Ask HN: What interesting problems are you working on? • Hacker News Board

Re: Ask HN: What interesting problems are you working on?

#531

Post by peterwoerner » Wed, Jan 29, 2020, 4:00 PM UTC

tobmlt wrote: post #482

peterwoerner wrote: post #449

Earlier quoted context omitted.

What library are you using for automatic differentiation. I am working on building code to optimize (and later build) high quality finite element meshes for structural analysis. For the initial proof of concept, I am simply doing finite differences, but would prefer to eventually add AD. I am unsure which packages are suitable (currently all numpy and scipy).

Both in the python version and so far in c++, I am using my own forward mode implementation in Numpy and Eigen, respectively. (Why? Well, it was easy, I wanted to learn, it’s been fast enough, and most critically, allowed me to extend it by using interval valued numbers underneath the AD variables) Here’s where I do something kind of funny In the AD implementation: Basically just write a class that overloads all the…

Thanks! I am pretty sure nobody does direct optimization on the mesh quality because it is hefty. I did come across a PhD thesis which was doing it for fluid structures interactions and his conclusion was it was inferior to other techniques. I have a few tricks which will hopefully make the problem more tractable.

I use FEMAP at my day job have found Laplacian smoothing and FEMAPs other built in tools have been wanting.

I am currently thinking that my goal is to try and use reinforcement learning to build high quality meshes. In order to do that you need a loss function and if you are building a loss function you might as well wrap an optimizer around it.

Re: Ask HN: What interesting problems are you working on?

#532

Post by jlevers » Wed, Jan 29, 2020, 4:02 PM UTC

syllable_studio wrote: post #475

dpflan wrote: post #209

Earlier quoted context omitted.

How large is a typical system - how much land you need to excavate?

If we attach our installation to an existing reservoir, we'll take up nearly zero land above ground. If we build a new self-contained upper reservoir it will be about 0.5 miles on each side and 40 feet deep. It can be built with material excavated from the lower reservoir. This may seem large, but it's for a huge amount of storage 20GWh - enough to balance the load of a large city. And keep in mind that it's about th…

Why would a new upper reservoir need to be so wide and shallow, rather than having much less surface area and being much deeper?

Re: Ask HN: What interesting problems are you working on?

#533

Post by sm001 » Wed, Jan 29, 2020, 4:09 PM UTC

I'm working on mobile apps to help researchers study dolphin acoustic communication, such as DC Dolphin Communicator 2019 which is free and open source on gitlab: https://play.google.com/store/apps/details?id=sm.app.dc&hl=e...

During a sabbatical trip in the Canadian Arctic in 1975, I came in close proximity with a beluga in Hudson Bay and was impressed with the unusual vocalizations which was in-air and about 3 feet from me. The beluga was tragically killed by Inuit a few minutes later. That's how my interest started. I later learned the basics from two people who were leaders in this field.

Re: Ask HN: What interesting problems are you working on?

#534

Post by syllable_studio » Wed, Jan 29, 2020, 4:13 PM UTC

JoeAltmaier wrote: post #498

syllable_studio wrote: post #467

Earlier quoted context omitted.

Thanks! Yes, and our research agrees that the short-life of Li-ion is a problem, and it's one of the reasons why we believe our solution has so much more promise than Li-ion for grid scale storage. The pump/turbine technology we use is the same that's been used for a hundred plus years for traditional pumped hydro dams, and the maintenance cost is very low. The life of a project is 40+ years. And in reality it can be…

Hydro has more loss, which is worrisome. Maybe 5X the loss per cycle than battery?

Hmm, I think what you're alluding to is that pumped hydro is about 70-85% efficient and Li-ion is sometimes quoted at 100% efficient (in theory). But here are some more details.

In reality, when Li-ion batteries are installed in a large system, I believe the round-trip efficiency is quoted much closer to PHS. Sorry, i can't find the best research to cite right now, but here are a couple sources i found with a quick search.

"lithium-based ESS rated for two hours at rated power will have an AC round-trip efficiency of 75 to 85%." https://www.windpowerengineering.com/how-three-battery-types...

https://www.sciencedirect.com/science/article/abs/pii/S03062... "Conversion round-trip efficiency is in the range of 70–80%"

This one says 90-95% https://researchinterfaces.com/lithium-ion-batteries-grid-en...

I've heard that Li-ion installations can get up to 90-95% round trip, which is fantastic, and better than PHS for sure. But it's not the most important detail in the equation. Here's why:

One thing to remember is that power is lost all over the system in conversion and transmission. So raw efficiency can be less important than getting the right capacity to the right place on the grid. And that brings us to cost.

Even though PHS is a little less efficient than Li-ion, 85% for PHS is still really good. (see other my other comment below about 70% vs 85%) And the math shows that investing in PHS is simply cheaper -- even after assuming that Li-ion will drop in price by 3x in the coming decades. This is partly because Li-ion has a much shorter life span and needs to be replaced about every decade.

Li-ion is still great and super important! But it's not looking like the best contender right now for grid-scale storage.

Re: Ask HN: What interesting problems are you working on?

#535

Post by jlevers » Wed, Jan 29, 2020, 4:13 PM UTC

davehcker wrote: post #329

I'm growing the freshest lettuce, iron-rich kale, and a lot of other leafy greens! While in college (CS & Math), I got heavily interested in growing food in the most efficient and healthiest way possible. I was a dreamer when I started so I thought more of how to grow 'earthly' produce on Mars, but then I realized that my own planet Earth is so massively underserved. It's basically like this- I mastered growing leafy…

Amazing! It would be awesome if people living far from traditional agricultural areas could access fresh greens without insane transportation costs (both financial and environmental).

Are your farming systems fully automated? If so, has that been more of a software challenge, or more of a mechatronics challenge?

Re: Ask HN: What interesting problems are you working on?

#536

Post by syllable_studio » Wed, Jan 29, 2020, 4:20 PM UTC

DoctorOetker wrote: post #528

syllable_studio wrote: post #475

Earlier quoted context omitted.

If we attach our installation to an existing reservoir, we'll take up nearly zero land above ground. If we build a new self-contained upper reservoir it will be about 0.5 miles on each side and 40 feet deep. It can be built with material excavated from the lower reservoir. This may seem large, but it's for a huge amount of storage 20GWh - enough to balance the load of a large city. And keep in mind that it's about th…

This would correspond to a height difference of about 1 km between upper and lower reservoir right?

yup! More fun facts about a 1 km head height... Off-the-shelf turbines are actually spec'd for a max head height of something closer to .5 km. So the design calls for a double-drop. This design approach is taken from the DOE research linked on our website.

Re: Ask HN: What interesting problems are you working on?

#537

Post by tmamic » Wed, Jan 29, 2020, 4:23 PM UTC

jlevers wrote: post #529

tmamic wrote: post #493

Earlier quoted context omitted.

I believe Atlantic had such an article recently. Good starting read is "How dentists rip us off" from Readers Digest. However, it's a bell curve. There are extremely moral and extremely immoral people. Some of them are dentists.

> There are extremely moral and extremely immoral people. Some of them are dentists. Absolutely true. However, it seems that other areas of medecine have better systems in place to prevent abuse, and dentristry would do well to follow suit.

Yes, dentistry is in many ways different from the rest of medicine. It's kind of separated from it. However, that is not the source of the problem. It's the fact that average person uses dental care more often than any other medical care, and dental treatments are in vast majority medical operations rather than drugs.

Let's focus on the second part of that statement. It means that majority of cost of dental care goes to the practitioner, rather than to drug makers. This means they have more reason to cheat. The payoff is higher.

Re: Ask HN: What interesting problems are you working on?

#538

Post by syllable_studio » Wed, Jan 29, 2020, 4:24 PM UTC

jlevers wrote: post #532

syllable_studio wrote: post #475

Earlier quoted context omitted.

If we attach our installation to an existing reservoir, we'll take up nearly zero land above ground. If we build a new self-contained upper reservoir it will be about 0.5 miles on each side and 40 feet deep. It can be built with material excavated from the lower reservoir. This may seem large, but it's for a huge amount of storage 20GWh - enough to balance the load of a large city. And keep in mind that it's about th…

Why would a new upper reservoir need to be so wide and shallow, rather than having much less surface area and being much deeper?

Good question, it doesn't really need to be, those numbers are partly just to visualize it. But we do have some reasons to keep it with more surface area: - less digging - less reinforcing needed - it's more stable - In some cases we're interested in floating solar on top of the reservoir which wouldn't work well if the reservoir was too deep.

But it's certainly not out of the question to go deeper instead.

Re: Ask HN: What interesting problems are you working on?

#539

Post by jlevers » Wed, Jan 29, 2020, 4:24 PM UTC

a_diplomat wrote: post #458

I'm a diplomat working on international norms for cyber and information warfare. I'm trying to get countries to agree on how to use and not use their capabilities, the influence on global democracy, the connection to armed conflict and the future of interstate relations. In practice, this means meeting a lot of people and spending a lot of time negotiating with other countries in scrappy conference rooms in the UN an…

> I'm a diplomat working on international norms for cyber and information warfare.

I didn't know any such norms existed. What are some of the existing agreements, and if you can talk about it, what are some of the new ones you're trying to push forward?

Your career sounds crazy...in a good way! Was your initial involvement with the UN in a technical role?

Re: Ask HN: What interesting problems are you working on?

#540

Post by jlevers » Wed, Jan 29, 2020, 4:30 PM UTC

tobmlt wrote: post #137

1.) A solver for the unstructured Euler equations. ...I was intending to volunteer time for an local university project investigating parallels between Holographic light with orbital angular momentum and hydrodynamics (in this case the Euler/Madelung equations). Not sure what happened as... volunteers get lost in the shuffle? Anyway, the solver is fun. 2.) Porting my Python code for nonlinear gradient driven optimiza…

Would you be willing to explain what the applications of (2) are? I'll admit that I only undersetand a fraction of what you said in that section, but I'm curious what you're using it for.