以有限元素法及類神經網路建構高速切削性能分析模型

Construction of an analytical model of cutting performance for high speed machining by finite element method and neural network

鄭松輝(SH Cheng)

指導教授 ： 林盛勇

國立虎尾科技大學/工程學院/動力機械工程研究所/碩士(2005年)

https://doi.org/10.6827/NFU.2005.00002

摘要

本文以有限元素法結合類神經網路來建構高速切削性能分析模型，
以不同刀具幾何形狀 (刀面斜角、刀鼻半徑) 及切削條件 (工件降伏強度、
切削速度、進給率) 來探討其對殘留應力、切屑形成型式和加工面表面粗
糙度等切削性能之影響，並進而推導出一量化的預測分析模型，提供相
關業界實務使用及指南參考。
首先應用 Third Wave Systems Advant Edge 軟體模擬高速切削，刀具
-工件逐步推移過程，計算工件和刀具應力、應變、溫度分佈及網格變形
等，經由自行撰寫程式將其等價至網格點上以供殘留應力、切屑形成型
式和加工面表面粗糙度等性能指標分析。同時結合 STATISTICA 及 AIM
類神經網路來建構此相關實用預測分析模型。最後藉由高速切削實驗加
以驗證，由理論計算所擷取網格之表面粗糙度與實際切削情形相吻合。
本文結果顯示，越接近工件次表面之殘留應力為最大且為壓應力；
切削進給率為殘留應力及表面粗糙度最大影響之參數。模具鋼 SKD11工
件材料於高速切削時，主切削區溫度將超過 600°C，使用碳化鎢刀具，
則刀具強度會急劇下降，且易產生磨耗。在相同切削條件下，表面粗糙
度值隨著進給率增加而增加，但若切削速度超過 450m/min 時，表面粗糙
度呈現較不規律之現象。

關鍵字

non

並列摘要

The main objective of this study is to construct an analytical model of
cutting performance for high-speed machining by finite element method and
neural network. Various combinations of tool geometric shapes (rake angle,
nose radius) and cutting conditions (yielding strength, cutting speed, feed rate)
are constituted to investigate the influences on cutting performances such as
residual stresses, chip types and surface roughness, etc. Furthermore, a
quantitative prediction model is established for the relationship between
cutting variables and their performances, and the model may offer some
practical references and guide consultants for industrial practice uses.
The Third Wave Systems AdvantEdge software is applied to simulate the
high-speed machining; the cutting tool is incrementally advanced forward
step by step during the cutting processes. The stress, strain and temperature
distributions within the workpiece and tool, and the deformations of the grid
patterns of the workpiece are thus determined. These data points are then
processed equivalently to the mesh nodes fulfilling the requirement of the
modeling construction. Moreover, the STATISTICA and AIM neural
networks are applied to synthesis the data obtained from the finite element
calculations. Finally, high-speed machining experiments are performed to
validate the accuracy of the model developed. The surface roughness obtained
from the calculation agrees quite well with that obtained from the
experiments.
The results show that residual stress is compressive and biggest near to
the workpiece machined subsurface. The feed rate is the most important
parameter influencing the residual stress and surface roughness results. The
temperature induced within the primary deformation zone is over 600°C in
high-speed machining of SKD11 mold steel, the Strength of tungsten carbide
tool will be reduced suddenly and hence accelerates the tool wear. Under the
same cutting conditions, the surface roughness is getting better when the
cutting speed is increased, but an irregular relationship between them existed
in this study when the cutting speed is greater than 450m/min.

並列關鍵字

non

參考文獻

4. Jeffrey D. Thiele and Christopher J., "Simulation and Analysis of Chip

9. T. Matsumoto, M.M. barash and C.R. Liu, "Effects Of Hardness On The

30. Wen-Tung Chien, Chung-Ti Chou, "The Predictive Model for

Computational Machining Method to Discontinuous Chip Formation",

Of Sintered CBN Tools, Cutting Tool Materials. Proceedings Of The