NVision develops new ways to visualise defects in turbine blades...

The Company:
A major manufacturer of power systems entered discussions with NVision to see how our excellent visualisation technologies could be deployed to assist them in identifying defects within turbine blades. As a closed entity, these components are difficult to test. There is no option to break into the component as it may be returned to the power system or alternatively there is a risk of disrupting the integrity of the test.

The Project:
The current method of evaluation involves requires the inspection of data from an x-ray machines and this is generally presented as a large series of tiff images. A technician works through each individual image and looks for potential defects (represented as white spots). It is not immediately obvious whether the observed white spot is part of a defect or data noise and thus it is necessary to look at the next few tiff images in the series to see if they are associative. Naturally, this is an onerous task and there is a risk of missing a potential flaw in the component..

To overcome this, with process guidance and sample data provided by the manufacturer, NVision has developed an exciting new prototype process to help NDT engineers read data from x-ray machines (or other internal cavity scanning tools) and display them in such a way that defects can be easily identified using GPGPU rendering techniques. This is achieved by turning the read data into a complex mesh, then associating it with a material such as glass in a professional modelling package (in this case 3DS Max) then importing it to our cloud based service called RealityServer™. The defects then are clearly visible as "bubbles" in the structure and that makes it easier for the NDT engineer to evaluate the component. As the application exists in the "cloud" the user just needs a broadband connection and an internet browser to access the session. Therefore, it is possible to collaboratively and globally share the experience with end-users, suppliers or consultants. Neither host nor the sharing community requires expensive native software or a high-performance workstation to access the session

The next step in process is to develop the application to include a quantitative tool to categorise the component in terms of a rule-based algorithm that would interrogate the test-sample based on the spread of defects or by individual defect size. A front-end application will permit NDT Engineers to securely login to a collaborative area where the sample can be viewed or manipulated as well as providing a means to advise and consult via communication threads. Moreover, we are working to reduce the complicated work-flow and convert the tiff-data to its transparent form in one conversion process. This will reduce the level of human interventions required and the tool will deliver its findings more quickly.

Read more case studies...