213 / 2024-04-17 18:51:56

Boiling temperature prediction model of environmental friendly insulation molecules based on machine learning

SF6 Substitutes,Boiling Temperature Prediction,Machine Learning

Sulfur hexafluoride (SF₆) is widely used in the field of electrical equipment because of its good electrical properties. However, sulfur hexafluoride has a strong greenhouse effect, and reducing the use of SF₆ is an important issue to be solved in the field of power equipment to achieve green development. The search for alternative gas of sulfur hexafluoride has attracted much attention.

Finding insulating gases with high dielectric strength, low boiling temperature and environmental properties is a challenging task. For the currently known compounds, the number of molecules that meet the high dielectric strength is large, but the molecules with high dielectric strength and low boiling temperature are extremely rare. With the increase of screening requirements, such as environmental protection requirements, the number of available molecules decreases exponentially. Elemental gases that can fully meet the requirements of all indicators have not yet become a reality. In order to break through the search strategy of new insulating gases, it is necessary to deeply study the characteristics of electric, boiling and environmental protection gases.

In this paper, a machine learning prediction model of insulation molecular boiling temperature is proposed. By learning the relationship between a large number of known molecular feature data sets and the boiling temperature, the structural characteristics and physicochemical properties of insulation molecules are deeply explored, and an environmentally friendly gas insulation molecular boiling temperature prediction model is constructed. This model can not only predict the boiling temperature of environmentally friendly insulation molecules effectively, but also provide a new way for the design and optimization of insulation materials.