3D modeling of Digital Mock-Up based on the integration of heterogeneous data – METIS

In terms of working methods and operations, the project focuses on various scientific topics as follows:
• Acquisition, processing and integration of large volumes of data, geometric or not, spatially localized: the data will be of such a heterogeneity (3D point clouds, photographs, reference catalogues, calculation results, maintenance notes, old version digital models, etc. . ) that the project will develop a methodology for sorting and matching and innovative SEO
• Identification of business components in a set of heterogeneous data: A library of components should be defined and created characterized by keys (signatures) for their identification in a large volume of data. The scientific contribution here is to find a way to «sign« complex components and store these signatures.
• Research in the large volume of data, the nomenclature of the studied product and the associated position matrices: when a component is recognized identification (signature), the first scientific contribution will be to characterize its dimensions, its position, and its orientation and so on using geometric and non-geometric data collected directly in the data set but also with the beforehand available knowledge about the component (ex. spoils if forged component etc.)
• Generation or modification of the digital mock-up: from the known bill of material, parameterized CAD components associated with different expert models can be generated

The first 18 months of the project allowed clarification of the vocabulary and identification of highly structuring concepts for the project. First, the concept of signature was identified as the keystone required to identify and integrate data even when they are highly heterogeneous (Bruneau et al., 2014). Signature typologies were created using the example of a motor piston (textual signature, a 2D signature ...). In addition, a mapping of heterogeneous data relative to their nature and their positioning in the product development cycle has been proposed.
The second choice concerns the infrastructure of the integration platform. The latter was designed with very modular in order to manage and integrate data using different technologies / algorithms, and without assuming any media / format of the data beforehand.

The project partners, in particular the industrial partners DeltaCAD and IFPEN, are active since many years in various industrial sectors (Transport, Energy , Mechanical, ...) who encounter problems of Reverse Engineering in various forms and who require dedicated IT solutions to address the issues. The need is manifest and has already been identified by many manufacturers who would like to have a software solution as METIS to handle Reverse Engineering in an optimal way by integrating all dimensions of the subject.

The potential impact of the METIS tool is largely related to its level of scientific and technological innovation, including the ability of the tool to leverage heterogeneous information sources and to reduce the need for interactive expertise. The research project aims to investigate in this direction, identifying and propose possible solutions, or propose new possible lines of research for the future.

M. Bruneau, A. Durupt, L. Roucoules, JP. Pernot, B. Eynard Towards new processes to reverse engineering digital mock-ups from a set of heterogenous data
International conference on Graphics Engineering June 19th –21th, 2013 Madrid, Spain INGEGRAF – ADM – AIP PRIMECA
M. Bruneau, A. Durupt, L. Roucoules, JP. Pernot, H. Rowson,
A methodology of reverse engineering for large assemblies products from heterogeneous data
TMCE 2014, May 19-23, 2014, Budapest, Hungary
M. Ouamer Ali, F. Laroche, A. Bernard, S. Rémy
Toward a methodological knowledge based approach for partial automation of reverse engineering
CirpDesign April 14th 2014, Milan, Italie

This project addresses the field of Reverse Engineering.

Today, Reverse Engineering is widely used by the manufacturing industry in order to capitalise knowledge which hasn’t yet been and which becomes of prime importance to ensure future evolutions of products.

Many applications are possible, like, Reverse Enginering existing products in order to modify/combine the components (this is the IFP Energie Nouvelle application case, with vehicles and motors) or else the maintenance of mechanical products which have a very long life cycle (planes, ships, nuclear power stations, offshore oil platform, …), when it is required to redesign and reproduce a component initially manufactured by a sub-contractor which has since disappeared.

Generally speaking, commercial solutions currently available on the market enable extraction of design intention information based solely on the parts geometry in order to rework it (RAPIDFORM XOR d’Inus Technology for example).

Scientific literature also includes articles aimed at handling reverse engineering of components or small assemblies. In this case, the geometry of the part is often obtained through 3D scanning or measurements. The surface of the part is hence sampled by a cloud of points and/or a surfacique mesh. Currently, it’s the analysis of this geometry through a knowledge filter which enables reverse engineering.

METIS, on the other hand, aims at providing reverse engineering solutions for larger and more complex mechanical assemblies (either by their dimension or their part count) like engines, vehicles (which are at the heart of IFP Energies Nouvelles issues)… For these assemblies, it is delicate and inefficient to sample the entire geometry. For example, for an automotive, performing a complete scan of all its components is extremely tedious, when not impossible. Indeed, to do so would require a full disassembly of the vehicle and a part by part manual scan : as the available 3D scanning systems when operating without CAD data (so no automatic path generation) have limited capacities when handling parts with complex shapes, the scanning process becomes mostly manual.

The idea underlying the METIS project is that for Reverse Engineering of large and/or complex assemblies, purely geometrical data isn’t sufficient. METIS aims at providing a solution which integrates all the information which is available for the considered product (including geometrical) in order to handle it and extract the digital mock-up.

Therefore the goal is to develop methodologies and associated tools which will allow users to create and maintain over time semantically rich digital mock-ups integrating the knowledge related to the considered product. This mock-up will be obtained from heterogeneous information, sometimes incomplete, like 2D drawings, photos, clouds of points, maintenance notes, simulation results … or even sometimes an older version of the CAD model.