Development Of Wood End-crack Evaluation And Testing System

Wood cracking is a common wood defect,which seriously affects the processing and utilization of wood.Therefore,it is necessary to study the law of wood cracking and explore the measures to prevent wood cracking.The traditional way to study wood cracking is generally to study the macroscopic shape of wood in the natural state,but this kind of experiment cycle is long,and it is impossible to cover all the climatic conditions that may appear in the nature in a certain period of time,so it is not easy to explore the complete law of wood cracking.As the main type of wood cracking is cross section cracking,this paper mainly studies the rule of wood cross section cracking,explores the wood cracking under different climatic conditions and the change of strain.Based on the maximum strain theory and correction coefficient method,an evaluation model for predicting wood cross section cracking is proposed.According to the verification results of wood cracking evaluation model,temperature and protective agent have a significant impact on end cracking,air humidity has a certain impact on cross section cracking,and ultraviolet radiation has no significant impact on end cracking.According to the model to evaluate whether the wood will crack under different environmental factors has high accuracy,which provides a reference for the rational use of wood in different occasions.In this paper,the end cracking of wood under long-term outdoor exposure was studied,and the protective effects of four kinds of protective agents,such as water-based paint,tung oil,wood wax oil and flexible agent,were compared.It was found that the untreated wood would crack most seriously,tung oil had the worst anti cracking effect,followed by water-based paint,wood wax oil and wood flexible agent had the best anti cracking effect.By using multiple classification and regression algorithms in machine learning,the cracking state of outdoor wood is analyzed and predicted,and the model is established.The accuracy of using the model to predict the cracking state is more than 80%.