CN-122021263-A - MBSE-based complex coupling scheme weighing method

CN122021263ACN 122021263 ACN122021263 ACN 122021263ACN-122021263-A

Abstract

The invention provides a MBSE-based complex coupling scheme weighing method, which comprises the steps of constructing a coupling matrix based on hierarchical elements and structure dependency relations of a MBSE model to realize static coupling quantification, constructing a conditional probability network model by taking the coupling matrix as a priori structure constraint and combining equipment simulation data to realize dynamic coupling quantification, determining a coupling degree threshold according to functional requirements and nonfunctional requirements of equipment and combining a preset architecture knowledge base and a design rule base to preliminarily screen candidate schemes, comprehensively grading by adopting a weight fusion algorithm combining a probability weight variation coefficient method and an entropy weight method based on the static coupling quantification result and the dynamic coupling quantification result, and screening out an optimal scheme.

Inventors

CAO WANGBIN
YUAN WENQIANG
ZHANG JIANHAI
NIU BIAO
Qin Feiwei

Assignees

杭州电子科技大学

Dates

Publication Date: 20260512
Application Date: 20260108

Claims (10)

1. A MBSE-based complex coupling scheme trade-off method, comprising: constructing a coupling matrix based on hierarchical elements and structure dependency relations of the MBSE model to realize static coupling quantification; the coupling matrix is taken as a priori structure constraint, and a conditional probability network model is built by combining equipment simulation data, so that dynamic coupling quantification is realized; determining a coupling degree threshold according to the functional requirement and the nonfunctional requirement of the equipment by combining a preset architecture knowledge base and a design rule base, and primarily screening candidate schemes; And comprehensively scoring by adopting a weight fusion algorithm combining a probability weight variation coefficient method and an entropy weight method based on the static coupling quantization result and the dynamic coupling quantization result, and screening out an optimal scheme.
2. The method for balancing complex coupling schemes based on MBSE according to claim 1, wherein: The construction of the coupling matrix comprises extracting elements of a system layer, a subsystem layer and a component layer from a MBSE model, constructing a coupling relation matrix of each level and a mapping matrix between the levels, identifying multi-disciplinary cross coupling conditions in the model, constructing a multi-disciplinary coupling matrix, fusing the coupling relation matrix of the levels, the mapping matrix between the levels and the multi-disciplinary coupling matrix, and forming a multi-level enhanced coupling matrix; The matrix fusion comprises the steps of constructing mapping matrices of a system layer-discipline, a subsystem layer-discipline and a component layer-discipline, adopting preset weights to carry out expansion fusion on the matrices, and carrying out normalization processing to obtain the multi-level enhanced coupling matrix.
3. The method for balancing complex coupling schemes based on MBSE according to claim 2, wherein: Initializing a Bayesian network structure based on the coupling matrix, acquiring module interaction data through multi-round system simulation, performing feasibility filtration, structure optimization and threshold pruning on the Bayesian network, simplifying the network structure, performing conditional probability learning based on the simplified Bayesian network, estimating conditional probability by adopting frequency statistics on discrete parameters, performing modeling and reasoning by adopting Gaussian network regression on continuous parameters, calculating the coupling degree between nodes, solving an average value, and obtaining a dynamic coupling quantization result; The feasibility filtering is based on a physical formula and causality relation to restrict the node connection direction and existence, the structural optimization is used for controlling the calculated amount by limiting the upper limit of the number of parent nodes, and the threshold pruning is used for eliminating network edges with the weight smaller than a preset threshold.
4. A complex coupling scheme weighing method based on MBSE is characterized in that the preliminary screening comprises the steps of extracting a main body from functional requirements according to an overall target, determining a key path for completing the target through functional decomposition, converting nonfunctional requirements into specific measurable indexes to achieve mapping from performance to coupling characteristics, determining the minimum coupling degree required by a model by combining the architecture knowledge base and a design rule base, and eliminating candidate schemes with the coupling degree exceeding the minimum coupling degree.
5. The complex coupling scheme trade-off method based on MBSE of claim 1, wherein the comprehensive scoring comprises constructing an original decision matrix based on the static and dynamic coupling quantization results and performing a normalization process.
6. The method for balancing complex coupling scheme based on MBSE is characterized in that the comprehensive scoring further comprises the steps of calculating objective weights of two quantized results through an entropy weight method, calculating a weighted average value and a weighted standard deviation of the two quantized results, taking the ratio of the weighted average value and the weighted standard deviation as a probability weight variation coefficient, carrying out weighted summation on the objective weights and the probability weight variation coefficient to obtain comprehensive weights, and finishing scoring.
7. The method for balancing complex coupling scheme based on MBSE of claim 1, wherein the equipment simulation data is interaction data generated by each module on a behavior layer or a data stream layer in the running, simulation or task execution process of the MBSE model.
8. The complex coupling scheme weighing method according to claim 3, wherein said initializing said Bayesian network structure comprises regarding parameters in MBSE model as random variable nodes, regarding system response as target nodes, and establishing initial association relationship between nodes.
9. A MBSE-based complex coupling scheme trade-off system, comprising: the static coupling quantization module is used for constructing a coupling matrix based on hierarchical elements and structure dependency relations of the MBSE model to realize static coupling quantization; The dynamic coupling quantization module is used for constructing a conditional probability network model by taking the coupling matrix as a priori structure constraint and combining equipment simulation data to realize dynamic coupling quantization; The scheme primary screening module is used for determining a coupling degree threshold value according to the functional requirement and the nonfunctional requirement of the equipment by combining a preset architecture knowledge base and a design rule base, and carrying out primary screening on candidate schemes; And the comprehensive weighing module is used for comprehensively scoring by adopting a weight fusion algorithm combining a probability weight variation coefficient method and an entropy weight method based on the static coupling quantization result and the dynamic coupling quantization result, and screening out an optimal scheme.
10. A non-transitory computer readable storage medium, characterized in that the storage medium has stored thereon a computer program which, when executed by a processor, implements the method of any of claims 1-8.

Description

MBSE-based complex coupling scheme weighing method Technical Field The invention belongs to the technical field of model-based system engineering, and particularly relates to a MBSE-based complex coupling scheme weighing method. Background The digital driving of high-quality development has become a core trend of the time development, the depth and breadth of digital construction are continuously expanded in the field of aerospace equipment development, and model-based system engineering (MBSE) has become a core methodology for supporting the whole life cycle development of equipment, and is widely applied to various links from concept design to test verification. Along with the fusion development of technologies such as generation type artificial intelligence and MBSE, the realization that a large number of differentiated MBSE models can be automatically generated through natural language description becomes reality, and how to accurately screen out an optimal scheme from the multi-dimensional and multi-scene models becomes a key problem to be solved in the current aerospace equipment development process. The development of the aerospace equipment has remarkable industry specificity, the disciplinary fusion range is extremely wide, the core disciplines such as aerodynamics, structural mechanics, thermodynamics, material science, electronic engineering and the like are covered, the knowledge fusion of a plurality of emerging crossing fields is also involved, and a complex multidisciplinary coupling system is formed. Meanwhile, the aerospace equipment is used as a highly complex integrated system, is composed of a plurality of subsystems and components which are functionally associated, and needs to operate in extreme environments such as high vacuum, strong radiation, extreme temperature and the like, the complexity of system operation is further aggravated by the uncertainty of dynamic environment, and strict requirements are put on the safety, reliability and stability of the equipment. The traditional scheme weighing method is highly dependent on experience judgment of engineers, is high in subjectivity, is difficult to comprehensively consider potential influence caused by multi-disciplinary cross coupling, is also difficult to accurately evaluate fluctuation rules of equipment performance in a dynamic environment, and is difficult to meet the requirements of refined development of modern aerospace equipment. MBSE support system requirements, design, analysis, inspection and verification of full-flow activities by applying a modeling method, and the core advantages of the method are three aspects of multi-science synergetic effect, information consistency and mathematical expression capability. By means of a unified modeling platform and language, MBSE can break communication barriers of engineering teams in multiple fields of machinery, electronics, software and the like, realize model sharing and real-time linkage, ensure consistency of data sources of all levels, provide support for whole-flow traceability, and support simulation calculation and analysis of a system by means of formal description modes. However, in practical applications, the current MBSE related tools and methods still have certain limitations, for example, the integration of functional architecture modeling and physical performance analysis is not smooth enough, and the data transmission between different modeling languages is easy to have the problems of information loss or format incompatibility, so that the complex coupling relation generated by multidisciplinary intersection is difficult to be precisely quantized. The conventional MBSE-based scheme evaluation focuses on performance index analysis of a single dimension, lacks comprehensive consideration on static structure coupling and dynamic operation coupling of a system, is difficult to objectively reflect the influence of a coupling relation on the overall performance of equipment, and cannot effectively and rapidly screen actual demands of mass models. Based on the core advantages of MBSE and the current application limitation, and combining with the development requirements of multidisciplinary coupling, high complexity and high reliability of aerospace equipment, the existing scheme weighing method has difficulty in meeting the accurate and efficient screening requirements. What is needed is a MBSE model scheme weighing method capable of fully adapting to complex coupling characteristics of aerospace equipment, which effectively solves the problems of difficulty in multidisciplinary coupling quantification, difficulty in collaborative evaluation of static and dynamic performances and the like, provides technical support for high-quality and low-risk development of the aerospace equipment, and promotes continuous and healthy development of the field of the aerospace equipment. Disclosure of Invention Aiming at the defects and shortcomings in the prior art, the invention p