Software & Web Applications

Architecture

Our offer is built on reusable components :

• numenGo|MPS is an advanced Media Property Server which is used to compute efficiently all media properties
• numenGo|FFC is a library of models for mass and heat transfer
• numenGo|FL is a high level library to be used in a system simulation platform (SimulationX, AMESim or simulink)
• the Python API allows to use all components in python scripting environment

On the top of these components, we can easily create business oriented solutions for our clients, to fit their specific needs in terms of modeling and ease of use.

Design methodology

We use Model-Driven Development process to create and maintain our solutions. In this process, we create a description of all our models, according to a metamodel that we established.

Thermofluid modeling is specific:

• Fairly causal approach
• Very limited possible symbolic reduction
• Many complex equations and algorithms, requiring efficient debugging

Our objectives were the following :

• High performance
• Limited maintenance
• Object-oriented approach
• Seamless cross-product integration, to promote our models in various environments

Since, we develop most of our models according to MDA (Model Driven Architecture).

This approach relies on describing all models in a specific and neutral “metamodel”. The implementation is then done in C++ (very good performance, portable, object-oriented, debuggable…).

From this description, we can automatically generate all kind of interfaces using “model transformations”. Thanks to this methodology, all our models have native interfaces for :

• C/C++
• Python
• FMI
• Modelica
• AMESim
• S-functions and simulink models

We can generate new bindings on demand, but most of all, each time we add a new component in our model libraries, it is automatically usable in all these environments, reducing drastically the maintenance.