“He said it multiple times throughout that meeting,” True recalled. “We heard him. From there, we decided we needed to find a better solution, and we overall just needed to get better.”
Bridger began working with IBM to develop a better leak detection system that relies on artificial intelligence, which “feeds off a tremendous amount of data, and we have plenty of it,” True said.
“Throughout our pipeline system, we have approximately 15,000 different sensors that via satellite put data into our control center,” he said.
That includes data such as flow rates, pressures and whether a pump is turned on or off. The information is fed into the control center in Casper, Wyo., where Bridger is based.
“It takes all these different types of data and says, ‘This is a normal situation,” True said. “Any deviation from that model is an exception, and that is what it is supposed to flag as a warning or an alarm.”
The more data input over time, the more the technology learns about the pipelines it monitors. The goal, True said, was for the new system “to be able to learn very, very quickly.”
True said the company’s existing leak detection system — and others on the market — had trouble adapting to changes along pipelines, such as added pumps and newly connected wells.