Powering the Future with Paul Browning of Mitsubishi Hitachi Power Systems
Paul Browning fell in love with gas turbines early in his career. Joining General Electric fresh out of Carnegie Mellon University, he was soon assigned to “one of the sexiest things that a materials engineer can work on.” After turbines got him interested in the power business, he rose through the ranks to become president and CEO of Mitsubishi Hitachi Power Systems Americas. We sat down to hear how AI is reshaping his industry.
What’s your vision for the power plant of the future?
We actually have a pretty clear vision, because we just broke ground on it … a power plant in Takasago, Japan, that’s going to be 65 percent fuel efficient. This is a natural gas power plant and is going to be capable of autonomous operation … it’s going to be smart enough to self-diagnose any kind of issue it has. It’s going to be smart enough to schedule its own maintenance … It’s going to be a power plant that’s, you know, you don’t need to have the control room at the power plant. You can have the control room somewhere remote. Really there’s not going to be a lot for people to do in that control room because the power plant’s going to know how to operate itself.
So what’s the role of humans going to be in that power plant of the future?
In a billion-dollar natural gas power plant, there are only about 35 employees. So it’s not like coal and nuclear that have hundreds of employees. We don’t have a really big financial incentive to try to displace human labor in a natural gas power plant, and the same is true for renewable projects as well … What we’re going after is much less unplanned maintenance, which is very expensive for customers. Also shorter maintenance intervals… less downtime and a more predictable operation of the power plant, better reliability. The other thing we get from our AI is a much more flexible power plant. That’s important in an age of intermittent renewables, where our natural gas power plants need to be the balance between supply and demand. When renewables are fluctuating due to the weather, the natural gas power plant fills that gap. So we can turn the power plants down to lower levels or we can ramp them at higher rates … During those dynamic periods is when artificial intelligence and machine learning can really help us understand where the edge of the operating envelope really is and help us operate closer to that, but safely and reliably.
What led you to this focus on artificial intelligence?
Energy is really fundamental to human prosperity. Think about what happened in Asia over the last two decades, where we lifted over a billion people out of extreme poverty into the global middle class. We also lifted a lot of people in the global middle class into higher levels of prosperity, and a lot of that came on the back of coal-fired and oil as our energy sources … At a great benefit for humanity, however, we also created a lot of CO2 emissions during those two decades, which are now creating a climate change challenge for the planet. We still have another billion people we have to lift out of extreme poverty, and several billion people that want to move from the global middle class up to higher levels of prosperity, and so we have to do something different in the next decade. What we think is going to happen is a combination of natural gas, renewables, and energy storage. We think, in order for that to happen, artificial intelligence is going to be critically important. Not just for the reasons I just explained—to make the gas turbines work better with renewables—but also because we’re going to have so many distributed resources in this new model … That grid is going to be a lot more complex than today’s grid, and it’s going to require a lot more intelligent oversight than we have today and probably an oversight that humans are not going to be capable of.
What challenges have you encountered going from strategy to implementation, especially with AI or automation?
We had to go into this one without a tic-tac-toe game plan that we just executed. We knew where we were going to start, but we had to figure it out along the way. We were plowing new ground. When we do something that’s like something we’ve done before—so for example, we make a brand-new gas turbine that’s a lot more fuel efficient than the last one we did—well, we actually know how to do that, right? We’ve done that before. We’ve had prior generations of gas turbines where we developed a more fuel-efficient one. We know how that innovation process works. But when we’re trying to do something new, like develop an artificially intelligent autonomous power plant, we’ve never done that before and so we had to find our way there. Having a willingness to change your plan along the way, that’s not an indication that you had a bad plan. It’s just an indication that you’re doing something really ambitious and challenging. If you actually understood everything you had to do on day one, you probably weren’t pushing yourself enough, and you weren’t being ambitious enough.
What are you most excited about right now that you’re working on?
I get a lot of personal satisfaction just out of this idea that prosperity and climate change are two problems that we need to solve simultaneously, and we’re one of the energy companies that can actually help the world figure that out. I really do get excited about that. We sponsored Carnegie Mellon University to do a carbon intensity index for the U.S. power sector, and our plan is to, over time, expand that around the world. Keeping ourselves focused on the idea that affordable electricity is really important to human prosperity, but that at the same time, we’ve got this huge challenge of climate change… What’s going to solve it is markets and technology, and that means companies like ours.