Tuesday 20 December 2016

Garage Harmony




I recently spoke to the team who won the US Wave Energy Prize– Alex and Max, to find out more about the AquaHarmonics buoy. I wanted to figure out what they did that made their concept come out on top. After careful consideration, I think what won them the prize was a combination of a smart design philosophy that had control at its core, along with being the dream-team for any kind of scrapheap-challenge-type competition. Let’s have a quick look at their concept and the control method. I’ll highlight the technical features that I think were really important for them. Then I’ll relate a little story Alex told about how he and Max met, and how they came to be tinkering with wave energy in Alex’s garage.

  

Buoy

 

The AquaHarmonics buoy is a small surface piercing float, reacting against the seabed with a pretensioned cable. Inside the buoy, the cable wraps around a horizontal drum, and the rotation of the drum relative to the buoy is impeded by an electrical generator. They had entered the competition intending to control their device using latching/declutching, but their control strategy soon evolved.

In the 20th scale tests there were however fewer restrictions on control. They made use of proportional integral control, using coefficients based on the peak frequency of the wave spectrum. What this means is that the force demanded by the generator was the sum of a constant times velocity and a constant times displacement.


Design philosophy


Proportional integral control is not new, but the design focus was. Instead of considering control as an isolated subsystem, in which case the goal becomes optimising capture width (the wave energy version of ‘efficiency’), control was integrated into the design objective of improving the cost of energy. There were several design choices that turned off the beaten track towards high capture efficiency, and in every case these resulted in lower design loads. For example, as they were damping the motion of a cable, power was only captured when the cable was in tension. They did not capture as much power as was available on each cycle, but as a result, they only needed to provide structural strength in tension.

Lower design loads translate into lower CapEx and OpEx, but it is difficult to quantify this at such an early stage. The Wave Energy Prize deserves credit here for providing benefit-to-effort metrics that take into account things that are often only considered an after-thought. These include device structural strength, mooring watch circle (footprint), peak mooring loads, capacity factor, end-stop events, and the time and energy required for adaptive control of the working surface (geometry control).

The AquaHarmonics control system included a form of geometry control, and specifically avoided extreme loads at the expense of power capture. In storm seas they had lower mooring loads than some of the sea states where they had chosen to prioritise power capture.



Not 'reinventing the wheel', and some good luck


Alex is aware that the international community will want to scrutinise how AquaHarmonics measures up to what has been tried before. Somehow their device got shuffled to the top of the pile – was this the luck of draw or have they got something special here? Alex didn’t think they’d done anything radical. In fact, they intentionally made sense and direction of other people’s work, using this as a springboard. They are also keen to put their ‘stuff out there’. During the competition, they shared their ideas and data with fellow engineers, who in return gave them valuable advice. Papers downloaded from the Pelamis website made them realise that for a heaving device they needed to reduce the weight of their original design. A review of existing concepts highlighted the importance of avoiding hard end-stops and snatch loads.

Alex also acknowledged that they would not have been able do what they did ten years ago. It would have been a major engineering project. Their first concepts were controlled with a Raspberry Pi (cheap minicomputer); their modelling used open source software WECSim. Alex offered their success as proof of what complete non-experts in the field of wave energy can do if they have access to open source tools and open access journals.

It also seems that there was some luck involved in Max and Alex meeting in the first place.


How Alex and Max came to be tinkering with wave energy in a garage


It was when Alex mentioned that before going to college, he had worked as a commercial fisherman in Alaska, a diesel mechanic, a machinist, and a fibreglass boat fabricator, that I started to suspect that some of the skills that had won him the competition had been picked up before his engineering degree at Oregon State. I asked Max whether the same thing applied to him.

‘Maybe. I was interested in computer games as a kid. That got me tinkering with computers and programming. It is really Alex who is driving this project though. I don’t do anything special. I’m replaceable’. Alex was suppressing a snigger. I asked him if Max was being modest. He said ‘Max is being modest, because that is Max. Let me tell you a little story’.

‘We were roommates in our first year at Engineering school. We were the only two students who hadn’t filled out the form for roommate preferences, so we got put together. Max didn’t take notes in lectures. He played video games when everyone else was studying. Then he finished his tests before everyone else, and aced every test.’

‘I really struggled. I went back to college at 23. Everyone else was 19. I had to teach myself how to learn again. I was really busy because I was working at the same time as a motorcycle mechanic; plus I was on the Oregon State Formula SAE team. Luckily I had Max as a roommate. He would talk me through equations five times before I understood.’

‘When we graduated, neither of our jobs were very rewarding. I started working on wave energy in 2010 in my garage as a hobby. I’d heard about wave energy at Oregon State. The resource is huge, and it seemed interesting as an unsolved problem despite all this money being pumped into it. At first we were very naive to the problem. We thought it was easy. We modelled waves using high school physics – as simple sinusoids. Then, when we started building things, and had control boards blowing up motors, we realised the practical challenges involved. We were embarrassed to tell people what we were working on.'

‘At about the time we were talking about getting it into a wavetank, we heard about the wave energy prize. We thought, “Well obviously that’s the thing to do.” The competition accelerated the technology development and our understanding. Because we are close friends, we have easy communication. I can trust Max with the programming; Max can trust me with the manufacturing and making things happen. Max and I are each other’s yes men.’

3 comments:

  1. Thank you Ally for keeping us updated about Wave Energy Prize. However, I cannot find any formal technical reports regarding WECs participated in this competition. I would like to find out how they calculate each performance metric and what are the scores for individual components. Have you seen any other related documents except Wave Energy Prize Rules?

    ReplyDelete
    Replies
    1. I couldn't find anything either, other than what's on Google, namely:
      as already mentioned, the rules: https://waveenergyprize.org/downloads/Wave_Energy_Prize_Rules_5.26.15_R1.pdf (metrics described in appendices),
      the FAQs http://waveenergyprize.org/downloads/FAQ_07142016_FINAL.pdf (Q12 & Q24 give a general description of the metrics),
      and a description of the devices at the point when they were entered into the competition: http://www.altenergymag.com/article/2015/09/technical-summaries-of-wave-energy-prize-official-qualified-teams/21560 (note that the AquaHarmonics technology had changed by the end of the competition).

      I agree it would be very interesting to see how different teams scored for different components, mainly because that would give an indication of which things were considered the most important influences on cost of energy. I don't yet know whether any more information is due to be released. The limited information on the web was the reason for interviewing the winning team.

      Delete
  2. I read in some 'blogging for dummies' manual about the best practice of acknowledging your personal connection to the subject matter. I try to do this but it feels a bit naff - very few people actually do this. However I need to now as someone has just congratulated me on the control system that won Aquaharmonics the WEP. So, just to be clear...

    I was just as surprised as anyone else when AH won - two regular guys I'd never heard of before made something in their garage that beat lots of more experienced, more qualified teams. I decided to track them down and find out how they won. After my first interview I thought 'they are good at making stuff, and they got lucky' and I wrote up my first draft. But some of their control choices didn't quite make sense, so we had a second interview. I realised then they were onto something and removed some of details from the first draft because it had become clear to me that they could be trade-secrets.

    When I published the version you see now I already knew I wanted to work with them on a control project of some sort. However the content for this blog was written before this realisation.

    I am very excited to now be collaborating with Aquaharmonics and Paul Stansell on a Wave Energy Scotland control project. We will be extending some work on reinforcement learning that Paul started a few years ago, and building algorithms that aim to reduce the cost of energy.

    ReplyDelete

Comments

Note: only a member of this blog may post a comment.