QMT Features: July 2016
Inspecting digital models
Digital inspection with 3D software reduces time to market and improves quality at a lower cost

Coordinate measuring machines (CMMs) are still largely used to measure the accuracy of manufactured parts. They can only assess what is visible and accessible on the surface of the part, and so they are used mainly to measure actual part surface deviation from an accurate nominal model. CMMs do not let you look inside the structure of a particular material, piece of rock, or mechanical part.

X-Ray tomography has been around for close to five decades. The detector and source technology improvements, in addition to reconstruction advancements during the past few years allow for acquisition of high-resolution, high-quality data that are able to deliver important information to characterise a material or detect indications inside a part. From medical CT, to micro-CT, nano-CT, and also Transmission Electron Microscopy (TEM), all ranges of resolution allow you to look at different levels of detail inside the sample. The data acquired from these techniques are digital samples onto which different analyses can be performed in order to better understand the structure of the material and to look at potential defects.

Whether you are looking at a ceramic filter and analysing its porosity and permeability, examining a piece of wood and characterising its strength based on fiber orientation and length, or studying a composite fan blade from a jet engine to find gluing and weaving defects or foreign object debris, all of these characterisations and detections are now possible on these digital models through advances in software solutions, such as Avizo Inspect.

The main CT revolution is the ability to acquire good data that lead to meaningful and accurate numbers that help researchers understand the material, or the manufacturer to more quickly integrate new materials in their production. This new understanding of the material or the part is of utmost importance to increase the development cycle of new materials, to more quickly understand and correct a defect during failure analysis, and to speed up the design cycle of a new part. It also helps control quality during the manufacturing process.

Reducing the development cycle decreases costs and allows for faster adoption of new technologies. The ability to monitor quality increases productivity and allows for delivery of higher yield at the end of the production line. It also provides greater assurance of delivering products that conform to quality standards, reduce the risk of failure, follow increasing safety regulation, and decrease the cost of defects.

The analysis process begins with a visualisation phase. The 3D data is loaded into Avizo Inspect, and different 3D visualisation techniques allow the quality of the acquisition to be assessed. Achieving good numbers relies on good data from the acquisition system, while optimising parameters of the CT acquisition is crucial for producing a good digital model. Type of material, source power, beam path, and phase contrast filter are all important and dependent upon what we are viewing. And, of course, resolution is highly important. Small cracks in a part will not be ‘visible’ if the crack width is smaller than the resolution of the CT. (It has been observed that an indication will be detectable in the digital data if its size is at least three times larger than the CT resolution).

The visualisation phase also allows for correcting artifacts linked to the acquisition process, if necessary. Algorithms can be applied to correct for charging or beam hardening, for instance, or to remove the infamous ‘ring artifact’ caused by a dead pixel on the detector. Looking at the 3D data allows you to begin understanding the conformation of the part and to detect visually obvious defects or indications.
Next is analysis, where meaningful numbers are computed from the data in order to characterise or quantify. Here, the application can perform the same kinds of measurements as produced by a CMM, but on the digital model. Distance, diameters, and angles can be directly measured on the 3D visualisation. This is also where pores are going to be quantified in the porous material, computing their location, size and shape, and distance to the surface of the part. In addition, clusters will be localised. All of this information helps characterise the material by assigning, for instance, properties of permeability or strain. These analyses are more often a succession of image processing algorithms, each brick of processing linked to the results of the previous, generating a workflow from the CT data that will produce a spreadsheet with the numbers to understand the data.

Avizo Inspect introduces a recipe mechanism that allows for automating such quantification or inspection workflow. Once a workflow has been created for a particular type of material and analysis, that workflow can be added to a library of recipes. Identical analysis on similar parts or materials can then be automatically re-played without further development.

Traceability is also of great importance in the inspection process. Tracing where a particular result is coming from, what processing has been applied in order to produce it, documenting the results of an analysis, and archiving and sharing them are important Avizo Inspect functions. The Recipe mechanism allows for automatic generation of reports, where the inspection workflow can export visualisation snapshots of the data and spreadsheets resulting from the quantification or analysis process. Each intermediate result and final result can be archived, so a full trace of the process is available for later review.

This inspection automation mechanism can then be integrated onto the production line. While Avizo Inspect can be used off-line or near-line, it can also be directly connected to the acquisition system. Each time new data are pushed from the CT acquisition, an acquisition service pre-processes the data, splitting it into individual parts in the case of multi-part acquisition, registering data to a reference model, and reducing some acquisition artifacts that may be present. Once a part has been pre-processed, it is sent to an Inspection application that will run a scenario that can be a combination of visual inspection and Recipes. An operator in visual mode can accept or reject a particular step of the inspection scenario, while a part can be automatically accepted or rejected depending on the numbers produced by the Recipe according to some nominal and tolerance information. The inspection scenario is created through a Designer application that defines all of the different steps of the inspection workflow and the combination of visual inspection and Recipes. A Reviewer application allows for reviewing rejected parts and confirming rejection.
Good CT data is nice, but achieving accurate and meaningful numbers is the true aim. Characterisation and inspection of these data through Avizo Inspect reduces the design cycle, shortens time to market for new material and new parts, and allows for early detection of defects, enabling you to look forward to increased yield and better quality out of the production line.
You can now view all QMT Magazine issues on your favourite tablet or smart phone.
Download the free Quality Manufacturing Today App from the Apple iTunes App Store or from QMT Magazine on Google Play.

Rob Tremain Photographer
Click above to see full page display and links to QMT articles.
Vision 2016 ad
Bowers logo
Nikon logo