Abstract
Surface roughness parameters Ra and Rt are mostly used as an index to determine the surface finish quality in the process of machining. Because of the strong nonlinear character of relationships between the process inputs and outputs, it is difficult to accurately estimate roughness characteristics by using traditional modeling techniques. In this work, accurate prediction of the Ra and Rt values during machining of reinforced poly ether ether ketone (PEEK) CF30 with TiN coated tools is achieved. The modeling is performed by using artificial neural network approach to represent the complex relationships between cutting conditions and surface roughness parameters. The input cutting parameters include cutting speed, depth of cut and feed rate. The network was trained with pairs of inputs and outputs datasets generated by machining experimental results that were obtained according to a full factorial design of experiment table. Predictions of the ANN based model were found to fit experimental data very well with a correlation coefficient as high as 99%. Complementary results that were not used during derivation of the ANN model have enabled one to assess the validity of the obtained predictions.
表面粗糙度參數Ra與Rt大多是用來決定在加工過程中表面光滑度的一個指標。 由於製程的投入及產出之間有強烈的非線性關係 , 所以難以使用傳統的建模技術來準確估計表面粗糙度的特性。 本研究在以TiN塗層刀具補強聚醚醚銅 CF30的期間可以準確預測Ra與Rt。 本研究模型使用類神經網路的方法來表示切削條件與表面粗糙度參數之間的複雜關係。 其中切削的輸入參數包含了切削速度、 切削深度以及進給率。 此模型的網路圖是由成對的投入及輸出資料集所生成 , 而這些成對的資料集是以加工實驗的結果所產生 , 加工實驗的結果是根據全因子實驗設計表所獲得。 以類神經網路為基礎的模型預測結果相當符合實驗數據 , 其相關係數高達百分之九十九。 在類神經網路模型的推導過程不考慮互補效果 , 能夠賦予此模型評估預測結果正確性的能力。
(*聯絡人 : [email protected])
Notes
(*聯絡人 : [email protected])