Mary Biddy
Mary Biddy’s assignment sounds like a task from
Mission: Impossible: understand
the inner workings of a “black box” mdash; a
device that takes a known input, does something mysterious to it,
and yields a different output.
The “black box” is third-party software with proprietary modeling algorithms that are hidden. Biddy feeds in data and scans the results.
“If I’m trying to run a simulation and the data is not converging, then I look at the variables and try to figure out why,” she says. “But there is always a part of the code that I can’t see.”
A year ago, Biddy received her doctoral degree in chemical engineering from the University of Wisconsin after researching vegetable-based lubricating oils. Today, she is a senior research engineer in the Aromatics and Acetyls Division of BP plc, one of the world’s largest oil companies.
The black boxes are part of mammoth models used to optimize the operation and profitability of nearly 20 aromatic chemical plants worldwide.
It’s a complex problem. Each plant operates somewhat different equipment. The cost of utilities fluctuates daily. So does the composition of chemical raw materials, which vary with the type of oil the refinery processes that day. Sometimes equipment breaks down, so plants need to find alternate routes to make the same products.
Most dauntingly, BP’s optimization model works in real time, adjusting to each new fluctuation in customer orders, prices, chemical inputs, and available equipment.
The model’s backbone was supplied by Aspen Technologies, which specializes in oil and chemical simulations. It also includes BP-developed models. Biddy has full access to BP’s models, but Aspen’s models are a trade secret.
When confronted with simulation problems, Biddy falls back on skills she learned in graduate school. “You can use logic to figure out why you’re not getting the right answer,” she says, “but you have to do it methodically and try to understand what’s important for each component of the model.”
Tickling the secrets out of massive models is a far cry from Biddy’s original plans. Although she studied chemical engineering at Texas A&M University in Kingsville, she planned to become an environmental lawyer.
Then she took an internship at Johnson Polymers, a subsidiary of SC Johnson & Son Inc. “My mentor thought I was naturally curious and would enjoy the research side of engineering,” Biddy recalls.
In graduate school, she pursued her environmental interests by looking at ways to make lubricating oils from renewable plants. “I found several ways to improve their physical properties, but the modifications reduced the oils’ biodegradability,” she says.
Biddy used molecular modeling to predict how modifications changed the physical properties of biolubricants mdash; an approach she might not have taken were it not for her DOE CSGF.
“The fellowship required that I take math and computer science courses, such as a class on algorithm development, that would not have been part of the normal curriculum,” she explains. “It gave me a more rounded education.”
Biddy eventually chose to join BP because of its strong commitment to green technology and the challenging work it offered mdash; even if her mission meant finding the possibilities in a black box.
