Ensuring the quality of foundry production requires the application of a complex system involving a large number of parameters of high variability. One of the main keys for the success of control and prediction systems developed for the industry lies on the ability to integrate and correlate as many variables as possible.
Among the control and prediction tools available in the market we find three different kinds:
- Those that apply to Design, such as numerical simulation.
- Those that apply to Production, like statistical analysis.
- Finally there are the predictive systems that combine the taking of process variables with expert knowledge through artificial intelligence systems.
When it comes to the tools used in the Design, it is known that even though numerical simulation is widely applied to assess the risk of defects it uses standard values to define its parameters. As a result, a same prediction is obtained for variable manufacturing conditions.
In predictive simulation, as a differentiating element compared to the numerical simulation, the metallurgical quality of the metal is also introduced as an input variable, using an special widget of Thermolan® software.
At engineering phase, the geometry of the part, the complete feeding system and the characteristics of the mold are defined. At production stage, working conditions, materials, temperatures and flow rates are determined. From these definitions and a right meshing of the system, the simulations of the filling are launched, verifying at the end of it the different temperature gradients within the cast part that serve as a starting point for the solidification calculation.
By integrating the different isolated masses that form during solidification, the component of the metallurgical quality of the moment, adjusts the risk of the appearance of microporosities with much greater precision. This allows in most cases to reduce the volume or number of feeders in the system, thereby achieving a considerable increase in patter performance and a notable reduction in shrinkage rejection.
Thus, a predictive simulation of this kind offers two complementary functions, for engineering, it is possible to predict precisely the future behavior of a new design by applying the actual conditions of each plant. And for Production, by integrating the geometry of the cast system and the quality of the metal used, immediate action can be taken when an unacceptable level of defect-risk is predicted. Both systems are easy to use and data displayed are easy to interpret, while both can also be adapted to every foundry.
