A high precision accelerometer-based sensor unit for the acquisition of ultra low distortion seismic signals

Ioakim, Panagiotis (2017). A high precision accelerometer-based sensor unit for the acquisition of ultra low distortion seismic signals. (Unpublished Doctoral thesis, City, Universtiy of London)

[img] Text - Accepted Version
Restricted to Repository staff only until 6 December 2018.

Download (9MB) | Request a copy

Abstract

Over 800,000 people worldwide lost their lives to earthquakes in the last decade and on average 171 people die every day due to earthquake related damage to structures and buildings. Precisely understanding the effects ground motion has on manmade structures is crucial to making them earthquake resistant. This can only be achieved by the precise measurement, recording, and analysis of ground displacement trends during a seismic event.

Although there is a vast amount of recorded seismological data available, current technology and processing methods fail to represent accurate ground displacement over time as the considerable technological challenges have yet to be overcome.

Raw seismic data has so far been primarily acquired with instruments utilising geophone or accelerometer based sensors. These instruments produce prominent time domain displacement errors due to the various system and sensor inaccuracies, and due to non-linear response. Since accelerometers provide acceleration over time data: whilst geophones are velocimeters, and therefore provide velocity over time data; in order to derive true ground displacement over time, a double, or single numerical integration is required respectively. During this essential numerical integration processes of data from such sensors, even small in magnitude errors accumulate to yield rather large displacement trend offsets over a typical event recording period of 60 to 120 seconds. In addition, the numerical integration process itself poses considerable challenges due to the theoretically infinite number of samples and the accurate determination of initial conditions required for an exact mathematical result to be obtained. The latter, is currently performed by averaging an up to 60 second pre-event data trend stored on the instrument.

Most post-integration data from current instruments appears to contain low frequency drifts amongst other noise artefacts, and generally requires baseline correction algorithms in an attempt to correct for these effects. Such corrections, although helpful, only aid to minimise the perceived effects of an assumed and collective source of error, and hence are largely unable to tackle the individual error contribution of each element within the system. Since individual element contribution is of a dynamic nature, the validity of these algorithms is limited by the accuracy of the initial assumptions made about a specific set of data. Faced with such a multivariable and uncertain dynamic behaviour, where even mathematical system modelling is of inadequate long term accuracy, a solution that aims to directly minimise these errors at source, rather than attempt to correct them postacquisition, is of immense importance when it comes to the recording, analysis, and understanding of earthquakes.

This thesis describes the design, implementation, and evaluation of a High Precision Active Gyro Stabilised (HPAGS) sensor unit of exceptional performance for the provision of highly accurate ground displacement data. Experimental results demonstrated that the device described herein, was able to diminish the inherent non-linear and environment-dependant effects of current sensors, and thus was able to provide highly improved time domain displacement data.

Publication Type: Thesis (Doctoral)
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
Departments: City, University of London theses
School of Engineering & Mathematical Sciences > Engineering
City, University of London theses > School of Mathematics, Computer Science and Engineering theses
URI: http://openaccess.city.ac.uk/id/eprint/19360

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year

View more statistics