Skip to main content


In the overall digitalisation context, i.e. having computers assisting all our tasks, it is difficult for a computer to understand the content of a document. To do so, the information that is in the document must be expressed in a different way, more structured (diagrams, tables), or with less expressive power but with stronger grammatical rules than the natural language. This way of expressing information is called a model in this context. The goal of MBSE is to replace most of the documents by models. And therefore to replace document editors with model editors, which are specific of the topic that the model is about. And finally to transfer these models from tools to tools, all along the project development, and to enrich them, e.g. from requirements models, to architectural models, to design models, to test models, etc.

The ability to create a flow of information transported from model to model, and from tool to tool, is called Digital Continuity. The fact that the tool can exchange models and understand the model of another tool is called interoperability of tools. The information included in the document has a meaning, which is called its semantic. Likewise, the meaning of models is also called their semantic. Semantic can also be represented by a model, in this case it is called a "conceptual data model" , or an ontology. An ontology is a representation of the concepts that are used by information, and of the relationship between the concepts. It gives a (semantic) structure to the information. It is not possible to create a simple and complete ontology able to describe the information contained in a document written in free natural language. But, as a model is a specific part of the information, it is possible to create an ontology for a particular (set of) models.

Tools can exchange information by pure item-to-item translation of the content. If we compare to language translation, the english (shooting in one's foot) will be translated in a foreign language by a gun sending a bullet in a foot: the tools will not be interoperable. Instead, if the two tools share the concept (in their own ontology) of "making a mistake that backfire on yourself", then the translation will go up to the ontology level to translate forth and back the right expression in the right language. This is called semantic interoperability.



ESA has set out strong ambitions for the European space industry in its latest Technology Strategy, including a 30% schedule reduction, and one order of magnitude of cost efficiency improvement. These targets cannot easily be reached solely through product technology improvement. Process change is needed!

Digital continuity is considered to be a major technique to enable:

  • a continuous flow of information to reduce source of errors throughout the life cycle;
  • the connection of all the engineering data to improve traceability, allowing change impact analysis and assessment of the impact of non-compliance to be performed more quickly; 
  • the availability of a single source of truth for each engineering data to allow system engineers to make accurate and up-to-date system trade-offs, avoiding late system changes and reducing cost;
  • the interface to many system design analysers to increase the proportion of verification by design, reducing the amount of testing required, and decreasing the overall project timeline.



It is System Engineering which is using preferably models rather than documents. It is a set of modelling techniques (methods, languages, tools). It is used to build a product, and describes all the facets of the product. Therefore, all the elements describing the product (its artifacts) can be manipulated by computers at a low level of details. The computer can therefore make all sorts of detailed relationships between the products artifacts. It can relate together two different disciplines using the same data (mass is defined by mechanical people and used by software people), it can relate a requirement with its implementation and with its test , it can relate a customer need with a supplier solution.

There are existing tools which are called Product Lifecycle Management (PLM) and Product Data Management (PDM). They  manage the products through their artifacts (“business objects”), but only globally. These tools do not enter into the artifacts, nor capture their relationship. For example, a functional model is identified as an artifact, but its relationship with e.g. the Computer Aided Design model, or the avionics architecture, is not addressed.

It is the role (and added value) of MBSE to make these links, and to look inside the artifacts. By doing so, MBSE transforms the product data into information.

Digital Continuity

Commercial tools start to address some of these problems through PLM and PDM, but in their own context, which does not offer the required industrial interoperability across tools, and may create unwanted vendor lock-in. The challenge is to implement digital continuity, which aims to create the above-mentioned links between the artefacts, across heterogeneous tools, by use of model-based techniques. These links need to be created in the three dimensions of system engineering:

  • across disciplines
  • throughout the life cycle
  • along the supply chain

It is important to note that, in MBSE, SE (system engineering) must remain the objective, MB (model-based) is ‘only’ an enabler. Therefore, the technology must be thought as ‘MB for SE’, considering model-based engineering as a lever for system engineering.



System engineering is the art to define, design, realise and verify a system in a three dimensional space: across disciplines, along the lifecycle, through the supply chain (ECSS-E-ST-10C)[1].

Mastering system complexity is at the root of our needs. Initially manageable by a single or a small team of system engineers, the complexity of some of our space systems starts to exceed what can be thoroughly encompassed by a human team. Moreover, some traditional assumptions (e.g. the a priori separation into two predefined space and ground segment), may not hold anymore, thus creating even bigger systems. System complexity includes architectural complexity (due to multiple configured, tightly coupled elements), functional complexity (due to a high number of interrelated mission requirements), organisational complexity (due to the number of distributed actors). This complexity cannot be handled any more via documents.



Because the functions become extremely complex, entangled, and full of exception and dependencies, or because the concept of operation has significant system implications, textual descriptions quickly become inadequate to describe the system behaviour completely and consistently.

->The system engineer needs formal representations beyond textual descriptions

The amount of information contained in the recent systems cannot be manually distributed across disciplines and along the supply chain.

-> The system engineer needs a controlled exchange of data supported by automated mechanisms.

Change impact, traceability and relationships between the various elements or aspects of a system cannot be achieved through an unorganised set of information.

-> The system engineer needs a structured knowledge of their system

The exchange of system data by documentation, email, or individual files makes it impossible to provide the right version of the right information to the right person at the right time.

-> The system engineer needs a single source of truth, consistent by construction, from which appropriate viewpoints may be extracted to communicate about the system.

The evolution of the system becomes more and more difficult to predict, as the evolution of key data may not be visible.

-> The system engineer needs, for progress monitoring, a dashboard with system engineering information such as main budgets (mass, power, link, etc.), key parameters, coverage statistics, trends, etc.

The history of the system elaboration is sometimes lost, and the lack of traceability does not allow, at the end of a project, requirement rationales or trade-off justifications to be recalled.

-> The system engineer needs traceability across the system of both formal derivation, satisfaction and justification of requirements versus design, as well as links to collaborative and informal aspects such as design and trade-off decisions at review milestones. 

The amount of interface with other stakeholders, customers and suppliers is increasing and cannot be handled only by loosely coupled documents.

-> The system engineer needs to create consistent baselines of system engineering data for communicating with customers (e.g. reviews) and suppliers (e.g. subcontracting a system element) and a mechanism to allow for the distribution of reduced models to external parties to protect intellectual properties.

The reuse of elements from previous projects is difficult when done from documentation, and the changes needed to adapt products to new requirements are spread over unidentified pages of documentation.

-> The system engineer must be able to initialise a system with a reusable set of consistent data (libraries), or add reused data during the project, with the capability to identify the required deviations and branches.

In the early stages of a project (i.e. Phases A, B1 and B2), the lack of an (as much as possible) exhaustive relationship model (dependencies) among all entities of the project and system, and the lack of the level of maturity of the represented/available information, quite often push the decision making process towards expensive and complex solutions. Such solutions might be even revisited and/or disregarded in later phases of a project (Phase CD).

->The system engineer must be able to grasp all relationships (even though they are not yet defined/consolidated) among all entities of the system (incl. entities of the project) and their maturity, in order to build a proper risk assessment and decision-making process.



The space domain has some particularity that can make the deployment of MBSE difficult:

  • The projects are performed by wide consortia involving various industries, themselves sometimes made up of other consortia (for example, the Galileo 2G project includes 11 consortia). This may be due to geographical return constraints, European commission context, ‘best practice’ constraints, industrial product policy for dual sourcing, etc.
  • Each space actor may or may not be trained in MBSE.
  • Each space actor may or may not have its own MBSE standard/formalism.

Integrators and suppliers do not necessarily share the same modelling levels, the same modelling goals, nor the same engineering constraints. This results in possible large differences and needs in terms of methods, semantics, model usage, modelling concepts, rules, practices, and tooling.

Sharing technology would be obviously the most efficient way, as all the engineering stakeholders would use the same methods and tools to fulfil their needs. Unfortunately, the diversity of engineering needs, and the heritage in technology investments of all the stakeholders, make such a situation unrealistic.

The next technical solution is to bring together formalism, languages, and even meta-models. But this is not sufficient, because it often happens that the model building and conformance rules are different, the semantics have discrepancies, and the process or the tools generate constraints. In addition, the space engineering community will never be able to converge on a ‘one size fits all’ metamodel that covers all concepts and is accepted by all space stakeholders.

Instead, semantic interoperability should be targeted, rather than shared formalism, standards or languages, sometimes leading to impoverishment of the engineering quality. It is ultimately more efficient to rely on automatic transformation tools than on modelling constraints on both sides. Native entity practices and methods should be preserved so as not to hinder engineering quality.