At 1404, a machine learning process associated with the 3D model is performed (e.g., by machine learning component 106) to predict one or more characteristics of the mechanical device. The one or more characteristics can include, for example, fluid flow, thermal characteristics, combustion characteristics and/or physics behavior. For instance, the machine learning process can perform learning and/or can generate inferences with respect to fluid flow, thermal characteristics, combustion characteristics and/or physics behavior associated with the 3D model. In an aspect, the machine learning process can also be performed based on input data. The input data can include fluid data, electrical data and/or chemical data associated with an input provided to a device associated with the 3D model. The physics behavior can be indicative of behavior related to fluid dynamics, thermal dynamics and/or combustion dynamics throughout the device associated with the 3D model in response to the input data.

At 1406, physics modeling data of the mechanical device is generated (e.g., by machine learning component 106) based on the one or more characteristics of the mechanical device. The physics modeling data can be indicative of a visual representation of the fluid flow, the thermal characteristics, the combustion characteristics and/or the physics behavior with respect to the 3D model.

At 1408, it is determined (e.g., by machine learning component 106) whether the machine learning process has generated new output. If yes, the methodology 1400 returns to 1406 to update the physics modeling data based on the new output. If no, the methodology 1400 proceeds to 1410.