During my time on the Verification Hardware team at Running Tide, I designed the mechanical components of our first-generation wave sensing buoy, the Accel Buoy. The goal of this project was to develop a small form factor buoy that would measure sea surface temperature and wave conditions. This goal required several new tech developments including building Running Tide’s first firmware-only electrical/software architecture, and—on my end—developing Running Tide’s first diecast hardware. The end product was the world’s smallest and cheapest wave sensing buoy, which the CTO of Microsoft called “world class”.
Introduction
In the Fall of 2022, the verification team at Running Tide set out to add a new buoy to our sensing fleet which would record sea surface temperature and wave condition data, to be fed into our algae growth and carbon removal quantification models.
Over the course of the development process, I designed three major revisions—a 3D printed proof of concept, an initial diecast "pathfinder" enclosure, and the final production model. Each progression served to isolate a section of the problem (e.g. seal design) and derisk any novel components. The design process involved a few key challenges, which I'll highlight here.
A progression of Accel Buoy prototypes.
Seal Design
I worked closely with the electrical team to select the diameter of the buoy such that they could fit all custom electronics, with a clear view of the sky for the GPS and cell antennas. Once we settled on 4.5" as the minimum feasible diameter, I got to work designing the o-ring seal. Using the Parker O-Ring Handbook as a guide, I designed a face seal that would keep the electronics safe over the buoy's months in the open ocean.
The challenge of the seal design was in its precarious balance between sealing robustness and weight minimization. Since the seal was all the way at the top of the buoy, any weight it added would shift the buoy's center of gravity up and make it less stable in the water (discussed more in the next secion). To minimize weight, I designed a small aluminum clamp ring to compress the seal, and used Solidworks simulations to shrink the mass as much as possible. To simulate the seal clamp, I estimated the spring constant of the o-ring and calculated the total distributed force it would apply to the clamp ring. The simulation estimated a minimum o-ring compression of ~20%, which was later confirmed on the real hardware.
The CAD simulation of the o-ring seal clamp ring.
The o-ring in the final seal design.
Enclosure Size & Shape
As hinted about in the seal design section, the Accel Buoy enclosure design process was a tedious game of tradeoffs between shrinking the overall size while retaining enough reserve buoyancy to keep the antennas safely above water and ensure the buoy was self-righting in any position. For the buoy to be self-righting, the center of buoyancy (i.e. the center of volume below the water) must produce a righting moment around the center of gravity in any position. Any changes to the buoy size or shape (e.g. increasing the overall diameter) simultaneously affects the center of buoyancy and center of gravity, making it hard to find an optimized enclosure shape.
To find the optimal enclosure design, I wrote a Python application that allowed me to upload a given CAD model of the buoy along with its mass properties. The program then analyzes the 3D mesh in all orientations and finds the volume of the buoy that would be underwater in each respective orientation, given it's total mass. The center of this underwater volume is the buoy's center of buoyancy. Analyzing the location of the center of buoyancy compared to the center of gravity allows the program to automatically calculate the magnitude of the buoy's righting moment (the buoyant force pushing the buoy into an upright position), and thus detect if the buoy design is self-righting or not. This program greatly accelerated the iteration speed in my design process and was later used for other buoys in Running Tide's fleet.
The results of the stability simulation, showing the waterline and forces acting on the buoy. The animation on the left sweeps through all angles and shows that the positions of the blue (upward buoyant force) and red (downward gravitational force) dots will rotate the buoy upright from all orientations. This is confirmed in the plot on the right, where a negative (i.e. clockwise) moment will push the buoy upright between 0-180 degrees, and a positive moment will do the same between 180-360 degrees.
Using my new buoyancy simulator, I was able to develop an optimized enclosure that incorporated a cone-shaped design, creating more buoyancy near the top. This curved cone shape allowed me to shrink the overall size while providing enough "high buoyancy" to keep the electronics above water and the buoy self-righting. To keep the center of gravity low, I used the battery as a ballast and tapered the thickness of the enclosure's aluminum walls, making the walls thinner up top and thicker at the bottom.
Throughout the enclosure design process, I worked closely with our diecast vendor. This involved many iterations to ensure the deep enclosure was compatible with the casting manufacturing process, and the walls were thick enough to inject molten aluminum but thin enough to allow the buoy to float.
Details of the enclosure design.
Lid Design
The buoy's plastic lid served a few functions—sealing the enclosure to water ingress, providing an RF-transparent top surface for satellite signals to pass through, and providing a non-metalic area for wirelessly charging the buoy.
A prototype of the Accel Buoy with the original flat lid. The wireless charging pad and the two satellite antennas are visible.
The original lid consisted of a flat piece of lasercut polycarbonate, which the clamp ring compressed onto the o-ring seal. However, after an extensive campaign to test the buoy's satellite connectivity under various wave and weather conditions, it was found that the flat lid created a place for water to pool up, blocking the RF signal.
Testing the satellite connectivity in a wave tank in the Maine winter.
Based on these satellite test findings, I decided to switch to injection molding, which would allow me to dome the lid so water could not pool up. This switch also made it possible to add more features to align the wireless charger, ensuring that the metal charge pad would also not interfere with RF signal.
A production buoy with the final injection molded lid.
Final Product
After a validation test campaign including pressure testing the seal, performing drop tests, and deploying open ocean drifters, the Accel Buoy went into production. The final product met all of the design criteria with flying colors, weighing in at 1.3 lbs and costing ~$300 at production scale—making it by far the smallest and cheapest wave sensing buoy ever developed.
In 2023, the Accel Buoy was deployed on multiple Running Tide carbon removal deployments, and was demoed at COP 28. When presented to the CTO of Microsoft, he called the buoy design "world class".
Accel Buoy being demoed at COP28 in Dubai.
Accel Buoys being deployed off a container ship in the North Atlantic.
An Accel Buoy being deployed for a carbon removal deployment.
Sea surface temperature data from an Accel Buoy drifting near Iceland & Greenland over the course of 9 months.