知識工程應(yīng)用于船舶結(jié)構(gòu)的優(yōu)化設(shè)計研究 精靈論文

1、知識工程應(yīng)用于船舶結(jié)構(gòu)的優(yōu)化設(shè)計研究楊和振，陳金峰（上海交通大學(xué)，海洋工程國家重點實驗室，上海 200240）摘要：針對船舶結(jié)構(gòu)設(shè)計變量是涉及多種設(shè)計和約束條件的離散變量，造成結(jié)構(gòu)優(yōu)化的高度5非線性、多峰性等問題，而且船舶設(shè)計過程中需要設(shè)計規(guī)范和專家經(jīng)驗等知識支持，結(jié)合其 具有很強的綜合性、模糊性等特點，本文提出了基于知識工程的船舶結(jié)構(gòu)優(yōu)化設(shè)計方法。該 方法利用知識工程與結(jié)構(gòu)優(yōu)化相結(jié)合，將獲取的設(shè)計知識構(gòu)建知識庫應(yīng)用于船舶結(jié)構(gòu)優(yōu)化設(shè) 計，并通過提取結(jié)構(gòu)關(guān)鍵參數(shù)實現(xiàn)參數(shù)化結(jié)構(gòu)模型與優(yōu)化數(shù)學(xué)模型的相互轉(zhuǎn)化，降低結(jié)構(gòu)優(yōu) 化設(shè)計對用戶知識水平的要求。水密橫艙壁結(jié)構(gòu)的優(yōu)化設(shè)計算例表明：滿足約束要求的情況

2、10下，其結(jié)構(gòu)重量在優(yōu)化后比優(yōu)化前降低了，保證了結(jié)構(gòu)性能合理的同時實現(xiàn)重量最輕的目標； 其次，將結(jié)構(gòu)參數(shù)化模型和數(shù)學(xué)優(yōu)化模型結(jié)合在一起，為設(shè)計經(jīng)驗少的設(shè)計者提供了一種結(jié) 構(gòu)設(shè)計的捷徑；總之，實現(xiàn)了從不同資源中獲取知識并應(yīng)用于優(yōu)化設(shè)計過程，促進設(shè)計能力 的提高，降低優(yōu)化設(shè)計過程對知識和經(jīng)驗的依賴。 關(guān)鍵詞：知識工程；船舶結(jié)構(gòu)優(yōu)化；知識庫；模擬退火算法15中圖分類號：U663.4Application of Knowledge-based Engineering for ShipStructure Optimization DesignYANG Hezhen, CHEN Jinfeng20(Sta

3、te Key Laboratory of Ocean Engineering, School of Naval Architecture, Ocean and CivilEngineering, Shanghai Jiao Tong University, ShangHai 200240)Abstract: Ship structural design variables are discrete variables involving a variety of design requirements and constraints, which leads to highly nonline

4、ar, multimodal and other problems in structure optimization. Besides, it needs design regulations and experts experience to support ship25structure design. Considering its highly comprehensive, empirical, fuzzy and other characteristics, this work presents application of knowledge based engineering

5、for ship structure optimization designmethod. This method combines the principle of knowledge based engineering design with structure optimization design. It constructs knowledge base including design regulations and expert experience for structure initial design. In addition, it can distil the key

6、design parameters of structure model to30achieve the mutual transformation of parameterized structure model and optimization mathematical model. The example of watertight transverse bulkhead optimization design shows the optimized weight of structure is lower than before on the condition of meeting

7、constrains required. It ensures the reasonable structure performance and lightest weight goal. secondly, it combines structureparameterized model with mathematical optimization model, which provides a shortcut for designer35with less experience of structure design. In conclusion, it gets knowledge f

8、rom different source and apply for structure optimization design, improve the design level, reduce the designers dependence on the knowledge and experience.Keywords:Knowledge-based Engineering; Ship Structure Optimization; Knowledge Base; SimulatedAnnealing400引言目前，結(jié)構(gòu)優(yōu)化設(shè)計領(lǐng)域已從航空擴展到船舶、橋梁、汽車、機械等更廣泛的工程領(lǐng)域

9、， 解決的問題從減輕結(jié)構(gòu)重量擴展到降低應(yīng)力水平、改進結(jié)構(gòu)性能和提高安全壽命等更多方面1-3456789101112131415 16171819202122橫艙壁列板 1 厚度109.0橫艙壁列板 2 厚度1111.5橫艙壁列板 3 厚度1214水平桁腹板高度14501362.5水平桁腹板厚度1010水平桁面板寬度400745.2水平桁面板厚度1513垂直桁T1 腹板高度14501362.5垂直桁T1 腹板厚度1210垂直桁T1 面板寬度550647.5垂直桁T1 面板厚度1511.5垂直桁T2 腹板高度14501362.5垂直桁T2 腹板厚度1510垂直桁T2 面板寬度550732.2垂直桁

10、T2 面板厚度1513扶強材 1 腹板高度150150扶強材 1 腹板厚度1010扶強材 1 面板寬度9090扶強材 1 面板厚度1010扶強材 2 腹板高度200200扶強材 2 腹板厚度99扶強材 2 面板寬度9090扶強材 2 面板厚度1212265270水密橫艙壁結(jié)構(gòu)優(yōu)化后的總重量為 57034.1kg，比優(yōu)化前的重量減少了 3.03%，既滿足了船舶設(shè)計規(guī)范、型材穩(wěn)定性和加工工藝等約束要求，又實現(xiàn)了知識庫中知識應(yīng)用，同時可 以將結(jié)構(gòu)優(yōu)化尺寸參數(shù)實時應(yīng)用于結(jié)構(gòu)參數(shù)設(shè)計模型。水密橫艙壁結(jié)構(gòu)重量目標函數(shù)變化曲 線如圖 9 所示，設(shè)計變量艙壁 1-3 列板厚度的優(yōu)化過程曲線如圖 10 所示，垂

11、直和水平桁材 面板寬度優(yōu)化過程曲線如圖 11 所示，優(yōu)化后的水密橫艙壁結(jié)構(gòu)模型如圖 12 所示。圖9橫艙壁結(jié)構(gòu)優(yōu)化過程重量變化曲線圖 圖10 橫艙壁結(jié)構(gòu)優(yōu)化過程列板厚度曲線圖Fig.9 Transverse Bulkhead Structural Weight CurveSchematic DiagramFig.10 One to Three Bulkhead Plate ThicknessOptimization Process Schematic Diagram圖11橫艙壁結(jié)構(gòu)優(yōu)化過程桁材面板寬度曲線 圖 12 優(yōu)化后水密橫艙壁結(jié)構(gòu)示意圖Fig.11 Transverse Bulkhead

12、 Structure Flange WidthOptimization Process Schematic DiagramFig.12 Optimized Transverse Bulkhead StructureSchematic Diagram2752802852905結(jié)束語設(shè)計規(guī)范對水密橫艙壁板厚的約束以及對扶強材、垂直桁材和水平桁材剖面模數(shù)的約 束，同時考慮型材的局部穩(wěn)定性和工藝性約束要求，建立水密橫艙壁結(jié)構(gòu)參數(shù)化模型并通過 知識工程技術(shù)轉(zhuǎn)化為數(shù)學(xué)優(yōu)化模型進行結(jié)構(gòu)優(yōu)化設(shè)計。建立橫艙壁結(jié)構(gòu)的參數(shù)化模型，通過 知識工程模塊建立標準的角鋼庫和板材庫以及將 CCS 設(shè)計規(guī)范要求、設(shè)計標準和專家

13、經(jīng)驗 等知識整理、歸納構(gòu)建簡單知識庫，并將橫艙壁結(jié)構(gòu)參數(shù)化模型轉(zhuǎn)化數(shù)學(xué)優(yōu)化模型,借助于 產(chǎn)品工程優(yōu)化模塊基于模擬退火優(yōu)化算法對水密橫艙壁結(jié)構(gòu)進行優(yōu)化設(shè)計，并將優(yōu)化設(shè)計數(shù) 據(jù)實時應(yīng)用于結(jié)構(gòu)設(shè)計模型，對結(jié)構(gòu)設(shè)計模型進行實時更新，避免了結(jié)構(gòu)優(yōu)化過程和設(shè)計過 程分開獨立操作，減少了由結(jié)構(gòu)優(yōu)化數(shù)學(xué)模型轉(zhuǎn)化為設(shè)計模型的重復(fù)工作。將知識工程應(yīng)用于水密橫艙壁結(jié)構(gòu)優(yōu)化設(shè)計，優(yōu)化設(shè)計結(jié)果表明：首先，橫艙壁結(jié)構(gòu)重 量在優(yōu)化后比優(yōu)化前降低了，保證橫艙壁結(jié)構(gòu)性能合理的同時實現(xiàn)了結(jié)構(gòu)重量最輕的目標， 降低產(chǎn)品設(shè)計的生產(chǎn)成本，產(chǎn)生一定的經(jīng)濟效益。其次，為船舶結(jié)構(gòu)優(yōu)化設(shè)計提供了極其豐 富的技術(shù)手段，將結(jié)構(gòu)優(yōu)化的數(shù)學(xué)模型和結(jié)構(gòu)

14、設(shè)計仿真模型結(jié)合在一起，為設(shè)計經(jīng)驗少的設(shè) 計者提供了一種船舶結(jié)構(gòu)優(yōu)化設(shè)計的捷徑，有助于設(shè)計者快速完成設(shè)計工作的同時，能夠快 速地積累經(jīng)驗。再次，實現(xiàn)從不同資源中獲取知識并應(yīng)用于船舶結(jié)構(gòu)優(yōu)化設(shè)計過程中促進設(shè) 計能力提高，幫助用戶在更短的時間內(nèi)、更精確地完成更多的設(shè)計工作，增強知識的重用性 和共享性，降低優(yōu)化設(shè)計過程對知識和經(jīng)驗的依賴，提高設(shè)計過程的有效自動化層次。參考文獻 (References)2953003051 Kodiyalam S., Sobieszczanski Sobieski J. Multidisciplinary design optimization-some formal

15、 methods, frame work requirements and application to vehicle designJ. International Journal of Vehicle Design, 2001, 25: 3-22. 2 Alexandrov N. M., Lewis R. M. Analytical and computational aspects of collaborative optimization for multidisciplinary designJ. AIAA Journal, 2002, 40(2): 301-309.3 Sobies

16、zczanski-Sobieski J., Haftka R. T. Multidisciplinary aerospace design optimization: Survey of recent developmentsJ. Structural Optimization, 1997, 14(1): 1-23.4 張科施. 飛機設(shè)計的多學(xué)科優(yōu)化方法研究D. 西安: 西北工業(yè)大學(xué), 2006.5 鄭幫濤, 魯海燕. 微分群體智能算法在船舶結(jié)構(gòu)混合變量優(yōu)化中的應(yīng)用J. 船舶力學(xué), 2009, 13(73):770-775.6 蔣如宏, 裘泳銘. 船舶主尺度方案的多目標模糊優(yōu)化設(shè)計J.船體與舾

17、裝,1999: 7-10.7 柳存根, 姚震球. 遺傳進化算法在船舶初步設(shè)計中的應(yīng)用J. 上海交通大學(xué)學(xué)報, 2000, 34(1): 41-45. 8 黃衍順, 陳斌, 涂躍紅. 雙殼散貨船結(jié)構(gòu)重量優(yōu)化設(shè)計J. 機電設(shè)備, 2007, 209(01): 3-6.9 陳寧,曹成勛,佘建國. 基于遺傳算法與臨境環(huán)境的機艙優(yōu)化仿真研究J.船舶工程, 2008,30(4):77-80. 10 劉遠. 基于遺傳算法的大型起重船拔桿式臂架優(yōu)化設(shè)計研究D. 2004.11 La Rocca G., Van Tooren M. J. L. Knowledge-based engineering to sup

18、port aircraft multidisciplinary design310315320325and optimizationJ. Proceedings of the Institution of Mechanical Engineers, Part G: Journal of AerospaceEngineering, 2010, 224(9): 1041-1055.12 Curran R, Verhagen W. J .C, Van Tooren M. J. L, et al. A multidisciplinary implementation methodology for k

19、nowledge based engineering: KNOMAD J. Expert Systems with Applications, 2010,37(11):7336-7350.13 Xing L. N., Chen Y. W., Wang P, et al. A Knowledge-Based Ant Colony Optimization for Flexible Job ShopScheduling ProblemsJ. Applied Soft Computing Journal, 2010, 10(3): 888-896.14 Zhou T. J., Xing L. N.

20、A knowledge-based genetic algorithm to the global numerical optimizationA. Computational sciences and optimizationC. Sanya, Hainan, China: IEEE Computer Society, 2009:513-516. 15 Jie J., Zeng J. C., Han C. Z., et al. Knowledge -based cooperative particle swarm optimizationJ. Applied Mathematics and Computation, 2008, 205(2): 861-873.16 Hu J., Peng Y.