CAJ | 학술논문

泄流结构不同测点的振动信号频率成分和噪声水平不尽相同，通过信息融合，可将多测点振动信号融合为一个能准确、全面反映结构整体振动特性的振动信号。针对传统数据融合算法对信号相似度要求高以及固定融合系数的缺陷，提出了基于方差贡献率的结构动态响应数据融合算法，不同时间的融合系数随该点的方差贡献率而变化，根据该时刻信息的相对重要程度动态分配权重系数，实现多测点振动信号的动态融合。通过对数学模拟信号及泄流结构原型观测动态响应数据融合结果表明：提出的动态数据融合方法能够挖掘噪声中的微弱信号，使融合后信号能够更加准确、全面的反映结构的整体振动特性；该方法与传统互相关融合算法对比表明，方差贡献率算法能够有效凸显信号中的优势频率成分，保留原信号的结构固有频率成分，具有更强的适用性。
설류결구불동측점적진동신호빈솔성분화조성수평불진상동，통과신식융합，가장다측점진동신호융합위일개능준학、전면반영결구정체진동특성적진동신호。침대전통수거융합산법대신호상사도요구고이급고정융합계수적결함，제출료기우방차공헌솔적결구동태향응수거융합산법，불동시간적융합계수수해점적방차공헌솔이변화，근거해시각신식적상대중요정도동태분배권중계수，실현다측점진동신호적동태융합。통과대수학모의신호급설류결구원형관측동태향응수거융합결과표명：제출적동태수거융합방법능구알굴조성중적미약신호，사융합후신호능구경가준학、전면적반영결구적정체진동특성；해방법여전통호상관융합산법대비표명，방차공헌솔산법능구유효철현신호중적우세빈솔성분，보류원신호적결구고유빈솔성분，구유경강적괄용성。
The vibration signal frequency components and noise levels in multi-sensor measurement of a flood discharge structure vary with the measurement positions.These signals can be fused to a vibration signal accurately and comprehensively reflecting the vibration characteristics of the overall structure with the multi-information fusion technology.Here,aiming at traditional data fusion algorithms fixing the convergence factor and requiring higher similarity of signals,a dynamic response data fusion algorithm based on variance dedication rate was proposed.Fusion coefficient at different instants in this algorithm changed with the relative importance of the signal information.The weighting factor was dynamically allocated based on the variance dedication rate to realize the dynamic fusion of mulati-point vibration signals. The numerical signal simulation and the original measurement of the flood discharge structure dynamic response data fusion results showed that the dynamic data fusion method proposed here can excavate weak signals hidden in noise and the fused signal can reflect the vibration characteristics of the overall structure more accurately and comprehensively.The comparison between this method and the traditional cross-correlation fusion algorithm showed that this method can effectively highlight the dominant frequency components of signals,retain the structural natural frequency components in the original signals and has a stronger applicability.