Benchmarking Surface Position from Laser Velocimetry with High-Speed Video in Impact Experiments. (2015)

• Record Type: Journal Article
• Authors: Howard, Marylesa; Luttman, Aaron; Machorro, Eric; Kelly, Rand; Blair, Jerome; Matthes, Melissa; Pena, Michael; Hanache, Michael; O'Toole, Brendan; Sipe, Nathan; Crawford, Kristen; Meehan, B.T.; Hixson, Robert
• Is Part Of: Procedia engineering. Volume 103(2015)
• Page Start: 221
• Page End: 229
• Publication Date: 2015
• Subjects: Photonic Doppler Velocimetry -- Light Gas Gun -- Data Analysis -- High-speed Video; Engineering -- Congresses; Engineering -- Periodicals; Engineering; Conference proceedings; Periodicals; 620.005
• DOI: 10.1016/j.proeng.2015.04.030
• Languages: English
• ISSNs: 1877-7058

Abstract

Abstract: Photonic Doppler Velocimetry (PDV) is a heterodyne laser interferometric technique for computing the velocities of free surfaces moving up to tens of kilometers per second. This information can be used to infer material properties such as equation of state and phase transitions, many of which are unknown even for common materials like steel. Broadly speaking, the methods of computing velocity from the voltages measured on an oscilloscope are either local or global in time, either frequency-based or phase-based, and formulated either statistically or deterministically. It is important to understand how velocities extracted using the different classes of methods relate to each other and how the results relate to measurements from independent diagnostics. In this work we present computed surface velocities of a flat plate of stainless steel impacted by a projectile traveling approximately 4 km/s from a light gas gun, using several different extraction methods, and we benchmark the results of the different PDV analyses against high-speed video captured at 5 million frames per second with a hybrid framing-video camera. The different extraction methods all show the same large-scale structures in the computed velocity profiles–and they agree with the high-speed video at the appropriate time scale–but they each show different small-scale features. We discuss the nature of these features, including descriptions of the numerical artifacts that one would expect with each of the different analysis techniques applied. Descriptions of the methods are provided, with a focus on the Local Polynomial Approximation method, and its uncertainty quantification, developed by the authors.