Flood Discharge Structure Multi-sensor Vibration Response Data Fusion Method And Parameter Identification Research

Modal parameter identification and damage diagnosis of flood dischargestructure is nearly20years hot topic, frequencies and mode shapes are the mostcommonly indicator of the structural dynamic damage diagnose, and the structurenatural frequency is the most easily measured dynamic parameter. The vibrationsignals of flood discharge structure, frequency components and noise level inmulti-sensor measurement varies with the measurement position., structural naturalfrequency identification result can only be accurate to a certain interval range insteadof an exact value, this is not conducive to damage diagnosis based on the frequencychange. These signals can be fused to a vibration signal that accurately andcomprehensively reflects the vibration characteristics of the overall structure bymulti-information fusion technology. Information fusion technology in structuralengineering is still in the initial exploration phase of signal simulation, In order toprovide accurate and comprehensive vibration signal, accurate identification of flooddischarge structural modal frequency, the paper mainly studies the following:(1) As traditional data fusion algorithms fixes convergence factor and requireshigh similarity for signals, this paper presents the “dynamic response data fusionalgorithm based on variance dedication rate”. And by numerical simulations, fromthree aspects (accuracy, comprehensiveness and concentrated frequency structureintegration) to explore the feasibility of the dynamic response data fusion algorithmbased on variance dedication rate, then use the Shu River pier structure and Ertanarch dam and Xinjiang pier structure prototype observation data to verify thefeasibility of the algorithm. Studies show that: the algorithm can effectively utilizerelevance, complementarity and redundancy of the multiple signals, and the fusionsignal can reflect the vibration characteristics of the overall structure, the dynamicfusion model is more flexible than the traditional weighted fusion algorithm.(2) Study the harmonic response modal excluding method based on theprobability density function. Flood discharge structure multi-sensor signals modalidentification is difficult to distinguish structural response and harmonic response. When the discharge structure vibration signal, mixed with harmonic vibrationresponse signal (such as riverbed hydropower generators or other mechanical devicesoperated generated harmonic signals), conventional de-noising method can not beexcluded. harmonic response modal excluding method based on the probabilitydensity function, analysis the diffirences of the harmonic response and structuralresponse from the statistical point of view. First to be signal frequency domaindecomposition, and then based on the probability density function fitting vibrationsignal of each frequency component. Finally, numerical Simulation a set of signalsmixed with harmonic response and structure response to see the feasibility of thismethod, and by pre-adding harmonic signal at a set of prototype observation signalsof flood discharge structure to study its effect.