Content

• Free Webinars
• Contact Form
• Why CT?
• Statistical Analysis of CT Inspection Results
• Hexagon’s Solution
  • More about VGSTUDIO MAX and VGinLINE
  • More about Q-DAS qs-STAT
  • Setting up an Inspection Plan
  • Automated Inspection and Data Transfer
  • Process Qualification and Statistical Analysis with Q-DAS qs-STAT
• Benefits
• White Paper Download

Do you have questions or need more information about Volume Graphics and/or Q-DAS software? Then simply fill out the form and we will contact you.

CT, a relatively new method for the quality inspection of industrial parts, has become a staple of many quality laboratories and inspection processes. Using an X-ray source to create 2D projections of a part from many different angles—which can then be reconstructed into a full 3D data set—it is well suited for capturing not only the outer surface of an object, but also inner geometries, material structures, and defects (e.g., pores).

This makes CT uniquely suitable as a basis for comprehensive inspections that both include dimensional metrology known from coordinate measurement machines or optical scanners and incorporate NDT methods, such as porosity analyses, wall thickness analyses, or calculating deviations from reference geometries on the inside of a part.

CT is also following the trend towards inline and/or near-line inspection: Advances in CT system technology—such as more powerful X-ray sources and faster, more sensitive detectors—have steadily reduced the amount of time necessary to acquire data of sufficient quality for analysis. With automatic handling and loading systems, it is often possible to scan a large number of parts per day.

Regardless of the inspection approach you select, CT will allow you to derive a wealth of data from your parts in a short amount of time. This data can not only be used to determine the quality of the individual part, but also holds great potential for insight into the behavior of the part-producing machine—and on a higher level into the entire production process itself. The tool required to achieve this potential is the statistical analysis of inspection results and their associated meta data, such as cavity number and production time.

CT-based metrology works on the determined surface of the scanned part and delivers the same type of results as other dimensional measurement methods, i.e., geometric dimensions and evaluations of shape and position tolerances. Statistical analysis of quality-relevant features, like the positions or straightness of a connector pin, or the diameter of a clearance hole, can be used during machine and process qualification to accurately predict whether parts of sufficient quality can be consistently produced using the selected method and parameters. And once production has been initiated, statistical trends in the same inspection results can be used to detect issues that may occur over time, such as tool wear or the accumulation of form coating in casting, so that mitigations can be implemented before part quality is actually affected.

Hexagon’s solution for the qualification and statistical analysis of CT-based quality inspection processes is the combination of the following software packages:

• VGSTUDIO MAX or VGinLINE from Volume Graphics for the quality inspection of individual CT data sets and
• Q-DAS qs-STAT from Q-DAS for the statistical analysis of the inspection results

Both software packages are leading products in their respective fields. In the quality lab, they can be combined into a semi-automated workflow in which CT scans of multiple parts are automatically analyzed in batches in VGSTUDIO MAX and the results are then manually imported into qs-STAT for process qualification or statistical analysis.

For inline inspections in production, a fully automated inspection and analysis workflow can be achieved using VGinLINE, which automatically analyzes each new CT data set as it is generated by a CT scanner on the shop floor, followed by an automated import of the inspection results with Q-DAS Q-DM (Quality Data Management) and subsequent statistical analysis in qs-STAT.

Inspections of your CT data can be set up using VGSTUDIO MAX, where individual scans of your parts can be visualized and examined. You can use GD&T to measure relevant features just like you would on a CMM. The extraction of an actual part geometry from a CT scan in VGSTUDIO MAX (known as “surface determination”) is a leading reference in the CT world, and its accuracy has been validated in several scientific studies. Furthermore, the algorithms for fitting reference geometries to an object’s surface in VGSTUDIO MAX and VGinLINE have been verified by both PTB and NIST.

Once defined, inspection plans can be applied to other scans of similar parts in an automated way. The automation capabilities of VGSTUDIO MAX can be easily used to inspect batches of scans, and inline inspections of an ongoing production process can be achieved using the VGinLINE automation system. Regardless of whether you are running semi- or fully automated inspections, inspection plans can be connected to your Q-DAS statistical analysis software. All available GD&T features can be exported in the Q-DAS ASCII transfer format. And all these exported results can be combined with customizable views of the assumed inspection results, which can be used to display easy-to-digest reports directly in Q-DAS qs-STAT. 

Q-DAS qs-STAT is a software package designed to statistically evaluate production-relevant quality information and to assess systems and processes. Significant and well-established statistical evaluations provide the basis for assessing and continuously improving processes in industrial production. Furthermore, the standards and guidelines included in the software provide the necessary basis for the interpretation of the results.

qs-STAT offers the right toolset for quality and production engineers at every stage of production, including

• machine qualification,
• process qualification, and
• statistical process analysis

Before production starts, production machines need to be set up. qs-STAT offers support through a wide range of tools and analyses and documents this ramp-up process.

A machine capability analysis is used to determine whether a machine is capable of producing parts so that the quality parameters (e.g., the dimensions of the measured features) are well within the tolerance limits in a stable manner. For this purpose, a sample of approximately up to 20 parts produced on one machine in one sequence without variations in tools, operators, materials, machine parameters, or ambient conditions is examined. For more expensive parts, it may be more efficient to perform pre-run studies with a lower number of parts beforehand to obtain an initial overview.

For each measured feature, the time series of measured values can be analyzed both visually and with statistical tests for undesired outliers or non-random trends. Variations in measured values with respect to the range of the tolerance band and the position of the measured values within the tolerance band is described by the potential and critical machine capability indices, Cm and Cmk. Cm and Cmk values are statistical capability measures, i.e., they describe whether a production machine is capable of producing parts in the quality required. In order to locate potential problems—such as a possible need for a final tool or milling path correction—it can be helpful to show the location of the measured features as well as the time series of their measured values in an overview image of the component, which can be exported from VGSTUDIO MAX and imported into qs-STAT using the Q-DAS ASCII data transfer format.

A pre-production process normally leads to a complete process capability study to evaluate the influence of all kinds of variables such as materials, machine parameters, operators or ambient conditions on the production process.

Based on the sample inspection, statistical measures are calculated to qualify the process. The results of an approved process capability study determine the warning and control limits on the basis of which a process control can be installed according to the statistical process control (SPC) approach. It does that in combination with other factors (e.g., the evaluation strategy), which is similar to a decision tree and can be used to determine both the best suitable distribution model and the most suitable Quality Control Chart (QCC).

The above method has two advantages over a 100% measurement and/or quality control based on a defined tolerance. First, the number of measurements to be taken will be reduced significantly, and second, statistical control via the control limits. A control limit violation may take place long before a tolerance limit is reached. The action can therefore be taken much earlier at a stage in which all produced products still are in top quality conditions.

Evaluation strategies are either based on standards, association and corporate guidelines (e.g., BMW, GMPT, Robert Bosch, or Volkswagen), or an individual approach. The selected evaluation strategy is included and used automatically, which leads to comparable and reliable results comparable. Aside from the pre-installed evaluation strategies based on standards and association and corporate guidelines, qs-STAT also offers the possibility of creating customer-specific strategies—for instance, to make results comparable across all products within the company. Thus, Q-DAS qs-STAT, as all Q-DAS software, is fully able to be validated according to the given standards.

Once the pre-production phase has been successfully completed, series production starts. To ensure a stable and reliable process within the required Cp and Cpk values, continuous requalification needs to take place, as even an assessed process may change over the time.

Whenever a process no longer meets its expected Cp or Cpk value, a deeper root-cause analysis may be required to determine the reasons behind deviations. qs-STAT supports quality engineers with a rich set of analysis options for determining root causes or correlations and uncovering dependencies that might not have been seen during the process capability study.

Besides standard supplementary information, such as machine (number) or operator, there are many process parameters that determine the results of a process. For instance, the temperature or the cooling time may have a significant impact on the output of an injection molding process. Analyzing the dependency of quality results on process parameters or ambient conditions such as pressure or humidity (when available) help quality engineers to identify the levers for process improvements. Furthermore, a part will usually not be produced on one machine, but on multiple machines at the same time. qs-STAT also supports quality engineers in determining whether the unexpected measurements results can be traced back to a specific machine based on the data.

Univariate correlation analysis can uncover interdependencies between two different characteristics, such as if characteristic 1 is increasing and characteristic 2 is also increasing. The correlation analysis checks for relationships between the two characteristics, as well as how significant and of what kind they are. Q-DAS qs-STAT offers different filters and selection criteria to detect trends and deviations. The analyses allow users to reach conclusions about significant influences caused by different factors, such as machine, batch, operator, tool, and temperature.