This paper presents a complete description of Virgo, the French-Italian gravitational wave detector. The detector, built at Cascina, near Pisa (Italy), is a very large Michelson interferometer, with 3 km-long arms. JINST 7 P03012In this paper, following a presentation of the physics requirements, leading to the specifications for the construction of the detector, a detailed description of all its different elements is given. These include civil engineering infrastructures, a huge ultra-high vacuum (UHV) chamber (about 6000 cubic metres), all of the optical components, including high quality mirrors and their seismic isolating suspensions, all of the electronics required to control the interferometer and for signal detection. The expected performances of these different elements are given, leading to an overall sensitivity curve as a function of the incoming gravitational wave frequency.This description represents the detector as built and used in the first data-taking runs. Improvements in different parts have been and continue to be performed, leading to better sensitivities. These will be detailed in a forthcoming paper.

The knowledge of the noise Power Spectral Density of interferometric detector of gravitational waves is fundamental for detection algorithms and for the analysis of the data. In this paper we address both to the problem of identifying the noise Power Spectral Density of interferometric detectors by parametric techniques and to the problem of the whitening procedure of the sequence of data. We will concentrate the study on a Power Spectral Density like the one of the Italian-French detector VIRGO and we show that with a reasonable number of parameters we succeed in modeling a spectrum like the theoretical one of VIRGO, reproducing all its features.We propose also the use of adaptive techniques to identify and to whiten on line the data of interferometric detectors. We analyze the behavior of the adaptive techniques in the field of stochastic gradient and in the Least Squares ones. As a result, we find that the Least Squares Lattice filter is the best among those we have analyzed. It optimally succeeds in following all the peaks of the noise power spectrum, and one of its outputs is the whitened part of the spectrum. Besides, the fast convergence of this algorithm let us follow the slow non stationarity of the noise. These procedures could be used to whiten the overall power spectrum or only some region of it. The advantage of the techniques we propose is that they do not require a priori knowledge of the noise power spectrum to be analyzed. Moreover the adaptive techniques let us identify and remove the spectral line, without building any physical model of the source that have produced them.

Below a few tens of hertz interferometric detection of gravitational waves is masked by seismic vibrations of the optical components. In order to isolate the mirrors of the VIRGO interferometer, a sophisticated suspension system, called superattenuator, has been developed. Its working principle is based on a multistage pendulum acting on seismic vibrations as a chain of second order mechanical low-pass filters. A complete superattenuator has been built and tested. This apparatus allows extending the VIRGO detection band down to a few Hz. A detailed description of the attenuation system and its performance are presented in this article. (C) 2001 American Institute of Physics

One of the goals of gravitational data wave analysis is the knowledge and accurate estimation of the noise power spectral density of the data taken by the detector, this being necessary in the detection algorithms. In this paper we show how it is possible to estimate the noise power spectral density of gravitational wave detectors using modern parametric techniques and how it is possible to whiten the noise data before they pass to the algorithms for gravitational wave detection. We report the analysis we made of data taken by the Caltech 40-m prototype interferometer to identify the noise power spectral density and to whiten the sequence of noise. We concentrate our study on data taken in November 1994; in particular, we analyze two frames of data: the 18nov94.2.frame and the 19nov94.2.frame. We show that it is possible to whiten these data, to a good degree of whiteness, using a high order whitening filter. Moreover, we can choose to whiten only a restricted band of frequencies around the region we are interested in, obtaining a higher level of whiteness.