Rapid recOnfiguration of manufacturing systems: a model-baseD software engineering and human Interaction Coupled approach – RODIC

Reconfigurable Manufacturing Systems (RMS) are currently considered as one of the most promising ways to make companies more adaptable to strong uncertainties. The main idea of RMS is to modify both the physical structure and the structure of the control system of a manufacturing system in order to allow it to adapt to an evolution of its environment or its objectives. To achieve this result, manufacturing systems must be designed in a modular form, and a configuration therefore corresponds to an assembly of these modules, whether they are software or hardware. Several successive phases are carried out during a reconfiguration: (i) detection of the need for reconfiguration, (ii) design of alternative configurations, (iii) choice of the configuration to be applied, (iv)production stop, (v) reconfiguration, and finally (vi) production restart.
The RODIC project aims at improving this process by proposing a tool allowing an operator in charge of the reconfiguration to evaluate more simply and quickly the configurations he wants to test (phase iii). The main strength of the RODIC project is to succeed in crossing three distinct scientific disciplines: cognitive psychology for the cognitive analysis of the operator's activity and the definition of the interaction to be set up with the tool, industrial engineering for the contextualization of the work, the design of the configurations and the definition of the performance indicators with respect to the enterprise architecture, and software engineering for the phase of verification, generation and evaluation of the proposed configurations.
The originality of the RODIC project is multiple: to succeed in proposing an instantiable model of a production system configuration, to allow the generation of a frugal and intelligent control system (as opposed to the current monolithic systems restricting the performances of the system by limiting the perimeter of the authorized configurations), to propose a set of targeted evaluation tools (as opposed to the traditional omniscient flow simulation models, which are cumbersome and time-consuming to develop), and to adjust this cycle as closely as possible to the expectations of the users in production situations.
The tool proposed in the RODIC project can be decomposed in five sections globally successive: the operator, through a first interface, will be able to choose, associate and parameterize a set of predefined modules in order to create his configuration. A model of the configuration will then be processed by Model-Based Software Engineering (MBSE) techniques in order to verify the consistency of the configuration and its compliance with the constraints of the enterprise architecture. A set of evaluation models will then be generated and used in a validation phase of the configuration against the expected performance indicators. These results are then displayed on a second interface in a form that allows the results to be explained, in order to facilitate the operator's decision making. When the configuration is validated, the configuration control system software is generated and ready to be applied on the real system.
The RODIC project will develop two complementary application cases. First, a virtual industrial case will be proposed, dealing with mostly automated systems, allowing to demonstrate the qualities of the tool on industrially validated performance indicators. A second case will be developed within a Teaching Factory on a totally manual system. The developments will be permanently integrated in the Teaching Factory in order to make it a showcase of the project.