SBIR-STTR Award

The casting process produces flaws of various types and sizes in turbine blades. Each flaw potentially degrades the life of the blade, but for economic reasons many initial flaw classes are permitted on the basis of acceptance criteria developed empirically by engine manufacturers. The problem with this approach is that it requires data obtained from experience. It cannot predict blade life in a new engine. Moreover, many blades containing a characteristic flaws can be in use before such a flaw is found to reduce substantially expected life. We propose to create a highly interactive, computer-aided design system to simulate the performance of a blade containing specific flaws and flaw sizes at arbitrary locations. The system will have the following innovative features: simulation based on a true geometric representation of the blade via solid modeling. A sophisticated, topology-based data structure to support linkage to the solid model, fast interaction, and accurate representation of evolving flaw shapes. The ability to specify flaws to arbitrary shape, including non-planar, size, and orientation at arbitrary locations in the geometric model. Automatic local remeshing to simulate flaw growth. Modualr encapsulation of fracture mechanics theories and growth-rate models for predicting the evolution of a flaw. State-fo-the-art techniques for scientific visualization via computer graphics

Keywords:
turbine blades flaws defects fracture mechan design simulation computer modeli computer graphi

The casting process produces flaws of various types and sizes in turbine blades. Each flaw potentially degrades the life of the blade but for economical reasons many initial flaw classes are permitted based on acceptance criteria developed by engine manufacturers. These criteria are developed empirically. The SBIR topic to which Fracture Analysis Consultants responded sought the creation of a turbine blade life-prediction capability, based on computer simulation of flaws. During Phase I, we created the prototype version of a highly interactive, comprehensive, and integrated system, TURBINE-FLAW, to simulate the performance of a turbine blade containing specific flaws and flaw sizes at various locations. The fundamental difficulty that had to be overcome was that the geometry of the object changes during crack propagation. A general predictive tool has to be capable of modeling the actual geometry of an existing flaw, predicting its growth, updating the model geometry in response to growth, and performing a new analysis. This cycle has then to be automatically repeated as the flaw grows. In this proposal we identify the Technical Tasks necessary to enhance TURBINE-FLAW from a prototype system to a reliable production quality software system, usable by practicing engineers.