TY - BOOK

T1 - Future Distributed Geomonitoring Analysis Methods, Algorithms & Technologies

AU - Schmidt, Roland

N1 - embargoed until null

PY - 2021

Y1 - 2021

N2 - The core task of geomonitoring is to find and employ suitable sensors and methods to monitor natural and artificial structures for changes and to trigger an alarm if necessary. In order to achieve this, reliable sensors and suitable evaluation methods are required. A reliable alarm process and long-term storage of the recorded data complete the requirements of such a system. The first part of the work is dedicated to discrete basis functions, which enable an advanced method for the reconstruction of structural displacements. Two independent measurements which follow physically different principles, e.g. 3D displacement measurements (total station) and slope measurements (inclinometer), are combined with each other by means of sensor fusion. Thereby more reliable results can be achieved. Due to the efficiency of the analytical data processing, an implementation on local battery-powered embedded systems is possible. In the second part the inclination sensors commonly used in geomonitoring are investigated in detail. Due to the differential measuring principle, the inclination sensors have a high temperature stability and the sensor noise is independent of the inclination. It was shown, that the noise of the inclination sensor following a differential measurement principle has a Cauchy Lorentz distribution. Therefore, different statistical methods have to be applied than for a normal distribution. By means of singular value decomposition it can be shown that the sensor elements have a different gain factor to each other. If orthogonalisation is applied on the data the confidence interval is improved by up to 34%. In the last part of this thesis, the "ASUA Urban Automation Reference Platform" and the associated devices, which were (co-)developed in the Celtic-Next project "Advanced Sensing for Urban Automation" (ASUA), are discussed. Due to the continuous change in the requirements for geomonitoring devices with regard to networking, integration into heterogeneous systems and data availability, these systems must also adapt. The developed devices enable local data processing for complex problems (reconstruction), local alarms and control via a cloud-based control system. The solutions found in the work provide the basis for a new generation of connected geomonitoring devices that enable intelligent local data processing. The model-based reconstruction allows a significant improvement of the results by data combination and tolerates failures of single measurement sensors. The detailed sensor analysis shows that the usual statistics fail for many sensors used in geomonitoring. Thus, other statistical methods and filters have to be applied to obtain reliable results.

AB - The core task of geomonitoring is to find and employ suitable sensors and methods to monitor natural and artificial structures for changes and to trigger an alarm if necessary. In order to achieve this, reliable sensors and suitable evaluation methods are required. A reliable alarm process and long-term storage of the recorded data complete the requirements of such a system. The first part of the work is dedicated to discrete basis functions, which enable an advanced method for the reconstruction of structural displacements. Two independent measurements which follow physically different principles, e.g. 3D displacement measurements (total station) and slope measurements (inclinometer), are combined with each other by means of sensor fusion. Thereby more reliable results can be achieved. Due to the efficiency of the analytical data processing, an implementation on local battery-powered embedded systems is possible. In the second part the inclination sensors commonly used in geomonitoring are investigated in detail. Due to the differential measuring principle, the inclination sensors have a high temperature stability and the sensor noise is independent of the inclination. It was shown, that the noise of the inclination sensor following a differential measurement principle has a Cauchy Lorentz distribution. Therefore, different statistical methods have to be applied than for a normal distribution. By means of singular value decomposition it can be shown that the sensor elements have a different gain factor to each other. If orthogonalisation is applied on the data the confidence interval is improved by up to 34%. In the last part of this thesis, the "ASUA Urban Automation Reference Platform" and the associated devices, which were (co-)developed in the Celtic-Next project "Advanced Sensing for Urban Automation" (ASUA), are discussed. Due to the continuous change in the requirements for geomonitoring devices with regard to networking, integration into heterogeneous systems and data availability, these systems must also adapt. The developed devices enable local data processing for complex problems (reconstruction), local alarms and control via a cloud-based control system. The solutions found in the work provide the basis for a new generation of connected geomonitoring devices that enable intelligent local data processing. The model-based reconstruction allows a significant improvement of the results by data combination and tolerates failures of single measurement sensors. The detailed sensor analysis shows that the usual statistics fail for many sensors used in geomonitoring. Thus, other statistical methods and filters have to be applied to obtain reliable results.

KW - Geomonitoring

KW - inclinometers

KW - basis functions

KW - Cauchy-Lorentz distribution

KW - model based reconstruction

KW - condition monitoring

KW - Geomonitoring

KW - Neigungssensoren

KW - Basisfunktionen

KW - Cauchy Lorentz Verteilung

KW - modellbasierte Rekonstruktion

KW - Zustandsüberwachung

M3 - Doctoral Thesis

ER -