A proof of concept for using microseismics as a site investigation tool has been developed and presented as a feasibility study utilising changes in the seismic wave Peak Particle Velocity (PPV) (m/s) and dominant frequency (Hz). The key significance of this thesis is the enhancement of near surface seismic imaging applications using a novel concept. Researchers have observed low frequencies when detecting geological features at depths greater than 100m. Mitchell, Derzhi et al. (1997) and Dilay and Eastwood (1995) have shown previously that the dominant frequency observed moved towards the low frequency range (<100Hz). Marfurt (1984), investigated how the dominant frequency varies with geological feature thickness and Marfurt and Kirlin (2001) used this concept to resolve geological features with a thickness of <20m. This thesis identified the following gaps in knowledge and identified that there is:- No study to demonstrate if the relationship between dominant frequency and geological feature thickness is observed in the near surface (i.e. depths less than 100m) at metre scale accuracy (i.e. <10m); - No study has used micro seismometers to apply this technique for near surface applications; and - No study which has considered if the dominant frequency and PPV characteristics can be used to develop a concept for a near surface site investigation tool deployed in the near surface. Considering the effect of medium properties there was significant effect on seismic wave characteristics such as PPV and frequency when utilising low frequency seismic sources in the range of 1-100Hz. The changes in the seismic wave characteristics during wave propagation through geological features characterised by different central feature widths and low Pressure (P) wave velocity zones were investigated. COMSOL Multiphysics Finite Element software modelled seismic excitations using the linear elastic equations that govern mechanical wave propagation. Dominant frequency was more responsive than PPV to material property changes. Considering the presence of a material property boundary, there was a significant effect on the PPV and dominant frequency characteristics, allowing a novel prediction methodology to be developed. The presence and width of a geological feature was detected at sub metre scale accuracy. Considering the presence of a geological feature surrounded by a low P wave velocity zone, the differentiation between the material zones can be detected numerically at sub metre scale accuracy, and this was validated in “blind” tests and pilot field trials with a systematic error of +0.4m and a random uncertainty of ±0.39m. Plotting PPV as a horizontal profile across the monitoring cross section allowed the visualisation of geological feature width, which inferred that geological feature location can be visualised with good accuracy. This research has confirmed that we can use the seismic wave characteristics i.e. PPV and frequency, to effectively map and locate near surface geological and manmade structures using a novel concept which can be deployed in the near surface. The range of validity of this novel concept is the detection and location of geological structures of a known type (i.e. a vertical dyke formation) for a range of different geological parameters such as, width and material properties in a low ambient noise environment. The effect of noise is important in terms of resolution and applicability of the method, and was investigated by adding noise to the sensitivity analysis. This research intentionally selected a field site that was characterised by a low ambient noise environment removing the requirement to utilise signal processing filtration methods as the impact from ambient noise was deemed insignificant. Consideration was given to sites that may be characterised by high ambient noise. When noise was increased to 2 x source PPV the “worst case” systematic error was - 2m and the random uncertainty was ± 1.6m. Both of which are greater than systematic error of +0.4m observed in the field trial. In high noise environment it would be advantageous prior to the experiment to establish if the PPV of the seismic source is powerful enough to overcome the effect of ambient noise. Future work could consider the application of filtration via various signal processing methods to minimise the effect of ambient noise. Preliminary simulations were conducted to consider the feasibility of future applications. There is potential to utilise the changes in the dominant frequency and PPV of the seismic signal as it propagates to locate voids and other subsurface features at depth. Future work will have to be conducted to determine subsurface feature location capabilities. Numerical simulations and pilot field trials demonstrate that this novel concept can be applied effectively achieving sub metre scale accuracy for a site with specific material properties and metre scale accuracy for site characterised by high ambient noise. These results are significant in forming the theoretical basis for the development of a novel microseismic site investigation tool.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:698539 |
| Date | January 2016 |
| Creators | Hooper, Chiara Mary |
| Publisher | University of Strathclyde |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://oleg.lib.strath.ac.uk:80/R/?func=dbin-jump-full&object_id=27338 |
Page generated in 0.0094 seconds