),

[0079]

[0080] Then, the proportional coefficient is logarithmically transformed according to the following formula to obtain the logarithmically normalized doping ratio m′: i (i=1,2,3…),

[0081] m′d =log(1+m * i ) (2)

[0082] Then, according to the following formula, the elements and proportions in the chemical formula are replaced by the element list ε in the periodic table. i (i=1,2,3…) for embedding representation,

[0083]

[0084] Get the numerical representation of the chemical formula, and convert the original chemical formula into a new vector v.

[0085] The vector v is calculated by the principal component analysis algorithm according to its dimension, and the low-dimensional matrix U k , and reduce it to the specified p-dimensionality using the following formula to obtain the reduced-dimensional input vector z p ;

[0086] z p =v j U k (4)

[0087] The temperature T in the original input data i According to the Min-Max normalization algorithm, the normalized temperature T′ is calculated as shown in the following formula:

[0088]

[0089] The input vector z p Combined with the normalized temperature T′ into a new vector, which constitutes the input vector Q of the first layer of the neural network, with a dimension of d = p + 1;

[0090] Each layer of artificial neural network extracts deep implicit features of input data, and calculates the feature h extracted by neurons in the sth layer through the weight matrix W and the bias vector b according to the following formula;

[0091] h s =ReLU(W (s) Q+b (s) ) (6)

[0092] Preferably, each layer of artificial neural network includes one or more neurons, activation functions, regularizers, batch normalizers, and discarders, which can improve the stability and generalization ability of the model; the hyperparameters in each layer of artificial neural network are grid searched to obtain the hyperparameters with the highest accuracy; the constructed artificial neural network module is a multilayer perceptron model for classification tasks, and can also be a model for other types of tasks; the last layer of artificial neural network uses an activation function to output the predicted category probability.

[0093] The output layer is the Lth layer. The output vector o of the output layer is calculated according to the following formula, and the activation function is used to convert the output vector to the numerical data corresponding to the crystal structure type;

[0094] o=W (L) h L-1 +b (L) (7)

[0095] The activation function used is the Softmax function, the total number of crystal structure types is C, and the probability of the predicted crystal structure numerical type is calculated by the following formula: Take the maximum probability is the crystal structure corresponding to the current input vector;

[0096]

[0097] The accuracy of the artificial neural network is evaluated by loss function, accuracy and other evaluation indicators; cross-validation is performed on the training set and the validation set, and the training set and the validation set are obtained by dividing the crystal phase transformation data set obtained in step (1) according to a certain ratio; the artificial neural network continuously optimizes the weight matrix W and the bias coefficient b as the number of training rounds increases. The accuracy of the crystal phase transformation artificial intelligence model obtained after optimization is greater than 90%.

[0098] (3) Set the chemical formula of the ferroelectric material with doping variables for which the phase diagram is to be generated.

[0099] The data format of the set chemical formula is a string form of one or more elements such as "element 1 [ratio 1] element 2 [ratio 2]", and this method enables the computer program to read the elements and ratios in the chemical formula; this structured string can also be a combination of other symbols, such as using parentheses, commas, periods, semicolons and other symbols to distinguish chemical elements and ratios, or it can be in the form of placing elements in symbols, or in the form of placing elements and ratios in different symbols.

[0100] Doping variable refers to doping a chemical element in the matrix material, and its doping ratio x is used as the doping variable, which is included in the character string of the doping ratio. The character string containing the doping variable can be converted into a computer-recognizable form and can perform mathematical operations.

[0101] (4) The upper and lower limits of the doping ratio are set to E1 and E2, and the step size is ΔR; the upper and lower limits of the temperature change are set to T1 and T2, and the step size is ΔT.

[0102] The range of variation of the doping ratio is the upper and lower limits of the component axis in the pseudo-phase diagram; the range of variation of the specified temperature condition is between 0 and 2000K.

[0103] (5) According to the set doping ratio range and step size, the proportional coefficient in the specified ferroelectric material chemical formula is calculated, and the total number N1 of the chemical formula is calculated using the following formula:

[0104]

[0105] After the calculation is completed, N1 chemical formulas are generated in batches, and the embedded representation method is used to obtain the numerical representation of the chemical formulas and optimize the data dimension.

[0106] The chemical formulas generated in batches are represented by an embedded periodic table of chemical elements. The character strings of chemical elements and proportions are converted into numerical lists to achieve numerical representation of the chemical formulas. A logarithmic normalization algorithm is used to enhance the impact of trace doping elements.

[0107] Preferably, the doping ratio m of each element in the chemical formula is firstly set according to formula (1) i (i=1,2,3…) to obtain the normalized doping ratio coefficient m / i (i=1,2,3…), and then logarithmically transform the proportional coefficient according to formula (2) to obtain the logarithmically normalized doping ratio m′ i (i=1,2,3…);

[0108] The dimensionality of the original data can be reduced by using methods such as principal component analysis and neural network-based encoders. Other algorithms can also be used to reduce the dimensionality of the original data. The processing of optimizing the data dimension can reduce the dimensionality of the original data and improve the feature representation of the data.

[0109] (6) According to the set temperature condition change range and step size, each temperature point is generated in batches to form a temperature list. The total number of times N2 that each chemical formula needs to be calculated at different temperature points is calculated using the following formula:

[0110]

[0111] The temperature is normalized by Min-Max and a normalized temperature list T′ (i=1, 2, 3…) is obtained according to formula (5).

[0112] (7) The numerical representation of the chemical formula after dimensionality reduction and the normalized temperature list are combined into a new numerical list as input data.

[0113] The input data is input one by one into the pre-trained crystal phase transition artificial intelligence model. The model can display the output results one by one or display all the results after execution. The output result of the pre-trained crystal phase transition artificial intelligence model is a numerical representation of the crystal structure type, which is converted into a string of crystal structure through a crystal structure type converter.

[0114] (8) Calculate the total number of times N to call the crystal phase transformation artificial intelligence model according to the following formula:

[0115] N=N1×N2 (11)

[0116] Within the range of N times, the crystal phase transformation artificial intelligence model is called one by one, the input data is read one by one, and the corresponding crystal structure under different doping ratios and temperatures is output.

[0117] (9) After executing N times, based on the output results, draw the component-temperature phase diagram of the current material under different doping ratios and different temperature conditions.

[0118] This embodiment is experimentally verified according to the above method. In the experiment, the crystal phase transition artificial intelligence model is a multi-layer perceptron model for classification tasks. Set 7 fully connected layers, which serve as 1 input layer, 5 hidden layers, and 1 output layer of the data respectively. The data received at the input layer is the processed form of the original data. After reading the chemical formula, the chemical formula is normalized, embedded, and reduced in dimension using the method described in the embodiment of the present invention to obtain a 20-dimensional chemical formula embedding vector and a normalized 1-dimensional temperature vector. In this embodiment, the amount of data contained in the crystal phase transition data set is greater than 10,000. Some chemical formulas in the crystal phase transition data set are processed as follows Figure 3 shown.

[0119] Combining the 20-dimensional chemical formula embedding vector Q and the normalized 1-dimensional temperature vector, the input vector in the input layer is 21-dimensional, in the form of

[0120] Q=[Q1 Q2 …Q 20 temp] T (12)

[0121] In the first hidden layer, the ReLu activation function is specified, so the output vector h of the first layer is 1 for

[0122] h 1 =ReLU(W (1) Q+b (1) ) (13)W (1) is the weight matrix of the first layer, b (1) is the bias vector of the first layer. In this layer of neurons, batch normalization and dropout operations are added to improve the generalization ability of the model. Therefore, the output vector h 1 Scaling and translation are performed by the following formulas,

[0123]

[0124] where μ batch and is the mean and variance of the current batch, γ, β and ∈ are trainable parameters. The output vector after batch normalization is obtained by calculation In the discard operation, the mask m is used to calculate

[0125]

[0126] Get the output vector after batch normalization and discarding operations This vector is also the input data of the next layer of neural network. In the subsequent second, third, fourth, and fifth hidden layers, the input vectors are all calculated using formulas (13) to (15) to calculate the output vectors. In the fifth hidden layer, the output vector is The last layer of neurons is the output layer, which is calculated according to formulas (7) to (8): The output crystal structure is obtained.

[0127] Increase the number of training rounds, divide the training set and validation set for cross-validation, and continuously optimize the weight matrix, bias vector and other hyperparameters in each layer of neurons, and finally obtain a crystal phase transition artificial intelligence model with an accuracy rate greater than 90% on the training set and validation set. Figure 4 It is the accuracy of the crystal phase transformation artificial intelligence model built in this embodiment in the training set and the validation set with the training rounds.

[0128] Set the chemical formula of the ferroelectric material with doping variables for which the phase diagram is to be generated. In the computer program, enter the chemical formulas of the three ferroelectric materials, as shown in Table 1. The three materials are:

[0129] (1) Material No. 1 is calcium zirconium doped barium titanate ferroelectric material:

[0130] Set the variable chemical formula of material 1 to "(1-x)Ba(Zr 0.2 Ti 0.8 )O3-xBa 0.7 Ca 0.3 TiO3", after standardization, the following form is obtained: "Ba[1-0.3x]Zr[0.2-0.2x]Ti[0.8+0.2x]O[3]Ca[0.3x]", where x is a variable for adjusting the doping ratio.

[0131] The upper limit and lower limit of the doping ratio range are set to 0.2 and 0.5 respectively.

[0132] Set the scale change step size to 0.002.

[0133] Set the upper and lower limits of the temperature variation range to 120K and 440K respectively.

[0134] Set the temperature change step size to 2K.

[0135] Table 1 Parameters for automatically generating three different material composition-temperature phase diagrams

[0136]

[0137] The computer program is executed to generate a list of chemical formulas with different doping ratios in batches by reading the chemical formula with doping variables, doping range, and step size information. In this material, 150 chemical formulas with different doping ratios are included. According to the generated chemical formulas, the doping ratios are normalized using the logarithmic normalization method, and then the elements and doping ratios are embedded using the periodic table. The embedded representation of the chemical formula is reduced in dimension using the principal component analysis (PCA) method to obtain a 20-dimensional numerical representation of the chemical formula.

[0138] Generate a temperature list and normalize it using the Min-Max normalization method.

[0139] The numerical representation of the 20-dimensional chemical formula and the temperature list after Min-Max normalization are combined as input data.

[0140] According to formulas (9) to (11), the total number of times the pre-trained crystal phase transformation artificial intelligence model needs to be called is 24,000. For each row of data in the input data, the crystal phase transformation artificial intelligence model is called one by one to output the crystal structure at different doping ratios and temperatures. After executing 24,000 times, the processing of all input data is completed.

[0141] According to the output results, the composition-temperature phase diagram of calcium zirconium doped barium titanate ferroelectric material is drawn when the doping ratio varies from 0.2 to 0.5 and 120 to 440 K. The results are as follows Figure 5 Compared with the phase diagram constructed by Acosta, Matias (2014) and others through experiments, it can be seen that in the phase diagram automatically generated by the embodiment of the present invention, the phase boundary change trend is consistent.

[0142] (2) Material No. 2 is tin-doped barium titanate material:

[0143] Set the chemical formula of material 2 to "BaSn x Ti 1-x O3", after standardization, the chemical formula is converted into the following form "Ba[1]Sn[x]Ti[1-x]O[3]", where x is a variable for adjusting the doping ratio.

[0144] The upper limit and lower limit of the doping ratio range are set to 0 and 0.3 respectively.

[0145] Set the scale change step size to 0.002.

[0146] Set the upper and lower limits of the temperature variation range to 150K and 440K respectively.

[0147] Set the temperature change step size to 2K.

[0148] Using the same processing method as material No. 1, according to formulas (9) to (11), the total number of times the pre-trained crystal phase transition artificial intelligence model needs to be called is calculated to be 21750 times. After each execution, the composition-temperature phase diagram of the tin-doped barium titanate ferroelectric material in the doping ratio range of 0 to 0.3 and 150 to 440K is drawn, and the results are as follows Figure 6 The phase boundary change trend in this phase diagram is consistent with the experimental results of Liu, Wenfeng (2017).

[0149] (3) Material No. 3 is lead zirconate titanate material:

[0150] Set the chemical formula of material 3 to "PbZr 1-x Ti x O3", after standardization, the chemical formula is converted into the following form: "Pb[1]Zr[1-x]Ti[x]O[3]", where x is a variable for adjusting the doping ratio.

[0151] The upper limit and lower limit of the doping ratio range are set to 0.35 and 0.46 respectively.

[0152] Set the scale change step size to 0.001.

[0153] The upper and lower limits of the temperature variation range are set to 500K and 690K respectively.

[0154] Set the temperature change step size to 2K.

[0155] Using the same processing method as material No. 1, according to formulas (9) to (11), it is calculated that the total number of times the pre-trained crystal phase transformation artificial intelligence model needs to be called is 10450 times. After each execution, the composition-temperature phase diagram of the material in the doping ratio range of 0.35-0.46 and 500-690K is drawn, as shown in Figure 7 The obtained phase boundary variation trend is consistent with the experimental results published by A. Bouzid et al. (2003).

[0156] In this embodiment, the intelligent generation method of the ferroelectric material component-temperature phase diagram is used to automatically generate the component-temperature phase diagrams of three different materials. Table 1 also lists that in this embodiment, the time required to construct the component-temperature phase diagrams of the three different materials is less than 1000 seconds. Compared with the time consumed by traditional experimental methods (usually 1 to 2 years), the above method greatly shortens the research and development time and improves the research and development efficiency. The use of high-performance graphics cards and processors can further shorten the execution time of computer programs.

[0157] Example 2

[0158] This embodiment provides an artificial intelligence generation system for a ferroelectric material composition-temperature phase diagram, such as Figure 2 As shown, including:

[0159] Pre-trained artificial intelligence module: used to obtain the crystal phase transition data set of electrical materials, train the multi-layer artificial neural network model, perform hyperparameter optimization and cross-validation during the training process, and obtain a pre-trained crystal phase transition artificial intelligence model;

[0160] Phase diagram automatic generation module: used to read the chemical formula of ferroelectric materials with doping variables, the range of doping ratio and temperature, the step size, and perform standardization processing, cyclically call the pre-trained crystal phase transition artificial intelligence model for processing, output the crystal structure type, and draw the component-temperature phase diagram of the chemical formula of the ferroelectric material with doping variables under different doping ratios and different temperature conditions according to the crystal structure type, wherein the chemical formula of the ferroelectric material with doping variables includes chemical elements and corresponding doping ratios, and the doping variables are included in the doping ratio.

[0161] The rest is the same as in Example 1.

[0162] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program code.

[0163] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0164] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0165] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0166] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0167] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0168] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An artificial intelligence method for generating a ferroelectric material composition-temperature phase diagram, characterized in that: The following steps are involved: Obtaining a chemical formula of a ferroelectric material with a doping variable and setting it in a string form, wherein the chemical formula of the ferroelectric material with a doping variable includes a chemical element and a corresponding doping ratio, and the doping variable is included in the doping ratio; Based on the chemical formula of the ferroelectric material with doping variables in string form, set the range and step size of the doping ratio and temperature; According to the set variation range and step size of the doping ratio, the chemical elements and doping ratios in the chemical formula of the ferroelectric material with doping variables in the string form are processed to obtain a list of chemical formula numerical vectors; According to the set temperature variation range and step length, a temperature list is formed and normalized to obtain a normalized temperature list; Based on the chemical formula numerical vector list and the normalized temperature list, they are merged and input into a pre-trained crystal phase transition artificial intelligence model as input data to output the crystal structure type; Based on the crystal structure type, a component-temperature phase diagram of the chemical formula of the ferroelectric material with doping variables under different doping ratios and different temperature conditions is drawn.

2. The artificial intelligence generation method of a ferroelectric material composition-temperature phase diagram according to claim 1, characterized in that: The chemical formula of the ferroelectric material with doping variables in the string form uses symbols to distinguish chemical elements and doping ratios, specifically: Only the chemical element is placed in the symbol and the doping ratio is placed on the side of the corresponding chemical element, or both the chemical element and the doping ratio are placed in different symbols, where the doping variable is included in the doping ratio.

3. The artificial intelligence generation method of a ferroelectric material composition-temperature phase diagram according to claim 1, characterized in that: The step of obtaining the chemical formula numerical vector list comprises: Based on the chemical formula of the ferroelectric material with doping variables in the string form, chemical formulas with different doping ratios are generated in batches according to the set range and step size of the doping ratio to form a list of chemical formulas with different doping ratios, wherein the number of chemical formulas generated in batches is: Where N1 is the number of chemical formulas generated in batches, R1 and R2 are the upper and lower limits of the doping ratio, respectively, and ΔR is the step size; Based on the chemical formula list of different doping ratios, the doping ratio is normalized using a logarithmic normalization method to obtain a logarithmic normalized doping ratio: The chemical elements and logarithmically normalized doping ratios of the chemical formulas in the chemical formula list are embedded using the periodic table to obtain an initial list of chemical formula numerical vectors, where the embedded expression is: Where v[j] is the jth initialization equation numerical vector, m′ i is the logarithm-normalized doping ratio of the ith chemical element, E j is the jth chemical formula containing n elements, ε i is a list of elements in the periodic table; Based on the initial chemical formula numerical vector list, a principal component analysis algorithm is used to perform dimensionality reduction to obtain a final chemical formula numerical vector list, wherein the dimensionality reduction is expressed as: z p =v j U k In the formula, z p is the numerical vector of the chemical formula after dimension reduction, v j is the jth initialization formula numerical vector, U k is a low-dimensional matrix.

4. The artificial intelligence generation method of a ferroelectric material composition-temperature phase diagram according to claim 3, characterized in that: The step of obtaining the logarithmically normalized doping ratio comprises: Based on the chemical formula list of different doping ratios, the doping ratio is normalized, and the normalized expression is: In the formula, m * i is the normalized doping ratio coefficient of the ith chemical element, m i is the doping ratio of the i-th chemical element; The normalized doping ratio is logarithmically transformed to obtain a logarithmically normalized doping ratio, and the expression of the logarithmically normalized doping ratio is: m′ i =log(1+m * i ) Where m′ i is the logarithm-normalized doping ratio of the ith chemical element.

5. The artificial intelligence generation method of a ferroelectric material composition-temperature phase diagram according to claim 1, characterized in that: The step of obtaining a normalized temperature list comprises: According to the set temperature range and step size, each temperature point is generated in batches to form a temperature list, where each temperature point represents the total number of calculations of each chemical formula numerical vector in the chemical formula numerical vector list. The calculation expression is: Where N2 is the total number of calculations of each chemical formula numerical vector at different temperature points, T1 and T2 are the upper and lower limits of the temperature, respectively, and ΔT is the step size; Based on the temperature list, the maximum and minimum normalization algorithm is used for normalization processing to obtain a normalized temperature list, wherein the expression of the maximum and minimum normalization algorithm is: In the formula, T′ i is the ith normalized temperature point, T i is the i-th original temperature, T min 、T max are the minimum and maximum temperature respectively.

6. The artificial intelligence generation method of a ferroelectric material composition-temperature phase diagram according to claim 1, characterized in that: The crystal phase transition artificial intelligence model includes an input layer, a hidden layer and an output layer. The input layer is used to input input data obtained by merging the chemical formula numerical vector list and the normalized temperature list. The hidden layer is used to extract deep-level implicit features of the input data. The output layer is used to perform classification prediction based on the extracted deep-level implicit features and output the predicted crystal structure type.

7. The artificial intelligence generation method of a ferroelectric material composition-temperature phase diagram according to claim 6, characterized in that: The hidden layer is provided with multiple layers, each hidden layer is added with a batch normalization operation and a discarding operation, and the execution steps of the hidden layer include: In the first hidden layer, the input data is processed to obtain the initial output vector h of the first hidden layer. 1 : h 1 =ReLU(W (1) Q+b (1) ) Where W (1) is the weight matrix of the first hidden layer, Q is the input vector, b (1) is the bias vector of the first hidden layer; The initial output vector h of the first layer 1 Perform batch normalization to obtain the initial output vector after batch normalization Where γ, β and ∈ are training parameters, μ batch and are the mean and variance of the current batch; Based on the initial output vector after batch normalization The discard operation is performed to obtain the final output vector of the first hidden layer, and it is used as the input of the next hidden layer. The above steps are repeated until the processing of all hidden layers is completed.

8. The artificial intelligence generation method of a ferroelectric material composition-temperature phase diagram according to claim 7, characterized in that: The final output vector of the first hidden layer is expressed as: In the formula, is the final output vector of the first hidden layer, and m is the mask.

9. The artificial intelligence generation method of a ferroelectric material composition-temperature phase diagram according to claim 1, characterized in that: The step of outputting the crystal structure type comprises: Based on the chemical formula numerical vector list and the normalized temperature list, the total number of executions of the pre-trained crystal phase transition artificial intelligence model is calculated, and the calculation expression of the total number of executions is: N=N1×N2 Where, N is the total number of executions, N1 is the number of chemical formula numerical vectors in the chemical formula numerical vector list, that is, the number of chemical formulas with different doping ratios, and N2 is the number of temperature points in the normalized temperature list, that is, the total number of calculations of each chemical formula numerical vector at different temperature points; The chemical formula numerical vectors in the chemical formula numerical vector list and the normalized temperatures in the normalized temperature list are merged as input data for one execution, and the pre-trained crystal phase transition artificial intelligence model is used for processing, and the crystal structure type of this time is output. This step is repeated N-1 times to obtain the crystal structure type of all chemical formula numerical vectors at each temperature point.

10. An artificial intelligence generation system for a ferroelectric material composition-temperature phase diagram, characterized in that: include: Pre-trained artificial intelligence module: used to obtain the crystal phase transition data set of electrical materials, train the multi-layer artificial neural network model, perform hyperparameter optimization and cross-validation during the training process, and obtain a pre-trained crystal phase transition artificial intelligence model; Phase diagram automatic generation module: used to read the chemical formula of ferroelectric materials with doping variables, the range of doping ratio and temperature, the step size, and perform standardization processing, cyclically call the pre-trained crystal phase transition artificial intelligence model for processing, output the crystal structure type, and draw the component-temperature phase diagram of the chemical formula of the ferroelectric material with doping variables under different doping ratios and different temperature conditions according to the crystal structure type, wherein the chemical formula of the ferroelectric material with doping variables includes chemical elements and corresponding doping ratios, and the doping variables are included in the doping ratio.

Images