Storage of Renewable Energy as Hydrogen on the Gas Grid. Is It a Workable Proposition?
Earlier this year an article appeared in Renewable Energy World quoting a report from Bloomberg which suggested that hydrogen produced from “surplus” electricity can be stored in the gas grid. This proposal is one of many which purports to provide an energy storage solution for “surplus” electricity generated by renewable sources. The driver for this proposal comes as the result of the non-dispatchability of renewable electricity generation as a method to accommodate the variability of, for example, wind power. While this proposal may appear to furnish a simple solution to the renewable hydrogen storage dilemma, it has quite a few hidden pitfalls.
The storage of renewable generated hydrogen in the gas grid is not a new idea. The concept has been proposed in Germany, reported by Florian Leucht at ICEPAG in 2014 at UC Irvine, and more recently in the Polish literature. These references are not to meant to an exhaustive recounting of the literature but to serve as gateway starting points for further investigation.
Many legal jurisdictions around the world, e.g., California and Germany, have legislated mandatory renewables content for their energy portfolios. California has an ambitious schedule of 25% by 2016 (end of & met) and 33% by 2020. Germany’s portfolio standards are just as aggressive with 33% renewable content for electricity and 20% renewable content for all energy components in place by 2020.
In the mix of capital equipment used for generation of renewable energy, wind power has garnered the lowest cost per kW of generation capacity. However, wind power is not dispatchable owing to the capricious nature of “every which way the wind blows”. The variance in wind energy production as a function of time of day can be seen in the figure below. The data are from the Tehachapi wind farm with a 5200 MW nameplate capacity and were recorded in 2010. The other phenomenon of note in this figure is the timing of the peak production of the wind power in California. Invariably, it occurs at night, in a pattern almost exactly opposed to the daily electricity demand curve.
Figure: Catalog of observed wind velocities over one month at the 5,200 MW wind farm in Tehachapi California
(See PDF Link for Figures.)
Owing to the mandated renewable’s acceptance requirements, the grid operator in CA, CA ISO, is required to purchase this wind generated energy. As the wind energy component is not dispatchable, the CA ISO must have some fossil fuel plants on standby in order to maintain the grid integrity, if and (surely) when the wind sources fade, as they will. Full integration of this wind resource should have some storage capability for the wind energy and/or more sophisticated weather (wind) forecasting ability to provide at least a 5-10 minute prediction window for wind performance to allow for initiation of the backup generation sources.
Storage Possibilities for Non-dispatchable Renewables
Of course, when ever the question of electricity storage arises, one’s immediate thoughts go to batteries. However, batteries will not be considered here as they are not yet ready for wide scale deployment as an economical grid scale storage solution, they are still too costly. Another grid scale storage solution is afforded by pumped storage. While an attractive solution to the problem, pumped storage is constrained by geographical constraints and is not considered here. Beginning in 2013 several companies in Germany, most notably, E.ON, proposed to use electrolytic hydrogen as a storage vehicle for this excess renewable energy. Energy storage built upon electrolytic hydrogen was the premise of the Bloomberg report and the main driver for this paper.
Hydrogen Generation for Storage of Excess Renewable Energy
The concept of excess renewable electricity is meant to describe the generation of renewable electricity in quantities that are unable to be used by the utility owing to lack of demand on its grid. One solution that has been used is that the utility offers this surplus power to other energy users outside of its own grid at often negative pricing, “Take my electricity, please”. Here we’ll look at electrolytic hydrogen generation, storage options, and, some applications.
See PDF link for full text.
Direct Conversion of Methane to Hydrogen and Elemental Carbon
In a recent paper in Science in November of 2017, Upham et. al., describe a catalytic process that employs a molten metal catalyst system to convert methane directly into hydrogen and elemental carbon. The reaction is very clean and occurs in 1 step with little or no product gas cleanup required, Equation 1, below.
For certain applications this process should favorably compete with the standard hydrogen production process of steam methane reforming (SMR). In the SMR process methane is reacted with water to produce carbon monoxide and hydrogen, Equation 2. The next step, the water gas reaction, is used to convert the carbon monoxide through reaction with water to hydrogen and carbon dioxide, Equation 3. The article also provides a good historical background to the SMR process.
The direct reaction:
The direct conversion of methane to hydrogen and carbon
CH4 (g) -à C(s) + 2H2 ΔH⁰ (298) + 74 kJ/mol (1)
The reaction is endothermic so external heat has to be applied to get any reasonable conversion. In fact, the reaction is conducted at 1000C and involves the use of a combination of Ni and a low melting metal (relative to 1000C) as the Ni carrying medium. Supporting metals tested included Bi, Au, Pb, Ga, In & Sn, all as melts.
The molten metal reactor system for the direct conversion of methane
The authors found that a 27% Ni – 73% Bi alloy exhibited the best reactivity for the process, equation 1. When operated at a temperature of 1065C a 95% conversion of the methane was observed. The reaction is very clean with the gaseous hydrogen easily separating from the reactor. The solid carbon product floated on the top of the metal melt. The initial carbon product contained a small amount of Ni + Bi (<4atom%) which could be removed by submersion of the carbon in the molten metal bath.
The reactor system is shown in Figure 1, in the PDF link. The figure shows Pt as the active metal catalyst but most of the work was done with Ni as it was more active than Pt. The rate data for all the systems studied are given in Figure S-12. The boiling point for Bi is 1564C with a low vapor pressure at 1000C.
Hydrogen from the Direct Methane to Hydrogen Process
The hydrogen from the subject process is extremely clean but may have some unreacted methane (a who cares?). (If the hydrogen is used for either ammonia syntheses or as a PEM fuel, the presence of methane is a “who cares”.) Another virtue of this process is that it is a single “pot” process which should scale with the volume of the reactor, a cubic function. Where could this technology be of use? See Appendix for clues.
 Another, hydrocarbon free process for hydrogen production is water electrolysis which produces both hydrogen and oxygen. The best electrolyzers run with about a 68% efficiency and scale with electrode area, a square function.
Is There a Market Potential for the Direct Methane to Hydrogen Process?
(Important background information is contained in Appendix A, see PDF link.) The US already has 2 extensive energy infrastructure networks in place, the electric grid and the natural gas pipeline grid. So, what if we set suitably sized (market determined) direct methane to hydrogen processing units along the major transportation corridors to produce the fuel hydrogen onsite. What could these service centers look like?
The hydrogen coming off the reactors would be fairly clean and dry but at low pressure (not too much above atmosphere pressure). The working pressure for hydrogen fueling stations for vehicle usage is typically 70 MPa (10,000 psi). The product hydrogen would have to be compressed to this level with, most likely, 2 stage compressors, not inexpensive. It would have to be stored in (very) high pressure storage tanks again not cheap. Finally, the byproduct carbon would have to be stored before pickup. This approach would obviate the need to develop a 3rd infrastructure of hydrogen distribution pipelines, probably a pipe dream, pun intended.
October 2018 Update: Anatomy of a Brandy Distillery Germain-Robin
An article describing a visit to Germain Robin, a world class brandy distillery, located in Redwood Valley, Mendocino County, can be found elsewhere on this website under the Services section. The Germain Robin brandy operation began it the 1980s and is operated as a craft distillery using the methods Mr. Robin learned growing up in Cognac, France. Germain Robin is a hands-on batch operation whose success was highly dependent on the skill of the master distiller. A picture of the alembic (pot) still is shown to the left, a full description is given on the “Sources” page. The large tank on the right is the condenser where the product is cooled and delivered. The pot still method is not high throughput process which is usually reflected in the cost of the finished product.
The original article ended with the report that the Germain Robin distillery survived the October 2017 Redwood Valley fire in Mendocino County. Most of the fire damages were concentrated on the east side of Highway 101, the main north-south artery. Fortunately for Germain Robin it is located on the west side of the highway so escaped any major damage.
The Sale of Germain Robin to E & J Gallo
The sale of Germain Robin to E & J Gallo of Modesto was noted in the original article posted on this website. A more detailed account of the sale of Germain Robin to Gallo was given in August of 2017 in a Wine Enthusiast article by Virginie Boone. This article noted that “Gallo acquired the brand, its inventory of aged and aging brandies, and other finished goods. Some of the brandies date back three decades and are made from a wide variety of grapes”. One can get a feel for the size of this inventory from the picture of the aging warehouse, above.
Consider the Coal Equivalency for a Record Large Quantity of Natural Gas
This past January (the 12th), eia reported a record week for withdrawal of 359 Bcf (billion cubic feet) of natural gas from storage facilities during the previous week ending on January 5. This withdrawal was about 80% greater than the previous record of 288 Bcf set in the week ending January 9, 2014. With an increase of this size it should not be a surprise that a withdrawal of this magnitude caught eia’s attention and mine(!), see the graphs below.
Driven by my curiosity, I wondered what the energy of such a large quantity, 359 Bcf, of natural gas would be in coal equivalents. In order to get an appreciation of the magnitude of this quantity of energy, I made some back of the envelope calculations to see a hypothetical coal equivalent. The calculations presented here are, mostly, in SI units; the values were taken from the US eia or the GREET 1.8 tables. Further, only the LHVs (lower heating value) for each fuel were used and values were rounded to whole numbers.
The Natural Gas:
Three hundred fifty-nine Bcf of natural gas is a large quantity of energy:
Moving to SI units: 359 Bcf x 0.0283 m3/ft3 = 10.16 x 10E9 m3 of gas
10.16 x 10E9 m3 x 34.64 MJ/m3 = 352 x 10E9 MJ for the energy content of the natural gas
For coal, the GREET 1.8 value for bituminous coal of 22,460,600 BTU/ton (22.460 MMBtu/ton) was used. (This GREET 1.8 value was also stated in SI units as 26.12 MJ/kg.) However, the readily available data for coal transport employs English units; I’m going to do something terrible, I’m going to mix units… (Oh, the horror!)
Coal energy requirement: 22.46 MMBtu x 1,055 MJ/MMBtu = 23.7 x 10E3 MJ/ton Using this value for the energy content of the coal and our calculated energy content for the natural gas at 352 x 10E9 MJ, we get a coal equivalent of:
352 x 10E9 MJ x ton/23.7 x 10E3 MJ = 14.8 x 10E6 tons of coal
The coal, used here in Colorado, comes out of Wyoming in unit trains of 100 or more cars. Wyoming coal is one of the most common coals used for electricity generation; even Excel Energy uses Wyoming coal for its coal burning power plants along the Front Range, even though coal is mined in Colorado. The coal is transported via these unit trains that can range from 115 coal cars on up. Each car can contain about 120 tons of coal. A unit car train of 120 cars with 120 tons of coal per car is moving about 14,500 tons of coal.
Q: How many such trains will we need to move our 14.8 x 10E6 tons of coal?
A: 14.8 x 10E6 tons x one unit train/14.5 10E3 tons = about 1,000 unit trains
A CXS publication noted that it takes about 5 hours to load a single 15,000-ton unit train. That time value indicates that it should take 5,000 hours to load the required number of unit trains. Since a week has 168 hours, the time window to move the coal energy equivalent of the 359 Bcf of natural gas says that we will need about 30 independent coal train loading stations to meet this record tally of energy transport. (No consideration was given here to transport times since sources and destinations are quite variable and distributed.)
When you crunch the energy numbers, it becomes apparent that we, in this country, use a very large amount of energy in all of its manifestations. If you are interested in lowering your energy footprint, let me suggest that energy conversation is the most cost effective and quickest route to implement any goal for energy use reduction.
October 24, 2018
Note: Dateline of article October 17, 1989, Article posted on blog October 21, 2017
My Most Stressful Day at Work Ever: Earthquake!
October 17, 1989, an exciting day with Game 3 of the World Series set to start at 5:35 pm at Candlestick Park. It was a purely local event pitting the San Francisco Giants against their cross bay rivals, the Oakland Athletics. The Athletics were already up 2 games to none for the Giants. Then it happened, striking at 5:04 PM on national television. EARTHQUAKE!
It was a 6.9 (Richter Scale) earthquake with the epicenter near Loma Prieta, a mountain in southwest Santa Clara County and the highest peak in the Santa Cruz Mountains. The intense shaking, lasting up to 15 seconds, was felt from Monterey County to north of San Francisco. The casualties totaled 63 killed and 3,757 injured. A majority of the casualties resulted from the collapse of the double decker section of the I-880 Freeway in Oakland, pictured above.
AT the time the earthquake struck I was on the floor of our plating shop in our new circuit board facility. I was the chief technical officer of the company and also had responsibility for overall safety and chemical handling. The shop had 45 50-gallon process tanks, of which over 50% were devoted to rinse operations; a 2,400 gallon copper plating tank; an 800 gallon solder platting tank; and, a 600 gallon tin-nickel plating tank.
With the earthquake, the process solutions began to slosh back and forth in their respective tanks with quite a bit of wave action. Yet, we experienced only a minor splash out from one 50-gallon tanks. Fortunately, our shop in Fremont was one Bay Area location that did not lose electric power. We were able to clean-up the spill and run it (legally) through our wastewater treatment unit. As such, it was contained and did not constitute a reportable spill.
We would have been in desperate shape had any one or more of the process tanks ruptured. Such spillage would have constituted a hazardous situation requiring evacuation of the (bermed) plating shop and the use of an outside HazMat clean-up company. A mandatory report to the sewer district could have closed our operation which could potentially have ended the company.
Earlier in the year when I was designing the process tanks with the outside vender, I insisted that the tank designs include seismic bracing. While this consideration added about 15-20% to the cost of the tanks, the investors came to agreement that such seismic considerations were probably prudent. Thus, the company was able to resume operations on October 18.After we did an orderly shutdown of operations, and confirmed that we did not incur any real damage to the facility, we called it a day, a very scary day. I headed home across the Dumbarton Bridge to the other side of the bay. It turned out that I crossed the bridge just before Caltrans closed it for a seismic inspection. Looking out the car window I saw some local damage to the KGO broadcast towers on the Baylands, Fortunately, the Dumbarton Bridge, the newest trans-bay bridge, came through unscathed.