Material data cleaning method combining nickel-based alloy prior knowledge

• Main Authors: HUANGFU WEI, XU BIN, LEI KAIXI, YIN HAIQING
• Format: Patent
• Language: Chinese; English
• Published: 09.03.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTEDFOR SPECIFIC APPLICATION FIELDS; PHYSICS

The invention provides a material data cleaning method combining nickel-based alloy prior knowledge, and belongs to the field of material alloy synthesis. The method comprises the steps: carrying outcurved surface fitting material data through a nickel-based alloy empirical formula; and simulating the material data through alloy simulation software based on nickel-based alloy priori knowledge, and cleaning the material data which does not accord with the nickel-based alloy priori knowledge according to a simulation result of the alloy simulation software and a curved surface graph obtained byfitting. By adopting the method, error data in the material data can be removed, the material data is further simplified, and data interference of the nickel-based alloy material data applied to machine learning is improved, so that the purpose of optimizing the machine learning effect is achieved. 本发明提供一种结合镍基合金先验知识的材料数据清洗方法，属于材料合金合成领域。所述方法包括：通过镍基合金经验公式对材料数据进行曲面拟合；通过基于镍基合金先验知识的合金仿真软件对材料数据进行仿真，根据合金仿真软件的仿真结果以及拟合得到的曲面图，对不符合镍