INTEGRADDE project aims to develop an end-to-end Digital Manufacturing solution, enabling a cybersecure bidirectional dataflow for a seamless integration across the entire AM chain in Direct Energy Deposition pilot lines.
The final goal is to develop a manufacturing methodology capable of ensuring the manufacturability, reliability and quality of a target metal component from the initial product design. Data will be captured in the Digital Thread and used to shorten AM process development times and enhance reproducibility and quality of part production in support of qualification objectives.
DataPixel contributes to INTEGRADDE by defining the strategy needed in the post-processing stage. It is of critical importance to identify which areas of the produced part need to be corrected. In order to perform this analysis, DataPixel makes use of its very precise optical sensors to obtain an accurate 3D representation of the shape of the produced part, which is then compared to the intended CAD model so disparities can be located. These deviations indicate whether there is an excess or lack of material, and define the strategy needed for correction. The dimensional information is stored so the process can be refined by learning from its own outputs.
Digital approach for Additive Manufacturing
The term “digital thread” refers to the information and information path that is gathered and stored when manufacturing a single part. Thus the ‘digital thread’ emphasises the required interoperability across the whole AM process. This digital thread will structure the information coming from all the stages in the additive manufacturing process in order to optimize the outcome.
The final stage in the additive manufacturing process is post-processing and verification of the part. The information required includes part tolerance, finish requirements, mechanical property requirements and material properties. Specifications for postprocessing of the part are conceived within the design and planning phase and implemented at the end of the execution phase. Following the testing and qualification step, it is needed a system capable of storing the compiled information from the postprocessing stage in order to be reintegrated in the early stages of the manufacturing process. Conventional postprocessing operations to finish the part and improve its properties can involve final machining of surface roughness to finish near-net-shape parts, heat treatment to improve properties or reduce residual stresses.
In order to perform these activities with the highest possible accuracy the system needs to integrate a dimensional inspection tool to extract which areas need to be corrected as well as other NDT techniques to identify the status of the material and verify if the produced components comply with the quality standards.
INTEGRADDE is establishing a solution for the post-processing stages of the additive manufacturing procedure that consists in the implementation of a digital thread where the information gathered downstream the manufacturing process can be used on the design and simulation stages. In this way the post-processing stage has a twofold approach, in one hand, it is used to correct the defects encountered after the analysis of the object and, on the other hand, this information is used to refine the design of the produced part to improve the results of the additive manufacturing stage and also get better predictions for the outputs.
In this context, DataPixel contributes to the post-processing activities by including their optical 3D scanning OptiScan sensors. Optical scanning systems allow very high-density 3D point clouds to be obtained at very high speed. The 3D point cloud is the basic element to initiate more complex calculation processes, like obtaining dimensional parameters and mathematical descriptions of the surfaces. DataPixel designs and manufactures the OptiScan family of high-speed optical sensors capable of digitising objects in a very wide range of materials and surface finishes, reaching accuracies up to 5 μm. With OptiScan optical sensors, it is possible to obtain high-precision 3D digitising for reverse engineering, virtual metrology, automated inspection and robot guidance applications.
Metrology sensors are the solution for this post-processing stage. DataPixel’s optical sensor included in the project is capable of digitising the produced components of the laser deposition process obtaining a high-density point cloud with up to 50 million scanned points. This allows to have a very precise approximation of the real shape of the produced object for its analysis.
In line with the above, with a booming market position and providing a high technological level compared to similar solutions, Innovalia Metrology markets a metrological software service called M3. In essence, this platform constitutes high-performance software for the capture and analysis of point clouds resulting from automatic scanning processes for the reliable and efficient acquisition of 3D information in different materials. Basically, this tool is made up of 3 components:
· M3 Gage: full 3D capture integration system capable of operating with different types of sensors and components.
· M3 Server: solution for highly efficient, secure and flexible management of digital component information for the storage of massive 3D point clouds.
· M3 Tablet: solution for workstations that allow mobile and portable virtual metrology operations to be carried out for convenient digital analysis of 3D measurement information and evaluation of the production procedures.
As part of the M3 platform and metrology assistant, the solution is also capable of providing integrated software-based automation tools for the customization of the metrology process, significantly reducing uptime, and increasing dimensional quality control productivity.
M3 software and its point cloud analysis capabilities provide enough potential to analyse and extract geometrical elements, as well as the ability to compare point clouds and CAD models. The tool provides the possibility of extracting geometric shapes and through its analysis engine, extract relevant information for decision-making in the manufacturing process of mechanical components.
Clearly defined standards and requirements make it easier for companies to meet what their consumers consider the quality and they improve the overall vision of what the company should work towards to. Modern quality operations consist of complex cyber-physical systems that are required to constantly gather and use digital manufacturing data. Often multiple players throughout the supply chain employ one or more proprietary software products, resulting in a critical communication break between supply chain partners. The data relevant to each partner must be intelligently organized and communicated. DataPixel works under Quality Information Framework (QIF): an XML-based, CAD-agnostic platform and standard. QIF seamlessly defines, organizes, and associates quality information including measurement plans, results, part geometry and product manufacturing information (PMI), measurement templates, resources, statistical analysis, etc.
The post-processing stage starts with the digitisation of the produced component in order to obtain a high-density point cloud.
Additionally, the CAD (STEP) is enriched by adding GD&T (Geometrical Dimensioning and Tolerancing) semantic data and automatic elaboration of the Measurement Plan. Based in GD&T, the point cloud is automatically analysed by the DataPixel virtual metrology module. The metrology digital thread is interoperable by means of the integration of metrology standards, in particular QIF (Quality Information Framework). Metrology results are then available for close loop optimization of the process.
Once the digitisation is completed, DataPixel uses their metrology software to align this scan with the digital model used for the design of the component. During the course of INTEGRADDE a new tool has been developed under the range of this software to obtain better results for this alignment. This tool is using [TV1] Reference Point System approach and substitutes the previous existing method based on a best fitting of all the points in the cloud. Instead, this new method is based on semantic elements, it extracts geometrical surfaces from both the CAD model and the point cloud which are, then, aligned with each other using a geometrical best fitting of these surfaces.
This new method developed for alignment delivers much better results than the previous existing method, obtaining a much more precise alignment.
To continue with the post-processing activities, once the model and the point cloud are aligned, DataPixel uses their metrology software to compute the deviations between both models. The output of this process is a disparity map, which represents the deviation of each digitised point of the object with respect to the model used for comparison. These deviations can be graphically shown as a 3D representation with a colour scale assigned to detect the areas where there is a lack or excess of material when compared with the reference model. But not only excess or lack of material is detected, DataPixel’s metrology software also analyses the geometrical surfaces in the point cloud to assess their dimensional characteristics, for example, in the case of cylinders it would analyse the diameter error, shape error, eccentricity or axis inclination.
The main purpose of this analysis is to detect and quantify the regions that need to be corrected. This is because a hybridisation of subtractive and additive processes is intended. In this way, the regions in red represent an excess of material that will have to be corrected via a machining process, and blue areas represent a lack of material that would have to be added using additive manufacturing. This is the main contribution of optical sensors to the digital thread. The information is used to define the machining strategy in working coordinates.
DataPixel in collaboration with ESI is also evaluating how accurate are the predictions of the outcome of the AM manufacturing process. For this, INTEGRADDE is developing a simulation methodology that estimates the final shape of the produced component based on the conditions that the object undergoes during the manufacturing procedure from its initial state.
Thanks to DataPixel’s evaluation of this simulation, using the method described above, the simulation model has gone through an iterative optimization process, adjusting the simulation parameters according to the information extracted from the physical part. After this optimization, the prediction given by the simulation software is able to produce a better estimation of the final shape of the physical object than the one obtained with the CAD model used for design.
Currently, DataPixel has analysed 4 additive manufactured components produced by AIMEN for INTEGRADDE project and 2 simulations produced by ESI to improve the quality of these procedures. In the following figures it can be seen the results obtained for the manufactured components and identify which areas needed to undergo a corrective procedure.
The first 3 components are a batch of T-coupons produced by AIMEN made of steel using LMD with powder. The dimensions of the 3 coupons are (110mm x 120mm x 112mm). To scan these components DataPixel has used and OptiScan sensor attached to a robotic arm. Prior to the scanning a path plan has been designed to perform the optimum scan of the component. After analysing the components, the deviations encountered do not exceed 0.2 mm in any case.
The jet engine component is also made using LMD with powder techniques. It is manufactured using Ti6242. The internal cylinder has a diameter of 200 mm. To scan this component DataPixel has used and OptiScan sensor attached to a robotic arm. Prior to the scanning a path plan has been designed to perform the optimum scan of the component. After analysing the components, the deviations encountered do not exceed 2 mm in general. Some deviations where encountered in the position of the lateral cylinders as well as in the lateral window. These deviations were associated to a misalignment in the robot coordinates. Also, thanks to DataPixel’s analysis of geometrical elements in the point cloud, it was detected that the internal cylinder’s axis was tilted. This inclination can be related to the thermal distortions suffered during the manufacturing process. In this way, dimensional metrology proves to be effective for detecting errors in the manufacturing process and helps to optimize and adjust the laser deposition techniques.
The next step in the project will be the application of these techniques on the industrial pilot lines included in the project, where this post-processing strategy could be adapted to the specific needs of each industry. The objective is to integrate a production line in which the information is shared between stations so that each station can be optimised gathering all the inputs from the whole cycle, in this way the whole process can be automatised and become more efficient.
Metrology data will help the industrial pilot lines achieve the desired final configuration for the manufactured parts within very thin tolerances, adding value to the additive manufacturing solution.
AM is a complex multistage process which can, if not properly controlled, result in defects and the risk of failure of parts in service. By including NDT methods in-process to find defects as they happen, parts can be inspected during the building process, it enables the AM process to be halted. Many of the existing non-destructive evaluation (NDE) standard procedures applied to conventionally forged and moulded metal components are generally applicable to parts made by AM. However, specific challenges must be addressed by newer AM-specific standardized NDE procedures.
INTEGRADDE develops and deploys the concept of digital thread enabling a seamless dataflow from product design to final manufactured metal part; linking design, simulation and modelling, building, online control and inline inspection, and post-processing for the manufacturing of metal AM components. It combines existing experimental manufacturing and inspection processes with on-development CAx software tools to create a multistage manufacturing concept enabling the DED of certified parts. The next steps in INTEGRADDE is to continue integrating the inspection systems in the pilot lines focusing on industrial target requirements to achieve a reliable, interoperable robust and zero-defect manufacturing. The objective is to convert the industrial cases into self-adaptive systems with in-line quality assurance and minimise the consumption of resources.
A crucial aspect in INTEGRADDE is to advance in the implementation of the metrology interoperability standards, particularly, QIF as well as CAD enrichment with GD&T. Thus, next step in the digital thread will be to validate a common frame for the interoperability of the corresponding inspection systems.
The implementation of a dimensional metrology stage in the post-processing activities enables additive manufacturing to improve the dimensional quality of the produced parts identifying which areas need to be corrected so that the final shape of the object can be achieved with a great accuracy. Also, it provides information to include on a feedback loop so that upstream processes in the production line, like design or simulation of results, can be refined and upgraded. This iteration will allow achieving a better output from the additive manufacturing stage.
Digital twin objects generated by DataPixel and ESI, and interoperability of the respective digital objects, are giving INTEGRADDE invaluable information for process preparation and monitoring.
The digital thread concept is validated by proving the ability of the systems to reintegrate information coming from the post-processing stage into the development activities. In this case, metrological information is helping to redefine the simulation models as well as to improve the precision of the laser deposition.
DataPixel S.L. was born in 1999 with the aim of designing, developing and manufacturing systems and solutions in the field of 3D computer vision and dimensional metrology for industrial and professional applications.
DataPixel focuses its activity on three main axes:
· Implementation of inspection and quality control systems on the production line using robotic systems and 3D vision digitized sensors.
· Software and applications for processing and analysing 3D point clouds and data from optical sensors.
· Consulting, design and specific development of applications and systems to measure dimensional and geometric inspection for the aeronautical, automotive and electronics sectors.
Its strategy focuses on the development of solutions adapted to the needs of companies, driven by the new requirements and challenges of the manufacturing industry in automotive, aeronautics, energy, rail and electronics. Thanks to their optical sensors, non-contact measurement and 3D digitizing systems, the dimensional information that will be generated and processed to use for different purposes. These non-contact sensors allow to work on shiny or dark coloured surfaces without the need for sprays. The exclusive laser, optical and electronic design provides the best synchronization with the machine resulting in high quality precision and repeatability.
DataPixel is aware that non-contact measurement and 3D digitizing systems are redefining the way industry generates, processes and uses the dimensional information throughout the entire product life cycle. By introducing digital twins based on digitized 3D point cloud and the application of Virtual Metrology and Inspection Automation Solutions, industrial processes are significantly improved. DataPixel helps production companies to reduce development time and overall production costs and improve product quality.