Models Exchange and Virtual Integration with MATLAB and Simulink
Giacomo Gentile, Collins Aerospace Applied Research and Technology
Alessandro Mignogna, Collins Aerospace Applied Research and Technology
See how Collins Aerospace Applied Research and Technology used Simulink® and the FMI standard to facilitate the internal collaboration between Collins Aerospace teams and streamline external collaboration between OEM and Collins product owners. You’ll see a comparison of the ability to export and import standalone FMUs with the coupling approach of the tools identifying difficulties and benefits. You’ll also explore multiple industrial and research use cases focusing on system and software validation and verification in a virtual environment, including exporting Simulink models towards different external simulation ecosystems and importing third parties’ virtual test benches in Simulink.
Published: 30 May 2022
[AUDIO LOGO]
I'm sure everybody here is aware of the complexities that emerge during the development of large scale cyber-physical system, especially when it comes to the final integration and verification stage when dealing with strict regulations, such as the aerospace one. In the last decades, we saw a growing adoption of model-based design and simulation to support our engineering process. But what does it mean to use model-based design and simulation to support systems integration and systems verification?
My name is Alessandro Mignogna. I'm part of the Italian side of the Collins Aerospace Applied Research and Technologies organization. And if you stick with me for the next 15, 20 minutes I'm going to show you how we are applying model-based solutions and simulation to enable early systems integration and verification leveraging MATLAB Simulink capabilities.
The organization I belong to is fully integrated with Collins Aerospace. That is a worldwide leader in advanced and intelligent solutions for the aerospace and defense industry, spanning among the entire set of systems that you might find today on airplanes as well as in airports.
Before we move to the technical content, let me try to set the scene. Despite of the COVID events, the aerospace industry is living a pretty exciting period. We are in that once-in-a-century moment, where we are called to completely reshape the concept of air-vehicles and airspace.
Think about electrification, hydrogen, autonomy-- novel airspace managements that target the four-hour door-to-door objective in European journeys-- completely new concepts that will require completely new designs and where system complexity will increase dramatically. At the same time, we have to deal with classical challenges, such as safety, quality, cost, time-to-market. In this context, it is clear that digital engineering is not anymore a plus. It is a must.
What makes even more challenging-- the overall system development-- is the increasing complexity of our business organization. Partnership among multiple companies with geographically distributed teams, working on parts of the system that often belong to different domains, such as mechanical, electrical, software, thermal, and so on so forward. Enabling an earlier integration among those parts not only permit to assess the integrated system performance, but also enable a smooth evolution of the overall system development.
And this is clear not only to us. Indeed, the overall aerospace community is asking for it. Here, I'm reporting a few highlights from just some of the most relevant initiatives ongoing today in the aerospace domain. On the top, we have the Airbus and Boeing initiatives, which are pushing more and more model-based systems engineering and virtual integration as the way to interact with their suppliers in future.
On the bottom, a few indicators from the US Department of Defense and the Air Force digital strategy that is strongly pushing model-based and virtualization technologies to support the entire product lifecycle-- from virtual prototypes down to digital twins. To enable this vision, the overall engineering community has been defining standards to facilitate sharing and interoperability of models with multiple partners and among different modeling frameworks.
In this area, we find standards like the FMI, the ED-247, the ASAM, and other standards. The FMI standards, which stands for Functional Mock-up Interface, is the standard we used in the work that I'm going to present in the rest of my presentation.
This is a MODELISAR standard, which aims to streamline the sharing of dynamic models among different tools and partners, contributing to the realization of a model-based, agile supply chain, permitting to create a smooth intercompany collaborative ecosystem and enabling for fast prototyping and pre-competitive assessment with A frames.
MathWorks provide full support to the FMI standard, enabling both an FMI import process that permit to integrate third parties models in both MATLAB and Simulink. And an FMI export process that permits to generate an FMU out of a MATLAB Simulink model, leveraging the Simulink Compiler product. In the remaining part of this presentation, I will show how we used such capabilities in one of our projects.
The work I will present is part of a Clean Sky 2 European project named MISSION, which has the ambition to define methods and tools to enhance aerospace system development, reducing testing costs, and paving the roads to certification by simulation. As Collins, we figured both as technology provider and use case providers, bringing to the project an activation system use case from the Collins Advanced Structures business unit.
The project aims to define a workflow and a tool ecosystem to cover the overall system development, from the early stages down to the final integration and verification ones. The objective is to define a solution to reuse engineering artifact and testing artifact, as well as to enable a smooth transition among the different validation and verification strategies.
From model-in-the-loop to hardware-in-the-loop, with the objective to achieve first-time-right tests for certification, eligibility of virtual tests for certification credits, and dependability and quality assurance. Our contribution in the project focused in the area of virtual-processor-in-the-loop type of verifications, where we used virtualized hardware models to run unmodified application flight code in closed-loop with models of the physical plant of the system.
In the first scenario, we used Simulink as the integration framework. In this context, the physical system plant model has been directly modeled in Simulink, while the other model running the system controller software has been imported as an FMU from an external tool named DESYRE, which is a Collins internal technology asset. The integrated system has been used to perform verification testing and to assess properties of the integrated system.
In a second scenario, we exercise the opposite workflow with a clear objective to assess equivalence among the two approaches. We exported the Simulink physical system model as a Functional Mock-up Interface unit and imported it in DESYRE. That now has the role of integration platform. We ran the closed-loop simulation between the system controller and the system plant and assessed the equivalence among the two different integration solutions.
The engagement with the MathWorks team has been extremely important in this, as well as in other, programs that we are running in Collins. I should say that the MathWorks team has been pretty responsiveness to any issue we reported, as well as pretty open and willing to discuss future needs and prioritization from the Collins point-of-view. This continuous engagement definitely contributed to the success of our programs.
To conclude, we saw how the aerospace community is betting on digital engineering and virtual integration solution to cope with the increasing complexity of our system. And we also saw how the MATLAB and Simulink products permits to realize efficient virtual integration workflow. The final remark is to underline the very good level of collaboration and interaction with the MathWorks team. That has been a key enabler for our success.
This concludes my presentation. Thank you.