Figure 1:Double Walled Telescope, L. Mode.
Figure 2: Double Walled Telescope, C. Mode.
Figure 3: Overview of the Current FE Model
Figure 4: Current FE Model: 1st Mode.
Figure 5: Current FE Model: 2nd Mode.
C. Del Vecchio
Arcetri Astrophysical Observatory
Florence, Italy
W. Davison, J.M. Hill
R. Gatti
A.D.S. Italia
Lecco (CO), Italy
Proceedings of the ESO conference on Progress in Telescope and Instrumentation Technologies, held in Garching, Germany, ed. M-H. Ulrich, p. 79 (1992).
Abstract
1. Introduction
2. The Structural Problem
3. Current Preliminary Results
References
Abstract
The FEA of the current elevation structure for the
Columbus Project Telescope is reported.
The purpose of the analysis is to determine the most
relevant static and dynamic characteristics of the
telescope as a function of the
elevation angle. The results of the study show that
a first resonant frequency greater than 8 Hz,
as requested by the specifications, can be achieved.
1. Introduction
The Columbus Project Telescope is a binocular telescope
equipped with two 8.4 m primary mirrors and with
secondary and tertiary optics permanently on board to
provide the maximum observing flexibility (see [4]
for a more detailed description of the project).
Three phases can be identified for the structural studies of the Columbus telescope. The first step was to develop simplified FEA models of the telescope structure and, separately, more elaborate models of important subsystems such as secondary and tertiary mirror spiders, primary mirror cells, etc.. The results of this phase (see e.g. 5 ) gave clear indications on a general geometry of the telescope that could meet the specifications, and therefore a more refined model of the structure was setup. This new telescope model was including some modifications that had occurred in the telescope baseline, first of all the change of primary mirror diameters from 8 to 8.4 meters, more accurate evaluation of added masses, and more detailed modeling of spiders, cells, and other subsystems. Although this model showed that we could meet our specifications with a further optimization, but at the cost of a considerable structural mass, it also made clear that there was ample space for improvements. In particular, we found that providing a sufficiently rigid support to the spiders was especially difficult at the rear wall of the telescope but not at the front side. This was the starting point of a significant revision initially of the supports of secondary and tertiary mirrors, and then of the geometry of the top of the telescope. We found that a swing arm type support can work as well as a more conventional spider, but only requires stiff attachments on one side, and this is what our geometry can provide. After testing the new approach on several simple models, that confirmed the viability of this concept and also provided preliminary dimensioning and balancing, we setup a new detailed model of the entire telescope. Here we report the first results obtained with the new detailed model, that was successfully run only at the beginning of April 92 and therefore is not optimized (see [1] ) and still suffers of some local problems, but essentially confirms the expectations. This model has been developed within the Algor Supersap environment, and we are going to reanalyze it by means of Ansys (an appropriate software tool allows us to translate any Supersap input file into its correspondent Ansys input file, permitting cross--verifications of static and dynamic results). It has been modeled with great accuracy (it contains more than 22,000 degrees of freedom), starting from {\em Autocad} drawings. It uses shell elements and beam elements, for the telescope structure and the cells, and brick elements to simulate the primary mirror.
2. The Structural Problem
The basic quantitative requirement for the Columbus telescope
structure is that all the global resonant modes of the telescope must
be above 8 Hz (10 Hz as a design goal). Local resonant modes causing
displacements of the optics that affect the optical pathlength should
occur at frequencies above 18 Hz as far as the primary is concerned,
and above 25 Hz for secondary and tertiary mirrors (see
[2] for a more detailed discussion about error
budget and specifications). Other requirements
are of course already embedded in the selected geometry, that takes
into account the optical configurations, the requirements of quick
switching of configuration, the on--board aluminization, the
instrumental volumes and masses, etc. Further requirements
are that the total thermal inertia and the local thermal time
constant, as well as the cross section for the wind, are kept to a
minimum.
The experience gained with the previous FEA of a variety of models of the Columbus telescope has shown that the chosen binocular geometry allows us to meet the above requirements, and that the following aspects are crucial for the performance:
The optimization can be done on the elevation structure alone disregarding the azimuth platform, provided that the adopted constraints are correctly representing the interaction of the two parts and the elastic properties of the hydraulic supports. In all the models we observed that the resonant frequencies of the three global modes tend to remain high going from zenith to about 40 °, and then tend to decrease significantly at horizon pointing. This is clearly related to the position of the supports on the C ring, but is not a serious inconvenient, as our error budget allows somewhat relaxed tolerances at low elevation angles. Concerning the local modes, a detailed analysis of the primary mirror modes shows that they depend mostly on the geometry and the rigidity of the six hardpoint actuators (see [3] for a description of the cell layout), while the situation is, as already mentioned, more complex for the secondary and tertiary mirror modes.
3. Current Preliminary Results
While Figs.1 and2 show the telescope structure still
representing the present baseline, in Fig. 3 the FE model
of the new model is plotted. The first two modes of the
new structure are shown in Figs. 4 and 5. Tab.1
presents a comparison of the main parameters of both structures.
It must be stressed that the results of the FEA of the new model
have not been optimized in any way and that previous experience
has shown that even considerable gains may be achieved by
proper optimization. The comparison shows that, with a lower
mass and a smaller wind cross section, the new model
seems to achieve higher resonant frequencies. Another potential
advantage of the new structure is related to the possibility
of inserting the aluminizing bell jar into the telescope along
the optical axis (as well as from the side, that is the only
possibility of the previous telescope). This allows us to consider
different handling schemes, employing bridge cranes instead of
a large external elevator.
| Previous model | Current model | |
| Total mass [Kg * 103] | ~ 320 | ~ 227 |
| Degrees of freedom | ~ 9700 | ~ 22000 |
| 1st res. frequency [Hz] | 7.3 (L) | 9.1 (L) |
| 2nd res. frequency [Hz] | 7.9 (C) | 10.8 |
| 3rd res. frequency [Hz] | 8.8 | > 11 |
References
[1] C. Del Vecchio, ``A FEA Optimization Procedure
for Improving Structural Performances'', Proceedings of this Conference.
[2] J. M. Hill, ``Optical Design, Error
Budget and Specifications for the Columbus Project Telescope'',
Proc. SPIE, 1236, pp. 86-107, 1990.
[3] L. Miglietta et al., ``Layout
Study of the Columbus Telescope M1 Cell'', Proceedings of this Conference.
[4] P. Salinari, ``Columbus Telescope and Enclosure'',
Proceedings of this Conference.
[5] B. A. Schrefler et al., ``Structural Study
of the Columbus Telescope'', Very Large Telescope
and their Instrumentation, ed. M.-H. Ulrich, pp. 225-235,
(Munich:ESO), 1988.
|
Figure 1:Double Walled Telescope, L. Mode. |
Figure 2: Double Walled Telescope, C. Mode. |
|
Figure 3: Overview of the Current FE Model | |
|
Figure 4: Current FE Model: 1st Mode. |
Figure 5: Current FE Model: 2nd Mode. |
