The architectural design and architectural transformation method proposed in this paper is aimed at the architecture design and analysis of the complex real-time embedded systems based on the models. The architecture models, as the authoritative data source, provide data for the work of other perspectives in the whole system design process. The functional architectures of the complex real-time embedded systems are described by SysML. SysML and extended FACE Profile are used to describe the logical architecture. Describe the physical architecture through SysML and extended MARTE Profile. Based on the created functional architectures, logical architectures and physical architectures of the complex real-time embedded systems, the method is implemented through model transformation to convert SysML functional architectures, SysML and FACE Profile logical architectures, SysML and MARTE Profile physical architectures into corresponding AADL architectural models automatically, which improves the efficiency and accuracy of subsequent complex real-time embedded system architecture analysis.
There is dense and heterogeneous information exchange between the SoS, System architecture model, and it is meaningful to build a unified SoS and system architecture modeling framework to bridge the gap between the two different architecture level modeling. The paper studied the grid-type SoS/system architecture modeling methods and the two different unified architecture framework building approaches. A unified SoS and system architecture modeling framework that aligns with ISO 15288 is proposed. The modeling framework takes full advantage of MBSoSE and MBSE to save time and cost in the engineering practice.
The planning of modern production systems faces unprecedented complexity challenges in terms of diverse stakeholder expectations, value chain integration, and the adoption of new technologies. Existing studies have shown that formal modeling of production systems at enterprise level has potential to alleviate such difficulties by enhancing stakeholder understanding and improving planning rigor. This paper proposes a domain metamodel for the production system architecture framework to enable formal architecture-centric production system planning. A domain-specific modeling language is further implemented through the SysML profiling mechanism to facilitate and formalize planning processes. The proposed modeling language is further validated using an exemplary aircraft production system.
In response to the increasing complexity of aviation equipment systems, the increasing urgency of the equipment pre-research process and the urgent need for rapid iterative optimization during the aircraft development process, this paper proposed a kind of model-based process for aviation equipment requirement demonstration. Starting from the operational concept, the requirement demonstration process for aircraft equipment is established from system of system layer to equipment layer, then to system layer, then to subsystem layer by modelling and simulation based on MBSE theory. The methods for each layer are proposed to support the agile requirement generation of aviation equipment from mission and task requirements to performance indicators. Through concept confirmation, logic verification and principle verification, the forward design of aircraft equipment complex system is practiced. Finally, this paper verifies the rationality of this method through a case, so as to provide a reference of requirement demonstration for aviation equipment in other scenarios.
The SysML language has demonstrated significant utility in the system engineering field as well as aerospace system design. This paper investigates the limitations of document-based methods encountered in design of a grid fin system and correspondent model solutions. The paper first introduces the fundamental architecture of the grid fin system. Subsequently, it identifies two principal problems encountered during the design process: the difficulty in managing primary components, and the problem of supporting control loop design. The paper proposes a viable solution to these challenges in the grid fin design with model-based method and SysML language. This research contributes to a better understanding to SysML language models on aerospace design in and provides practical insights into addressing the challenges encountered in the design of similar complex systems.
The flight test plays an important role in aircraft development. For flight tests, the flight test schedule directly determines the flight test duration (FTD) and flight test cost. To generate an effective flight task schedule, the task scheduling problem for flight tests, which refers to searching for the optimal arrangement solution for a given flight test task requirement, is proposed and studied in this paper. First, a description and classification of the task scheduling problem for flight tests are presented. Based above, the mathematical model for the basic problem is established. Then, the solving algorithms including the exact and approximate methods are reviewed. Finally, numerical experiments on the task instances of different scales are conducted.
As modern aircraft systems become increasingly integrated, they exhibit a high level of fusion and interconnection between systems, requiring multiple systems to work together to effectively achieve the functionality of a single system. This presents a challenge for validating product functionality in ground tests during the aircraft integration phase. In the face of highly integrated and complex aircraft systems, decoupling the inter-level interconnections within the aircraft is critical to establishing effective product functionality validation. This study draws on principles and tools from Systems Engineering (SE) to establish an inherent inheritance relationship between aircraft product design and aircraft process design of ground tests. This enables effective recognition and decoupling of internal interconnections within the product, and establishes a connection between the Hardware-In-the-Loop requirements and ground tests hardware during the aircraft integration, forming a general method framework for the decoupling “V” model of the aircraft system tests. In the application of ground tests on aircraft for aircraft fuel system, this method effectively analyses the requirements of system test simulation and, combined with the test environment characteristics of the Aircraft Final Assembly Stage, forms a corresponding fuel system test facility.
The development of the system architecture study, especially using the Model-Based System Engineering (MBSE), provides an opportunity of a better modeling and analysis of a system-of-systems (SoS). The Unified Architecture Framework (UAF), developed from the UPDM and the NAF, is a framework which is designed for the modeling of a system-of-systems (or an enterprise). The present study focuses on the application of Model-Based System-of-Systems Engineering (MBSoSE) in a typical military system-of-systems engineering based on the UAF domain meta-model (DMM), in order to solve a typical problem in the military domain. Four major layers are analyzed and modeled: the strategic layer, the operational layer, the services layer and the resources layer. Using the model-based method, the study provides a top-down design of the military system-of-systems, which is an available and flexible solution for the typical military problem, and which leads to the inputs of the system level design.
The current system design approaches for humanoid robots are mostly platform-based, making it difficult to consider the operational scenarios comprehensively. This leads to challenging modifications to the complex humanoid robot before meeting real-world task requirements. To boost the effectiveness and efficiency of task-oriented humanoid robot design, a novel robot design method based on Model-Based Systems Engineering (MBSE) is proposed. Firstly, a top-down model-based humanoid robot development method is presented, which integrates the development process and model to design and describe the requirements, functions, the logical architecture and the physical architecture of the humanoid robot. Then, the development of the search and rescue humanoid robot is taken as an example to illustrate the effectiveness of the method.
This paper introduces the model-based embedded software development and verification method. Taking radar system software as an example, the development process of model-based embedded software is introduced, and the use of models formed in different stages in the development process and its related verification methods are discussed. Model-based embedded software development can make full use of the advantages of model simulation and verification, analyze the correctness of the model intuitively and comprehensively, realize virtual design verification, and promote the deepening of design optimization. Not only can the requirements and design defects be discovered as early as possible, but also the correctness of the design can be guaranteed under the premise of satisfying the requirements.
As a new type of flying car with both land and flight functions, flying car require a complete and scientific development framework to analyze and design its functions and architecture in detail. Based on the MBSE methodology, define the stakeholders and usage scenarios of the flying car from the perspective of the human-machine-context, and use the modeling approach to develop the requirements, functions and architecture of the flying car at the concept stage. The flying car MBSE development framework combs and defines is completed by using model-in-the-loop simulation and human-in-the-loop test environment according to the system of system architecture design thinking, which verified the requirements, functions, interactions, and performance of the concept stage of the vehicle.
This paper presents a method for generating a domain ontology-based model of the logical architecture of a radar system in the context of MBSE. The method consists of constructing a domain ontology for the logical architecture of the system and using it as a framework for a Model Knowledge Base, and then devising a method based on this to automatically generate a radar system logical architecture model from an existing radar system functional architecture model. The method aims to store and reuse domain knowledge of the logical architecture model of a radar system in the form of an ontology and to use radar domain expert knowledge to assist the user in the decision making process during the logical architecture model generation. This approach could improve the efficiency of modeling, speed up the design process of the radar system logical architecture model and assists the designer in trade-off exploration during the model generation process to find a better solution.
This paper proposes an effective approach for applying Model-Based Systems Engineering (MBSE) to radar system design. Specifically, a cross-domain system ontology model is first developed to capture radar design domain knowledge, enabling a common understanding between system engineers and radar designers. Furthermore, a domain-specific architecture framework is defined to provide a big picture view of radar system design (including operational analysis, system analysis, logical/physical architecture definition) and meanwhile organize model-based radar system architecting processes accordingly.
Traditional assembly process modeling methods often exhibit limited adaptability to changes in the production environment, leading to inflexibility when confronted with material shortages, equipment failures, and other production disturbances. In this paper, we introduce a novel adaptive assembly process modeling approach specifically tailored for aircraft manufacturing. Our method differentiates between product-specific assembly constraints and optimal assembly sequences derived from process expertise. This distinction enables the dynamic determination of subsequent assembly stages during actual production while accounting for production disruptions, such as material shortages and equipment malfunctions. By ensuring that product-specific assembly constraints are satisfied, the proposed method permits a controlled degradation of optimal assembly sequences to accommodate fluctuations arising from production disturbances. The efficacy and adaptability of our approach are demonstrated through its successful implementation in real-world aircraft manufacturing scenarios, yielding enhanced flexibility and resilience of the assembly process in the face of unpredictable production conditions.
The model-based systems engineering is gradually applied to the development of complex systems, it has the advantages of accurate transmission of design output, rapid response to design changes, and undemanding reuse of design achievements. In this paper, MBSE is evaluated from three perspectives, including the theoretical system of model-driven research and development (R&D), the requirements of product R&D as well as the requirements of complex system R&D, and the shortcomings of MBSE are sorted out. On this basis, this paper proposes the R&D process of complex system hybrid-driven by model and data, and analyzes the key technologies to be solved.
For airborne computing platform, fault diagnosis model is a very important part of Prognostics and Health Management System. Because of the lack of labels, unsupervised learning is commonly used in fault diagnosis model. However, unsupervised learning method requires too much computing power and storage, so it’s hard to be deployed to airborne computing platform with limited computing power and storage. In this paper, we proposed a model compression method named Optimized Adversarial Distilling Model Compression (OADMC) that combines Knowledge Distilling and Adversarial Learning to solve this problem. In our method, we trained an Artificial Neural Network (ANN) model to imitate the unsupervised model, and use Simulated Annealing (SA) method to find the difference between them. OADMC can made required computing power reduced to less than 25
The traditional avionics system architecture design process of civil aircraft has encountered many problems, such as inconsistent design process, poor traceability, and difficult interface matching. Based on the design characteristics of avionics system and the system engineering design ideas of COMAC, this paper designs a method suitable for civil aircraft avionics systems, and develops an avionics system modeling tool that conforms to this method. The results of model practice indicate that this method can achieve collaborative modeling between systems, real-time interface synchronization, and automation of avionics network design. The modeling results meet the requirements of model design.
To develop a complex, safety-critical system, it is of great importance to identify the safety requirements at the earlier stage of system development. Functional Hazard Assessment technique is commonly used in order to identify top-level safety requirements based on the system functions. In this article, we present a systematic approach to conducting FHA starting from the list of system function, based on which the functional Failure Conditions are identified, then each Failure Condition is classified based on its Failure Effect. Different Failure Effects are categorised to different level of Severity, to which a safety goal is associated. By studying the chain from Failure Condition to Failure Effect, to Severity, and to the associated safety goal, the safety requirements for each Failure Condition can be determined at the end of Functional Hazard Assessment.
A unified Model-based SoSE and SE Tool-Chain is significant and necessary for the dense and heterogeneous information exchange between the SoS, System, and domain engineering models. This paper studies data exchange mechanisms between the SoS architecture tool, System architecture tool, and other engineering tools. A unified Model-based SoSE and SE Tool-Chain framework is given for reference. Compared with other frameworks, this solution uses the SoS and System architecture tool as a hub in the center. It is flexible and fully utilizes the existing modeling tools, which will save costs and time significantly. Ultimately, a helicopter system model as a case study demonstrates the proposal's feasibility.
Model Based System Engineering (MBSE) methodology still exists shortcomings on quantitative requirements. There is no rigorous analysis process. Indexes are loosely connected with other system requirements and the determination of indexes is over-reliance on designer subjective decision. The aims of this article is to study the method for defining performance indexes so that the indexes can link with functional requirements and system general design data can be added into MBSE model to support the analysis. The decomposition process of multi-level indexes is proposed. The requirements of quantitative indexes are verified in MBSE simulation by means of scaling up CPU time of some actions to make the simulation logic and integrating the subject design models into MBSE operational logical model. The feasibility of new analysis and modeling methodology are demonstrated on the MagicDraw software with helicopter system.
