Repairing an e-bike battery

Door mux op donderdag 19 oktober 2017 20:46 - Reacties (7)
Categorie: Videos (PowerElectronicsBlog), Views: 1.421

Eerder deze week postte ik een video waarin ik een Bosch mid-drive e-bike motor 'repareerde', en vandaag is het tijd om de accu onder handen te nemen. De originele accu is 288Wh, maar houdt duidelijk een stuk minder capaciteit vast. Daarom upgrade ik hem met nieuwe cellen, zodat de capaciteit naar 500Wh gaat

Repairing a Bosch e-bike (part 1)

Door mux op maandag 16 oktober 2017 09:30 - Reacties (12)
Categorie: Videos (PowerElectronicsBlog), Views: 2.808

Recent heb ik voor 150 euro een elektrische fiets met Bosch middenmotor gekocht, voor gebruik in de heuvels van de Eifel (=Ardennen, maar dan in Duitsland). Voor de kenners: ja, 150 euro is veel te weinig geld, maar de fiets kwam met een aantal gebreken. Eťn daarvan was een motor die een vreemd geluid maakte, de ander was een bijna kapotte accu. In deel 1 repareren we de motor!

Let's look at nuclear power. Part 3: Fission reactors

Door mux op maandag 9 oktober 2017 11:34 - Reacties (14)
Categorie: Techniek, Views: 2.536

Nuclear power is a contentious subject, for a bunch of reasons. But really, what are those reasons? Even if you have an opinion on the subject, more often than not that will mostly be based on qualitative arguments, not an in-depth understanding of the subject. Today I would like to start to change that.

Nuclear power plants

Today we'll be putting the theoretical knowledge gained in part 1 and part 2 into practice and see how we can convert that nuclear binding energy released from nuclear decay into electricity. Well, we do it by heating up water into steam and running it through a big turbine, but why? There are big drawbacks to that.

Energy conversion
Energy is part of a model of reality we use, but energy exists in wildly different forms. Light has energy, it is expressed as the wavelength of photons. Thermal energy fundamentally exists as the momentum of the atoms that make up a mass of material. Electric energy is a combination of electric and magnetic fields. Don't even start about gravitational potential energy, it's crazy when you think about it. Just the position of matter seems to affect energy levels there.

More fundamentally speaking, types of energy exist as quantum mechanical properties of particles, and if you want to convert one type into another, it has to undergo a compatible interaction. So let's look at one of the easiest types of interaction: the photovoltaic effect.

Photons of visible light - the most abundant type of electromagnetic radiation by energy content on earth - have an energy of between 1.8-3eV. That means that if one of these photons interact with an atom, all that energy is transferred to, most likely, the outer electrons on that atom. A couple electronvolts is not a lot of energy, so if that atom were in a vacuum, the electron would simply go into an excited state and after a while re-radiate an identical photon when it fell back down to its ground state. In a semiconductor, the electron would actually be able to free itself from its parent atom and move around, with some trickery we can direct this flow of electrons and generate an electric current. That's how solar panels work - at least the photovoltaic ones.

But there is more nuance to it than that. If you give an electron one eV of energy, it gains an electrical potential of one volt. That's why we call it electron-volt. So you would expect blue light falling on a solar panel to cause it to output about 3V, or less with green or red light. But in reality, solar panels output only about 0.6V, less than one-third of red light, and five times less than blue light. What is that all about? And where did the other energy go?

Well, the first part is still correct: a photon hit an electron, and it went walkabout. But not before shedding its excess energy - anything beyond 0.6V - as heat. This interaction is actually pretty complex, as multiple energy conversions take place. And here am I telling you that this is a simple example... However, for the part that we are interested in: light energy is directly being converted into electric energy. That's direct conversion.

Nuclear reactors on the other hand almost never have this luxury. First of all; the energy levels are literally a millionfold higher. Even if you would be able to make some kind of direct conversion device, we're talking about particles with about a MeV - that's a mega-electronvolt, or million electronvolt - of energy each, and quite a lot of spread too (some are 1MeV, some 10MeV, some 100keV). You wouldn't just lose two-thirds of your energy - you would only be able to do direct conversion at a fraction of a per cent efficiency. And let's say you can make a 'nuclear panel' that converts these high-energy particles directly into electricity like a solar panel would. You would be dealing with 100kV-10MV - something that for many reasons is not practical or even possible to deal with on semiconductors. More importantly, even if you could deal with it you run into another problem: the types of particles produced by nuclear reactions are generally not electrically charged. Most energy goes out in neutrons or neutrinos, and those do not interact with the electromagnetic force. There is no direct pathway to electricity, which is what we eventually want. Direct conversion is out of the question.

There is another problem. Whatever your energy receiver is - it is made out of atoms. The energy levels present in fission and fusion products is often in excess of the binding energy of atoms, so when you allow these particles to just slam into your energy receiver it would probably break these atoms apart by the dozens, eroding the material - hell, transmuting it even - and quickly deteriorating the usability of your reactor. You have to gradually 'slow down' these fast neutrons so they can be absorbed as relatively harmless heat without devastating your reactor internals. And, well, this fact combined with the lack of uniform interaction methods means the only way you can really get any energy from a nuclear reaction - fission or fusion - is thermally. Boiling water and running it through a turbine.

The plant

It's incredibly important to understand that a nuclear power plant is much, much more than a nuclear reactor. Even though we will be almost exclusively talking about the reactor in the coming chapters and blogs, the vast majority of a nuclear power plant is all the other bits.

As all current and in all likelihood most future nuclear power plants are thermal plants, the biggest features in a power plant are going to be the turbine and cooling facilities. You see, just making something very hot by itself is not conducive to producing electricity. You really need a temperature difference to create power. The absolute maximum amount of energy you can extract from a difference in temperature is given by the Carnot equation:

In this equation, η is the total efficiency, TL is the absolute temperature of the cold side of your heat engine and, you guessed it, TH is the hot side. You can see immediately that if your cold-side temperature would be zero, the efficiency would be 1, engineer speak for 100%, perfect. Unfortunately, that temperature is absolute - you would need to make the cold side colder than the temperature of outer space. In reality, on earth you can only really hope to cool something down to a little over room temperature: 300 K. This fundamentally limits the efficiency of a thermal engine, and you really want that hot end of the engine to become as hot as possible to get the most electricity out of your heat.

So those giant cooling towers you see are really essential to the power plant. And they are so big for good reason; they need to dissipate massive amounts of thermal energy. Think about it: a so-called boiling water reactor heats water to only about its boiling point - 100C. Using the equation from earlier, this yields a thermodynamic efficiency of only 20%. If a power plant makes 1GW of electricity, it then has to remove 4GW of heat. To do that, cooling towers evaporate lots and lots of water. Not everywhere in the world is it easy to come by large amounts of water, so in rare circumstances you will instead find this:

This is what is known as a dry heat exchanger, and it is really not much more than a massive CPU heatsink. A maze of pipes containing hot liquid - usually water - and massive amounts of fins and fans to transfer that heat into the air. This approach does require much more land area, mainly because the alternative - evaporating water - is just so good at dissipating heat.

Beyond the cooling towers, we find the actual power-generating beast. This is usually a turbine generator - essentially an aircraft engine in reverse. The hot steam from the reactor, which is under quite a lot of pressure, really wants to expand but there is a big fan with lots of rotors in the way. By expanding, these fans are rotated and when you are really clever about the amount of blades, their pitch and their overall geometry, you can get the majority of that theoretical efficiency from such a generator. This turbine, in turn, drives an electric generator that actually moves the electrons into the wires.

Not unimportant to mention is those wires. Because nuclear power plants, for various reasons, are generally giant beasts outputting power in the gigawatts [GW], you need a lot of on-site transformers, distribution switches and electric safety equipment. One of these plants can power a small country, so it's also not a bad idea if the power plant is situated in a convenient location with respect to the nation's power grid.
One of dozens of transformers for the Beaver Valley nuclear power plant

Lastly, whereas most other types of thermal plants - e.g. coal and gas - are usually fairly compact, nuclear power plants often have a lot of sprawl associated with them. This is for good reason; a NPP generally has a bunch of duplicate infrastructure and failsafes installed. A well-situated NPP is located not just near a high-capacity water mains, but near a natural source of water in case not only the water mains fails, but the emergency pumps do as well. On-site facilities like chemical reprocessing are often duplicated at least in part, and many facilities have generous safety zones associated with them, so that emergency services have quick and easy access to any part of the plant that could go wrong, even though there are often multiple failsafes already.

So, now that this is out of the way, we can finally talk about nuclear reactors.

The fuel cycle

The design of a nuclear reactor is mostly dominated by the type of fuel it has to run on. You see, all nuclear fission reactors have to run on radioactive material, but not all isotopes are created equally. Generally speaking, isotopes can be divided into two categories: Fissile and fissionable. Fissionable isotopes are - for all intents and purposes - all radioactive materials that undergo prompt decay, i.e. that decay when they capture a neutron. Fissile materials are a subset of fissionable materials that only undergo prompt deacy after capturing a low-energy neutron, or slow neutron.

At first, this may make no sense. But look at it from an energy perspective: fissile materials undergo fission at much lower energies, so they must be less stable, more prone to decay, i.e. it's much easier to sustain a chain reaction with fissile material. This is why fissile materials are more suited to nuclear power generation. You need a controlled reaction to get significant power from nuclear fuel; without a chain reaction all you have to go on is natural decay, and this is incredibly slow at natural rates. That is why uranium and thorium ores aren't glowing white hot all the time.

Natural radioactive isotope ores have very low densities of actual uranium, thorium and other radioactive species - in the hundreds of ppm or about 0.05% range. The first step in making nuclear fuel is therefore simply separating out all the desirable elements. The radioactive isotopes, then usually in extraordinarily pure form, have to be further refined to get the right amount of fissile isotopes in the fuel. Also, generally the isotopes are oxidized - both to prevent accidental chemical oxidation during use (i.e. a fire) and because the oxides of uranium, plutonium and thorium are more stable and have a higher melting point than the base metal.
One of a couple of uranium recovery methods from mined ore

Up to this point, generally this process of refining has been done in solution, i.e. a series of chemical reactions that purify and chemically alter the state of the fuel. At this point the fuel is ready to be packaged into fuel rods, which means the uranium oxide or whatever other isotopes are used get pressed together and sintered (the small granules of material get molten together at high temperature) and the resulting pellets are packed together into a large tube known as the fuel rod. The tube is made from materials that pass through thermal neutrons very easily, while also being tolerant to heat, water and most fission products.

Fuel rods are surprisingly small - typically about 1cm (half an inch) across, and about half a meter to a meter (1 yd) long. This is because as you can imagine, when you make fuel rods, you do not want them to melt from prompt decay while grinding them down to size. The fuel rods by themselves are of course much more radioactive than raw ores, but surprisingly not by too much - they produce very little energy by themselves. Most neutrons produced by radioactive decay inside the rods are way too fast to be captured by the fuel, so they fly out of the rod. The real power is produced when multiple rods are placed close to each other, in bundles. Neutrons from one rod then pass through a moderator, for instance carbon rods, and get slowed down enough to cause prompt decay in another rod. That is how you produce power. The moderator is again something you have to consider carefully; it should slow down neutrons but not completely block them, it should not transmute to problematic isotopes when hit, it should be tolerant of high temperatures and water, etc. Water and carbon - specifically graphite - are the most common choices in nuclear reactors worldwide. Typically, we name reactors by the moderator; if light water (H2O) is used we call it a light water reactor (LWR), if heavy water (deuterium oxide, D2O) is used it's a heavy water reactor. If the moderator doesn't reduce neutron energy (but some other mechanism is used) we call it a fast reactor or fast neutron reactor.

And this gets into the really fun part: disposal. At the end of the fuel cycle, when the nuclear fuel rods are depleted to a point where they cannot generate enough power anymore, they have to be disposed of safely. This is also where you find out a lot of interesting details about nuclear reactors. One of the main reasons for fuel rods to be so thin, is because uranium has very poor thermal conductivity; similar to other heavy metals like lead. This means that the very inside of a fuel rod gets significantly hotter than the outside. If fuel rods were badly designed, this is one of the biggest reason for cracking and uneven burn-up. Of course, burn-up doesn't literally mean the fuel has been burning - it is typically already an oxide - but that it has transmuted into other species. These species build up inside the fuel, 'clogging up' the process and causing fission to become less efficient the longer it goes on. And because some parts of a fuel rod have been facing more other rods, those sides of the rods tend to burn up quicker than the 'outsides' of a bundle. Oh, and some fission materials are gases, so they produce voids in the material. Really, this is a fascinating part of nuclear reactors.
Yep, this is actually perfectly safe. That bundle will be producing over 100MW of power, yet now it is not dangerously radioactive or hot

So disposal typically really starts with inspection. Fuel bundles are carefully disassembled and inspected in the correct orientation, checking for possible cracking. Small parts of the fuel that have come off are a major hazard, even if it is just the cladding (the tube that the fuel sat in, or other not initially radioactive material). After that, depending on where in the world you are, the fuel rods are just kept the way they are and put into a spent fuel pool or caskets, or the fuel is disassembled further, ground up and chemically reprocessed. Often, only a tiny fraction of the total nuclear binding energy present in the fuel has been consumed, and the only reason the fuel rods output lower energy is those fission products that are in the way. So with a lot of chemical reprocessing, the fuel can be reassembled, combined with new fuel and loaded into so-called mixed oxide or MOX fuel rods. Unusable fission products - which is much less than the initial amount of waste - are casked and stored underground for a couple millennia.

Of course, this is not all. During operation, some water will have transmutated to contain tritium (radioactive hydrogen) as well as one of the daughter isotopes of oxygen and nitrogen. The carbon moderators, as well as plenty of other parts of the reactor will have been irradiated and have to be replaced.

But for some perspective on the amount of actual waste in this entire endeavour: the entire fuel pack is something a very strong person could probably carry. The full reactor is surprisingly small. There is not a lot of waste, it is just very hard to process waste. And these fuel bundles last quite long, typically a year or more. Compared to chemical energy, nuclear energy is extremely dense.


I hope you enjoyed this third part. We are not nearly done yet. Today you learned about how a typical nuclear power plant works - it pretty much has to use thermal conversion because there is no other way to get energy out of nuclear reactions directly. Also, most of the nuclear power plant has very little to do with nuclear fission; it's mostly cooling towers, chemical reprocessing, power transformers and that kind of stuff. The actual reactor is tiny. The fuel even more so - it's just a couple dozen or hundred really thin fuel rods facing each other, with a moderator in between. And when they run out, you can reprocess them into new fuel if you want. Saves a lot on that pesky waste.

This blog was entirely about water/carbon-moderated nuclear power plants, the most abundant type in the world. In the next blog, we will be talking about Thorium, that much-hyped element that seems to solve all of nuclear's problems. Does it, though?

[video] E-bike conversion Flevobike Basic (051)

Door mux op dinsdag 11 juli 2017 17:35 - Reacties (8)
Categorie: Videos (PowerElectronicsBlog), Views: 1.884

Ligfietstijd! Het is al even geleden dat we het over m'n elektrische-fietsprojecten hebben gehad, en in deze video toon ik de conversie van mijn ligfiets naar elektrisch. Ook bekijken we in detail hoe een goedkope chinese motor controller er van binnen uitziet

[video] A deep dive into ADCs

Door mux op dinsdag 4 juli 2017 16:15 - Reacties (9)
Categorie: Videos (PowerElectronicsBlog), Views: 2.094

Microcontrollers hebben bijna allemaal een ADC erin zitten. Maar hoe werkt die, en hoe zorg je ervoor dat je hem goed gebruikt? Dit is een opvallend complexe vraag, en ondanks dat deze video meer dan 20 minuten is kom ik niet eens in de buurt van dit uitgebreid genoeg uitleggen. Maar hopelijk geeft het wel een inzicht in de complexiteit van dit onderwerp. Enjoy!