Distributed Acoustic Sensing using straight, sinusoidally and helically shaped fibres for seismic applications

Abstract

Distributed Acoustic Sensing (DAS) is a versatile dynamic strain sensing method that has been adopted for a wide range of seismic applications. In DAS, optical fibres are interrogated and used as sensors, where a strain or strain-rate measurement is made along a specific length of the fibre, called the gauge length. Its main appeal is the spatially dense data over long distances. The main limitations of DAS, however, are that it is mainly sensitive along the axial direction of the fibre and that the signal-to-noise ratio is worse than that of standard geophones. The first issue limits its adoption in surface reflection seismic when the fibre is deployed horizontally. Also, due to the very nature of the measurement (i.e. elongation and contraction of the fibre), it is commonly considered as a single-component measurement, therefore it lacks the information from the other components.

This thesis studies the potential of obtaining multi-component information from DAS as well as investigating the use of combined fibre configurations for surface-seismic applications. We approach this by examining several fibre-shaping approaches with static and dynamic strain measurements. First, the concept of the sinusoidally shaped fibre is examined to make a directional strain sensor in a direction other than the fibres’ axial direction using a static-strain approach. Secondly, the combined use of straight and helically wound fibres for obtaining multi-component information from DAS data as well as assessing the usefulness of using such a combination is investigated in a surface-seismic setting.'

Using the sinusoidally shaped fibre, two approaches are investigated. The first approach involves the use of the sinusoidally shaped fibre embedded in a homogenous material. An analytical model is presented to describe what happens to the deformed fibre in three main directions, which was validated via a finite-element model. Along with the model, loading experiments were performed on a sinusoidally shaped fibre embedding in a polyurethane-type (i.e. called Conathane®) strip in the following directions: in-line (i.e. transversal in-plane with the sinusoidal fibre), broadside (i.e. perpendicular to the sinusoidal fibre), and along-strip (i.e. along the strip’s longest dimension). We saw that the fibre is mainly sensitive to the in-line and broadside directions, and it is slightly more sensitive in the in-line direction relative to the broadside direction. We also saw that the geometrical parameters of the fibre, as well as the mechanical properties of the embedding material, affect its directional sensitivity. This is exploited in the second approach where the embedding material is now adapted to a low Poisson’s ratio metamaterial as well as further adaptations in the geometry of the fibre, aiming to create a unidirectional strain sensor. Experimental results showed improvements in the sensitivity but not as much as predicted by the analytical or numerical modelling.


Using DAS in field settings, multiple configurations of straight (SF) and helically wound fibres (HWF) with different wrapping angles (α) were buried in a 2-m trench in farmland in the province of Groningen in the Netherlands. Significant amplitude differences are observed between the straight and helically wound fibres. It is observed that shaping the fibre into a helix dampens the amplitude inside the surface wave significantly. Also, a polarity flip is observed with the use of HWF with a wrapping angle of 30◦. This hints that there is a contribution of the vertical component on the response measured by the HWF as also supported by the theoretical models. The reflection response is also examined using a set of engineered SF and HWF fibres. The main seismic reflections are present in both fibres with higher amplitude in SF compared to HWF, contrary to what was expected. Also, using post-stack images we see that the SF and HWF provide reflection structural images comparable to surface-deployed geophones but with an (expected) lower signal-to-noise ratio. We show that the combined use of SF and HWF is useful, as reflections were better shown for the shallow section, unlike HWF which provided better reflections in deeper sections. Furthermore, we discuss the effect of gauge length on the retrieval of surface waves along with the use of different fibre shapes using active and passive sources.

With the active-source data, we show that the gauge length plays an essential role in the retrieval of surface waves depending on their wavelength range, as it might cause distortions in the waveform which appears as notches in the (frequency, horizontal-wavenumber)–domain, as well as complicates picking the dispersion curves of these waves. On the other hand, the helically wound fibres might require a longer gauge length to retrieve the surface wave properly. This decreased sensitivity of the helically wound fibres is also shown from virtual shots obtained by passive interferometry as well as a recorded earthquake in the area.