A picture of a construction crew somewhere deep underground, knee-deep in water that shouldn’t be there, staring at a stationary tunnel boring machine is one that you can practically picture without actually seeing it. This isn’t speculative. Something very similar occurred in Sri Lanka for almost ten years on a project now known as the Uma Oya Multipurpose Development Project, costing hundreds of millions of dollars in emergency engineering, lost energy revenue, and delays. Water was not meant to consume the tunnel; rather, it was meant to transport it.
A USD 514 million EPC contract was approved for the 15.3-kilometer Headrace Tunnel at the center of that project, with planners estimating an internal rate of return of more than 13%. These figures were predicated on the cooperation of the geology. It didn’t. In order to prevent the walls from collapsing, engineering teams had to switch to emergency grouting, both before and after excavation, as groundwater poured through rock fissures at volumes that no one had accurately modeled. The project had lost an estimated USD 700 million in power generation revenue alone by the time the delays were tallied. Nine years. By comparison, that number has the ability to make other project difficulties seem doable.
The failure itself—groundwater issues arise in tunnel projects with unsettling regularity—is not what makes this story worth revisiting today, but rather what it makes clear about the direction infrastructure engineering is taking. In particular, into the sea. Subsea mining operations are steadily progressing from concept to construction, focusing on seafloor massive sulfides and polymetallic nodules that are located hundreds or thousands of meters below the surface. Furthermore, once you’re underwater, the water problem doesn’t get any easier. It becomes much more bizarre.
In its most basic form, the mechanics of dewatering involve maintaining workspaces sufficiently dry to safely extract material. This entails controlling surface runoff during rainy seasons, process water used for drilling and ore washing, and groundwater that seeps through rock pores and fissures in a traditional underground mine. Every source acts in a unique way. The slurry that remains after mineral processing, known as tailings water, frequently contains high concentrations of solids and fine particles that, in a matter of weeks, destroy conventional centrifugal pumps. With wear-resistant alloys, sealed motors, and integrated agitators that keep compacted sediment moving rather than hardening into something unmanageable, the machinery built to handle it must adopt a completely different design philosophy.
For the worst of these circumstances, submersible slurry pumps have emerged as the most dependable equipment, and for good reason. They operate directly within the material; there is no priming cycle, no suction piping, and no exposed mechanical parts that must contend with pressure gradients. The only method that works in a pond full of mineral tailings or a deep sump at the bottom of a mine shaft is direct immersion. Applying this reasoning to subsea infrastructure is difficult because every variable is multiplied by the environment. At depth, the ambient pressure is oppressive. Timelines for corrosion speed up. Remotely operated vehicles and days of coordination on the seafloor may be necessary for maintenance access, which could take hours in a surface mine.
The subsea mining sector may still be undervaluing this. The lessons learned from projects like Uma Oya indicate that poor geological assessment during the planning phase leads to compounding effects, including political issues in addition to engineering ones. The construction delays caused water supply disruptions and subsidence in the communities surrounding the Sri Lankan tunnel, turning a technical setback into a social crisis. It’s not totally plausible that subsea operations carried out far from populated coastlines could somehow avoid that dynamic. It is not necessary to be close for environmental disruption to have an impact.
Angled hole injection of cementitious material prior to and following TBM excavation, the grouting techniques that ultimately stabilized the Uma Oya tunnel, are one aspect of a larger technical toolkit that is currently being researched for adaptation to deep-sea environments. As this develops in real time across engineering journals and mining conferences, it seems as though the industry is looking backward into hard-won tunnel-construction knowledge, realizing that the ocean floor and a Sri Lankan mountain have more in common than anyone first realized. fractures in rocks. Water finds its way. Both get worse under pressure.
Before being deployable at significant subsea depth, open sump pumping, deep well dewatering, and mechanical thickening systems that function dependably at surface-level atmospheric pressure will need to be reengineered. Although the physics are manageable—the oil and gas sector has been running machinery at these pressures for decades—the material handling requirements of mining are very different from those of drilling a wellbore. The relatively clean fluids that petroleum engineers design around will not behave at all like slurries rich in nodule fragments and seabed sediment. Whether any existing dewatering platform is truly prepared for that combination of depth and solid loading is still up for debate.
The Sri Lanka case unequivocally demonstrates that ignoring groundwater as a secondary planning concern is a mistake with very specific and expensive consequences. The accounting line that was recorded was the USD 39 million in direct added costs. Every infrastructure engineer should include the USD 700 million in lost generation revenue in their next project briefing. Dewatering is not a backup plan. It is becoming more and more the project.
