Search

In this article, ultrasonic phased arrays are deployed as an imaging tool for industrial process analysis. Such arrays are typically used for sonar, medical diagnosis, and nondestructive testing; however, they have not yet been applied to industrial process analysis. The precise positioning of array elements and high frequencies possible with this technology mean that highly focused images can be generated, which cannot currently be achieved using ultrasound tomography. This article aims to highlight the potential of this technology for the measurement of bubble size distribution (BSD) and to demonstrate its application to both intrusive and noninvasive process measurements. Ultrasound images of bubble reflectors are generated using the total focusing method deployed using a 32-element, 5-MHz linear phased array, and an image processing algorithm for BSD determination is presented and evaluated under stationary and dynamic acquisition conditions. It is found that the sizing accuracy is within 10% for stationary reflectors larger than $4\lambda $ in diameter and that the algorithm is stable across the expected spatial variation of reflectors. The phased array is coupled to a six-axis robotic arm to scan a solid sample containing bubble reflectors at velocities up to 500 mms−1. The sizing accuracy is within 45% for bubbles larger than $4\lambda $ in diameter and at velocities up to 300 mms−1. However, above this velocity, the algorithm breaks down for reflectors smaller than $9\lambda $ in diameter. The ultrasound system is applied to a stream of air bubbles rising through water, which is verified via photographic analysis. Images were generated both intrusive and noninvasive, via a 10-mm Perspex barrier, to the process stream. The high bubble density in the process stream introduced scattering, limiting the measurement repeatability and the sample size in the measured distribution. Notwithstanding, this result demonstrates the potential of this technology to size bubbles for intrusive and noninvasive process analyses. [ABSTRACT FROM AUTHOR]

Many industrial processes involve particles in a carrier fluid, and it is often of interest to monitor the of these particles noninvasively. The aim of this paper is to develop a theoretical model of multiple particle-wall impact vibrations that can be used to recover the particle size from experimental data. These vibrations have been measured by an ultrasonic transducer attached to the exterior of a vessel containing a stirred-particle-laden fluid. A linear systems model is derived for the response of the piezoelectric ultrasonic transducer, which has a single matching layer. The acceleration power spectrum of these vibrations has been shown experimentally to contain information on the size of the impacting particle. In particular, the frequency of the main spectral lobe is inversely proportional to the particle size. We present a theoretical model that agrees with this empirically observed phenomenon. The theoretical model is then used to simulate multiple particle-wall impacts, with each particle impacting at a randomly chosen location. A set of theoretical vibration spectra arising from multiple particle-wall impacts are integrated and compared with the experimental data. The ability of this approach to distinguish between different particle sizes is clearly shown. [ABSTRACT FROM AUTHOR]

A method of obtaining particle size and concentration information, from ultrasonic transducer measurements of particle-wall interactioas in a particle laden fluid, is presented. A mathematical model of the flexural vibrations of the vessel wall due to the particle impact is constructed. The key component of this model is the derivation of an analytic expression for the impact force amplitude. An analytic expression for the power spectrum is then obtained that shows its explicit dependency on the system parameters. This spectrum is then integrated over a specific frequency range and a comparison with experimental results is reported. [ABSTRACT FROM AUTHOR]