Using simulation in product development is nothing new. There are many simulation tools that help users solve practical issues.
Beyond achieving success with important singular issues, engineers and their companies have been trying to achieve a global solution at a system level. This type of solution encompasses subsystems or components from domains such as mechanical, electrical, thermal and controls, and includes embedded software.
The new design paradigm for complex systems is referred to as systems engineering. The goal is to identify a technical baseline to design, verify, validate, qualify and manufacture the targeted system. ISO 15288 defines as many as 25 processes involved in systems engineering. Unifying these diverse processes has become the role of model-based systems engineering (MBSE).
Complex systems always require more functionality, components and models. This leads to a couple of questions: Do we need a larger variety of languages to represent different models? Do we need more tools to handle the variety of models? The answer is not necessarily. The best solution is the oft-used phrase less is more.
What does less mean in this context? It implies that we need an engineering-converged solution. Instead of using multiple models from various tools, a unified model representation enhances knowledge capture efficiency across multiple disciplines. Modelica, VHDL-AMS, and Simscape are languages that have been developed along this direction. The open-standards-based Modelica language has especially attracted industry interest since its emergence in the 1990s.
Deploying MBSE based on open standards requires consistent support of modeling and simulation techniques. But, many tools and technologies that enter the MBSE domain early in the development process hinder the further application of simulation technology at a higher level.
Successful model-based design (MBD) deployment at one division usually generates a “film” wrapping the model, leaving it difficult for other divisions to leverage. This leads to isolated islands of technology within the company. As a result, there are lots of processes needed to transform the model and data into something that can be used at a system level. Usually, this implies mountains of work to overcome the many walls between models; most often, companies fail at completely removing those walls.
The Functional Mockup Interface (FMI) offers the possibility of doing less transformation between models and data, resulting in high model reusability. Models with 0D, 1D or 3D content can be modeled in various languages and still integrated into a lean simulation process using FMI under the MBSE paradigm. Here are some of the permutations of the FMI:
A standardized model unit such as an FMU (Functional Mock-up Unit) could easily be shipped to different tools that support the FMI standard. Then FMUs could be connected at the model exchange (ME) and co-simulation (CS) levels according to FMI 2.0.
Mature technologies such as robust design or parameter tuning on the specific model representation can be migrated to an FMU, allowing engineers to access the model and apply their skills and expertise. There are already some successful trials importing FMUs into Excel and coupling FMUs with Matlab/Simulink.
The FMU can be used for verification and validation, such as in a hardware-in-the-loop simulation environment.
With the rapid increase of smart device applications, there is also a push to get the FMU running on portable terminals. In gas and oil plants, for example, maintenance engineers can use a mobile device to monitor a sensor's readings and compare them with the model output. This will definitely transform the way engineers work.
As systems continue to become more complex -- with more and more information and content -- open standards with their efficient new technologies provide us with a leaner way to make it all work, making less is more much more than a cliché.
Rui Gao holds a PhD in Mechanical Engineering of Kyushu University in Japan and an MSc in Aeronautics Automatic Control from Northwestern Polytechnical University in China. Rui has over 20 years experience in system modeling and simulation, control system design and reliability analysis. He is active at Modelon's office in Tokyo. Rui is also member in the steering committee of the System Modeling and Simulation Working Group (SMSWG).