This is the second article in a two-parter on pumped storage hydro and the US Department of Energy’s FAST competition, intended to find ways to speed development and deployment of this key technology. The first article dealt with the basics, including the inhibitors to speed of delivery. This article spends more time on Team Livingston’s innovative approach to rapid tunneling.
Most pumped storage hydro proposals assume the use of tunnel boring machines, and indeed, I made the same assumption in proposing that Elon Musk join the Boring Company and Tesla Energy to build pumped hydro. Livingston’s insight after digging into the subject was that tunnel boring machines were a key source of expense, a single point of failure, were economically unsalable after a single use, and need to come out the other end of tunnels as they don’t back up. They have the advantages of relatively continuous tunneling and automation, but still represented an unreasonably high portion of the cost and appeared too slow. That’s two strikes on the economics.
Livingston decided to find an alternative, and he thinks he has: oscillating disk mining equipment.
That’s an oscillating disc cutter (ODC) in action, chewing through hard rock at a much faster pace than other mining and tunneling technology. The solution has been in development since at least 2006. It features a hard metal disk with a somewhat conical cross-section that provides a wedge shape at the edge. The wedge is pushed into the rock face at an angle and vibrated into and out of the rock rapidly. The major advantage is that it flakes off much larger pieces of rock with each cut than a typical pressure wheel with rotating cutters pushed into the rock face. Bigger flakes equals much faster cutting.
While tunnel boring machines will cut through anything from hard rock to sand, they aren’t typically used where the the rock is fractured or heavily sheared. They also have a tendency to fail when underground conditions fail, with expensive and long-lasting tunnel construction delays with the machines stuck underground for a year or more. They continuously tunnel an unfinished hole, but the hole must be finished and fortified behind them and they can’t go back. In many cases, they are simply entombed forever at the end of job. Because of their lack of blasting, they are often used under urban areas for the necessary tunnels.
Oscillating disk technology is mostly being developed in the mining sector for the room and pillar method mining of ore bodies, where vast rooms are cut through the ore and pillars are left in place to hold up whatever is above.
The mining industry doesn’t consider itself in the same way as the tunneling industry. The economics are different. Among other things, every ton of ore extracted in mining is a source of revenue, while every ton of rock extracted in tunneling is just another cost in a year’s long expenditure.
Livingston’s insight was that the oscillating disk cutter technology being developed by Komatsu and Sandvik would be excellent in tunneling. He says:
”Oscillating disc technology has been demonstrated to reduce the breaking energy required for rock excavation by as much as 90%. This force reduction is being utilized to increase tunnel excavation rates by as much as 400% over road header machines. The further benefit of ODT’s quick deployment and lower cost, is the simultaneous deployment of many machines at multiple portals to lower construction time risk. The technology enables longer tunnels so typical in higher head PSH.”
Other vendors are making oscillating disk technologies, but not ones that are suitable for pumped hydro, while other ones are further behind in development. Livingston is under NDA to a couple of companies for their technology, but is able to say that he believes that at least two suitable technologies will reach the market commercially within two years, plenty of time given the long lead times for regulatory portions of the development.
No more entombed or otherwise unsalable machines. Machines that can be pulled back when they run into conditions that require the human touch. Machines that are much cheaper. Machines that will likely chew through softer rock even faster than they chew through hard rock. So what’s the catch?
Well, they aren’t continuous tunneling machines. Room and pillar construction makes it easy to push a machine into a rock face and pull it back after a bit to get the ore out. No room for that and no sense building lots of side tunnels if you need to keep going for 13 kilometers (8 miles) across the upper and lower tunnels in the proposed site.
A big part of Livingston and Conroy’s pitch to the DoE was to leverage soon-to-be-commercialized oscillating disk technology in a mostly continuous tunneling model. This would entail not only having the oscillating disk cutter, but also automated ceiling bolting and shotcreting for tunnel stabilization, both of which exist in various places, but currently aren’t attached to the expected commercial machines. It would also require narrow gauge automated mining trucks passing each other in a stream going to the face and then back out potentially for kilometers.
One of the conversations I pulled together with Livingston and Conroy was with my occasional collaborator, David Clement. Pulling together two threads, David is the deepest machine learning expert I know personally and co-author with me of the machine learning tutorials based on the Plastic Dinosaur, defining its body, its neural network brains, attention loops, bias in neural networks, the virtual to physical loop of learning, and the core machine learning concept of salience.
In discussions with David on other projects, we discussed Livingston’s approach, and realized that there might be a play for a machine learning extension and improvement. Initially, the discussion centered around using salience approaches with machine learning looking at and highlighting aspects of video feeds from the face to assist a human remote operator or geologist to spot changes or potential risks more quickly. We also discussed the geolocation of autonomous mine carts underground with fairly simple physically visible markers and what sensor sets would be required for autonomous carts, based on David’s and my backgrounds in the spaces. It’s still early days, but it will be interesting to see how that would work.
Livingston and Conroy received a cash award from the FAST competition as well as assistance from national labs to develop it over the next year. If successful, they could accelerate the delivery of pumped hydro not only in the United States, but globally. The approach is highly transferable technology. Among other things, I’ll be brokering a discussion with the Scottish developer, Mark Wilson, with the Americans early in the new year to explore the potential for Scottish use.
The US DOE recognizes the value of pumped hydro in our future decarbonized grid. Now to get banks and politicians all over it.
Have a tip for CleanTechnica? Want to advertise? Want to suggest a guest for our CleanTech Talk podcast? Contact us here.
Latest CleanTechnica TV Video
CleanTechnica uses affiliate links. See our policy here.