Improving robustness of ITER maintenance schedules
Independent thesis Advanced level (degree of Master (One Year)), 10 credits / 15 HE creditsStudent thesis
In the experimental ITER fusion reactor, currently under construction in Cadarache, France, components close to the plasma will get activated and contaminated over time. Therefore, these components will need to be maintained remotely in a so called hot cell facility. The maintenance will be executed according to step-by-step procedures, with every procedure meticulously validated and formally described in an Operation Sequence Description (OSD). To support formal validation, OSDs are to be captured in an ITER-dedicated Structured Language (SL), a form of domain specific programming language.
The Remote Handling Study Centre (RHSC), located at FOM institute DIFFER, Nieuwegein, The Netherlands, has an ITER Remote Handling Control Room representative setup which connects to a virtual ITER Hot Cell back-end environment. In this setup, Remote Handling maintenance scenarios and procedures can be tested under realistic conditions.
In this thesis, a Structured Language for ITER maintenance is proposed, implemented as a prototype to capture several realistic operational sequences from existing projects, and integrated in the RHSC workflow. The Structured Language programs are converted to schedulable tasks, by inferring their resource usage and time duration from the structured task descriptions.
These tasks are then scheduled using a proposed set of scheduling heuristics. The resulting schedule is subsequently submitted to many simulated executions, with random variation in the execution time of individual tasks. The variation that is applied in the execution time, is based on representative task execution times was measured for a recent and repetitive set of maintenance tasks performed at JET.
It is shown that inserting a certain amount of slack in the schedule improves its robustness. ITER maintenance scheduling should account for uncertainty by including slack in the resulting planning.
Place, publisher, year, edition, pages
2013. , 35 p.
Remote Handling, robustness, domain specific languages
IdentifiersURN: urn:nbn:se:lnu:diva-30529OAI: oai:DiVA.org:lnu-30529DiVA: diva2:665340
Subject / course
Software Technology Programme, Master Programme, 60 credits
Lundberg, Jonas, Professor