隨著科技產業的精密化生產而日趨複雜,各行各類的人為操作系統要求下,其架構與功能都伴隨著許多非線性、線性、類比、模糊不確定、及競爭的環境因素。當人操作界面及控制系統分析時,常會面臨到無法有效使用及更精確完整的數學模型來完整描述在研究分析上,使複雜系統之組織與特性無法即時監控輸出精確的數值來分析,因而在開發控制模組時讓研究開發產生了阻礙導致難以實現,造成生產時效浪費。 當前研究已經證明,模糊派翠網路已被研究成為一種圖形化介面模擬工具,也有多研究人員將模糊派翠網路對離散事件的系統進行類比、控制和性能分析用來模擬和控制製造流程的所有活動。故本文提出整合智慧流程控制系統的設計技術來推論分析模擬洗衣機所花費的時間及監測即時表面粗糙度在複雜自動化工廠生產流程銑削加工為例驗證模糊理論(FuzzyTheory)派翠(Petri net)網路法、類神經網路法與圖控程式軟體LabVIEW 2011所開發出的一種圖形化介面技術模式,將理論推演到實際應用與重新的架構整合建模來詮釋該研究。本文提出整合以模糊理論派翠網路法、類神經網路法將模糊理論的推論法則置入於派翠網的圖形化介面建模預測以及類神經網路的模擬邏輯控制方法,能夠在LabVIEW2011軟體下完整開發出一套可具備出基本控制理論定義模型的圖形控制介面軟體模擬分析工具。 本文現階段成果是採用兩個模糊輸入變量,及一個輸出。此軟體將來可以擴充到多輸入、多輸出的架構。本文是以前饋式類神經網路控制輸出表面粗糙度來要求精度和性能分析。得到結果顯示LabVIEW 2011軟體是可以多用途程式架構下輸出可應對高、中或低品質的表面粗糙度,故此開發的技術可發揮一定的作用。
The demands for complex and precision technology in a variety of industries have resulted in a functional framework that requires non-linear, fuzzy, non-determinant, and competitive environmental factors. Presently, operators are constrained by the limited analytical capabilities in the control systems because complex and precision mathematical and theoretical models are unable to capture and describe the real-time output in an organized and analytical manner, thus the limitations in the research prototype control models. Recent researches have proven that Fuzzy Petri-Network may be utilized as a development environment for a graphical modeling tool. Some researchers utilize Fuzzy Petri-Network to compare, contrast, control, and to provide functional analysis in the modeling of manufacturing activities. This paper proposes an integrated design methodology based on the Intelligent Process Control System, together withwashing machines The time it takes inference analyze and CNC-milling machining, that demonstrates the development of a graphical modeling tool which integrates the Fuzzy Theory Petri-Network, Neural Network, and LabVIEW 2011. This graphical model tool represents an integrated real-world developmental application derived from theories. This paper proposes an integrated Fuzzy Theory Petri-Network and Fuzzy Neural Network Inference Rules, within a Graphical Model Interface, that may predict and model a control network. At this time, this paper can model two input variables within a framework with one output. Additionally, this paper utilizes the surface roughness of Feedforward Neural Network for precision control and analysis. Preliminary results showed LabVIEW 2011 can accommodate a multi-level quality control model with the surface roughness as input variables, therefore this developmental model may be useful