Intelligent MBSE design approach based on retrieval augmented large language model

doi:10.11996/JG.j.2095-302X.2024061188

Abstract

Abstract:

Model-based systems engineering (MBSE) is one of the most important methods for today’s digital design of products. However, due to the high specialization of systems engineering and the complex interrelationships within products, the application of MBSE to complex products has proven challenging. To address this problem, an intelligent design method based on retrieval-augmented large language model was proposed for the first time. The method first established an object-oriented multi-modal vector representation for models, leveraging retrieval-augmented generation techniques that incorporate domain knowledge and modeling rules to guide the model in more accurately generating MBSE model diagrams. Secondly, a diagram optimization method based on the MBSE model relations was proposed, cross-validating the model accuracy through the results of contextual interaction. Thirdly, the large language model was employed to call modelling APIs and to select the proper materials to generate design models and eBOM. Finally, a dataset containing 24 scenario models was constructed for method validation. Experimental results showed that the approach possessed high accuracy and usability. A case study with water jet propulsion as the modelling object further demonstrated that the approach can effectively enhance the modelling efficiency while maintaining usability, marking an important step toward intelligent application of model-based systems engineering.

Key words: model based systems engineering, large language model, intelligent design, prompt engineering, computer-aided design

CLC Number:

YU Han, CHEN Zhiyuan, XIONG Xirui, DAI Yuanxing, CAI Hongming. Intelligent MBSE design approach based on retrieval augmented large language model[J]. Journal of Graphics, 2024, 45(6): 1188-1199.

Figures/Tables 15

References 27

[1]	BOLSHAKOV N, BADENKO V, YADYKIN V, et al. Cross-industry principles for digital representations of complex technical systems in the context of the MBSE approach: a review[J]. Applied Sciences, 2023, 13(10): 6225.
[2]	APVRILLE L, DE SAQUI-SANNES P, VINGERHOEDS R. An educational case study of using SysML and TTool for unmanned aerial vehicles design[J]. IEEE Journal on Miniaturization for Air and Space Systems, 2020, 1(2): 117-129.
[3]	闫佳宁, 张安, 黄湛钧, 等. 基于SysML的民机系统功能设计方法及应用[J]. 图学学报, 2024, 45(2): 277-283. DOI
	YAN J N, ZHANG A, HUANG Z Y, et al. Function design method and application of civil aircraft system based on SysML[J]. Journal of Graphics, 2024, 45(2): 277-283. (in Chinese) DOI
[4]	CHEN L, ZHOU Y Y, HUANG C L, et al. Parametric optimization of vortex generator configuration for flow control in an intake duct for waterjet propulsion[J]. Ocean Engineering, 2023, 281: 114908.
[5]	PEER J, MORDECAI Y, REICH Y. NLP4ReF: requirements classification and forecasting: from model-based design to large language models[C]// 2024 IEEE Aerospace Conference. New York: IEEE Press, 2024: 1-16.
[6]	DARM P, MARCHETTI F, GARCIA G, et al. Leveraging language models semantic similarity capabilities to facilitate information reuse in system engineering[C]// The 74th International Astronautical Congress. Baku: IAF, 2023: 1-9.
[7]	JIN Q, YANG P, WANG T, et al. Papyrus SysML-based model validation extension[C]// 2022 2nd Conference on High Performance Computing and Communication Engineering. Harbin:SPIE, 2023, 12605: 406-411.
[8]	TAO F, MA X, LIU W, et al. Digital engineering: state-of-the-art and perspectives[J]. Digital Engineering, 2024, 1: 100007.
[9]	SHOSHANY-TAVORY S, PELEG E, ZONNENSHAIN A. Conceptual design support by MBSE: established best practices[EB/OL]. [2024-07-01]. https://link.springer.com/referenceworkentry/10.1007/978-3-030-93582-5_84.
[10]	ZHANG X, WU B, ZHANG X,, et al. An effective MBSE approach for constructing industrial robot digital twin system[J]. Robotics and Computer-Integrated Manufacturing, 2023, 80: 102455.
[11]	ABUAL-HAIJA Q. SysML-based design of autonomous multi-robot cyber-physical system using smart IoT modules: a case study[M]// Machine Learning Techniques for Smart City Applications:Trends and Solutions. Cham: Springer International Publishing, 2022: 203-219.
[12]	MONTORI F, TATARA M S, VARGA P. Dynamic execution of engineering processes in cyber-physical systems of systems toolchains[EB/OL]. [2027-07-01]. https://doi.org/10.1109/TASE.2024.3362132.
[13]	KLYKKEN T E. A case study using SysML for safety-critical systems[D]. Oslo: University of Oslo, 2009.
[14]	PARTHASARATHI H, RAMACHANDRA S, SRINIVASAMURTHY P N. Model‐based systems engineering for aero gas turbine engine subsystems[J]. INCOSE International Symposium. 2016, 26(s1): 70-82.
[15]	BRUGGEMAN ALMRM, LA ROCCA G. From requirements to product: an MBSE approach for the digitalization of the aircraft design process[J]. INCOSE International Symposium. 2023, 33(1): 1688-1706.
[16]	马彦丽, 张继忠, 郝冀斌, 等. 基于MBSE的柴油机系统设计建模方法[J]. 图学学报, 2024, 45(2): 355-362. DOI
	MA Y L, ZHANG J Z, HAO J B, et al. Modeling method for design of diesel engine system based on MBSE[J]. Journal of Graphics, 2024, 45(2): 355-362. (in Chinese) DOI
[17]	粟华, 方施喆, 田昆效, 等. 基于MBSE的地空导弹总体方案设计[J]. 图学学报, 2024, 45(2): 268-276. DOI
	SU H, FANG S Z, TIAN K X, et al. Conceptual design of surface-to-air missile based on MBSE[J]. Journal of Graphics, 2024, 45(2): 268-276. (in Chinese) DOI
[18]	朱伊明, 张玉珠, 陈斌, 等. 基于MBSE的空间引力波探测系统指标管理及其探测灵敏度分析[J]. 图学学报, 2024, 45(2): 259-267. DOI
	ZHU Y M, ZHANG Y Z, CHEN B, et al. Space gravitational wave detection parameter management and performance analysis based on MBSE[J]. Journal of Graphics, 2024, 45(2): 259-267. (in Chinese) DOI
[19]	李戈, 彭鑫, 王千祥, 等. 大模型: 基于自然交互的人机协同软件开发与演化工具带来的挑战[J]. 软件学报, 2023, 34(10): 4601-4606.
	LI G, PENG X, WANG Q X, et al. Challenges from LLMs as a natural language based human-machine collaborative tool for software development and evolution[J]. Journal of Software, 2023, 34(10): 4601-4606. (in Chinese)
[20]	蒋灿, 郑哲, 梁雄, 等. 大语言模型驱动的交互式建筑设计新范式——基于Rhino7的概念验证[J]. 图学学报, 2024, 45(3): 594-600. DOI
	JIANG C, ZHENG Z, LIANG X, et al. A new interaction paradigm for building design driven by large language model: proof of concept with Rhino7[J]. Journal of Graphics, 2024, 45(3): 594-600. (in Chinese) DOI
[21]	BASHATAH J, SHERRY L. Prompt engineering to classify components of standard operating procedure steps using large language model (LLM)-based chatbots[C]// 2024 Integrated Communications, Navigation and Surveillance Conference. New York: IEEE Press, 2024: 1-8.
[22]	CÁMARA J, TROYA J, BURGUEÑO L, et al. On the assessment of generative AI in modeling tasks: an experience report with ChatGPT and UML[J]. Software and Systems Modeling, 2023, 22(3): 781-793.
[23]	LONGSHORE R, BELL R, MADACHY R. Leveraging generative AI to modify and query MBSE models[R]. Acquisition Research Program, 2024.
[24]	FUCHS J, HELMERICH C, HOLLAND S. Transforming system modeling with declarative methods and generative AI[C]// AIAA SCITECH 2024 Forum. Orlando: AIAA, 2024: 1054.
[25]	FENG Z, GUO D, TANG D, et al. CodeBERT: a pre-trained model for programming and natural languages[C]// Findings of the Association for Computational Linguistics:EMNLP 2020. Kerrville: Association for Computational Linguistics, 2020: 1536-1547.
[26]	BI X, CHEN D, CHEN G, et al. DeepSeek LLM: scaling open-source language models with long termism[EB/OL]. [2024-05-21]. https://arxiv.org/html/2401.02954v1.
[27]	刘建国, 张志远, 戴原星, 等. 喷水推进技术发展综述[J]. 船舶, 2023, 34(6): 1-13.
	LIU J G, ZHANG Z Y, DAI Y X, et al. State of the art review of waterjet propulsion technology[J]. Ship & Boat, 2023, 34(6): 1-13. (in Chinese)

语法元素	描述
bdd	块定义图，用以说明系统结构的信息，其中显示的模型元素(模块、执行者、值类型、约束模块、流说明、接口)都是其他模型元素的类型
package	包，用以说明系统中的包结构
block	模块，用以说明系统结构的基本单元，包含模块名称、约束属性、值属性
constraints	约束属性，代表模块值的约束，通常是数学关系

语法元素	描述
bdd	块定义图，用以说明系统结构的信息，其中显示的模型元素(模块、执行者、值类型、约束模块、流说明、接口)都是其他模型元素的类型
package	包，用以说明系统中的包结构
block	模块，用以说明系统结构的基本单元，包含模块名称、约束属性、值属性
constraints	约束属性，代表模块值的约束，通常是数学关系

方法	视图平均
方法	元素数量	Recall	Precision	F1
LLM	11.7	0.362	0.268	0.279
LLM+case	12.6	0.542	0.417	0.449
LLM+case+object	30.8	0.787	0.633	0.673
MbseLLM	28.3	0.786	0.663	0.697

方法	视图平均
方法	元素数量	Recall	Precision	F1
LLM	11.7	0.362	0.268	0.279
LLM+case	12.6	0.542	0.417	0.449
LLM+case+object	30.8	0.787	0.633	0.673
MbseLLM	28.3	0.786	0.663	0.697

方法	项目元素数量	Recall	Precision	F1
LLM	31.3	0.096	0.085	0.090
LLM+case	41.1	0.469	0.316	0.378
LLM+case+object	52.2	0.816	0.432	0.565
MbseLLM	65.6	0.967	0.407	0.573