反應曲面法在無人搬運車系統之應用分析

APPLICATION AND ANALYSIS OF RESPONSE SURFACE METHOD IN AN AGV SYSTEM

張金泳(Chin-Yung Chang)

指導教授 ： 陳雲岫

元智大學/工程學院/工業工程與管理學系/碩士(1992年)

摘要

物料搬運在一彈性製造系統(Flexible Manufacturing System,FMS)中整
合了不同的生產作業型態，其中又以無人搬運車系統(Automated Guided
Vehicle,AGVS)為最常採用的自動化物料搬運系統之一，是以模式設計之
優劣對整個系統績效影響甚巨。本研究則是運用實驗設計
(Experimental Design)中之反應曲面法(Response Surface Method)，在
系統績效極值化之前提下，探討AGVS中各要因對系統之影響及其數學模式
。針對一彈性製造系統中，無人搬運車系統在運作時工件平均等待搬運及
實際搬運的時最小時，分析各因子的最佳水準組合，並在迴歸模式
(Regression Model)模擬得知下，可配置不同之因子水準而預測反應值。
經由ProModel模擬軟體之系統模擬及統計套裝軟體SAS之數據分析，吾人
建立模擬迴歸模式及最佳化之水準組合。

關鍵字

彈性製造系統 ； 無人搬運車系統 ； 實驗設計 ； 反應曲面法 ； 迴歸模式

並列摘要

Material handling plays an important role in a Flexible
Manufacturing System(FMS). Automated Guided Vehicle System
(AGVS) has been widely used to handle materials in a FMS.
However, several factors succh as buffer size, part arrival
rate, speed of AGVS, number of AGVS, dispatching rule and
flowpath in an AGVS have significant effects on the performance
of a FMS. In this reserach, we examine the effects of these
factors by the Response Surface Methodology(RSM) and nested
design. Result show that part arrival rate, speed of AGVS and
number of AGVS affects the mean time of parts in a system
significantly. The behavior of response variables is studied
through the simulated regression model and the linear
combinations of levels of significant factors for the minimum
value of mean time of parts in the system is obtained.
Simulation and data analysis are finished by ProModel
simulation software and the statistical software package SAS,
respectively.

並列關鍵字

Automated Guided Vehicle ； Experimenal Design ； Response Surface