| High-sensitivity quality sensors are essential for public safety.For example,in airports,railway stations,subway stations and densely populated areas,the timely detection of leakage of trace toxic and harmful gases will greatly reduce economic losses and protect the safety of the public.However,traditional gas sensors have limitations such as low sensitivity,low efficiency,large volume and high cost,which cannot be applied to the above scenarios on a large scale.Therefore,it is necessary to find a new sensitivity mechanism in order to realize high sensitivity identification of a variety of substances.Among them,the modal localization phenomenon was proposed by Nobel Prize winner Mr.Anderson.It has been proved that the phenomenon can effectively improve the sensing sensitivity by several orders of magnitude,which provides ideas for solving the above problems.Three methods for multi-material recognition was proposed in this paper based on mode localization with a coupled beam array.The mechanism of multiple mass disturbance sensing of the three-degree-of-freedom system based on mode localization principles was discussed and analyzed.Vibration characteristics were analyzed in order to find the best metric to achieve the detection of high sensitivity and wide linear range.The specific research contents are as follows.An equivalent physical model of a multi-degree-of-freedom weakly coupled cantilever array was established,and the general form of eigenvalues and eigenvectors was analyzed without mass disturbance.It was found that the eigenvalue had a maximum value.Moreover,the first-order eigenvectors were always the in-phase vibration state.Based on the first-order perturbation method,the expressions of the eigenvalues and eigenvectors when two mass disturbances existed in a three-degree-of-freedom system were solved,and the manufacturing variatiion was equivalent to the stiffness disturbance.Besides,the functional relationship between the variation and the resonance frequency was derived.The relationship between the eigenvalue sensitivity and the degree of freedom was studied,and the conclusion that the three-degree-of-freedom eigenvalue sensitivity was the highest was found.The differences between the six metric expressions characterizing the localization degree were analyzed and compared.After a comprehensive comparison in terms of the linear range and sensitivity,it could be concluded that the metric expression proposed from the perspective of energy accumulation was the optimal method.Three multi-substance identification methods were proposed,which were divided into a qualitative identification method and two quantitative identification methods.Among them,the qualitative identification method was based on mode-shape identification,and the quantitative identification method was by detecting the mode change amount and frequency deviation.The displacement was recognized,and the limitations or expansibility of the three methods was analyzed from the perspectives of linear range,sensitivity,and misjudgment.The types of errors that might exist in the detection process were analyzed.Based on the Monte Carlo method,the impact of process variations on quality detection was simulated.It was found that the effect of process variations on the eigenvalues of the system matrix was about 3 orders of magnitude less than that on the eigenvectors.From the perspective of structural optimization,two ideas for reducing the impact of variations on detection were proposed.They were appropriately increasing the coupling factor and reducing the number of cantilever beams.Finally,an experimental test system was established.The three characteristic modes and pre-localization phenomena appearing in theory and simulation were confirmed,and so was the effectiveness of three proposed methods.Some possible reasons for discrepancies in numerical values were analyzed as well.The research work in this paper provides a guiding significance for the detection of trace multiple substances with high sensitivity and wide linear range to some extent. |
