ABSTRACT

The optimum adjustment the CNC (Computer Numerical Control) processing parameters is extremely important, especially in finishing processes such as coating, painting, and varnishing where surface quality is required. This work aimed to determine the CNC processing parameters for the best wood surface quality by ANN (Artificial Neural Network). For this aim, the surface roughness values of intermediate values not used in experimental studies were also estimated and the effects of parameter variables for each wood species were revealed. Surface roughness measurements (Ra) were made according to the DIN 4768 to determine the surface quality of wood materials. The prediction model with the best performance was determined through statistical and graphical comparisons. It has been observed that ANN models achieve quite satisfactory results with acceptable deviations. As a result of ANN analysis, the optimum values of tool diameter, spindle speed and feed rate for spruce wood were determined as 2 mm, 10000 rpm and 5 m/min, respectively. These values for beech wood were determined as 4 mm, 12500 rpm, 5 m/min, respectively. The findings of this study can be effectively applied in the furniture industry to reduce time, energy, and cost for experimental research within the range of experimentation conducted.

摘要

CNC（Computer Numerical Control）加工参数的最佳调整极为重要，尤其是在涂装、喷漆、清漆等对表面质量有要求的精加工工序中。这项工作旨在通过 ANN（人工神经网络）确定最佳木材表面质量的 CNC 加工参数。为此，还估计了实验研究中未使用的中间值的表面粗糙度值，并揭示了每种木材种类的参数变量的影响。根据 DIN 4768 进行表面粗糙度测量 (Ra)，以确定木质材料的表面质量。通过统计和图形比较确定具有最佳性能的预测模型。据观察，ANN 模型在可接受的偏差下取得了相当令人满意的结果。作为 ANN 分析的结果，云杉木材的刀具直径、主轴速度和进给速率的最佳值分别确定为 2 mm、10000 rpm 和 5 m/min。山毛榉木的这些值分别确定为 4 mm、12500 rpm、5 m/min。本研究成果可有效应用于家具行业，在所进行的实验范围内减少实验研究的时间、精力和成本。