Is That Fuel Cell Economic? Forget cents/kWh or GGE

It’s very easy to find graphics that compare the cost of several forms of emerging energy technologies.  In the case of fuel cells, since they make electricity to perform a task, the metric used is the cost of a unit of electricity: (pick your currency) per kilowatt-hour or cents/kWh.  In vehicular applications, an attempt is made to equate hydrogen fuel with gasoline, generating the absurdly fictitious “gallon of gasoline (or liter of petrol) equivalent” or GGE.  If you see these metrics used, you are witnessing one of the greatest failures of the fuel cell industry: its inability to articulate its worth.

Allow me to demonstrate how daft and self-defeating these comparisons can be.  We all use batteries in our daily lives.  Batteries generate electricity, which we put to good use.  The going rate for AA batteries is broad, but to pick a number, a 4 pack of Energizer batteries is $4.01, or about $1/battery.

Each battery provides a little less than 3 watt hours of electricity.  That means that if you compared AA batteries with other forms of generation, its output would be about $330/kWh.    Now let’s take that value and do the conventional comparison with other generation forms, shown below.  Note that this uses a logarithmic scale.

Clearly it looks horribly expensive.  But here’s the thing- we don’t value the AA battery for its costs of electricity, we value it for its convenience and size.  Relative to other kinds of electricity generation, it costs a fortune, but only because we are using the wrong metric.  A friend and mentor reminds me frequently that what is being sold is the product of the product, not the product itself.  And so it goes with fuel cells.

This is readily apparent in three applications where fuel cells are deemed commercial and where sales are happening every day: remote telecommunications backup power; uninterrupted storage and as replacement drives for battery systems in materials handling equipment.  In each case, what make the fuel cell commercially viable is its context in an economic system that values the attributes of its output, not the actual output itself.  Consider any of the following factors:

  • Avoided costs of equipment to achieve desired level of enhanced:
    • Reliability
    • Power Quality
  • Avoided costs of battery maintenance and disposal
  • Value of quiet operation
  • Avoided fees and penalties for emissions for certain industries
  • Longer operational periods in EPA non-attainment areas
  • Much longer operation in battery replacement applications
  • Portability

Each one can be valued and monetized in the overall economic equation.

Total Cost of Ownership (TCO)

A far better metric and methodology for evaluating the commercial viability of fuel cells is total cost of ownership (TCO), a form of life cycle cost analysis.  Investopedia defines TCO as “The purchase price of an asset plus the costs of operation. When choosing among alternatives in a purchasing decision, buyers should look not just at an item’s short-term price, which is its purchase price, but also at its long-term price, which is its total cost of ownership. The item with the lower total cost of ownership will be the better value in the long run.”

Materials Handling Equipment

Use of TCO is best illustrated by the business case for fuel cell replacements in electric drive systems for materials handling equipment.  Virtually all indoor forklift trucks are battery powered electric units. A recent DOE report provides a detailed analysis, summarized below.[i]

The relevant costs of ownership include the following items:

  • Cost of the bare forklift
  • Cost of the required battery or fuel cell systems
  • Cost of battery changing and charging or hydrogen fueling infrastructure
  • Labor costs of battery changing or hydrogen fueling
  • Cost of energy required by the forklifts
  • Cost of facility space for infrastructure (indoor and outdoor)
  • Cost of lift truck maintenance
  • Cost of battery or fuel cell system maintenance.

When each of these cost elements are evaluated, the following table results for a hypothetical Class I or Class II forklift:

Total Annual Cost of Ownership
Cost Element Battery Drive Forklift Fuel Cell Drive Forklift
Amortized cost of truck $2,800 $2,800
Amortized cost of Battery or Fuel Cell Unit $2,300 $2,600
Per Truck Cost of Charging Battery/ Fuel Cell Hydrogen Fueling Infrastructure $1,400 $3,700
Labor Cost for Charging or Refueling $4,400 $800
Cost of Electricity/Hydrogen $500 $2,400
Infrastructure Warehouse Space $1,900 $500
Forklift Maintenance $2,800 $2,800
Battery/ Fuel Cell Maintenance $3.600 $2,200
Total Cost of Annual Ownership $19,700 $17,800


If one only looked at the total costs of the forklift plus drive unit, the fuel cell unit is more expensive.  Add in the cost of charging batteries, the cost of hydrogen and the charging/fueling infrastructure, the fuel cell looks that much worse. Now factor in the productivity costs or gains: the labor cost for refueling a fuel cell is far less than that of a battery unit; the warehouse costs are far less, as is the maintenance costs of the drive units.  At this level of analysis, the fuel cell drive is the clear winner, although it represents a discrete set of values.  An analysis which looks at the sensitivity of each of these parameters underscores the clear choice of the fuel cell drive unit.

Source: NREL

There are other increases to productivity for the fuel cell that these analyses do not capture.  Over a normal shift, the battery loses capacity such that near the end of a shift it cannot lift the same amount of material to the same heights.  Fuel cell units retain their power as long as they have fuel available.  In some cases, more electric drives need to be purchased to compensate for long battery recharge times and the diminished capacity of units over shifts.  Finally, while the costs of battery maintenance are included above, the environmental cost of battery disposal are not.  Fuel cell units do not have that problem.

A similar analysis can be performed for remote telecom site backup.  In this case, the fuel could either be a liquid (operating a methanol fuel cell), or compressed hydrogen.  Diesel generators or battery units require maintenance, with very high labor costs attributed to simply getting to the site and returning.  A fuel cell unit, once installed and made ready for operation, eliminates a considerable amount of potential labor costs.  In addition, it does its work without emissions and without noise.

Gallon of Gas Equivalent

The GGE metric, when applied to hydrogen fuel cells, is one of the most inappropriate attempts to compare one type of fuel system with another.  Clearly it was created on the assumption that it would be easy for the general public to understand.  All it does is obfuscate reality.  GGE works as a comparison basis when we are talking about fuel use in an internal combustion engine which resides in a vehicle that was designed with an internal combustion engine.  Period.  The only time GGE would apply to a hydrogen fueled vehicle would be if hydrogen was being combusted in a conventional engine.  Further, GGE by definition relies on petroleum economics to set price and petroleum market dynamics have nothing to do with the cost of hydrogen.  A fuel cell vehicle is designed to accommodate a fuel cell and make best use of its attributes.  It is far more efficient than combustion engines and makes best use of the available electricity to incorporate features not found in today’s vehicles.

The introduction of fuel cell vehicles offers the opportunity for end users to think about their transportation quite differently from what is on the road today.  Instead we persist in talking about these new vehicles in terms that do not fit.  Most egregious is the continued use of GGE as a filter to determine where R&D funding should go for hydrogen vehicles and refueling infrastructure.  For several years, if someone guessed that a particular project or concept could not beat $3.50 GGE it was rejected.  Achieving lowest cost is certainly a goal but using the wrong yardstick is just plain stupid.

A far more appropriate metric for comparing new and emerging vehicles (and the R&D associated with them) is cost per distance.  This metric would apply to fuel cells and to electric vehicles.

Emerging Energy Metrics

As new forms of energy emerge in this new energy economy as much care needs to be taken in their evaluation to assure that we take fully into account the change in perspective that comes with them.  The risk in using old methods to consider new concepts is that we miss altogether their potential.

[i] Ramsden, Ted USDOE “An Evaluation of the Total Cost of Ownership of Fuel Cell-Powered Material Handling Equipment” NREL/TP-5600-56408. April 2013

Contrarian Approach to Fuel Cell Powered Forklifts and Scooters in Taiwan 燃料电池发展的逆向操作:台湾叉车及摩托车

The worldwide population of motor scooters is approaching 130 million. China alone produced over 40 million gasoline powered motor scooters in 2011. Many of these engines emit 8 to 30 times the hydrocarbons and particulates emitted by automobiles. Several companies are developing fuel cell powered scooters to reduce these enormous emissions. Fuel cells are devices that make electricity from hydrogen and oxygen, emitting water vapor as the exhaust. When hydrogen is produced from renewable sources, or even from natural gas the emissions are far less than those resulting from oil refining and combustion. Fuel cell powered scooters run on that electricity.


Two years ago I wrote about a very forward thinking fuel cell technology company in Taiwan (, Asia Pacific Fuel Cell Technologies, Ltd. (APFCT). The company had just rolled out its first major demonstration of fuel cell powered scooters.

两年前我在台湾为一家非常具有前瞻性的燃料电池公司 – 亚太燃料电池科技公司写了一篇文章(。该公司当时刚举行其首次大型燃料电池摩拖车的示范运行。

What was unique about the company and its scooters was the approach APFCT took to fueling. APFCT designed their system with simplicity and consumer convenience in mind. Instead of taking the path of nearly all fuel cell transportation devices that require the refilling of an onboard cylinder with highly compressed hydrogen, the APFCT units use small canisters that store hydrogen in metal hydride powder. Instead of driving the vehicle to a fueling station and waiting for a cylinder to be filled the user simply takes their empty canisters to a vendor who exchanges them for filled canisters (with about the same internal pressure as a racing bike tire).


In its first demonstration APFCT put 80 scooters on the road at a beach resort in southern Taiwan. Tourists were permitted to use the scooters for free. When they ran out of hydrogen all they needed to do was to take the empty canisters to any 7–Eleven convenience store, repair shop or police station for exchange. Why 7-Eleven? Taiwan has the fifth largest number of 7-Eleven stores in the world, behind the U.S., Japan, Thailand and South Korea. There is a 7-Eleven within walking distance of almost any place in Taiwan.


APFCT has continued to build upon this hydride storage fueling model over the last two years. It has tested a number of different vehicles, all of which use identical canisters. Those with larger hydrogen demands simply require more canisters for operation.


Scooters 2.0


Last November, APFCT began a second scooter demonstration in Taiwan with the city government of Taipei. In this demonstration 20 scooters have been deployed for use in environmental auditing site inspections and surveying by city officials.


101 scooters

APFCT’s current scooter model has a range of approximately 80 km.


glacier color scooter

Fueling costs can be very economic – in the Taipei demonstration, the local cost of electricity to generate the hydrogen results in a canisters exchange cost of NTD 30 (about USD 1).

充氢成本可以是经济实惠的 – 在台湾示范运行中储氢罐交换价格为新台币30元 (约一块美金),此价格包括当地产氢所使用的电力及物流费。

rear canisters

chen fueling

APFCT says this current model would sell for about NTD 90,000, about USD 3,000. That’s not quite a commercial price, but getting close. Assuming a successful demonstration, orders from city governments and the public could generate sufficient volume to get the price down, which would make APFCT fuel cell scooter be competitive with gasoline powered scooters.


Fork Lifts


APFCT has migrated its consumer friendly fueling system to a forklift application. They recently completed a demonstration of 5 forklifts in a distribution center in Taiwan operated by the RC Mart chain.


forklift front


Forklifts are an area of significant growth for fuel cells and one of the few applications that are commercially economic. Globally, there are at least 5,000 forklifts in operation at large distribution centers. These forklifts were all originally electric drive battery units. Their electric drives were all replaced with a fuel cell power system. The fuel cell systems themselves are somewhat expensive, however when one compares their total cost of ownership of the swapped out system with that of an electric drive, fuel cell systems are cheaper to operate and increase worker productivity.


The following chart provides a quick cost comparison between the two systems. In this case, a Class III forklift is used, which is a smaller unit where the operator rides on the truck.


TCO chart

Source: APFCT, Worthington Sawtelle LLC, National Renewable Energy Laboratory

Annualized cost

图片来源: APFCT, Worthington Sawtelle LLC, National Renewable Energy Laboratory

The bottom line here is that even though their fueling infrastructure and electricity costs are less, the battery driven units require significant labor for charging and refueling. What the chart does not show is that more battery units are required for a three shift day than fuel cell units; one battery unit must always be charging.


Virtually all fuel cell options for forklifts use high pressure hydrogen storage linked to a fuel cell with high internal pressures. Notice that in the high pressure bar above, all but the cost of hydrogen are likely to be relatively constant. The economics of the system depend almost entirely on the cost of hydrogen fuel. All systems currently in operation get their hydrogen delivered to a dispensing station in the distribution center from tube truck deliveries. The cost of that hydrogen increases with distance from the hydrogen production facility. Because of these high costs, a few operators are considering the installation of small natural gas reformers to generate hydrogen on-site from natural gas, which is relatively inexpensive in today’s market.


APFCT, characteristically, has developed a much different solution to this application, one which enhances its already winning cost analysis. The APFCT unit is shown as the third bar in the chart above, labeled “Low Pressure Fuel Cell.” This forklift design uses four fuel canisters that are identical to the ones used in the scooter. But unlike most other fuel cell forklifts, the APFCT unit uses a low internal pressure fuel cell. Lower internal pressures are less susceptible to membrane failure and have less moving parts. In the picture below the cabinets by this unit are the refuelers. Fuel canisters are placed in a rack in the unit and refilled with hydrogen being released from water through electrolysis.

亚太燃料电池一如既往地为叉车发展出与众不同的解决方案,图表成本分析中也显示了此解决方案的优势。亚太的叉车系统列在图表中的第三条,标有“低压燃料电池” 。该叉车设计采用四个储氢罐,与燃料电池摩托车所使用的储氢罐是相同的。和大多数燃料电池叉车不同的是,亚太使用了低内压的燃料电池。较低的内部压力使质子交换膜较不易损坏且运动部件较少。在图中位于叉车旁边的柜子内置有充氢机。充氢机可透过电解水制氢将氢气充填至氢气罐。

Next Steps


The best technology does not always make it in the marketplace, however. APFCT’s fueling approach offers a number of clear advantages over what is now regarded as conventional. Nonetheless, a number of alternative methods to store and dispense hydrogen in transportation applications have been attempted and then largely abandoned – usually due to the fact that such commercialization decisions are heavily influenced by the automobile manufacturers. It remains to be seen if APFCT can overcome the momentum already gained by others who are thoroughly invested in the high pressure cylinder on-board hydrogen storage model.

有时最好的技术并不容易商业化,亚太的低压储氢方式提供了许多明显的优势。尽管如此,目前已有部份储氢和配氢的替代方法都已经尝试过后并放弃 – 通常是由于此类替代方案的商业化主要是由汽车制造商所决策。让我们拭目以待,看亚太是否能克服多数使用高压储氢罐的主流,让低压金属储氢成为通往氢经济的快捷方式。

Bloom “Box”: Reverse Engineering the Economics

Last July Bloom Energy announced the placement of a 200 kW “Energy Server” (Bloom’s preferred language for “generator”) at Keio University in Japan, another at a large Softbank building in Fukuoka (Softbank is a major Japanese mobile phone service provider), as well as the creation of a 50/50 joint venture with Softbank to establish Bloom Energy Japan, Ltd.  These announcements were all part of the steady drumbeat of Bloom unit installations.  About the same time, however, Bloom issued a white paper where the President of Bloom Energy Japan, Miwa Shigerumotu, provided some insight into their pricing structure.  Below is a picture of the Bloom installation at the Softbank building.


Source: Bloom Energy Japan, Ltd.

As with other installations, Bloom is selling the product of the product, rather than just a piece of equipment.  The customers receive full output of the units with no upfront costs, paying a fixed rate of 25 yen/kWh, or about USD 0.21/kWh.  This rate is fixed at 10 years, with no fuel adjustment clauses.  In the white paper Mr. Miwa acknowledges that 25 yen is high relative to current prices, but goes on to say that the days of predictable electricity prices are over.  From April 2010 to April 2014, electricity prices in Japan rose an average of 10% annually.  He argues that paying a premium of about 5 yen now will be more than balanced out longer term when compared with the volatile and escalating conventional electricity cost.  His bottom line: a Bloom Energy Server provides price hedging and risk mitigation.

Fair enough, but is Bloom likely making money here or just initiating a loss leader program to pave the way for future sales?  We can get a very approximate sense of the implied cost of the unit with some reverse economic calculations.

The most significant variable cost to Bloom is fuel price.  The Bloom units are running on liquefied natural gas (LNG), which is not an inexpensive commodity in Japan.  The chart below gives some perspective to Japanese LNG pricing relative to the US and UK (Blue is US Henry Hub; Green is the UK and Red is Japan).  Last winter Japan LNG was at about $19/MM Btu.

Rice Univ

Source: National Gas Price in Asia, Rice University

The figure above overlays the timing of the Fukushima disaster and the closure of the Japanese nuclear fleet, which clearly had a major impact on price (and on the price increases noted above by Mr. Miwa).  Longer term, however, most analysts do not foresee a return to UK- or US-like pricing in Japan.   The Economist forecasts a decline in Japanese LNG price over time.


Whereas the IMF foresees a relatively constant price for the next several years.



Forecasting LNG prices in Japan is further complicated by the fact that, at least historically, Japanese LNG prices have been strongly correlated to world oil prices.  It remains to be seen if OPEC’s production announcement keeps oil prices stable in the short term but resulting in an increase when many now uncompetitive shale projects fail to survive, reducing supply.

For the purposes of this approximation, we’ll assume a fuel cost range for our Bloom units between $10 and $19/MM Btu.  We’ll also assume that Bloom internally financed the capital cost for these units at about a 2% interest rate.

Along with some other assumptions about O&M costs and taxes, the following break even maximum costs of capital as a function of fuel cost assumptions can be calculated.

Fuel Cost, $/MMBtu $/kW installed to achieve 10 year levelized cost of electricity equal to USD 0.21/kWh*
10 $9,000
15 $6,800
19 $5,000

When the first Bloom units were sold, industry estimates for their cost was on the order of $30,000 kW to $40,000/kW.  Perhaps 100 have been sold in the interim.  Given that starting point, it seems very unlikely that 100 units could result in economies of scale that would reduce cost by a factor of 5 or 6, as would be necessary for these two examples.

Even given all of the necessary caveats about the very approximate nature of the estimates and assumptions made above, loss leading is the market introduction strategy for Bloom in Japan.


*21 cents uses the current exchange rate – at the time the transaction was completed earlier this year the rate would have been nearly 25 cents US.  Perhaps this is why there was a report last month that the current rate is 28 yen/kWh.

Fuel Cells in the (Japanese) Home!

Many in the US are unaware of the fact that residential fuel cells are being routinely sold in Japan, especially in new home construction.  They are called ENE-FARM or “energy farms” that produce about 750 W of electricity with heat recovery.  The ENE-FARM fuel cells are all Proton Exchange Membrane (PEM) technology, but solid oxide based units are en route.They are especially popular in homes with radiant floor heat.  24,000 units were sold in 2012; nearly 18,000 were sold in the first half of 2013.  At the end of 2013, about 50,000 units were in operation across the country.   Some of that more recent demand was fueled by Fukushima, but more on that later.  The national government has a goal of 1.4 million units by 2020 and 5.3 million by 2030. If the economics continue to achieve economies of scale, those goals will easily be exceeded.  Let me walk you through the promotional material used by a gas company near where I live.


Shizuoka Gas’s offering happens to be a Panasonic model that came out in 2013.  It consists of two boxes: the fuel cell, which includes the reformer, fuel cell stack and inverter; and the hot water unit which includes waste heat recovery, storage and a backup heat source.  The figure below shows the boxes and their control panels.



Fuel Cell

Electrical Output 750 kW
Exhaust heat output 1.08 kW
Electrical Efficiency (HHV/LHV) 35.2% / 39.0%
Heat Recovery Efficiency (HHV/LHV) 50.6% / 56.0%
Dimensions 1.85 m x 0.4 m x 0.4 m (6 ft x 1.3 ft x 1.3 ft)
Gas consumption (HHV / LHV) 2.1 kW/1.9 kW
Noise level 33 dB
Weight 95 kg (209 pounds)

Hot Water Storage Unit

Hot water temperature 60 o C (140 o F)
Storage Capacity 147 liters (39 gallons)
Dimensions 1.85 m x 0..56 m x 0.4 m (6 ft x 1.8 ft x 1.3 ft)
Weight 209 kg (460 pounds)

Hot Water Supply and Backup Unit

Heat source Instantaneous latent hear recovery
Hot water supply capacity 41.9 kW
Heating capacity 17.4 kW
Maximum gas consumption 64.8 kW
Dimensions 0.75 m x 0..48 m x 0.25 m (2.5 ft x 1.6 ft x 0.8 ft)
Weight 44 kg (97 pounds)
Noise 49 dB


Shizuoka Gas offers free maintenance for 60,000 hours or 10 years.

“Learning” Operation

This particular unit has the ability to analyze the demand pattern for hot water and electricity in the home, and then adjust its operation accordingly.  In the event that an unusual call for hot water occurs the backup water heater engages.  The picture below shows the demand for electricity at the top and the demand for hot water at the bottom.  The middle, pink section, shows the learned state of the water storage unit that anticipates need.

learning storage


Overall Efficiency

Everyone appreciates the rationale for a gas company to sell gas appliances, but it goes farther than simple demand creation.  Japan has no source of natural gas and relies on imported LNG instead. Combusting LNG for power generation yields the typical mid-thirties efficiencies.  Reforming LNG at the end user location however, gives much better results.  This is how Shizuoka Gas explains it:

Conventional Generation Efficiency

con eff

Adding the ENE-FARM to the Equation

fc eff


The suggested retail price for the system is around ¥2 million installed.  That’s about $20,000 – and about 2/3 the 2009 cost.  After subsidies, however, most consumers end up at or less than ¥1 million, or $10,000.  Annual savings are on the order of $600/year, so the simple payback is around 17 years.  Clearly this is still for environmentally conscious upscale consumers, but there are plenty of them in Japan.

Taking it all a step further..

SHIZGAS (as Shizuoka Gas likes to be called) is promoting a solar PV/fuel cell cogen system that clips the peak that the fuel cell can’t handle with an installation like this:



Here is the schematic of the house:




But they’ve gone much farther than a simple concept and built a 22 unit subdivision called Eco Life Square Mishima Kiyozumi.  It’s been completed since 2011.

4-11-2014 5-26-26 PM



Eleven years ago I was involved with a small fuel cell development company that was pursuing a residential fuel cell.  When I joined it became very clear that such a first product for the technology was a bridge too far, way too far, and we reoriented ourselves to a commercial scale unit, the first prototype of which was a naphtha fueled PEM unit that operated at a gas station to make hot water for car washes and provide backup power in the case of an emergency.  Ironically the Japanese firm that tested that unit, ENEOS, now offers an ENE-FARM system (not derived from our technology however).

Residential fuel cells for the US may still, indeed, be a bridge too far, but they are becoming well established in Japan, and in part, because of Fukushima.  Residential customers, albeit wealthy residential customers, want the ability to have more predictable electricity costs and the ability to self-generate, having experienced power shortages, outages and higher costs.  Also, unlike the US, there is a very strong ethic here to mitigate carbon.  While not carbon neutral, fuel cell efficiencies do mitigate LNG fired central generation emissions.