As a kind of natural high quality composite material,wood is widely used in all aspects of human life from construction to furniture.With the development of modern industry,a variety of processing equipment has been developed,and human processing means of wood is becoming more efficient and automatic.Due to the structural and mechanical properties of wood itself,there are many factors affecting the quality of wood processing.In a processing process,the traditional machining always maintains a fixed power or processing parameters,only to ensure the processing quality as the premise,and if always maintain a maximum machining force is bound to cause a waste of energy.Therefore,this paper studied the transverse cutting process of spruce,established a wood spruce model based on spiral bundle structure,and used Abaqus for simulation analysis.Combined with the data of actual tests,the changes of cutting force and energy in the transverse cutting process of wood were studied.At the same time,the sawing processing data of spruce were analyzed in this paper.Based on multi-physical domain information and LSTM neural network algorithm,the intelligent sawing parameters of spruce were selected to ensure processing quality and reduce energy consumption.Based on the tracheid theory of wood,this paper proposes a spruce spiral bundle model,which consists of six spirally wound tracheids at a certain Angle forming a RVE(Representative volume element).In order to simplify the research process,the spruce wood studied in this paper can be regarded as transversely isotropic material,ignoring the different mechanical properties caused by different growth stages.After analyzing the wood microstructure and transverse cutting mechanism,this paper determined the stress-strain relationship of spruce in elastic stage and the yield criterion in inelastic stage,and determined that spruce needs five different engineering elastic constants to describe its mechanical properties.At the same time,the VUMAT subroutine applied in Abaqus wood transverse cutting simulation was written.In this paper,Abaqus extended finite element method(XFEM)was used to simulate the fracture of spiral tube bundle RVE model of spruce.After the model is established,the initial crack control position is preset on the model,and then the pre-processing is carried out on it,and the task is submitted for solving.In the post-treatment,the stress and energy changes in the crack propagation process can be obtained,which can be the basis for the simulation of transverse cutting of spruce wood.When the spruce wood is simulated by transverse cutting,the wood model is homogenized,and the written VUMAT file is selected when the solution is submitted.In the post-treatment process,the cutting force and energy changes of spruce wood in the transverse cutting process can be obtained.Combined with the data curve obtained in the actual test,the change of force and energy consumption are analyzed.It is concluded that the mechanical properties of spruce can be well demonstrated by the spruce model based on the spiral bundle structure,and there is always a maximum force and a certain energy consumption in the transverse cutting process.When using joinery band sawing machine to cut spruce wood,different processing parameters(saw wheel speed,feed speed)correspond to different cutting forces,that is,different processing power.In this paper,the sawing test of spruce wood was carried out,and the machining errors(flatness error and invalid thickness error)under different combinations of machining parameters were obtained.After analysis,it was concluded that there was always an optimal combination of machining parameters,which could satisfy the machining quality and avoid invalid power at the same time.At the same time,an intelligent saw-cutting equipment based on multi-physical domain information fusion and LSTM neural network algorithm was established in this paper to predict the processing quality of spruce wood saw-cutting.The results were basically close to the actual test data,so the optimal processing parameters that could meet the processing quality and reduce energy consumption could be automatically selected. |