14, Aug 2014 / rebound


And just like that, Phase I of the IcePoint™ NSF project is complete! Well, perhaps time has slowly dragging by for those of you hoping for new blog posts, but it flew by for us. Since we last touched base with you, the project rapidly developed from a warehouse with little more than potential into an operational prototype system.

Although it came down to the wire, the testing turned out well. On the last night of testing before the report was due, we generated 2 kW of refrigeration below -20°C, squarely nailing our big goal for this system. We’ve got all the components operating but are still working out some of the kinks of getting the whole system to play together reliably.

As planned, we learned a lot about IcePoint™. Just as beneficial, however, was the prototyping and development experience we gained along the way. Some primary take-aways:

  • The ice tank, cornerstone of the IcePoint™ concept, works largely the way we predicted. Some engineers we spoke with were scared that we would just be making impenetrable blocks of ice that provided no refrigeration capacity or melted at 0°C instead of -20°C.  On the contrary, the ice inside the tank performed just as modeled, forming a well behaved packed bed that we could easily interact with.
  • Other elements of system performance did not quite live up to expectations. For example, to improve the separation system, we tried using an intermediate fluid to help draw water out of the main brine loop. Unfortunately, the literature values used, based on a similar approach at a different concentration, did not extrapolate well to our test conditions and the result was an underperforming separation system.
  • I learned how to plastic weld!
  • We reinforced the idea that perfect is the enemy of good. We spent a few weeks on really interesting modifications to improve the separation system performance (after the first idea mentioned above didn’t pan out), but when push came to shove, achieving the proper validation in a timely manner required a quick and dirty, low tech solution

Above: Luke welding and a sampling of materials we tried for improving the separation system performance- they’re all naturally colored!

  • It’s important to “fail fast”. On the success side, we started testing the ice tank early in the project. We had to rebuild it a few times since the early versions had problems with ice flow as it fell into the brine. Thankfully, due to the early start, the tank was working great by the time the complete system was ready for testing. On the more-to-learn side, we could have done a better job with the waste heat portion of the system. We made the ever-so-common mistake of believing components would function per design just because they were “off-the-shelf” commercial products. It turns out the water heater used to simulate our waste heat source had a tolerance of 20°C on its set-point temperature. Given that few things work well the first time, we’ll keep getting better at failing fast and fixing quickly
  • Really cool material discovery. Here is one of our favorites – fun to play with even though it ended up being a poor fit for our needs. 

Moving forward, we are spending a couple months refining the 2kW design and prototype operation. Our goal is to get the system operating reliably and smoothly to prepare for the next phase, where we will be designing, building, and testing a system 5X larger than will be connected to real freezer loads. Yum, ice-cream!


Leave a Reply

Your email address will not be published. Required fields are marked *