Seamless model integration with flow engineering

Flow Engineering is a modern software for model-based requirements management. I have finally found some time to get to grips with this browser-based tool. What stands out with Flow is the seamless model integration. As it deals with design models, Flow follows Model Driven Development (MDD) rather than MBSE.

Update 2025This topic has become so important that I am writing a book about it. You can read here, why it is so important and here is the book website where you can actively participate. Product Velocity can contribute.

Flow Engineering

Pari Singh originally founded Flow Engineering to develop rocket engines more efficiently. This worked so successfully that Flow made a pivot. Instead of offering engineering services, Flow Engineering became a tool manufacturer.

Flow is based on several core ideas:

• Seamless integration of models and simulation results into requirements management
• Fast and intuitive Decomposition of systems
• The tool is designed for collaboration, combined with a high level of user-friendliness

At the end of this article I have included a 5-minute video that shows the model integration. In particular, you can see how with one click in seconds:

• Flow verified using stored model requirements (weight from the CAD model)
• Flow visualizes the dependencies between models,
• A new value triggers the recalculation of three concatenated models,
• Among other things, this chain changes the geometry in the CAD model,
• Flow extracts the new value (weight) from the CAD model in order to verify the requirement again.

We were the fastest design consultancy in the world for rocket engines. (...) The fastest in the world was 8 weeks and we could do it in two. And the reason we could do it so quickly is we basically built tools that automated and accelerated the design process.

Systems Engineering Podcast

Requirements management in Flow

At first glance, the module for Requirements management to the typical tabular view of a requirements tree for similar tools:

• The tree structure can be seen on the left
• A request has a text and various fields

But a closer look reveals a few details. In particular Nozzle Mass Flow Rate has a value highlighted in color ≤ 700 kgwhich is displayed in the Value is repeated once again.

The following screenshot shows what is behind this:

• The upper area shows the editor. When writing, Flow recognizes the text "less than 700 kg".
• If I accept the suggestion by pressing Enter, the text is transformed into a parameter consisting of three parts: (1) Interval (less than or equal to), interval limit (700) and unit (kg).
• The lower area shows the saved and edited element in which the transformed text can be seen
• Flow compares the recognized value from the request with a concrete value (actual value). I will explain where this comes from in a moment
• Flow performs real-time verification, which is indicated by the green checkmark.
• Flow has also recognized that the units of measurement of Value and Actual Value do not match (kg vs. kg/s) and displays a corresponding warning.

Where does the "Actual Value" come from?

Flow can compare values from the requirements with design values. The example just shown is a Target Check and has been 600 kg/s is set. This value can be set explicitly in the project under Design Values can be set. A target check can also be compared with a value from a model, more on this in a moment.

In addition to Target Check, there are a number of other options, as can be seen in the screenshot on the left.

Integrate models

Flow currently has seamless integration with the six tools shown on the left. In addition to these integrations, Flow also enables the import of other formats.

The integrations shown on the left use either API keys or OAuth to communicate directly with the APIs of the corresponding tools. In many cases, the connection only requires a few clicks.

A simple example of model integration is Google Sheets. Simple models can be created quickly in a spreadsheet. The following table shows a simple lever system that amplifies the force via a pivot point. Lines 2-4 are the input values from which the spreadsheet calculates the force amplification in line 5 using the formula =B4/(B3-B4).

In Flow, the model must be linked once and values selected. We can then use the above-mentioned Actual Values Select values from the model.

4-minute video: Feedback loops with model chains

The following short video shows how the whole thing works using a realistic example. From minute 4 onwards, it is particularly impressive to see how the CAD model changes based on the modified parameters:

Conclusion

Model integration is not new, but all too often integration is not integrated to the same extent as can be seen with Flow. When experimenting with Flow, it was clear that usability must have been the top priority. This is a key advantage for teams where simulation is a high priority.

For some teams, the cloud-first approach could pose a problem: Flow runs in a web browser and seamless integration can only be realized with a small number of tools that are also web-based. Flow does not only support Simscalebut also Matlab/Simulink, but only Simscale has a similarly seamless integration.

You can tell that Flow Engineering comes from the aerospace industry, where component-based thinking still prevails. It is not yet clear to me how well interface management works with Flow and whether functional decomposition also scales. These are topics that I would like to investigate further in the future.

But at least when it comes to linking requirements and design models, Flow is the best I have seen on the market so far. For users who recognize themselves in this problem, Flow is definitely one of the best solutions I know.