Mixed Lots are to be sampled at Enable Small Series - App-based Effort Reduction during adaptive Testing in the Production of Variants (APProVe)

03/05/2019

Contact

Name

Jonathan Greipel, M. Sc.

Workinggroup Leader

Telephone

work Phone
+49 241 80 28383

E-Mail

In variant production, a 100% inspection of all inspection characteristics of manufactured components due to small batch sizes is common. This is associated with a high testing effort and leads to bottlenecks in the production of testing machines. This results in high production costs and time.

  Group Picture of the research team Copyright: WZL (f.l.t.r.) Hüttemann (WZL), Dr. Schultz (Q-DAas), Huber (iqs Software), Dr. Kellermann-Langhagen (FQS ), Dr. Gutensohn (PFW Aerospace), Greipel (WZL), Herzogenrath (Lasuscher Präzisionstechnik), Radeck (Q-Das), Chishnjak (Tebit ), Kerbl (TCG UNITECH)

In large series production, this effort is reduced by a sampling inspection that only refers to key characteristics. To use the sampling inspection, the key characteristics, the inspection scope, and the inspection frequency are determined using statistical methods and high personnel costs. This is not possible in variant production due to the small lot sizes and the increased personnel costs caused by the variants.

To solve this problem, the new research project APProVe (App-based effort reduction in adaptive testing in the production of variants) is developing an app for automated effort reduction in test planning for variant production. The aim is to create mixed batches of variants and thus achieve a sufficient batch size for random testing.

The project is supported by the AiF, specifically the FQS Forschungsgemeinschaft Qualität e. V. For two years the Chair for Production Metrology and Quality Management at the Laboratory for Machine Tools and Production Engineering (WZL) of the RWTH Aachen - under the direction of Prof. Dr.-Ing. Robert Schmitt - will develop algorithms based on machine learning for the identification of key characteristics and the definition of mixed lots. These are implemented in an intuitive app that allows companies producing variants to perform adaptive random sampling.

We want to achieve that even SMEs, which often manufacture in small quantities, can reduce their inspection costs by means of a combination of new machine learning methods and traditional sampling inspection plans. The competitiveness of SMEs will thus be significantly increased while quality remains the same.

Dipl.-Ing. Guido Hüttemann, Senior Engineer and Head of Department at the Chair of Production Metrology and Quality Management and responsible for the subject area Model Based Systems

In the research of algorithms and development of the app, a project-supporting committee consisting of industrial partners who are active in the areas of consulting, the provision of CAQ software and the production of variants in small series supports the chair. Members are Q-The | Hexagon, iqs Software GmbH, Transfact GmbH, Tebit GmbH & Co. KG, GFE Präzisionstechnik Schmalkalden GmbH, Lauscher Präzisionstechnik GmbH, PFW Aerospace GmbH, TCG UNITECH GmbH and PFW Aerospace GmbH.

During the kick-off meeting in April 2019 at the FQS Forschungsgemeinschaft Qualität e. V. in Frankfurt am Main, Germany, it was discussed with the companies, which requirements exist for an app for cost reduction in adaptive testing in the production of variants. It was pointed out, for example, that the algorithms should take up human input, e.g. when certain characteristics of a product must always be tested at the customer's request. In addition, the algorithms should always indicate the risk that a reduction of the testing effort is associated. This gives the company the opportunity, firstly, to decide whether this should be entered into and, secondly, to discuss its acceptability with its customer. An orientation to current standards such as DIN EN 9138 Aerospace - Quality management systems - Statistical product acceptance requirements and ITAF 16949 was explicitly requested.

The app to be developed and the research results will be made available to participating companies. The research project will make a significant contribution to the development of a VDI Guideline VDI/VDE GMA Expert Committee 1.21: VDI Guideline 2600-3 "Adaptive Test Planning".

Further information can be found on the website of the project APProVe.