Date of Award


Document Type

Open Access Dissertation


Mechanical Engineering

First Advisor

Sourav Banerjee


Primary objective of this work is to introduce the multi-scale computational model for the bio-inspired acousto-ultrasonic band pass sensor that are capable of mechanically sense and/or filter wide range user defined frequencies. Selecting a particular and/or a distinct band of frequencies is essential for many applications in science engineering and technologies. For example design of sensors in chemical, biomedical and biological applications; device application for acoustic modulation by breaking the acoustic reciprocity and the sensors used in precision manufacturing applications requires sensing and/or filtering of wide range of acousto-ultrasonic frequencies. Presently, electronic devices are widely employed in commercial applications for selecting the target frequencies. Concurrent to the electronics sensors mechanical sensors with smart materials are significantly contributing to the sensing technologies, especially where electronic sensors are not compatible. Mechanical sensors are traditionally made of cantilever beams and use the resonance phenomenon to select the target frequencies. Considering the required size of the sensors, the above physics limits the design of these sensors for only the high frequency (> ~3 KHz) applications. Hence, to employ such sensors, it is apparent that for the low (sonic) frequency operations humungous geometrical size will be required. Thus, in order to sublime the wide applicability of the mechanical sensors, in this work, a physics bases mechanical band pass frequency selection mechanism is proposed that is universal can be adopted for selecting extremely wide range of frequencies with controlled geometric configurations. To model the envisioned band pass frequency sensors, in this work, principles and the mechanics of the human cochlea are studied. Human cochlea is the most advanced and sophisticated band pass frequency sensor in nature, where it selects the sonic frequency band (20 Hz – 20 KHz) and filters all the infrasonic and ultrasonic frequencies using a device length of only ~ 35 mm (sub-wave length scale device). Inside the cochlea, the Basilar Membrane (BM) is naturally designed based on the variable stiffness model, starting from the base to the apex of the cochlea. During selecting and filtering the desired frequencies, the BM performs four major operations; (a) it create local resonances; (b) it captures only the chosen frequencies and remain unresponsive to the other frequencies; (c) it senses the input frequencies with a sensory medium (called hair cells); and in turn (d) it spatially selects the frequencies. Inspired by the cochlear mechanics, mimicking the functionalities of the basilar membrane, in this PhD dissertation, a mechanical frequency selection mechanism is proposed exploring two diverse innovative designs (1) Acousto-Elastic MetaMaterial (AEMM) model and the (2) Basilar Membrane (BM) model. Two approaches are adopted in designing the AEMM based mechanical sensor; (a) stop band technique, (SBT) and the (b) band pass technique, (BPT). The proposed AEMM consists of a heavy core mass encapsulated in a matrix inside a stiff frame. AEMM’s are recently proposed for stopping the acoustic frequencies and create the acoustic band gaps. Using SBT method, several AEMM models are studied to create very large stop band, such that, all unwanted frequencies in the environment can be filtered and user defined frequencies can be passed through the device, automatically. However, it was found to be challenging. After several unsuccessful attempts using SBT, the new BPT method is adopted where local resonance is the key in selecting a specific frequency. Using BPT, by filtering the other possible frequencies, automatically, it is intended to develop a model which is only able to select the target frequency. Using BPT, it has been found that the proposed AEMM structure is able to mimic the functionalities of the basilar membrane and a distinct frequency can be selected by efficiently placing a smart material capable of electromechanical transduction (e.g. piezoelectric material) inside a unit cell AEMM. It has also been reported that a broadband frequency is possible to be sensed using a multi-cell structure with a systematic selection of model parameters. Comprehensive studies with analytical, numerical and experimental approaches are performed to establish the hypothesis. AEMM model uses geometric configuration and the physics of local resonance by mimicking the functionalities of the basilar membrane. However, the mechanical frequency sensor based on exact BM model is not available. Hence, in this dissertation a real geometric configuration of the basilar membrane is adopted to serve the central objective. Using BM model, two designs are proposed; the plate model and the beam model. The plate model is preferred over the beam model, where a continuous frequency band is necessary to select without losing the intermediate frequencies. Alternatively, beam model is preferable for the precise selection of the discrete frequencies within a target frequency band. In the plate model, a trapezoidal membrane is designed, whereas, in the beam model, a series of beams supported at the ends with linearly varying lengths are proposed to fit the trapezoidal basilar geometry. In recent years, notable attempts were made to fabricate the broadband frequency sensors. Although, few experimental studies have been reported to fabricate band pass sensors mimicking the mechanics of the basilar membrane, a true predictive model to design these sensors is missing. An ultra-fast and versatile model is necessary such that it could be used for the optimization of the model parameters. Non availability of such predictive model hinders the optimized design of the cochlea type sensors tailored to specific applications. Hence, in this research, two novel predictive models (plate type, beam type) for the band pass frequency sensors are proposed, mimicking the tapered geometry of the basilar membrane. It is aimed in this dissertation to develop the most flexible/versatile predictive models with all possible variable parameters that contribute to the frequency selection process. The models are developed in such a way that they can be employed for the optimized design of the sensors for wide varieties of scientific applications, respectively. Hence, the predictive models developed herein not only capable of handling the homogeneous model parameters but also capable of managing the functionally graded model parameters. This study reports that using the proposed predictive models, it is also possible to manipulate the attributes of the target frequency band using the functionally graded model parameters. The model flexibility based on the functionally graded parameters will allow the used to alter the geometric configuration of the envisioned sensor for a selected specific designed frequency band. Studies, using the finite element method (FEM) confirm the outcome of the proposed predictive models and prove that the innovative proposed model presented in this dissertation is even couple of orders (~ at least 3 times in a conventional personal computer) faster than its counter FEM model. In addition to the introduction of bio-inspired mechanical band pass sensor, in this research, few novel applications of the proposed sensors are identified and envisioned, discussed herein. Two major applications in Mechanical and Biomedical engineering are identified, respectively. Mechanical application is in the realm of energy harvesting using the AEMM model and the biomedical application using the BM model is identified in the realm of pathogen identification where it is possible to sense and detect the presence of mycotoxins, a carcinogenic metabolite excreted by the fungal pathogens. . In this work, very promising power densities were recorded using the AEMM energy scavengers. This motivates the harvesters to be employed for powering the low power electronic gadgets. On the other hand the characterization and the genus identification of the fungal pathogens can be achieved by classifying their secondary metabolites called mycotoxins. A BM based cantilever beam design is proposed to detect the presence of the type of the mycotoxins in the environment.