CAD Model

O-Seal Torus Design

Full 3D model of the O-Seal device. Drag to rotate.

O-Seal CAD model, drag to rotate 360 degrees 360° · Drag to rotate
Engineering Views

Inside the device

Cutaway view of the O-Seal: quadrant removed, exposing the hollow inflation chamber, flow bore, iris stack and turbine
Cutaway. A quadrant removed: the hollow inflation chamber inside the outer shell, the open flow bore, and the micro-turbine hub.
Exploded view of the O-Seal subassemblies along the flow axis
Exploded. Top to bottom: inflatable shell with sensor pods and acoustic modem, strut cage, inner ring with electronics bay, iris stack, micro-turbine.
O-Seal underside with the iris fully open
Transit: iris open. The full aperture keeps water moving through the device while it rides the flow.
O-Seal underside with the iris closed and the shell inflated
Sealing: iris closed. The blades throttle the channel while the shell inflates against the pipe wall.
O-Seal inflated inside a pipe section, seated beneath a red fracture
On station. Shell inflated against the pipe wall, seated directly beneath the fracture.
Components

Engineering breakdown

Rigid Inner Ring

Maintains an open flow channel for normal water throughput while providing structural support for the capsule.

Flexible Outer Shell

Durable polymer housing that protects internal components and allows the capsule to navigate through pipe bends and junctions.

Pressure Sensors

Miniaturized sensors continuously monitor local pressure differentials to detect leak signatures with high precision.

Acoustic Communication

Short-range underwater modems coordinate multiple units for complex repairs. Through-wall transmission is in development, with bench validation underway.

Autonomous Control

Onboard microcontroller regulates navigation, sensor processing, and iris deployment decisions independently.

Iris Mechanism

When a fracture is detected, the iris closes to throttle the flow channel while the outer shell inflates against the pipe wall, anchoring the capsule and creating a secure long-term seal at the leak site.

3D Model

Iris Visualization

An O-Seal flows in with the water, reaches the crack, and inflates to seal it. The leak stops and the fracture goes green.

Live 3D · Seal sequence Iris mechanism of the O-Seal device
Prototype Testing

Internal testing results

97%
Water retention rate achieved in internal bench testing
<1s
Full outer-ring inflation upon detecting a leak signature
90 min
To seal 3 unknown breaches across 2 km of pipe with a 3-unit swarm
$400
Per-repair cost at scale. Traditional excavation runs $10,000–$15,000.
Sealing

97% Water Retention

The functional prototype generated sufficient pneumatic pressure via CO² cartridge to halt fluid loss, retaining over 97% of water that would otherwise have escaped through the breach.

Speed

Sub-Second Inflation

High-frequency pressure sensors and algorithmic pneumatics trigger full outer-ring inflation in under one second upon detecting a leak signature. Faster than any human-dispatched repair crew can respond.

Coordination

Autonomous Acoustic Swarming

Multiple capsules coordinate without human intervention or cellular signals, using localized acoustic sound-wave communication to hand off tasks and systematically clear a pipe segment.

Academic Context

Physics-Informed Neural Networks in pipe mechanics

O-Seal's detection engine is grounded in peer-reviewed fluid mechanics research. Here is why PINNs outperform conventional AI in this domain.

01 / Physics Edge

Not a black box

Unlike traditional AI that requires millions of real-world failure examples, a PINN embeds the exact mathematical laws of physics (including the Navier–Stokes equations for fluid dynamics) directly into its loss function. The model cannot violate physics.

02 / Data Efficiency

80% less training data

Academic research shows PINNs can accurately model hydraulic states and identify anomalies like pressure drops using up to 80% less training data than purely data-driven models, making them ideal for lightweight, real-time deployment on microcontrollers inside a moving capsule.

03 / Inverse Problem

Backtracking the leak

Identifying the size and location of a leak from only internal pressure readings is a classical inverse problem in fluid mechanics. A study in Computer Methods in Applied Mechanics and Engineering demonstrated that PINNs excel at solving these inverse problems in high-velocity fluid flow, accurately backtracking a subtle pressure wave to its exact physical origin.

04 / High-Frequency

Hundreds of reads per second

O-Seal’s neural network is trained on extensive synthetic datasets mimicking real-world fluid-structure interactions. It processes pressure readings hundreds of times per second to differentiate a dangerous pipe crack from normal operational fluctuations such as a pump turning on.

Specifications

Technical parameters

<1 s Inflation Speed ▾
Upon detecting a leak signature, the CO² cartridge system triggers full outer-ring inflation in under one second. High-frequency pressure sensors and algorithmic pneumatics enable this response time, which is validated in internal bench testing.
97% Water Retention Rate ▾
In rigorous internal development and testing, the functional prototype successfully generated the necessary pneumatic pressure to halt fluid loss, retaining over 97% of the water that would otherwise have escaped through the breach.
16 bar Pressure Rating ▾
The O-Seal can withstand operating pressures up to 16 bar (232 psi), which exceeds typical municipal water system pressures of 3 to 8 bar. This ensures the seal remains intact even during pressure surges or water hammer events.
5+ years Seal Lifespan ▾
Once deployed and anchored at a leak site, the O-Seal’s inflatable torus and sealing materials are engineered to maintain their integrity for over 5 years, providing a long-term repair solution without requiring excavation or replacement.
Acoustic Communication ▾
O-Seal units communicate via localized underwater acoustic modems, enabling real-time coordination between capsules without cellular signals or human intervention. This enables swarm-like task hand-off for complex multi-breach repairs.
Flow Power Source ▾
A micro-turbine harvests energy from the water flow and recharges the onboard Li-Po buffer, so missions are not limited by a single charge. The harvester keeps the unit powered for long deployments inside the network.