3D Printing Data Management Platform

What you get

The FITMAN “3D Printing Data Management Platform” is a web-platform aiming at alleviating the process of 3D printing through using 3D scanning for supporting the design of 3D products and improving the quality control by analysing available data (generated in the whole process).

The platform targets the Smart and the Digital Factory domains. It is based on four Enablers (FIWARE Generic and FITMAN Specific) integrated and delivered as a virtual machine instance per network by cloud in order to ensure its easy (web-based) and scalable (big data driven) usage.

Why to get it

The quality management in the 3D Printing process is still an open issue in additive manufacturing. There are many parameters that can be measured during the process of printing. Since it can last for hours, the data produced during that process is extremely huge. On the other side, if some problems during the printing happen they can be very costly since a) many hours of work are lost and b) the material is lost as well (not to talk about unsatisfied customers).

Therefore, traditional concepts in quality management, like SPC, cannot be applied very easily in additive manufacturing, esp. because of the big data nature of the produced data so that traditional data processing methods e.g. sampling, cannot be applied efficiently.

The FITMAN “3D Printing Data Management Platform” provides a set of tools and web-based User Interfaces in order to overcome the previously mentioned barriers by providing novel methods for big-data driven quality management. Particularly, the Platform enables the collection and advanced analysis of the data generated during the manufacturing process (i.e. parameters of the process) and from the printed product. Note that both cases are very relevant for the quality since the first set (measurements from the manufacturing process) reflects how stable the process is itself (influencing the costs) and the second set concerns the quality perceived by the customer (influencing customer satisfaction).

The second case is extremely important for scenarios where CAD models are not available for 3D Printing, but 3D Scanning is rather used for the creation of models used in printing process. In this situations we assume that printed product can be scanned after printing using 3D Scanners. We assume that in these situations scanned models are used as golden standards and the quality of the product can be defined by comparing printed products (through scanning them) with golden standard.

Documentation

Preface This document is the Open Specification of the FITMAN "3D Printing Data Management Platform" that is released as Virtual Machine and composed by the integration and expansion of four components coming from FIWARE Generic Enablers Catalogue and FITMAN Specific Enablers Catalogue.Specific Enabler (FITMAN-DIPS). FITMAN Open Specifications of the “3D Printing Data Management Platform” are released under the same terms and conditions of FI-WARE Open Specifications - see FI-WARE Open Specification Legal Notice (essential patents license) and FI-WARE Open Specification Legal Notice (implicit patents license).For more information in the FITMAN project itself, please refer to the FITMAN website
Copyright Platform and integration components: Copyright © 2015 Nissatech, DFKI and DataPixel About Generic and Specific Enablers please refers to the ”Terms and Conditions” section.
Overview The FITMAN "3D Printing Data Management Platform" is a web-based, integrated platform for the advanced processing of the data collected in the 3D printing process in order to enable a better quality management in the process as a whole. The problem of managing quality in modern manufacturing is related to the traditional view on the quality as sampling and abstraction, which was quite acceptable for the era of statistic process control (SPC) where only a few selected parameters were selected and monitored for the variations (in the process). In the 3D Printing process the problem has become very emerging since the process can be very long and many parameters influence the quality. The functionality provided by the Platform is based on two important principles: 1) for all well-established processes, it can be learned from the past what is the normal behavior/quality of the process/product (it doesn’t mean that it is an optimal behavior) 2) normal behavior/quality of a process/product is described by all parameters that somehow influence the process (although it is possible to preselect features that influence the process mostly) The quality (of a process/product) is then defined as the closeness to the normal behavior/state (process/product). The FITMAN "3D Printing Data Management Platform" supports the realization of this notion of the quality by supporting both above mentioned principles through: - learning the model of the normal behavior from the huge amount of past data (application specific), - comparing the model with the real-time data in order to cluster the current production/product wrt. desired/expected quality and - visualization of results. It means that our methods will enable collection of past data, support learning models (it is application specific) and real-time comparison of process data with the model.
Detailed Architecture and Specification As visible in the following picture the FITMAN “3D Printing Data Management Platform” is composed by four integrated components:
  • FITMAN Specific Enabler "Dynamic CEP" (FITMAN-DyCEP)
  • FITMAN Specific Enabler "Dynamic Visualization & Interaction" (FITMAN-DyVisual)
  • FITMAN Specific Enabler "3D Scanning Storage & Visualization" (FITMAN-3DScan)
  • FI-WARE Generic Enabler “COSMOS”
In the figure below the overall architecture of the FITMAN "3D Printing Data Management Platform" is reported. The user interacts with the system through the 3DXML interface, where the data will be visualized in the context of the initial (printed) object. As explained in the figure, the data from 3D Scanner (3D Scan SE) contains the deviation (map) calculated as the difference between the printed object and golden standard. This data is used for the (real-time) determination of the quality of the produced object (DyCEP SE), based on the previously defined patterns. Usually the patterns should be based on the models that are learned using some analytic methods which are application specific. The Platform supports the calculation of the models in cloud-based environment (COSMOS GE). The results of the comparison are presented as 3DXML visualizations (DyVisual SE). Beside the data about the product, the Platform supports the processing of the process data that are generated in the 3D Printing process. Main features of FITMAN "3D Printing Data Management Platform" from a user’s point of view are the following:
  • Add new training files to project
  • Remove training files
  • Set-up training parameters
  • Add metadata to the training data
  • Set up training schedule or run training at the moment
  • Monitor training execution and logs about previous executions
  • Review models and previous versions
  • Add detection files, files upon the anomaly detection should be executed
  • Get visualized results
 

3dPM

For more information about “3D Printing Data Management Platform” components refers to the following addresses:
  • FITMAN Specific Enabler "Dynamic CEP" (FITMAN-DyCEP)
    • http://catalogue.fitman.atosresearch.eu/enablers/dynamiccep
  • FITMAN Specific Enabler "Dynamic Visualization & Interaction" (FITMAN-DyVisual)
    • http://catalogue.fitman.atosresearch.eu/enablers/dynamic-visualisation-and-interaction-dyvisual
  • FITMAN Specific Enabler "3D Scanning Storage & Visualization" (FITMAN-3DScan)
    • http://catalogue.fitman.atosresearch.eu/enablers/3dscan/documentation
  • FI-WARE Generic Enabler “COSMOS”
    • http://catalogue.fiware.org/enablers/bigdata-analysis-cosmos

Information

For more information about “3D Printing Data Management Platform” components refers to the following addresses:
  • FITMAN Specific Enabler "Dynamic CEP" (FITMAN-DyCEP)
    • http://catalogue.fitman.atosresearch.eu/enablers/dynamiccep
  • FITMAN Specific Enabler "Dynamic Visualization & Interaction" (FITMAN-DyVisual)
    • http://catalogue.fitman.atosresearch.eu/enablers/dynamic-visualisation-and-interaction-dyvisual
  • FITMAN Specific Enabler "3D Scanning Storage & Visualization" (FITMAN-3DScan)
    • http://catalogue.fitman.atosresearch.eu/enablers/3dscan/documentation
  • FI-WARE Generic Enabler “COSMOS”
    • http://catalogue.fiware.org/enablers/bigdata-analysis-cosmos

Terms and Conditions

Experimentation/testing within the scope of the FI-PPP FITMAN “3D Printing Data Management Platform” asset is available to the Parties signed to the FI-PPP program under the conditions established in the FI-PPP Collaboration Agreement. External availability
  • FITMAN Specific Enabler "Dynamic CEP" (FITMAN-DyCEP) is released under the Apache License, Version 2.0
  • FITMAN Specific Enabler "3D Scanning Storage & Visualization" (FITMAN-3DScan) is realized under the FreeBSD License
  • Following components are required for realizing the analytics part
    • Apache Hadoop, open source project under Apache v2.0 license
    • Apache Oozie, open source project under Apache v2.0 license
    • Apache Cassandra, open source project under Apache 2.0 license
    • PostgreSQL, open source project under PostgreSQL license
    • FastDTW, algorithm open source implementation by Stan Salvador, under MIT license
    • JFreeChart, open source, free software, under LGPL v3.0 license