Deformation analysis is a procedure used for detecting the height of point stability, which plays the
main role in determining displacements. The obtained displacement and deformation values are
represented by using a deformation model, which describes the most probable deformation course. In
the past classic procedures of deformation analysis were based on congruent model. Due to
computing, measuring techniques and evaluation algorithm development also the deformation analysis
method has changed. Besides basic determination of the object’s geometry it now also includes
displacement timeline, as well as potentially the causes for their formation. We focused on kinematic
deformation model in which we assumed that the given subject is constantly moving. Using that, we
are also able to incorporate time component in the model. At the beginning of the thesis, we set a
hypothesis that the use of kinematic deformation model in control measurements of geotechnical
objects allows the geometric changes to be considered in connection with time functions. On the
chosen object, where several epochs were made, we checked the reference point stability and
determined the statistical significance displacements of control points or their stability with the help of
congruent deformation model. We were considering the time frame in which severe movements at
some points had already been detected before. Kinematic deformation model is based on adjustment of
all the observations in all epochs where besides unknown coordinates also speed and acceleration
displacements are unknown. Since this kind of system is hard to calculate, we chose a Kalman filter
technique, which allows the evaluation of the unknown quantities. These unknown quantities may
have some noise, such as time and miscalculated observation function. This kind of algorithm is
commonly used in geodetic engineering, where it is not possible to ensure redundant measurements. It
is based on the determination of the system condition using the least squares method, but at the same
time it also evaluates the reliability degree. As part of the master’s thesis we made a programme for
processing kinematic measurements by using the Kalman filter for the case of discreet kinematic
model. We tested the stability of control points and displayed the results numerically and graphically.
As expected, the used approach offered some additional information (speed and acceleration of the
displacements) about the present object’s condition in space and time, which represents an alternative
approach to the congruent evaluation model used in deformation analysis so far.
|
