Daily Image
29-10-2018
Non-Iterative Radio Telescope Design
| Submitter: | Albert-Jan Boonstra and Ronald Nijboer |
| Description: | How to design a radio telescope such that it can deliver the most science for a fixed budget? Traditionally one would start with requirements and a concept of operations, and after initial designs one would check to which extent the designs can meet the requirements, including non-functional constraints such as cost. Discrepancies would then lead to design changes or changes in non-functional constraints, and the process would start again with a next loop, aimed at detailing the designs further. This design loop: requirements, design, verification, followed by adaption, usually is sequential and iterative. But does it always need to be so? What if one would include costing models explicitly in the designs at the same level as the scientific figures of merit such as sensitivity? This was what we have looked into after the idea was coined by Jaap Bregman in 2005 (in SKA: an Engineering Perspective, Springer). Both scientific figures of merit (F) and costing (C) depend on telescope design parameters (p) such as number of telescopes and observing bandwidth. Optimal values for these parameters can be found using an elegant optimization approach: Lagrange multipliers (see picture for the basic formulas). Details can be found in our paper (Oct. 2018, doi: 10.1029/2018RS006624), where we conclude that the LOFAR 2010 designs are very close to the optimal ones, and where we have assessed the 2017 SKA1-Low designs. As it probably will remain difficult to include all science requirements, all non-functional constraints, and all design options in one tractable model, the design process will remain iterative. However, by including cost explicitly in the first design steps, big de-scoping actions late in the design cycle can hopefully be avoided. |
| Copyright: | ASTRON / Top Foto |
| Tweet |  |