Determining the dynamical behaviors of engineering structures by geodetic measurements

Planning, design and realization of engineering structures have high prime costs. So behaviorsof these structures under temperature changes, wind and traffic load must be monitored bycontinuous or periodical measurement methods.In this study, Bosporus Bridge which is a dynamic system had been identified by non-parametric methods in time and frequency domain dependent on cause-response quantities.In the first section, necessity of monitoring and control of engineering structures, studiesabout this subject, development process of deformations on structures and classification ofdeformations according to system identification had been discussed.In the second section, theoretical information about suspended bridges which are dynamicsystems, basic components and properties of structures, behaviors under temperature change,wind and traffic load, determination of natural frequency and periods by Finite ElementsMethods had been discussed.In the third section, parametric and non-parametric methods which are used for systemidentification, ARX (Autoregressive with Exogeneous) model which is a non-parametricmethod dependent on cause-response quantity, regression analysis, multiple regressionanalysis and cross correlation function which explains the response of the system, had beendiscussed.In the fourth section, identification of system by time series dependent on response quantityof system, determination of trend, periodic and stochastic components, and time seriesanalysis which represents the system properties had been explained. Determination offrequency, period, phase angles, amplitudes of long and short time changes by spectrumanalysis had been discussed.In the fifth section, basic properties of Bosporus Bridge had been explained and loads whichaffects the bridge and sampling frequencies had been determined. And then reference andobject points had been constructed around the structure and time series measurement had beentaken. These measurements had been analyzed both by cause-response quantities like, ARX,regression analysis, multiple regression analysis and by non-parametric methods. Whenanalyzing the time series dependent on cause quantity, trend component in the series had beeneliminated and high and low frequencies of Bosporus Bridge had been determined by powerspectrum in periodic component analysis. Before determination of high and low frequencies,series had been multiplied with Hanning Function to reduce the spectral leakage in signals.And then high and low frequencies had been determined by Fast Fourier Transform after thesmoothing and differencing process. By using the low frequencies, significant periodicmovements which exist in measurements statistically had been determined and amplitude andphase angles of these measurements had been calculated. And also stochastic section of timeseries had been determined by AR (Autoregressive) and ARMA (Autoregressive MovingAverage) models.In the last section, process and analysis result of the study had been presented and somesuggestions about the subject had been given.Keywords: Bosporus Bridge, Time Series Analysis, HFD, ARX, AR, ARMA