Study On The Mechanism Of Dynamic Stress Modal Learning Method

For the safety of Bridges and buildings,the dynamic strain is one of the important indexes.Traditional dynamic strain measurement methods,such as strain electrical measurement,digital speckle interferometry,etc.,are limited by contact measurement and local measurement,which cannot be used for global measurement in real environment,and can only be used for local discrete point measurement or off-line measurement.The displacement strain transformation matrix can be used to better understand the essence of dynamic strain and is the basis of many new strain testing techniques.However,the transformation matrix is usually given directly by the theory and does not conform to the natural environment of the structure,and its inherent properties are affected by the surrounding environment.In order to obtain the displacement-strain transformation matrix of the structure in the natural environment,this paper proposes a method to obtain the displacement-strain transformation matrix with the transformation of the external environment factors and the use of the internal physical characteristics of the structure combined with the machine learning method.The research work mainly includes the following contents:Around the theory of modal analysis,the physical relationship between the discrete point and the global is introduced,and the possibility of predicting the global point according to the discrete point is introduced.The relation between displacement and strain is verified by strain modal theory.The simulation software is used to build a clamped beam model,and the experimental research is carried out on the different excitation positions,modes and orders of different excitation points,and the results show that the position and different excitation points have little influence on the experiment.Under the condition of using the first six modes,the calculated displacement and strain values can reflect the actual displacement and strain changes of the structure to be measured with high accuracy.Particle swarm optimization is used to study the convergence accuracy,convergence and efficiency of the algorithm under different population sizes and iteration times.The fitting effect of different modes on dynamic strain before and after using particle swarm optimization algorithm is discussed,and the influence of different modes on dynamic strain identification is revealed.This method is applied to the dynamic strain measurement of experimental beams,and the transformation process of displacement and strain is analyzed by deducing the expression of transformation matrix,and the transformation mechanism is revealed.The results of experiments with different excitation on experimental beams show that the displacement strain transformation matrix provides a new development direction for the measurement of dynamic strain.Under sinusoidal and random excitation,the measurement accuracy of strain reaches 99.82% and 99.70% respectively.The experimental results show that the displacement strain transformation matrix obtained by this method is more consistent with the actual surroundings.The method can identify the dynamic strain by introducing the displacement strain transformation matrix,which provides an encouraging impetus for the development of the strain measurement method.