Data Analysis Of Material Failure In Atmospheric Environment Based On Machine Learning By Python

Improving the corrosion resistance and aging resistance of materials,and predicting the time when materials fail due to corrosion/aging,thereby ensuring the safety and efficiency of engineering,has always been an important topic in material failure research The combination of machine learning technology and material failure research is a research hotspot that meets the development needs of the times.This article reviews the current status of machine learning in material failure research,and summarizes the general steps and methods for analyzing material failure data using machine learning algorithms.Methods,the relationship between material failure in the atmospheric environment and its actual use environment was discussed,and the impact of different environmental factors on material failure was studied from the perspective of macro data.It has gradually become a trend to use Python,which is easy to learn and use,and has an open source code ecology and many mature integrated models as a computing tool to provide assistance to materials science research.This article first uses the Extra Trees algorithm integrated in Python to analyze the material performance parameters(mechanical properties of the PC sample and the gloss of the polyester coating)collected in the actual exposure experiment of polymer material samples in different atmospheric environments,climate environment Data analysis of parameters(temperature,humidity,irradiation,rainfall and rainfall duration,etc.)and environmental parameters of pollutants(sulfation rate,sea salt particles,and dustfall,etc.),and then the model is established through a multi-layer perceptron neural network algorithm.The weather resistance of the material is evaluated,The analysis results show that:1.The Extra Trees algorithm can be applied to the study of the influence of environmental factors on the appearance of the coating and the mechanical properties of the PC sample.The calculation results are reasonable and interpretable.The results of the simulation calculations believe that:(1).Ultraviolet radiation,water-soluble dust reduction and sulfate rate are the key parameters for the deterioration of the polyester coating's loss of light;(2).The length of rainfall is the key parameter for the thinning of the polyester coating;(3).Ultraviolet radiation and rainfall duration are also the key factors that deteriorate the tensile properties,impact properties and bending properties of PC specimens;2.When using the Extra Trees algorithm to study the impact of environmental factors on the properties of materials,the experimental period can be used together with the environmental factors Enter to participate in the calculation,increase the difference between the data and at the same time,you can also get the tolerance of the material properties to the environment;3.The importance parameters obtained by the Extra Trees algorithm can guide the more reasonable use of data in subsequent calculations and improve The performance of other machine learning models;4.For the data scale of the general case of material failure in the natural environment,you can find the best hyperparameters of the machine learning model by performing a loop experiment on all possible combinations with high performance;5.Developed data analysis program for environmental corrosion and aging of materials,and realized two main functions of quantification of environmental factors and regression model of material performance.