Why fuel cell cars don't work - part 2

Door mux op donderdag 26 februari 2015 22:39 - Reacties (36)
Categorie: -, Views: 16.055

This is the second part of my blog series about hydrogen fuel cell cars, and why I think they don't work. If you haven't read the first part, I highly recommend it as it lays the foundations for understanding what exactly constitutes a fuel cell vehicle (FCV) and how it is different from battery electric vehicles (BEVs). Today I will talk about production, storage and transportation of hydrogen.

This is a continuation of a blog series, here is Part 1. If you're done reading this part, I've also written Part 3.

This is an extremely long, in-depth blog series, so I'll start by giving you a summary. This summary will exist at the top of every part of this series. If you're interested in the technical details, please do read on and make sure to come back for the next parts.

First of all, HFC cars are perceived to be a good bridge between fossil fuels and full electric because:
  • You can still fill up like you do with a gasoline or diesel powered car
  • The mileage you can get out of hydrogen is perceived to be more adequate than what you get from batteries
  • Hydrogen fuel cells are thought not to wear out as quickly as batteries (or conversely, batteries are thought to wear out very quickly)
  • Hydrogen as a fuel is perceived to be a relatively small infrastructural change from gasoline and diesel
  • Hydrogen is perceived as a cleaner solution than gasoline, diesel or natural gas
In reality,
  • You cannot fill up like you do with gasoline or diesel. It is actually pretty ridiculous how hard it is to fill up a HFC powered car
  • You won't even go 100 miles on current tech hydrogen tanks that are still safe to carry around in a car
  • Fuel cells wear out crazy fast and are hard to regenerate
  • Hydrogen as a fuel is incredibly hard to make and distribute with acceptably low losses
  • Hydrogen fuel cells have bad theoretical and practical efficiency
  • Hydrogen storage is inefficient, energetically, volumetrically and with respect to weight
  • HFCs require a shit ton of supporting systems, making them much more complicated and prone to failure than combustion or electric engines
  • There is no infrastructure for distributing or even making hydrogen in large quantities. There won't be for at least 20 or 30 years, even if we start building it like crazy today.
  • Hydrogen is actually pretty hard to make. It has a horrible well-to-wheel efficiency as a result.
  • Easy ways to get large quantities of hydrogen are not 'cleaner' than gasoline.
  • Efficient HFCs have very slow response times, meaning you again need additional systems to store energy for accelerating
  • Even though a HFC-powered car is essentially an electric car, you get none of the benefits like filling it up with your own power source, using it as a smart grid buffer, regenerating energy during braking, etc.
  • Battery electric cars will always be better in every way given the speed of technological developments past, present and future
All of this is written from the perspective of somebody who helped make these hydrogen-powered racing karts happen. You have to admit, that looks super awesome!

The math behind gas compression

One of the biggest problems with hydrogen is that it is certainly a great fuel judging from energy-to-weight ratio, but it is also a very non-dense gas. The energy density of gasoline is about 32MJ/L, whereas uncompressed hydrogen only stores 0.013MJ/L. In other words, about a 2500-fold discrepancy in volumetric energy density. This can only mean one thing: we need to compress this down to fit it inside a car. Or any vehicle.

Compressing hydrogen is lossy and hard
I very briefly touched upon hydrogen compression being lossy and causing all kinds of temperature effects, but why? Well, compressing and decompressing a gas - any gas - doesn't come for free. You need to put in a considerable amount of energy.

Intuitively, you need to put in effort (work) to compress a gas. That should be obvious; it's not something that happens automatically. This energy you put in effectively gets converted into heat. We have formulas for this process, which is called adiabatic compression:

P(1- γ) x Tγ = constant

with γ = 1.4054 for hydrogen (Adiabatic index of hydrogen)

where P is pressure (in Pascal) and T is absolute temperature (in Kelvin). So if we take hydrogen at 1 bar (100kPa) and room temperature (298K), this constant value that pops out of the formula is 27.54. Now we increase the pressure to 200 bars, what is the resulting temperature?

Tγ = constant / (P(1-γ)) = 25105

This yields T = 1351K

That's scorchingly hot. Actually, it is so hot that you can't just say that we have compressed 200 liters of hydrogen into the space of 1 liter; gas at higher temperatures in a confined space will have a higher pressure. So to find how much gas is actually in there, we need to use another formula in the adiabatic process:

PVγ = constant

at 25 degrees let's say we have 200 liters or 0.2 cubic meters. The constant then becomes:

100 x 103 x 0.21.4054 = 10415

At the resulting pressure of 200 bar:

10415 / (20 x 106) = V1.4054, so V = 4.61 liters

That's much larger than 1/200th of the original volume. 4.61 times as much, in fact. The reason for this is, well, hot gas actually takes up more space than cold gas. In order to 'fit' this amount of hydrogen into a 1-liter tank at 200 bar, we need to actively cool the gas as it is pressurized. Otherwise it will go in the tank 'hot', cool off and actually be much less than 200 liters unpressurized.

You can breathe now; you already survived 33% of the math in this blog post. Here is a really pretty 4K panoramic picture of the racing circuit in Aragon, Zaragoza where we raced in the summer of 2009 (click for 4K version).


Now to the energy issue: We know that 200 liters of hydrogen are about 16.53 grams and the specific heat (i.e. the amount of energy you need to put in it to get it to heat up by 1 K) is 14310 J/kgK, so this equates to 250kJ of heat energy we put into the system. Now comes the interesting part: that 200 liters of hydrogen initially contained 1.6MJ of energy. Just compressing it to 200 bars cost us 250kJ, which is 15.7% of the energy content of the hydrogen. Let this just sink in: just the act of compressing it requires a significant proportion of the energy actually contained in the hydrogen!

To make matters worse, when we decompress it inside a car to use it as an energy source, the exact same amount of energy needs to be added to the gas, otherwise it will enter the fuel cell way too cold and cause damage. This principle is how the superconducting magnets in nuclear magnetic resonance imaging (MRIs) are cooled. In a vehicle, we can use this cooling effect to our benefit though; because fuel cells are not that efficient, the ingoing cold hydrogen can be used to cool the stack or the waste heat from the fuel cell can be used to heat up the expanding gas. But this nevertheless poses a big problem with the total (well-to-wheel) efficiency of hydrogen. If you need to compress hydrogen more than once, for instance when you need to transport it with a tanker vehicle, then store it at a fill-up station, you are going to need a significant amount of additional energy.

We're just talking about compressing up to 200 bars, which is the limit for metal containers at the moment. If you would use composite (carbon fiber, etc.) or special lined steel containers, pressures of up to 700 bars are used. Again, realize that we are trying to put a 2500 times less dense fuel into our tank, we need to compress it as much as possible. This pressure is for instance used at the California quick-fill stations for the Honda FCX Clarity. At 700 bars, if we do the math, this compression energy loss amounts to 389kJ/L or 24.3%.

Just this compression energy loss isn't all. Three paragraphs ago i said that as you pressurize the gas it becomes hot and takes up more space. In order to fit inside a tank, you need to actively cool down the gas as it's pumped into the tank. Or, conversely, if you're transferring hydrogen from a high-pressure tank in a quick fill station to the lower-pressure tank inside a car, you need to continuously heat the gas, otherwise you risk overpressurizing as the gas heats up inside the tank to ambient temperatures. The compression and decompression cycles required for distribution make up a significant amount of the inefficiency in transporting hydrogen from place to place.


Also, with the pressures and precision involved, there is no way you can actually 'fill up' by yourself without a lot of technical assistance. A lot of engineering has gone into robotic hydrogen filling stations that make sure you don't risk blowback, leakages or freezing and/or burning your hands during the filling process. It's really a marvel of technology that (recently, since about 2008) Shell has been able to build these great human-operated filling stations using high tech filling nozzles. Talking about pressure...

Storing hydrogen is probably even harder
The central issue to this subchapter is the simple fact that hydrogen molecules are extremely tiny. So small, that they can fit through the holes between atoms and molecules of any solid material. Ever heard of gas-tight fittings? Not if you're talking about hydrogen. Every pipe, every tank, every container you make for storing hydrogen will slowly but surely leak hydrogen. Even the best materials will leak it pretty fast actually, and this of course has to do with the ginormous pressure differential between the outside world and the inside of the tank, as well as the need for hydrogen tanks to be somewhat portable, so wall thickness is bound to practical limits.

Leakage rates can be calculated with Fick's law of diffusion:

J = D x p / t

Where J [mol/m^2s] is the mass flux of hydrogen, D [m2/s] is the diffusivity coefficient, p [mol/m3] is the molar density difference over the wall and t [m] is the wall thickness. Hydrogen has a molar weight of about 2g/mol and volumetric density of 89 g/m3, which makes for about 44.5 mol/m3. For best case steel, the diffusivity coefficient of hydrogen through steel is about 10-9 [m2/s] at room temperature. This means that for a steel vessel with 1 square meter surface area, 1cm thick walls and 200 bar pressure difference, the diffusion rate is:

J = (1 x 10-9 x (200 x 44.5 / 0.01) = 0.00089 [mol/s].

This is about 75 mol/day or 0.15 kg/day - in other words, such a vessel would be empty in a month or so. Again, let this sink in for a minute: a 1-cm thick, 1 square meter area steel tank that weighs as much as a human (77.5kg) will leak hydrogen so fast that it's unusable as a fuel tank. But this is not the main issue (see below, there are kind of ways around this)

The diffusivity becomes a lot higher at higher temperatures too, climbing about 2 orders of magnitude to D=10^-5 at 1000K. Why is that relevant? Let's think about distribution for a second. Turbulence and friction causes the gases in large natural gas pipelines to heat up considerably. Also, to save on additional cooling costs and power, currently gas pipelines pipe out hot gas and let the pipes themselves do the cooling. This is not feasible for hydrogen, as it would diffuse out so quickly. Really, any plain steel vessel is not usable for hydrogen storage or transmission. Worse still, all eligible metals have similar, often worse (aluminum) diffusion coefficients. Metals are not a great choice for hydrogen containers.


So why do I dwell on steel so much, can't we use something like polymer linings (~1 order of magnitude better)? The thing is, steel really is the only affordable option for long-haul pipelines. We already have really good oil and gas infrastructure in place, trillions of dollars worth. It's the wet dream of futurists to reuse this for the hydrogen economy. This is why some universities are doing studies into the exact loss rates of steel pipelines, as well as ways to increase the gas flow so that the actual losses become insignificant as compared to the amount of gas that is transported (which of course leads to the heat problem again). Because even if such a pipeline leaks a couple kg/s of gas; if it transports thousands of kilograms per second, that is an acceptable loss. Natural gas pipelines also lose an appreciable amount (about 0.5%) during transportation in these pipelines.

But, this is not actually practically feasible at the moment because one of the biggest hurdles towards hydrogen pipe infrastructure is that...

Hydrogen embrittlement makes logistics even harder!
A big problem with pipelines is a phenomenon called Hydrogen Embrittlement; as hydrogen diffuses through the container material it occasionally forms bonds with the iron atoms, transforming the iron into a different molecular structure and in the process creating additional internal stresses and even different crystal structures with additional boundaries or dislocations. Over a long period of time this lowers the maximum elastic stress and thus toughness of the steel. Contrary to LNG or oil pipelines, hydrogen pipelines need to either be significantly overengineered or replaced every 5-10 years because of this reduction in strength.

Actually, hydrogen embrittlement is not just a problem for steel, although it is by far most pronounced in steel. Aluminum and other metals as well as some polymers are adversely affected by hydrogen seeping into the microstructure. As it turns out, hydrogen fairly easily falls apart into single hydrogen atoms. Atomic hydrogen, much like any atomized material, is very reactive and will force itself into the microstructure of any material over time.

So if steel isn't usable, how do we store hydrogen?
Hydrogen storage canisters on a hydrogen-powered folding bike

But that still leaves us with the question of how to store hydrogen in a car. Steel is not really a good option. Aluminium cartridges is how we did it at Formula Zero, but these also leaked measurably and only held about 100g of hydrogen. As it turns out, this is a fundamental problem. No materials exist with less than about 10^-9 diffusion rates, which would still mean your vessel is essentially empty after a month - or a year if you store it at lower pressures. This is even worse with more modern high-performance composite tanks that are much lighter than steel or aluminum for the same amount of storage, but have even worse leakage properties. So we have to look elsewhere.

Two options are in use right now. The first is to liquify hydrogen. Hydrogen boils at 20 K - very near absolute zero - so if you cool it down to below that temperature you can carry it around practically unpressurized in liquid form. This is how most long-term storage containers worked for many decades, like NASA's Space Shuttle. The downside to this is that you need really good thermal insulation as well as some way to keep refrigerating the hydrogen tank. This is, as you might imagine, really hard. Not necessarily because insulation and refrigeration is particularly difficult, but because you can't make the tank a closed, isolated chamber like a thermos. It needs an entry/exit point that can resist high pressure and very large temperature differentials, and that is almost only possible with ceramics and metals - which conduct heat way too well. So there has to be a constant power source attached to your car, or a large battery, that can make sure that the hydrogen doesn't boil off when you are not using your car. Worse still: if something goes wrong and your unattended car loses power, the hydrogen starts boiling, pressurizing the tank and causing an explosion. So for maximum security you need either a safety valve or a tank that can still withstand these super-high pressures. This is one of those fundamental problems that make hydrogen storage inherently unsafe, whereas gasoline and battery storage can be inherently safe.

The other method is a combination of pressure, adsorption and high-tech interior pressure vessel liners. Hydrogen in gaseous form has its molecules very far from each other; pushing them together to increase the density creates a high pressure. However, some materials like activated carbon and metal hydrides can form weak but relatively stable bonds with hydrogen, causing the hydrogen molecules to stick to the surface of the liner in a very compact fashion. You still need some pressure to keep the hydrogen adsorbed - and when you release the pressure (or heat the liner), the hydrogen removes itself from the surface again - but for the same density we are talking an order of magnitude lower pressures. The downside to this method is of course the fact that any adsorbant is pretty heavy, because hydrogen is such a small atom compared to all the others and activated carbon can only adsorb at most one hydrogen molecule per carbon atom. This makes current vehicle-class adsorption hydrogen tanks about the same weight as pressure tanks for the same amount of fuel. A couple hundred kilograms of tank to store 5kg of hydrogen.

The pressure tank of a modern Honda FCX Clarity (2011+ models)

This is one of the reasons why the newest generation of Honda FCX Clarity still uses pressurized hydrogen gas tanks; at a pressure of about 350 bars.

Making hydrogen? Downright impossible

Now comes by far the most controversial part, and the one that I think gets most overlooked because it is not actually happening on a large scale at all yet. Making hydrogen is a very environmentally damaging process and although 100% clean methods exist, they will take too long to set up and are currently too expensive to make hydrogen cars affordable in the foreseeable future. Compare making hydrogen to making ethanol fuels; even though ethanol as a fuel is very 'clean' compared to other vehicle fuels, you need an unsustainable amount of cornfields to produce bioethanol. Or you need to extract it from oil, which defeats the purpose (and is less efficient and more polluting than just extracting gasoline and diesel from oil and using that).

Roughly the same problem exists with hydrogen production. There are very roughly speaking four major methods of hydrogen production:

Hydrogen as a by-product
At the moment, if you buy a hydrogen canister from Linde or DOW or some other chemical company, you get hydrogen that is most likely not actually produced with the intention of making hydrogen. Hydrogen is a common by-product from a bunch of chemical reactions used to make everyday chemicals; ammonia, nitrogen oxide, carbon dioxide for carbonation of drinks, alkali metal production, etc. etc. . Some continuous process chemical plants that make these substances are essentially already setup to produce hydrogen; at some point in the process there is an easy way to tap off some very pure hydrogen. For a while, this was actually very cheap to come by as supply was high (especially with the relatively large hydrogen flows in ammonia plants - a big component of fertilizer) and demand was low. The downside is of course that this isn't sustainable as a production method for a true hydrogen economy. Besides, the energy efficiency is very low if you really only look at the amount of hydrogen produced. Lastly, especially in the case of ammonia production, the source of the hydrogen is actually natural gas which means this process is always worse (energetically) than standard steam reforming (I'll get into that a couple subchapters down).

Hydrogen from electrolysis

This is an awesome video. Not relevant to the text, but worth viewing anyway

The big promise of hydrogen is electrolysis from water. It is very easy to do; just put two metal plates in water, add a little bit of salt to make the water conduct reasonably well, put a current through the plates and hey presto, you're producing oxygen and hydrogen from water. This is the exact reverse reaction from a hydrogen fuel cell. It's even so exact, that a perfect electolysis setup is exactly as efficient as a perfect (cold cathode) fuel cell - roughly 80%. In practice, of course, things get a lot harder.

Plain electrolysis as you might have done in high school is out of the question of course. Most plate materials oxidize within minutes, especially on the oxygen side. Efficiency is also really, really bad. This has a lot to do with the fact that it's very hard to move those ions around in water at a reasonable speed. Most of the energy you put into the water gets lost as ions formed at one electrode bounce their way through the water towards the other side (high effective internal resistance). This also limits the volumetric/gravimetric efficiency of such a setup.

Another big problem with electrolysis from water is, like in fuel cells, the effect of electrode poisoning. Any contaminants in the water will most certainly react with either the hydrogen or oxygen - or both - to eventually form a thin and slimy material on the electrodes, which further reduces the speed of electrolysis.

Don't worry too much though; these problems aren't something that really stops anyone from using electrolysis. People have found ways around this; by using various better electrode materials and catalysts (platinum group materials work best), higher temperatures, higher pressures and additives that make sure that any contaminants present go away from the electrodes to some other collection point. Electrolysis is a fairly well-understood process. We have been doing it for ages now.

There have even been developments to reduce or eliminate the required subsequent pressurization of the hydrogen, with fancy high-pressure electrolysis systems or by carrying the reaction components away in special high-pressure wicks. As we've calculated before, this can improve the total cycle efficiency by up to 25% at 700 bar. Unfortunately, as far as I know there has been close to zero development on this technology since I started being technically interested in fuel cells and the highest pressure HP electrolyzers still only operate at about 20-80 bar, negating most efficiency gains because of some inherent downsides to high pressure electrolysis (like the hydrogen and oxygen permeating through the proton exchange membrane, causing them to mix instead of stay neatly separated). Practical electrolyzers at the moment get about 25% efficiency for fuel cell grade hydrogen.

Images like this simply don't work. Nope.

Still, it's very important to grasp that hydrogen from electrolysis isn't as clear-cut a solution for hydrogen production as it's sometimes portrayed. It is actually pretty expensive, not necessarily efficient and it is fairly preposterous to for instance claim (like I've seen many times) that you can 'essentially just put the current from an offshore wind turbine into the sea water and catch the bubbles'. That sounds awesome and if it were really that simple it would be a good production method (invisible power generation as viewed from land, can easily meet the demand, etc.). But you actually have to do extensive pre-treatment of the water before it can be electrolyzed, and the electrolysis doesn't 'just work' on any input, you need a big power conversion platform and hydrogen electrolysis plant as well - adding quite significant cost. This makes electrolysis typically by far the most expensive and resource-intensive method, though it's the cleanest and most sustainable truly scalable method in the end.

Hydrogen from natural gas
The methane gas reforming process

The most promising technique for generating hydrogen currently is steam reformed natural gas. If you heat up water and natural gas to almost pyrolysis temperatures (in excess of 1000K), the natural gas naturally reacts with the water to produce carbon dioxide and hydrogen gas. These gases have wildly different densities, so the mixture quickly stratifies and very pure hydrogen gas can be scooped off the top. With a controlled inflow of gas, water and a controlled removal of carbon dioxide and hydrogen, this process can be made into an extremely fast continuous reaction. This is helped by certain catalysts. Theoretically steam reforming of methane and other n-anes (ethane, propane, butane, etc.) can be done at acceptable speeds with about a refrigerator-sized machine. Because of the long start-up time (the machine needs to heat up before every use) it's not quite possible to put a steam reforming machine into a car, but it is certainly possible to use them at fuel stations. This eliminates the need for (leaky) hydrogen pipelines and storage tanks. From a practical point of view, steam reforming is awesome. The wikipedia article still speaks of the need for complicated carbon monoxide filtering, but this is a solved problem. Multiple companies have announced or are already selling these small-scale hydrogen plants. And even though we're probably past peak oil, there is still a crapton of methane gas in the earth's crust to help us move around for the next couple of decades.

So what's the downside? Well... obviously, it still outputs carbon dioxide. A lot of it, actually. Also, the gas still needs to be compressed. If we ignore the energy cost of pumping up, refining and transporting the natural gas, as well as startup energy cost of the machine, the best case cycle efficiency of steam reformed hydrogen looks a bit like this:
  • Steam reforming process: about 70% efficiency (see below)
  • Compression to 200 bar: 85% efficiency
  • Decompression into the fuel cell: 95% efficiency using decompression cooling for the fuel cell
  • Fuel cell efficiency: 60%
This makes for a total methane-to-fuel cell efficiency of 34%. At a currently typical 125Wh/km, this means we require about 369Wh/km from the natural gas. Natural gas produces 1 mole of CO2 per mole of gas, and one mole of gas contains 891kJ of energy, or 164Wh. In other words, per kilometer we produce 2.25 moles of CO2, or 99g. that is exactly the same as a compact petrol or diesel powered car. Environmentally, you win absolutely nothing with this approach. And this calculation is fairly positively biased; I'm ignoring most other sources of inefficiency in this chain, and I'm assuming a state of the art electrical drivetrain.

Also, methane is the best possible situation: methane contains 1 carbon atom for 4 hydrogen atoms. The less carbon, the better. Any other fuel contains relatively more carbon, and thus will yield worse CO2 emissions.

As for the efficiency of any chemical process, in general it's very easy to analyze this; you simply take the enthalpy of formation for all molecules on both sides of the reaction and subtract them from each other. For instance, if we take steam reforming:

CH4 + 2 H2O <=> CO2 + 4 H2

Note that this is the reduced version of the steam reforming process; in reality this is a two-step process with carbon monoxide being involved somewhere. Now, the enthalpy of formation for these molecules is:
  1. CH4 (gas): -74.87 kJ/mol
  2. H2O (gas): -241.818 kJ/mol
  3. CO2 (gas): -393.509 kJ/mol
  4. H2 (gas): 0kJ/mol
Doing the math, this yields a difference of 164.997 kJ/mol of CH4 converted into 4 moles of H2, or 29.5% more energy on the right hand side compared to the left side. This means the reaction is endothermic: you need to *add* energy to keep it going.

This is not necessarily the entire story, though; you can also look at it from the point of view of heating value. The HHV of methane is 889kJ/mol, and for hydrogen is 286kJ/mol. But, we make 4 moles of hydrogen for each mole of methane, which means that the actual usable energy - ignoring the fact that we need to heat up and cool down some water in the process - goes up! We extract 1144kJ worth of hydrogen fuel from just 889kJ worth of methane fuel. Of course, this energy needs to come from somewhere and this is exactly the reason why the reaction is endothermic. So far so good, right?

Well, the higher heating value isn't necessarily a good measure for the amount of actual energy you can recover from both methane and hydrogen gas. In practice, the lower heating value - which takes into account that you need to heat up and/or cool down the water consumed/produced by a reaction - is a better approximation of what you can actually expect to get. The lower heating value of natural gas and hydrogen are 801 and 242kJ/mol respectively, yielding about 121% actual usable fuel efficiency from this process, for 129.5% energy consumption - or a total theoretical maximum process efficiency of 93.4%.

This is roughly how you do these calculations. In practice, the reason for the much lower than theoretically possible efficiency of steam reforming is the fact that this is actually a two-step process. A lot of energy is lost in the carbon monoxide reaction, reducing the theoretical maximum to about 85%. I'll leave the math to this as an exercise to the reader.

A rough illustration of the effectiveness of hydrogen from the gas reforming process. I'd say this is NOT an effective way to win the race

Hydrogen from plant matter, coal and other heavy fuels
There are other ways to do steam reforming though, and another big possibility is to use more abundant fossil fuels as well as biofuels. But... if you've read the previous chapter already, you'll be asking: how is this better? Short answer: with heavier fuels we can do carbon stripping and carbon capture.

Carbon stripping is a slight modification to the steam reforming process. Methane is the ideal input to steam reforming, because it has relatively uncomplicated chemistry and it's 'easy' to control the process such that very little undesired compounds get created. When dealing with larger carbohydrates - for instance biodiesel typically has a mixture of molecules 12-20 carbon atoms in length. The different molecules have different optimum temperatures, pressures and partial pressures for steam reforming (or any chemical reaction, really). The solution? Break down complex hydrocarbons to methane, and then do steam reforming. This is called carbon stripping. In general, the chemical reaction is:

CnHm => n CH4

Of course, this doesn't always fit. Actually, it rarely fits. There are very few if any pure hydrocarbons larger than methane that have exactly 4 times as many hydrogen atoms as carbon atoms, so like in steam reforming, a bit of H2O is added on the left side to 'add' hydrogen in the mix, and unbalanced carbon is ejected as CO and CO2. The methane is then reprocessed in classic steam reforming.

This is a great process because most heavier hydrocarbons are incredibly rich in hydrogen; for both weight and volume, classic liquid fuels like gasoline and diesel have more hydrogen atoms than liquefied hydrogen by itself. Also, this allows for production of hydrogen from any hydrocarbon: coal, oil but short-cycle fuels as well, e.g. wood, waste ecological matter, even sewage. Of course, at some point you start to run into the issue of energy balance. As you decrease the amount of energy in the source matter, you have to add more and more energy to 'get the hydrogen out', until you actually have to add more energy than the heating value of the resulting hydrogen. This is also the answer for everybody who has been asking 'well, why don't you just steam reform pure water?'. Apart from the fact that it doesn't work chemically, it would cost much more energy to get hydrogen from water than other methods. By the way, steam reforming water into hydrogen would technically probably be pyrolysis? Not sure about that one. Maybe someone in the comments can help me out with that one.

Carbon stripping is a good way to get one unified fuel type (hydrogen) from any input fuel, and one of the big underlying lures of the hydrogen economy. Much like electricity, hydrogen is a big common denominator for all chemical energy sources. With steam reforming, we can make use of that fact. But, like with everything I have talked about before, the issue of efficiency is a big one. Carbon stripping is an energy intensive process that always yields fundamentally lower total process efficiency than just steam reforming from methane. To the point where at the moment, no hydrogen producer can beat the efficiency of a well-tuned Atkinson cycle internal combustion energy just in fuel to hydrogen conversion. In other words: even just production of hydrogen in this way is - at least with current technology - less efficient than running cars directly on gasoline or diesel. Also, the carbon emission argument still stands, but with the use of short-cycle fuels (plant matter that absorbs the same amount of carbon from the atmosphere as it grows) this may be said to be a moot point.

And then there is Carbon capture. Carbon capture in general means that you don't release CO2 emissions into the air, but you find some other place for it. For instance pressurizing it into empty gas or oil wells, using it to grow plants under a very carbon rich atmosphere or using it directly for other chemical processes. Of course this can be done for any process that generates gaseous carbon emissions, but it is especially well-suited to large stationary plants. It wouldn't be feasible to add a big exhaust gas bag to the tailpipe of your car. Economies of scale and the ability to position a hydrogen production plant right on top of a suitable carbon capture location make this an attractive option. It is quite expensive though, and the cost is one of the big reasons for the failure to find investors for two big carbon capture projects in the Rotterdam and Delfzijl.

But what about bacteria? Algae? I've read so much about that recently!
That's right, Jay.*

Conclusions and next up...!


In part one we looked at how fuel cell vehicles work, in this part we've looked at a lot of the math behind logistics of hydrogen. By now it should be clear that hydrogen is a very complicated fuel to use for vehicles and even if we would like to use it on a large scale, it would require a complete retooling of our existing infrastructure. But there are other ways around all of this. In the next part I will touch upon alternative fuels for fuel cells, alternative fuel cell types and alternative ways of using hydrogen.

Do read on in part 3 of this blog series

*I'm being quite dismissive here, mostly because yes: these methods do exist in laboratories but also no: these methods won't be available in a any kind of short term and have very little backing. Most of what you hear about them in media is just a cry for attention and money to develop these fabrication methods more. Therefore, I'm obviously not in a position to critique or analyze them, but I also feel like it's doing them a disservice to put them on the same level as established, scalable methods of hydrogen production

Why fuel cell cars don't work - part 1

Door mux op zaterdag 21 februari 2015 21:00 - Reacties (51)
Categorie: -, Views: 76.911

This is an extremely late article. Hydrogen fuel cells (HFC), especially in the context of personal vehicles, are and have been part of the yearly news cycle for more than 10 years now. Multiple car companies have attempted and only very recently has one sort of production car been announced. Meanwhile, electric cars have taken off like nobody's business, despite the big downsides the public (and car companies) think electric cars have compared to HFC cars. I have been involved in quite a few ways in the nuts and bolts of electric cars and fuel cells, so I pretend to know a thing or two about why this is.

Finished reading this part? Go on to part 2!. Also, Part 3 just went live.

This is an extremely long, in-depth blog series, so I'll start by giving you a summary. This summary will exist at the top of every part of this series. If you're interested in the technical details, please do read on and make sure to come back for the next parts.

First of all, HFC cars are perceived to be a good bridge between fossil fuels and full electric because:
  • You can still fill up like you do with a gasoline or diesel powered car
  • The mileage you can get out of hydrogen is perceived to be more adequate than what you get from batteries
  • Hydrogen fuel cells are thought not to wear out as quickly as batteries (or conversely, batteries are thought to wear out very quickly)
  • Hydrogen as a fuel is perceived to be a relatively small infrastructural change from gasoline and diesel
  • Hydrogen is perceived as a cleaner solution than gasoline, diesel or natural gas
In reality,
  • You cannot fill up like you do with gasoline or diesel. It is actually pretty ridiculous how hard it is to fill up a HFC powered car
  • You won't even go 100 miles on current tech hydrogen tanks that are still safe to carry around in a car
  • Fuel cells wear out crazy fast and are hard to regenerate
  • Hydrogen as a fuel is incredibly hard to make and distribute with acceptably low losses
  • Hydrogen fuel cells have bad theoretical and practical efficiency
  • Hydrogen storage is inefficient, energetically, volumetrically and with respect to weight
  • HFCs require a shit ton of supporting systems, making them much more complicated and prone to failure than combustion or electric engines
  • There is no infrastructure for distributing or even making hydrogen in large quantities. There won't be for at least 20 or 30 years, even if we start building it like crazy today.
  • Hydrogen is actually pretty hard to make. It has a horrible well-to-wheel efficiency as a result.
  • Easy ways to get large quantities of hydrogen are not 'cleaner' than gasoline.
  • Efficient HFCs have very slow response times, meaning you again need additional systems to store energy for accelerating
  • Even though a HFC-powered car is essentially an electric car, you get none of the benefits like filling it up with your own power source, using it as a smart grid buffer, regenerating energy during braking, etc.
  • Battery electric cars will always be better in every way given the speed of technological developments past, present and future
So now, on to the proper reasons and math. Even if you don't have much interest in electric or, well, any type of car, this is still good stuff. Today we will be looking at what a fuel cell car is and why people think they are cool.

Reasons why people like HFC cars

There are actually a bunch of reasons why people seem to historically like hydrogen fuel cell cars. There are a lot of differences pertaining to the age of the individuals asked, level of education and of course political leanings. But wait, first I'll talk a bit about myself.

Tinkering with the Forze II in Zaragoza

I've been involved in the first international hydrogen racing championship. Started in 2007, it was called Formula Zero with 'zero' pertaining to being zero-emission. I was in one of the teams, known as Formula Zero Team Delft ('Forze') and my main raison d'être was designing and assembling electronics for the race kart. Yeah, they weren't actually full-fledged formula one cars; it was just small class racing karts. Top speed around 110-120km/h, 0-60 in about 3.5-4sec, nothing especially interesting about them from a racing perspective. But they were hydrogen powered, which was extremely cool back then and kind of still is - at least from a technical point of view. The Formula Zero championship eventually merged with Formula Student, so if you're interested in more info, take a look at their website. I am no longer involved with any of this - I've gone on to do my Master's thesis about power conversion in electric cars and I now have a couple of businesses that mostly do electronic system design for optimizing power conversion. In computers. Not quite the same anymore.

Nevertheless, I do have a lot of hands-on knowledge from my time at Formula Zero and I know what goes into building a hydrogen fuel cell powered car. I've done a lot of literature research and kept up with the technology. And: during all this time that hydrogen fuel cells have been in the news, I have never come across any kind of public article that properly explains WHY things are the way they are with these enigmatic machines. I've been meaning to write about it since at least 4 years, and thought I would have been beaten to the punch many times already. But no... so, here we go.


Whoo! A hydrogen fuel cell car! (courtesy of goodsense.nu)

People don't like change
People don't like change. I don't actually believe that statement, but almost always the first reason that people give me when I ask them 'what do you like about the concept of a hydrogen fuel cell car?' is: well, I can just stop at a gas station and fill 'er up! No worries about having to wait for 6 hours for the stupid battery in an electric car to charge. People imagine it's basically the same as a gasoline-powered car. People imagine the same kind of filling stations with the big gas trucks distributing hydrogen to them.


The next reason people give is range, or the distance you can drive on a single tank. They imagine for various reasons - hydrogen being very light is an often heard one - that you an fit a bunch of hydrogen in there and just go for 500 miles on a tank. I'm of course deliberately saying 'they imagine' and not using any definite statements, because I'll tell you later why this is not (entirely) true. But it's true in the eyes of a lot of people!

Of course the obvious reason people like it is because it is much like a normal car *and* it is clean. It solves the CO2 problem without needing to ride a bike everywhere. Hydrogen can just be made from water, and when it reacts in a fuel cell it becomes water again, right? Absolutely no carbon emissions. This is much the same argument that is given for nuclear power plants.

When presented with the ecologically responsible alternatives, electric cars are closest to home. But batteries are of course horrible. Batteries are no good. I mean, look at phone batteries. They die all the time.

The funny thing is that basically all of these perceived advantages are... well, they are correct! Most of them at least - except for the cleanliness argument, although even that has some merit. The big problem here, though, is that technology isn't ready and probably more importantly: the environmental and financial cost of a switch to this kind of a situation is basically insurmountable.

The contrast here is that battery electric cars can be introduced en masse today, no problem. And they are much cleaner and better for the environment, both in the short and long run. The argument against hydrogen fuel cells is not that they can't be a good alternative to fossil fuel cars or that they 'just don't work', but that there are many technological and even fundamental physical problems (i.e. the laws of nature are against us) that need to be solved to get there. And you can prove that even if you get there, batteries will be better in every single way possible. This is the central theme of this blog series and what I will be trying to prove. So where to start?

Let's start with the fuel itself. What is hydrogen?

What is hydrogen and why is it so tantalizing?


Hydrogen, the atom, is the smallest atom you can make in our universe. It's just one proton and one electron, nothing else. When you combine two hydrogen atoms together, you get H2 - hydrogen molecules, normally a gas at room temperature and pressure. Hydrogen is very light: a cubic meter of hydrogen only weighs 90 grams. As a comparison, air at sea level weighs about 1300 grams per cubic meter. At sea level, hydrogen is 14 times lighter than air. So, like a bubble of air in water, in nature hydrogen gas will rise up into the stratosphere if you don't keep it in a container. Yes, this is a problem but no, we're not at the part of the blog where I talk shit about hydrogen.

If you take a hydrogen molecule and combine it with one oxygen atom (from the oxygen molecule O2), and you add a little bit of activation energy, the oxygen and hydrogen combine to form a water molecule. This is called combustion. Well, actually this is a reduction-oxidation, or redox reaction. Redox reactions happen because one part of the equation, the hydrogen, kind of has excess electrons while the other part is - again kind of - missing electrons. When combining, the electrons around the atoms merge together. The fact that there is electrons involved means that if you can somehow separate those two reaction compounds (reagents) and redirect those electrons, you can get a little bit of electric current from the reaction. This is what fuel cells do - with untold gazillions of these reactions happening per second to get enough power that we can actually use it for something.


Normally, if you want to perform a chemical reaction, you have to supply all the reagents. In this case: you'd need a canister of hydrogen and a canister of oxygen. But... oxygen is already really plentiful in the air we breathe. The big advantage of hydrogen in a fuel cell is the fact that it uses oxygen from the air and thus only needs to supply the hydrogen. This is, by the way, exactly the same with fossil fuels. They also need oxygen to combust inside the engine, but you can just pull that from the air. A regular car would need a ginormous oxygen tank if the atmosphere didn't exist. And so would humans. We breathe air just like our superior mechanical masters.

About that, hydrogen is really light. For the reaction from hydrogen+oxygen->water, you need about 8 times as much weight (well, mass, but we're not doing science here) in oxygen as you need hydrogen, so it's a giant weight saving you do if you just get oxygen from the air. And because the reaction product is just plain old water, you can expel that into the environment with zero danger. But there is another part to this equation. This all sounds great, but how much energy do you actually get from hydrogen? How much do you need to carry around?

To give you some perspective: a liter of gasoline contains about 46MJ of energy. A lithium ion battery contains only about 0.7MJ per kg, or about 60 times as little. A kilogram of hydrogen? A whopping 146MJ/kg, more than three times as much as gasoline. Crazy. This is an awesome fuel. Hydrogen gives you, by far, the most energy out of any chemical energy source known to man - if you don't have to carry any oxidizer. The only 'better' energy sources use fundamentally different types of energy, e.g. nuclear fuel. This has all to do with the nuclear forces vs. electromagnetic forces, but I won't go into that here.

How do electric and fuel cell cars work?

A lot of people talking about fuel cell and even electric cars either assume that whoever is reading their arguments knows how these things work, or they themselves actually never looked at how they work. This is a big, big problem because you really need to know what is going on to understand the (dis)advantages of either drive system, as well as the routes towards optimization. I'll give a brief overview of both technologies from a systems perspective.

Electric cars
Let's start with electric cars. An electric car system looks like this:

Battery electric vehicle system overview

The most important and by far biggest component in an electric car is the battery; electric cars require much less energy to work in the same way as a gasoline powered car, but batteries store so much less energy per kg that this technical advantage is completely lost. To get what is considered adequate range, you need hundreds of pounds of battery to get there. With the energy consumption of modern electric cars hovering between 90-160Wh/km, an adequate battery is around 50kWh of usable capacity (or about the equivalent of 4.5L or just over 1 gallon of gasoline). If you would make the battery pack exactly 50kWh with the highest gravimetric (energy/mass) density batteries around, such a battery would weigh 200kg (450lbs), but in reality in order to extend the life of the battery pack the batteries are slightly oversized. Also, as the really high density battery chemistries are prone to fire and explosions and all that, car manufacturers like to use slightly less energy dense but overall much safer types of batteries, e.g. LiFePO4. This means that these batteries are usually in the range of 275-350kg.

Besides the battery, the biggest component(s) by weight is/are the motor(s). Current generation electric cars use still fairly heavy central motors with an actual axle to the wheels. Upcoming generations will use in-wheel or near-wheel motors that have almost no drive train associated with them: no transmission, no gearing at all, no weight lost to things that aren't necessary. This is one of the reasons electric cars can actually become lighter than traditional gasoline-powered cars; the chassis can be reduced because there are almost no driving forces on it anymore. Note though: this is an advantage of both battery and fuel cell electric cars!

The other two major components in an electric car are the motor controller and some kind of cooling system. Motor controllers, as the name implies, regulate the power going into the motors. The cooling system is necessary to keep the motor and controller, but mostly also the battery at a reasonable temperature. Batteries don't particlularly like high temperatures and as efficient as lithium chemistry batteries are, they still generate some heat when you discharge them rapidly. As a rule of thumb, the cooling system needs to remove about 1/10th of the rated power of the car in heat at 50 degrees C. For some perspective: a gasoline powered car needs to be able to remove about 2x the rated engine power at 95C. That's 20x as much energy, but at a higher temperature which means it's about 3x as easy to do. Do the math and you get a cooling system that should be about 1/6th the size of that of a regular car.

Emile's Motor Controller - or EMC for short
This is an image of a 2x125A 150V motor controller I once made for Formula Zero, but never finished the firmware for. Still an awesome piece of high performance electronics

However, altogether the weight of the motors, cooling system and motor drive are nowhere near the size of the battery. In some current production hybrids, these non-battery components are actually pretty heavy and large (Toyota Prius II: altogether about 110kg), but this is a transient phenomenon. In the future this will all go down to a couple tens of kg, if that.

There is a question mark in the block in the overview schematic, marked: 'charger/dcdc'. As it stands, most cars incorporate their battery charger into the car, sometimes combined with a dc/dc controller that regulates the voltage coming from the battery into the motors. This is because every car is different, and you cannot really design one type of charger (at least not at the moment) that can charge any car optimally. So charging stations - e.g. the ones in parking lots you see often these days - are not much more than a three-phase outlet, and the actual charging algorithm necessary to properly charge the battery is entirely inclusive to the car.

This may very well change in the (near) future with direct charging, where the charging outlets actually have the conversion built-in and all the car does is identify itself and tell the charger 'hey, can you give me 400V 200A DC?'. This saves on a lot of cost mostly, as well as a little bit of weight.

Fuel cell cars
Something a lot of people only barely realize is that fuel cell cars are just electric cars, but with a fuel cell and hydrogen tank instead of a battery. Well, almost. Fuel cell powered cars look a bit like this:


The standard parts - There's the motor, controller and cooling system much like an electric car. But as fuel cells (at least portable ones) are appreciably less efficient than batteries, they need to get rid of a lot more heat they produce. As a rule of thumb, a vehicle fuel cell just on its own is about 40% efficient*. So, you need about as much cooling as the rated power of the vehicle, maybe a bit more, at 60 degrees. Do the math and the cooling system should look about the same as a gasoline powered car.

Portable fuel cells are, at the moment at least, almost all Proton Exchange Membrane or PEM fuel cells. This is the type of fuel cell with the highest power-to-weight ratio. As far as I know, the current concept fuel cell cars all use PEM fuel cells at about 60 degrees C and close to 1 bar operating pressure. Now, here's where all that other stuff in the diagram comes in.

Pressure reduction and heat exchange in the hydrogen path - A kilogram of hydrogen takes up a couple cubic meters. In order to take some reasonable amount of fuel with you, you need to compress it down a *lot*. A couple hundred atmospheres of pressure is what you need to take the equivalent of a full tank of gas with you. But the fuel cell won't accept this directly; you need to reduce the pressure. And you can't just do that willy nilly. As you relieve a gas of its pressure, it cools down. If you would use a reduction valve to go directly from 200 to 1 bar, it will freeze to close to absolute zero, become brittle and shatter. This reduction is usually done in steps; first from 200 to about 25 bar, then to 5-10 bars and then to the final pressure. Other thermal and gas flow effects necessitate some additional pipework and in-between stages.

When doing static testing (i.e. without driving wind), just the normal fan wasn't enough to cool the fuel cell. So we... added some extra fans

Cooling system - PEM fuel cells use hydrogen and oxygen either directly from the air or in solution with a proton exchange membrane in the middle. Using a catalyst, the hydrogen and oxygen bond to form water and give off an electron to conveniently placed electrodes in the fuel cell. I won't bother illustrating this here; wikipedia can help you with that. This process generates lots of heat and it's actually so much heat that it is not enough just to 'cool' the fuel cell by piping off the excess water that has been produced. You absolutely need additional cooling. This is done in one of three ways:
  1. by putting a closed loop water cooling system on the hydrogen side of the fuel cell,
  2. bydoing the same on the water side and
  3. by using a separate cooling system that uses heat exchangers to interface with the 'hot' parts of the fuel cell.
All of these have problems, and there is no 'best' solution, or a clear end solution for that matter. the issues are, in summary:
  • Cooling using the oxygen side of the circuit means you need a way of relieving pressure, as the water produced by the reaction ends up here.
  • Cooling using the oxygen side means the exhaust hydrogen scrubber needs to be in the cooling circuit, which complicates things a lot
  • Cooling the hydrogen side means you will lose a lot of hydrogen through the walls of your cooling system
  • Cooling using the hydrogen side means you cannot use steel and a lot of other materials that deteriorate under the influence of hydrogen (hydrogen embrittlement and other effects)
  • Cooling with a separate circulation increases weight a lot as you need to build heat exchangers into the stack
I've seen fuel cells with all three methods, and ones that combine everything: all reagents are put into solution and a heat exchanger extracts heat from the water into a separate system. Very small fuel cells are simpler still; they don't need nearly as much cooling and can just rely on the evacuated water to get rid of the heat.

I'm telling you all this because this is what the 'extra stuff' in the water cooling path represents. There is a lot to it! This is not a job for a standard radiator and water pump. And this is why the hydrogen, oxygen and cooling paths in fuel cells are a prime target for optimization.

Oxygen input - Air isn't perfectly 100% oxygen. Actually, it's mostly nitrogen, about 20% oxygen and a lot of trace elements as well as aerosolized liquids and solids. A hydrogen fuel cell will 'clog up' with crud, ranging from small particulate matter that physically blocks reaction surface area to certain molecules (nitrous oxides for instance) that over time 'poison' (make inactive) parts of the fuel cell stack. These unwanted parts of the air need to be scrubbed out, after which the remaining oxygen-rich mix needs to be compressed down and brought into solution so it can enter the oxygen side of the fuel cell.

This is one of the easier extra subsystems that fuel cell vehicles need to deal with. However, it is one of those parts that is still fairly expensive as current-tech scrubbers use consumable porous materials (e.g. Millipore reverse osmosis scrubbers) to do the job. This requires regular maintenance. Future tech, some of which is actually pretty far along in academia, allows for more or less lifetime guaranteed scrubbers, much like catalytic converters in Diesel cars.

The exhaust - PEM fuel cells still exhaust a little bit of hydrogen in their exhausts - partly because the reaction surfaces can't be made perfect and partly on purpose to make the reaction go a certain way. The fuel cell we used for the races exhausted about 15% of the ingoing hydrogen flow, which is a waste. This is why the hydrogen is purged/regenerated from the exhaust flow and put back into the system.

Those blue cylinders in the bottom-center of the image are very large ultracapacitors - our racing kart had 36 of these to keep up with peak demand.

Batteries and/or capacitors - And now to the next big disadvantage of fuel cells; they don't throttle well. Going from zero to full power, or even from 10% to full power requires the entire cell chemistry to rebalance, which is fundamentally limited by diffusion speeds. It takes in the order of seconds to throttle. In the meantime, either you have insufficient power to do what you want (and believe me, even 2 seconds is a very long time to wait before your car starts accelerating from traffic lights) or you have excess power coming from the fuel cell which can't go anywhere, so it needs to be thrown away.

This is why basically all fuel cell vehicles use either a small traction battery - larger than a normal starter battery but much smaller than an EV battery - or a bank of ultracapacitors like the Maxwell Boostcap. Both technologies have their advantages and disadvantages, for our racing kart we chose ultracapacitors because regulations didn't allow for batteries at the time.

So that's the technical differences between battery electric vehicles and fuel cell vehicles!

A great view from the side of Imperial College of London's entry in the Formula Zero championship; essentially, the drivers were sitting on top of a solid block of technology

* By the way, if you're ever interested in researching these numbers: don't believe what manufacturers tell you without checking the type of efficiency they mean. There's a big difference between theoretical energy content (what I talked about before, about 150MJ/kg), higher heating value, lower heating value and even some other ways of calculating fuel cell efficiency


Today we've looked at the absolute basics of hydrogen fuel cell cars; who am I to talk about this, why do people like them and what are they exactly, on a technical level?

I've put off publishing this for a long time, actually. This exact post has been sitting in a backwater of my NAS for about a year because I didn't like how it was so opinionated. Every sentence oozes my dislike of fuel cell technology in cars. This is not a good way of doing science. The reason I'm still publishing it now is because I think there is still a bit of a knowledge void on the internet to be plugged, and I just couldn't rewrite it in a well-structured manner without my opinions.

Secondly, it really is time to look at these things. Even Elon Musk of Tesla made some off-handed remarks about fuel cells being 'silly'. Well, why? EVs are taking off and people are genuinely wondering why there aren't any FCVs they can buy and own. Batteries aren't a particularly well-trusted technology.

Read about hydrogen production, storage and transportation in part 2..

Supermarkten follow-up: uitgaven 2014

Door mux op woensdag 11 februari 2015 09:10 - Reacties (43)
Categorie: -, Views: 5.375

In m'n vorige blog over supermarkten heb ik ruim uit de doeken gedaan hoe wij supermarkten benaderen - met name op het gebied van prijsstelling en winkelgemak. Ook heb ik uitgebreid uit de doeken gedaan hoe je zo gunstig mogelijk gebruik kunt maken van aanbiedingen en afprijzingen. In deze post bekijk ik hoe dit invloed heeft op ons daadwerkelijke winkelgedrag

Ons jaaroverzicht

Al onze boodschappen worden gepind van één rekening, die wij bij de SNS-bank hebben. De SNS-bank heeft een handige feature waar je al je transacties kunt downloaden als .csv-bestand. Dit is gewéldig voor mensen als ik die graag alles analyseren en proberen trends te vinden. Dus, laat ik deze data eens met jullie delen.

In heel 2014 hebben wij in totaal ¤1844,27 uitgegeven aan niet-nutsvoorzieningen en vakantie, waarvan ¤1602,46 boodschappen en de rest uitgaven bij non-foodzaken en non-foodartikelen bij supermarkten. De winkels waarbij we dit geld hebben uitgegeven zijn:

Gezamenlijke uitgaven 2014 per winkel(type)

Duidelijk is hier te zien wat het effect is van een winkel dichtbij: veruit de meeste uitgaven doen we bij de C1000 omdat het voor het grootste deel van 2014 simpelweg de dichtstbijzijnde winkel was. Wat dies meer zij: er zaten drie filialen binnen 5 minuten van elkaar, met alleen een AH en Plus die hiertegen moesten concurreren. Duidelijk dus.

Interessant voor ons was om te zien dat Albert Heijn zo laag scoort. AH is geen dure winkel, en heeft een assortiment dat prima bij ons aansluit. Maar het zit een paarhonderd meter verderop dan de C1000, dus we gaan er niet voor kleine/snelle boodschappen heen. De enige reden om er naartoe te gaan is voor aanbiedingen of specifiek assortiment, en dan koop je niet 'nog even' wat ander spul erbij.

Omgekeerd is het interessant om te zien dat we zoveel bij Dirk en Lidl hebben gekocht. De dichtstbijzijnde Dirk en Lidl zijn 7km verderop; bijzonder ver om zomaar even heen te gaan. Dit komt omdat het op de route naar de universiteit is - en omdat ze allebei goede aanbiedingen en een assortiment hebben dat bij ons aansluit.

Plus scoort ontzettend laag, voor een winkel die nauwelijks verder zit dan de AH. Hoge prijzen voor artikelen die wij willen kopen en slechte aanbiedingen zorgen hiervoor. We gaan er heel af en toe heen voor een goede aanbieding en pakken dan vaak nog wel wat impulsaankopen mee, maar het is geen winkel waar we automatisch naartoe gaan.

Uitgavenontwikkeling door het jaar heen

En dan nu het leukste. Ik heb onze uitgaven per per maand uitgerekend en in een grafiek gezet. Hierin is te zien in welke maanden we bij welke winkel het meeste uitgaven, en hierin zijn interessante trends te zien. Dus, hier is-ie:

Uitgaven in 2014 als functie van tijd

Wat een hoop lijnen zeg. Laat het even op je inwerken, de structuur wordt vanzelf zichtbaar. Iedere lijn in deze grafiek is een winkel. De rode lijn die het hoogste zit is de C1000; daar hebben we structureel het meeste uitgegeven. De andere lijnen... ze zijn interessant. Dit zijn overigens alleen grafieken voor supermarkten waar we meer dan 50 euro hebben uitgegeven dit jaar. Andere bovengenoemde uitgaven zijn weggelaten.

Allereerst algemene trends. Aldi ziet een flinke uitgavenpiek in Juni, Augustus en Oktober-November. In Juni hebben we er een duur non-fooditem gekocht, in Augustus/Okt/Nov waren we op vakantie (jaja, twee vakanties!) en kochten we vrijwel uitsluitend bij Aldi en Lidl onze boodschappen.

Over Lidl gesproken: die lijn is opvallend stabiel vergeleken met de rest. Dit komt simpelweg doordat Lidl heel weinig interessante aanbiedingen heeft en we er vrijwel alleen voor groente/fruit naartoe gaan. Dat is niet iets dat bijzonder veel in prijs varieert van maand tot maand; we kopen altijd groente en fruit in seizoen en dat is altijd 50 ct à 1 euro per kg.

De C1000-lijn is heel interessant. Er zitten twee gigantische pieken in; dit zijn de C1000 Euroweken. Veruit de beste aanbiedingen van het jaar, en hoopjes verschillende aanbiedingen die we allemaal kopen. Afgelopen jaar waren de Euroweken blijkbaar in April en September.

Hoogvliet en C1000 hebben een prachtige wisselwerking in deze grafiek. In Oktober 2014 ging de Hoogvliet Berkel open, en sindsdien doen we onze dagelijkse boodschappen vrijwel uitsluitend daar vanwege de zelfscanners en vergelijkbare prijzen, maar ruimer assortiment. De C1000-lijn daalt dus terwijl de Hoogvliet-lijn navenant stijgt. Overigens waren er in juni blijkbaar bijzonder goede Hoogvliet-aanbiedingen, want toen hebben we daar een hoop geld uitgegeven. Dat was echter bij de Hoogvliet Zoetermeer.


Allereerst is het leuk om je eigen financiën en gedrag te bestuderen. We zijn aanbiedingen/koopjesjagers, dus het is logisch dat ons uitgavenpatroon divers en inconsistent is; we wachten op aanbiedingen en als die er zijn kopen we groot in. Alsnog zijn we mensen en dus gevoelig voor luiheid (dichtstbijzijnde winkel = meeste uitgaven) en winkelgevoel (Hoogvliet = fijner winkelen dan C1000). Ook is het duidelijk te zien dat sterke aanbiedingen ons toch uit de luie stoel weten te lokken (Dirk) en dat een speciaal assortiment werkt om klanten te binden (Lidl).

Daarnaast is dit een geweldig voorbeeld van hoe huishoudboekjes helpen inzicht te geven in bestedingsgedrag. En het is vrijwel moeiteloos; uitgaven worden door de bank automatisch ingedeeld en in de mobiele app weergegeven als staafdiagrammen. Uitgebreide overzichten zijn te downloaden als .csv-bestand. Dit is een goede reden om consequent pinnen te verkiezen boven andere technieken zoals het boodschappen doen van een vast bedrag per week.

Maar, hoeveel ik ook analyseer: er is bij ons weinig meer op boodschappen te besparen. In 2014 hebben totaal 16 euro méér uitgegeven aan het huishouden dan in 2013. We zijn luxer gaan eten en hebben meer variëteit, dus in die zin is er winst, maar ik verwacht niet dat we de uitgaven kunnen terugbrengen naar bijv. 100 euro per maand of minder. Waar wél op te besparen is, zijn onze uitgaven aan nutsvoorzieningen, belastingen en abonnementen. Daarover in de toekomst meer!