A Data-centric engineering workshop held at the Samberg Conference Centre, MIT, Cambridge Massachusetts, followed by a visit to Northwestern University to celebrate Greg Olson's change of purpose. The hotel in Cambridge (Kendall) used to be a fire station.
 Giant slide rule - accuracy scales with length |
 Fire buckets - two of the buckets have odd shapes (rounded bottom and conical) in order to prevent them from being used for other purposes. |
 Snowed on the second day in Cambridge MA, but otherwise quite a warm spell |
 The building ahead is the Samberg Centre. |
 |
 |
 The Charles River adjacent to MIT |
 |
 |
 |
 The sun then came out |
 |
Methods such as thermodynamic calculations of phase diagrams, irreversible thermodynamics, kinetics, first principles calculations, finite element analysis etc. are now routine in research on new metallic alloys. They do not on their own enable a complete solution because real materials have extreme complexity. For example, there are thousands of possible solid-state phase transformations that can be exploited to engineer the properties of multicomponent steels, leading to myriads of possibilities for novel combinations of properties.
Remarkably, there is no method that enables the mechanical properties to be estimated, given a complete description of the chemical composition, structure and processing, reducing the entire alloy design procedure into one of trial-and-error.
This is where machine learning founded on a Bayesian framework comes to the fore in exploiting the vast quantities of accumulated mechanical property data. It will be demonstrated that with careful design, the method has physical foundations and is able to be used predictively, well beyond the bounds of the original learning set. This will be illustrated with specific examples of successful alloy design that has been scaled to commercial components that are safety-critical.
 |
 |
 |
 |
 |
 |
 |
 Steel building under construction |
 |
 No idea what this is about, but the plastic bags seem to be tied with deliberate intent |
 Phil Withers from a distance |
 |
 |
 |
 Close-up views of the steel-framed building being constructed |
 |
 |
 |
 This is Chun Te Wu, a student of Professors Hung-Wei Yen and Jer Ren Yang |
 |
 Charles Kuehmann and Greg Olson |
 |
 Carelyn Campbell from NIST |
 Peter Voorhees |
 |
 Jason Sebastian and Greg Olson |
 Aziz, from Questek |
 |
 |
 Zi-Kui Liu |
 Johnathan Montgomerey, Watertown Arsenal |
 Johnathan Montgomerey, Watertown Arsenal |
 |
 |
 |
 |
 Alex Umantsev |
 |
 |
 |
 |
 Amit Behara and Aziz from Questek |
 4th edition, 2017 |
 Free download |
 Free download |
| PT Group Home | Materials Algorithms |