Noyron is presented as a foundational computational model for engineering created by Leap 71, designed to encode expert domain knowledge, physics and manufacturing rules in a single coherent framework. The model incorporates thermal behavior, logic and related data so that it can support general engineering tasks rather than a single niche application. It is proprietary software that is developed internally and is described as expanding in capability as new insights are gained from designing and manufacturing highly elaborate machinery.
The system relies on PicoGK, which is an open source kernel that provides robust tools for generating complex geometry. Noyron forms the base layer for several more specialized computational engineering models, including Noyron RP for rocket motors, Noyron EA for electromagnetic actuation and locomotion, and Noyron HX for heat exchanger design. These specialized models sit on top of the same core framework, indicating a modular approach in which one core model supports multiple domain specific tools.
A central function of a computational engineering model is to generate the geometry of an object, but Noyron explicitly goes beyond geometry by attempting to predict performance comprehensively. Given a defined set of input parameters, Noyron will construct a design that is expected to perform correctly by inferring mechanical movements, thermal behavior and other performance characteristics. It aims to behave like a human engineer by using distilled expert knowledge to match the requested functionality, then feeding back observed behavior from real world tests or simulations to refine future iterations. The output from Noyron includes detailed geometry, manufacturing process parameters and associated files, such as post processing paths and guide geometry, along with summaries of predicted performance parameters, numerical fields and physical data suitable for direct transfer to simulation systems. The model is described as being under constant development and training as part of an ongoing effort to align its predictions more closely with real world behavior.
