Modern astronomy makes use of optical telescopes for the observation
of sky objects in the wavelength range from ultraviolet to infrared,
that is from about 300 nm to 30 
m
.
Although the recent years have seen the development of space astronomy
from automatic telescopes carried by satellites, the huge cost of these
satellites and some of their inherent limitations will mean that
ground based telescopes will be still for many decades and perhaps
centuries the main instruments of astronomers.
Indeed this last decade has seen the start of a number of new
projects for telescopes larger than any in operation today.
These telescopes, thanks to the high quality of optical systems and
their electronics, aim at being almost as performing as satellite
instruments, with a much lower cost.
Two main developments are presently being pursued in this field. One aims at the manufacturing of larger primary mirrors, with diameters from 6 to 10 meters, which constitute a significant technological leap with respect to the 4-m class mirrors used in the best telescopes built until the mid 80s. The advantages sought with larger mirrors are the increased light collecting performance which is proportional to the mirror area and the improvement of the theoretical resolution which is proportional to the ratio between wavelength and mirror diameter.
The other development line aims at decreasing the disturbances caused by the atmospheric environment on the optical performance of a telescope. This objective is sought on the one hand by locating new telescope on high mountain sites with favorable atmospheric characteristics, and on the other hand by reducing local effects, mainly of thermal origin, caused in and by the observatory itself. Ground based telescopes must of course observe through the atmosphere, which has two main consequences on the quality of observations. The first consequence is a degradation due to the turbulent variations of the index of refraction. This causes the image of a star to appear as a randomly moving patch with an angular size which is often quite larger that the theoretical limit size due to diffraction from the telescope optics. This degradation is called the seeing and is often quantified as the mean angular apparent diameter of a star image. Although contributions to this seeing effect come from all atmospheric layers, it is now known that some of the largest disturbances are generated close to the telescope itself and are ultimately caused by differences of temperature between air and the observatory structures. These negative effects tend to disappear when the telescope is exposed to wind, which, however, creates then another problem.
This second disturbing effect is caused by vibrations of the telescope due to the wind mechanical turbulence. During the observation the telescope must track the star with very high accuracy. If the telescope is even partly exposed, the wind will tend to shake it. The guiding accuracy required even for short integration times is such that the telescope oscillations due to the wind load cannot be absorbed by the structural rigidity of the telescope alone and must be corrected by an active control loop. There is nonetheless a limit to the amplitude and frequency bandwidth of the wind disturbance that can be corrected by the control system. Moreover the primary mirrors of the large telescopes of the newest generations are much thinner and less stiff, for a number of reasons, than those of predecessor telescopes, and therefore may also have the figure of their optical surface deformed by the wind loads.
Thus the wind flow has a twofold action on the overall telescope performance: on the one hand it improves the seeing quality but on the other hand it degrades the guiding accuracy; conversely if the telescope is well shielded from the wind, guiding will be very accurate but seeing will inevitably worsen. It is important to note that both seeing and guiding inaccuracies have similar effects on the image quality of the telescope inasmuch they both cause an enlargement of the apparent size of a star image recorded during the exposure time.
The engineer designing the enclosure of a telescope has therefore the difficult task of finding a compromise in the exposure of the telescope to the wind, which will maximize the overall quality of the observation. Telescope enclosures are a very particular type of buildings, which must fulfill an unusual set of requirements. This situation has produced many different technical solutions, depending on a variety of parameters such as the size and type of the telescope, its geographic and meteorologic location, the type of observations pursued at and various other requirements concerning maintenance, access and operation.
It is important to underline that the problem, such that it is set
nowadays, is quite new.
Until a few years ago, the design of a telescope enclosure was
deemed to a quite straightforward matter and was based on a tradition
of dome building which had been inherited since the time of
refractive telescopes. These traditional concepts were applied
uncritically to make domes for
reflective telescopes with primary mirror sizes larger than 2 
3
meters, and constituted an established practice of the art which
was in fact
hiding a basic ignorance of the interaction between a telescope
and its local atmospheric environment, particularly with respect
to the seeing problem. Moreover objective difficulties,
such as the
great variability of the atmospheric environment, hindered
comparisons and benchmarks among the best and largest
telescopes with respect to the real physical limitations to ground
based astronomy.
It was then only when particular circumstances proved unequivocally that some of the concepts traditionally used for the design of telescope enclosures were counterproductive, that innovative research work was undertaken on the subject.
Most of the work presented in this dissertation was performed in the
framework of the development of the new Very Large Telescope (VLT)
observatory by the European Southern
Observatory (ESO). The VLT observatory will be constituted by four 8-m
telescopes which will eventually
be able to combine their light beams into a single image. When
completed near the year 2000, the VLT will be the largest astronomical
observatory in the world
.
Because the VLT project was started without prejudiced ideas, the development of the VLT covered a wide spectrum of conditions and design solutions. As a consequence this work should have applications well beyond the scope of the VLT project, and has the ambition to become a reference for the design of telescope enclosures.