Vibration Response Analysis And Measurement Of Railway Hub Station Buildings Under Multiple-Source Excitations

With the rapid development of economy, in the past decade, a large number ofmodern railway hub stations have been built in our country. In order to realize“zero-distance or short-distance transfer” among railway, subway, light railway andbus, spatial multiple-layer structural systems are widely adopted in practice. Therefore,the structural systems have been excited simultaneously by multiple-sourceexcitations including human walking, train moving and equipment operating. Thiswill significantly affect the service and performance of the station buildings.Therefore, it is of theoretical and practical significance to study the dynamicalresponses of railway hub station buildings under multiple-source excitations in thedesign and vibration control.Focusing on the Tianjin West Railway Station building, this dissertationsystematically studies the random excitation models of human walking and trainmoving and their applications, the combination problem of multiple-sourceexcitations and the analytical approach on the dynamical responses of railway hubstation buildings under multiple-source excitations. The main research work includesthe following assignments:(1) An improved stochastic walking excitation model considering variations ofperiods, amplitudes and impulses is proposed to accurately describe the individualdifference of walking excitation among pedestrians and the similar cycle difference ofa single walking time histories. The normalized cycle curve of walking excitation isextracted based on Zivanovic model and optimized using non-linear least-squarecurve fit method. The improved model is verified superior to conventional models forsimulating both the time and frequency characteristics of walking excitations, andagrees well with the real measured time histories. Focusing on the Tianjin WestRailway Station building, human body is equivalent to be a mass-spring-dampermodel with two degrees of freedom. The human-structure coupled dynamic equationsare established using the improved model, and the pedestrian-induced structuralvibration responses are simulated by direct integration method. The results indicatethat, by employing the improved model, the frequency distribution of the structuralresponses is wilder and the high frequency excitation components obtained throughthe model has non-negligible effect on the structural responses, which agrees well with the measured data. In addition, the coupled effect between human and structurereduces the structural responses, but with little impact.(2) A ten-degree secondary suspension train model is established in term of thetheory of vehicle random vibration. The intact track irregularities and correlationamong wheels on rails are considered, and the displacement time histories of eachwheel set on rail are simulated. The MATLAB program is implemented to obtain thestate governing equation of the ten-degree secondary suspension train model. Thetrain random excitation time histories are finally obtained from the state equations.The effectiveness of the train model is verified in both time and frequency domain bycomparing with measured data and several other train moving models. The dynamicalresponses of the Tianjin West Railway Station building under train moving randomexcitation are simulated and measured. The results indicate that the time andfrequency characteristics of simulated dynamical responses agree well with theexperimental data. The predominant frequency of the waiting hall is at the mostsensitive frequency band of human body during vertical vibration, and the humancomfort should be considered.(3) The combination problem of multiple-source excitations acting on railwayhub station buildings is investigated by considering stochastic walking excitation andtrain moving random excitation. Probabilistic method and Monte-Carlo simulation areadopted to obtain the probability distributions of the extreme value of the two kinds ofexcitations during structural design duration. The study considers the probabilitycharacteristic of combined excitations and introduces The Extreme I Distribution withtwo parameters to describe the probabilistic properties of the extreme value of thecombined excitations. The combination coefficients are determined, which have aguarantee rate of95%. The K-S test shows that the probability characteristics of thecombined excitations are well evaluated by the Extreme I Distribution compared withthe empirical distribution given by Monte-Carlo simulation.(4) Considering the frequency overlap among structural responses under eachexcitation including human walking, train moving and equipment operating, aweighted power coherent superposition method, which is based on Root Mean Square(RMS) of acceleration, is proposed for the dynamical responses of station buildingsunder multiple-source excitations. The rationality and applicability of the weightedpower coherent superposition method are also discussed. The proposed method anddirect superposition method are both adopted in the dynamical analysis of the station buildings simultaneously excited by the three excitation sources. The results indicatethat the RMS interval of structural responses calculated by the proposed method canproperly envelope the RMS value of acceleration by the direct superposition method.The direct superposition method and the proposed method are also used to analyze thedynamical responses of the station buildings excited by random vibrations both intime domain and in frequency domain. The overall responses of the station buildingare obtained in terms of the interacting superposition effects on the station buildingdue to the individual excitation.(5) The acceleration time histories of measure points on the waiting hall andplatform layer are obtained through the measuring experiments on the Tianjin WestRailway Station under multiple-source excitations. Spectral analysis and RMSevaluation are both introduced to analyze the effects of power coherent superpositionto the overall dynamical responses with respect to human walking, train moving andequipment operating. The measured data is also compared with the numerical ones.The results show that the RMS interval of structural responses is well calculated bythe weighted power coherent superposition method. The agreement will be betterwhen the distance to excitation sources increases.