SRMP Seminar: July 17

[AEM Communications]

Good morning,

We invite you to attend the upcoming SRMP seminar.

Friday, July 17

Akerman 225

2:30–3:30 p.m.

First Talk:

Title: Data-Driven Model-Free Robust Performance Refinement of Dual-Stage Hard Disk Drive Controllers

Abstract:  Hard disk drives (HDD) continue to be essential for modern long-term data storage. The state of micromagnetic technology has vastly increased data density, yet external disturbances (e.g. vibrations, cooling fans) push the read/write head positioning off track. Modern high-capacity HDDs use a dual-stage actuator setup: voice coil motors for coarse motions, and piezoelectric actuators for fine control. Historical controller design methods are insufficient for the increased demands of next-generation, triple-stage HDDs (and beyond), which require sophisticated analysis techniques and less human-in-the-loop iterative design. This work focuses on designing an algorithm for dual-stage controller performance refinement through best replicating the ‘black box’ intuition- and experience-driven analytical methods used by control engineers. The main objective of this algorithm is to enforce in-HDD robust stability to all forms of uncertainty while pushing performance, which is proven through prescribing desired behaviors that are representative of common controller designs.

Speaker Bio: Erik Lehner is a 4th-year Ph.D. student in the ARDC Lab, advised by Prof. Ryan Caverly. His research focuses on learning methods and control theory applied to hard disk drives. Outside of grad school, he enjoys swimming, rowing, anything geography related, learning about languages, and mediocre $10 mochas from a cafe on a Saturday morning..

Second Talk:

Title: Freestream Reconstruction in Hypersonic Shock Tunnels

Abstract: Hypersonic shock wind tunnels replicate high speed flows in a controlled setting, and are essential in validating hypersonic vehicle designs. Accelerating flow to mach 12 and above is no easy feat, and these tunnels are themselves novel technologies. Extreme operating conditions preclude direct observation of the flow. Thus, free stream conditions in the test section are indirectly inferred from other measurements, introducing significant uncertainty to experiments. This work uses onboard sensor data from the test article itself, paired with Bayesian inference techniques, to find the most likely free stream conditions that explain the data. The test article in question is a geometry particularly sensitive to small changes in free stream conditions called a double cone. Double cones are sensitive because they exhibit a hypersonic flow phenomena called a Shock Wave Boundary Layer Interaction (SWBLI). SWBLI’s occur in a wide range of hypersonic circumstances such as control surface deflections, and scramjet inlets. The sensitivity of double cones to freestream conditions has resulted in repeated failures of validation between wind tunnel and CFD. By building out the freestream flow field from sparse onboard measurements, we will be able to both better explain SWBLI phenomena, and gain deeper understanding into a shock tunnel’s behavior.

Speaker Bio: Leonid Scott is entering his second year as a PhD student advised by Dr. Candler and Dr. del Val. His work combines both hypersonic computational fluid dynamics, and uncertainty quantification, to make better use of hypersonic wind tunnels. Prior to joining University of Minnesota, he developed the hypersonic CFD code, US3D, and worked as a software engineer within various non-engineering related startups. Next week, Leonid will give a similar talk in Munich Germany for the World Congress on Computational Mechanics.

--

AEM Marketing and Communications 

Aerospace Engineering & Mechanics | aem.umn.edu

College of Science & Engineering | cse.umn.edu 
University of Minnesota Twin Cities | umn.edu

110 Union Street SE, Minneapolis, MN 55455

612-625-8000 | aem-...@umn.edu