On Fuel Cells

I've been meaning to write about the challenges and opportunities with fuel cells (for power generation, not transportation) for a while now. Here is a note I had written about a year ago.

(1) The core issue with any energy source is COE (cost of electricity), which is composed of 3 parts
      a. capital cost
      b. efficiency
      c. lifetime

(2) Capital cost is high because of high operating temperatures (800 - 900C). This affects

1. fuel cell cost
   
2. Metallic interconnect cost - Coming up with an oxidation resistant metal for 5 year  life @ 850 C is generally accepted to be the showstopper.
   
3. Balance of Plant (BOP) cost - again, high temp requires more expensive pumps, blowers, valves etc. BOP forms > 70% of system cost, the other 30% being the stack itself Bloom claims that they have a regenerative SOFC (which generates oxidant internally) which can operate at slightly lower temperature, may be 700C. Typically 800C is a big break point in terms of metal cost/life, so this is an advantage for them. The issue is - I am not sure how much improvement they are actually reaching.
   

      Fuel cell cost itself is generally high because of the long sintering process which is a batch operation. Just like going to 12" Si wafers in semiconductor industry, the ability to make larger area fuel cells in one sintering batch is key. GE was able to demonstrate a 12" planar fuel cell.

      Siemens went the other way. They raised the temperature and replaced metals with ceramic conductors. They made good technical progress but could never sell the product - it was simply too expensive. Even they are rewinding the clock and looking at alternatives.

    Several companies (GE, Siemens included) also believe that they have to do away with the sintering process, and have a "spray-on" fuel cell to minimize cost. You spray a ceramic just as you spray thermal barrier coatings onto turbine blades.

Some numbers: $1500/kW would be a good starting point. US DoE, under their SECA program, set a long term target of $400/kW for stack cost, and $800/kW for system cost. Having been involved in costing, I found that this requires some extreme assumptions on how thin your metals will have to be, and how large your individual cells will have to be.

(3) Efficiency

1. This can be improved at higher temperatures, but I have listed the issues above, under cost. Durability also drops with higher temperature
   
2. Hybrid fuel cells are the dream. You take the exhaust hot air at 900C and send it into a gas turbine. At GE we decided that the whole thing wouldn't make $$sense unless we hit a combined efficiency of > 75%. Siemens has demonstrated a 250kW hybrid system with I think > 55%. This is pretty good... you will need to go to 5MW to reach 75%.
   
3. Regenerative SOFC may help improve efficiency - this is where I don't have enough data. It is likely that it is an advantage. However, I don't know if it can go close to 70%... they claimed it can.
   

 

(4) Durability

 The accepted standard is 5years (40k hrs), with 10 shutdown cycles each year.

(a) Again, here, some companies claim that they can operate at a lower temperature and thus extend life. This may be possible.

    (b) Another aspect of life is thermal cycling. Each shutdown-restart of a system takes a heavy toll on the stack. This is why many (myself included) believe that SOFCs form good base base load machines, or generators for utilities/substations, rather than as portable devices, because they involve too many shutdown cycles.

 

Other business models

 

One option for fuel cells is Combined Heat and Power (CHP). Some PEM fuel cells manufacturers worked on that for a long time -with the idea of setting up a 5kW fuel cell heat generator either in your home, or a bank of these near an apartment complex for "district heating"... they eventually ended up in the back-up generator space. They tried selling CHP systems to European customers, but the efficiency requirements were rather high, and it was challenging. Now - SOFC can do better at heat generation because as such operating temps are much higher. 85% efficiencies are possible if you account for both heat and power.

Another concept is about making H2 in a reformer along with the fuel cell (from propane or natural gas) and fueling your car with it. Plug Power & Honda are working on a home-refueling system... But again you know my thoughts about the whole H2 economy.

Now... thinking of SOFCs - the application for SOFCs that makes sense to me is remote power generation, specifically for cellphone towers etc. The reason is that conventional PEM fuel cells may not operate very well in India's hot environment. I'm taking off from your Acme Green Shelter example. Some US PEMFC companies started with the idea of remote power generation using propane tanks... but found it tough to break into. Again, installed cost was an issue, and as such there are only a few farms and ranches in the US that don't have grid connection... market was not large enough.

So one can explore this further. What should be kept in mind while working on these is that in the space of flexi-fuel sub-100kW, your competitor is not grid, but diesel gensets. They can also run on fossil fuel, and they are cheap as hell...

Plug Power (www.plugpower.com) sells their fuel cells to telecom companies as back-up. PEM Fuel cells are good power back-ups because they can electronically come on instantly, and in one package replace the need for battery-back-up+Diesel-genset-kicking in. It made their life much easier to go telecom because they would need to produce only 48V. In fuel cells, the voltage is proportional to stack height. Tall stacks are hard to build. FCE's SOFC stack is I think 80V or something - 100 cells tall (@ 0.8V/cell).  Inverters are fine, but not there could be issues at high power ratings.

 

On Solar Thermal Energy

The Sun is one of the most promising sources of energy for our future. Energy from the sun may be captured in one of two ways. The first is through the direct conversion of sunlight into electricity, using a solar cell, technically known as Photo-Voltaic Module (Solar PV, in short). These are not very different from the solar cells that run many calculators. The second method of extracting the sun's energy is by collecting the heat we receive from the sun, and using the same to perform useful work.

In order to get the reader oriented, is useful to look at how much energy we get from the sun. Each square meter on the surface of the earth receives about 1 kW of solar energy. Of course, this number varies with location on the planet, time of day, and season. Taking this further, 1 sq. m of land on a sunny day in India can easily receive as much as 5-6 kWh of energy (8 hours of sunlight * 0.75 kW /m2 average flux). For more information on amount of energy received from the sun, see http://www.oksolar.com/abctech/solar-radiation.htm. How much of this we can actually extract between 8% and 25%, depending on the technology used. At 10%, 1 m^2 can generate at least 0.5Wh per day. Compare this with the typical household energy usage of about 3-5 kWh.

Solar cells or solar panels are definitely a very convenient way to put sunlight to use - they directly provide electricity, which can then be fed into the power grid. However, they are still 4-5 times more expensive to install than conventional coal-based plants. http://www.solarbuzz.com/index.asp keeps track of solar panel costs, and as of May 2008, it costs about $4,800 to install 1 kilowatt of solar PV capacity. Over the next 5-8 years, these costs are expected to come down with efficiency improvements, volume manufacturing and durability improvements. In the meantime, opportunities for solar panels exist in areas where conventional power is too expensive, or where the grid has not reached.

Over the last few months, though, I have started to get more excited about solar thermal energy. The central idea here would be to use a concentrator such as a parabolic dish to focus sunlight onto a small receiver, and thus heat up a thermal fluid such as water. There are several concentrator technologies, ranging from simple dishes, to long parabolic troughs, to foot-ball-field-sized arrays of reflectors concentrating sunlight onto the top of a tower (http://www.solucar.es/sites/solar/en/tec_termosolar.jsp). The process of transferring heat to the fluid can be very efficient. The next step would be to put this energy to use.

One option is to use this heat to boil water to steam, and drive a steam turbine to generate power. A 35 MW solar thermal steam power plant is being set up in Rajasthan to study the feasibility of this option. Given that the net efficiency of such plants is likely to be below 20%, it is possible that Solar PV will become a comparable and more convenient alternative in just a few years.

Another option is to focus sun's heat directly onto a "Stirling Engine", which is an engine that runs from heat supplied externally. Stirling Engines have been around for several years, but can be expensive to build to last. Stirling Energy Systems http://www.stirlingenergy.com/ is working on building such power plants. For a country like India, I suspect that this technology will be too expensive.

The option that excites me the most is to use concentrated solar power directly as a source of heat rather than as electricity. There are several process industries that could save on costs incurred in burning fuel for low-grade heat. For example, Pepsi Co. has installed a solar thermal plant in Modesto, CA to heat up the oil in which it fries its Frito-Lays chips (http://www.modbee.com/1618/story/259206.html). A group at IIT Bombay has developed an indigenous solar concentrator dish, which is being used in dairies in Anand, GJ to pasteurize milk.
There are several challenges with solar thermal energy. These include

(1) The source is limited to daytime. However, one needs to look into the economics of whether substituting fuel consumption during daytime alone is enough to justify the cost of solar thermal collectors.
(2) The land requirements could be significant. 1 MW heat generation capacity, assuming 40% land coverage and 50% heat transfer efficiency would require 5000 sq. m or 1.23 acres of land - which could be rather expensive.
(3) The cost of solar concentrators themselves is currently rather high. IIT Bombay's 160 sq. m dish is priced at Rs. 30,00,000, and has a heat generation capacity of about 80 kW.

In my next post, I plan to go into more detail on the costs and potential applications of solar thermal energy. For now, I will retire, and utilize this relatively light evening during my MBA to catch up on some sleep.

On Energy: The Hydrogen Economy

I have been meaning to write on this topic for a while – Energy. In the interest of full disclosure, my experience in this field is from working on hydrogen production (from natural gas) and fuel cell design and development (both PEM or low temperature, and SOFC or high temperature flavors) for four years. In the first part of this series, I will review hydrogen.

1. Is the Hydrogen Economy really feasible?

I once attended a talk on evaluating the economic aspects of using hydrogen as our primary fuel. The talk referred to the 1970’s oil crisis, and how natural gas emerged as a potential replacement for gasoline. Natural gas burns cleaner than, is just as widely available (cheap at that time, though presently expensive), just as safe as gasoline. Besides there is tremendous potential to produce natural gas from farm waste. Most importantly, the existing gasoline engines can burn natural gas with very little modifications. Yet it did not “stick”. For anything to stick, Americans have to buy into it – and they simply couldn’t give up the comfortable feeling of a liquid gushing into their fuel tanks. A simple google search landed me at this article from last summer:

http://www.usatoday.com/money/autos/2007-05-08-natural-gas-usat_N.htm

Excuses include safety concerns or the lack of refueling stations. If we can not replace the entire refueling network of the country with natural gas to service an existing engine (the same ICE that you have today), what makes us think that we will replace the entire refueling network with hydrogen filling stations, and develop a new type of engine to burn it (fuel cells)?

2. Where does the hydrogen come from?

Let’s say we actually do accept hydrogen as a fuel. Where does the hydrogen come from?

The best known way to make hydrogen today is from natural gas!! The best and largest natural gas reformers today work at about 75% efficiency, meaning that the total amount of hydrogen produced from 100 megajoules worth of natural gas is worth about 75 megajoules of energy only. We lose upwards of 25% of the energy content just to split the hydrogen out of the methane, propane, or what have you. That fuel cell based cars have higher net efficiency (> 45%) when compared with an internal combustion engine (~ 30%), doesn’t make the economics much better. (Natural gas à Hydrogen à Fuel cell = 70% * 45% = 31.5%, is not much better compared with Natural gas à IC Engine @ 30%). By the way, making hydrogen from natural gas makes sense for industries that require hydrogen as an input.

For those others who want to produce hydrogen from coal gasification, or diesel reforming, or from other carbon-based fuels, I will not continue the discussion. It does not make sense at the very high level.

Now – the more relevant hydrogen production methods are: (1) electrolysis from wind power and (2) electrolysis from power generated by solar panels, or (3) direct solar electrolysis – using sunlight and catalysts to directly split water. In each of these cases, we are essentially using hydrogen as a means to store and carry energy from the place where it is available (i.e. where the wind blows or where the sun shines), to where it is used (your car). These techniques make sense to me. Barriers to these methods include cost and storage technology.

The cost of hydrogen produced by electrolysis is rather high. Let’s start with gasoline, which costs $0.36/kWh at $3/gal used at 25% efficiency. You need as much energy to split water to hydrogen as you get from burning it. So if wind energy costs $0.10/kWh, the hydrogen produced, used at 40% efficiency (including cost of transporting the fuel), will cost $0.25/kWh. Add to that the cost of (1) hydrogen compression (2) transportation and (3) fixed capital – which all may double this number to $0.50/kWh. (Consistent with DOE’s findings of current costs of $10/kg H2, or $0.35/kWh.) Cost reduction in electrolyzers and other equipment could see the cost of hydrogen produced from wind electrolysis drop to at or below gasoline equivalent prices.

Solar photovoltaic energy is still not cheap enough to beat these numbers ($0.10/kWh is not here yet).

3. Summary

The biggest challenge in realizing a hydrogen economy may be people’s acceptance of the system, and the costs and inertia against transitioning to such a system. Wind based electrolysis is perhaps the most economical and sensible method of producing hydrogen, but electrolyzer cost and technology is still catching up.