Elastic Data Modelling Of Multi Component Space And High Throughput Design Of Nb-Ti-V-Zr Alloys

Metal is the most important materials in human's life,and it is widely used in structural application.In structural materials,the mechanical properties are the most important.Among the mechanical properties,the elastic constants are the basic but fundamental property.Many bizarre properties,such as gum metal and biomedical applications in refractory alloys,can be predicted by elastic constants.At present,the second-order elastic constants(SOECs)at 0k can be easily calculated by first principles calculation.However,the calculation of elastic constants under external conditions,such as high temperature and high pressure,is complex and expensive.Therefore,how to quickly evaluate the elastic properties of materials under extreme conditions is of great significance to accelerate the design of structural materials.In the past,alloys are always designed based on a matrix and then adjusting the properties by adding some trace elements.However,the candidate composition space is limited in traditional alloy design.Recently,the high-entropy alloys(HEAs)are developed,in which the constrain of matrix is released in the alloy design.Under this concept,the candidate composition is extended to the full composition space,which provide more possibility for excellent alloys.However,the HEAs brings both opportunities and challenges to alloy design.Due to the sharply increase of composition space,it becomes a major difficulty how to efficiently design alloy composition with target properties.In this paper,refractory alloys were chosen as the research object.Firstly,the relationship between compositions and phases in refractory alloys was investigated and a machine learning model of compositions and phases was built.Based on model,a quaternary system was determined considering the density and the range of BCC phase.Then the lattice constants,SOECs and third-order elastic constants(TOECs)across the full-composition space of the determined system in BCC phase were calculated by new algorithm proposed in this paper and then a fitting database of composition-properties was created.According to the database,the high-throughput design of the system was carried out.In addition,the nonlinear effect was introduced by TOECs,and then the elastic constants of the BCC phase under high temperature and high pressure were evaluated.All in all,this paper presents a set of high-throughput design paradigm for multicomponent alloys from phase distribution to mechanical properties,which provides a new idea for the design of multicomponent alloys.Regarding the phase distribution of refractory alloys,consisting with Cr,Hf,Mo,Nb,Ta,Ti,V,W,Zr refractory elements,a large number of compositions and phases data are extracted from the published papers and phase diagrams,and then used machine learning(ML)method to mine the data and built the ML-model of compositions and phases.Then the phases of different compositions were predicted by the model.The results show that the support vector machine(SVM)model with alloy elements as descriptors is the most accurate in reflecting the composition-phase relationship in refractory alloy,and the corresponding accuracy of training and testing are 92%and 88%respectively.Furthermore,the reliability of the composition-phase ML-model is validated by comparing with the experimental results of ternary phase diagram and the phase of several multicomponent alloys.Based on the ML-model,the phase distribution of low-density Ti-V-Zr-X(X is other refractory elements)quaternary systems are predicted.The result shows that the BCC phase region of Nb-Ti-V-Zr is the largest,as a result,Nb-Ti-V-Zr is selected as the research object in the following sections.After determining the research system,this paper focuses on the ab-initio prediction of the mechanical properties.With the help of higher-order elastic constants(HOECs),such as TOECs,the SOECs under high pressure and high temperature can be analyzed.However,the process of traditional method for calculating TOECs,such as strain-energy method,is complected and the cost is quite expensive.Therefore,this paper firstly studies the method of calculating HOECs,and proposes a new method for calculating HOECs,namely dimension reduction algorithm of strain modes(DRASM).In this method,the reduction of dimension of required number of strain modes is achieved by carefully selecting the strain modes.As a result of reduction of the number of strain modes,the efficiency of DRASM is improved significantly.At the same time,t he corresponding software named as Elastic3rd was developed to automate the process of calculating HOECs.In addition,the criterion for evaluating the strain mode was proposed.Taking the calculation of TOECs of diamond as an example,the traditional methods were compared with DRASM.The result shows that the computational efficiency of DRASM in calculating TOECs of cubic system is about 3-5 times of that in the traditional method.Moreover,the DRASM is excellent in accuracy,robustness and scalability.In addition,an error of the TOECs in diamond measured by shock wave experiment is fixed.Using the DRASM method and Elastic3rd software developed in this paper,t he SOECs and TOECs of Nb-Ti-V-Zr quaternary system in BCC phase across the full composition space are calculated.Firstly,the reliability of the method and result is verified by comparing with the experimental and calculation results in the literature.Further,the relationship between properties and composition is described by the formula under the CALPHAD framework,and the database between the composition and elastic constants across the full composition space is constructed.According to the database,the influence of interaction parameters between multi-components on elastic constant is studied.The result shows that the contribution of interaction parameter to elastic constant decreases with the increase of the number of components and the effect of interaction parameters can be omitted when the number of components larger than three,which provides a guidance for building the elastic constants database for other multi-components alloys.Based on the composition-properties database and the ML-model of phases in refractory materials,the high throughput design in Nb-Ti-V-Zr system is further carried out.A series of single-phase BCC alloys with bizarre elastic properties were designed in Nb-Ti-V-Zr system,including gum-metal,partial auxetic material,single crystal isotropic material and bio medical-application materials.The reliability of prediction was confirmed by comparing with some literature results.In addition,the SOECs of BCC phase under high pressure and high temperature across the full composition space were investigated using the composition-TOECs database.The result shows that the born instability region decreases while increases in shear instability region in Nb-Ti-V-Zr system with the increase of pressure.With the increase of temperature,the stability decreases slightly.The derivative of SOECs against temperature exists minimum around X_0.3Y_0.7Zr_0.1 and X_0.3Y_0.7Zr_0.1,which means that such alloys will show high stability at high temperature.In addition,this paper designed several multifunctional alloys,including Nb_0.2Ti_0.7Zr_0.1 and Nb_0.16Ti_0.62V_0.08Zr_0.14,showing partial auxetic,gum-metal and biomedical application and Nb_0.35Ti_0.63V_0.02,showing single crystal isotropic and biomedical application.