Alternate Propulsion Plant Using Fuel Cell Technology

Since the marine industry continually laments restocking engineering talent to power its future, Maritime Reporter & Engineering News decided to hit the road and find where the future lies. The journey wasn’t far, approximately 45 miles from our office on Long Island at the Webb
Institute. Presented here is the recent work of two Webb upper classmen, Dusty Rybovich and Michael Cariello.

It is our responsibility as engineers to design a better world. Currently, the world is moving toward more environmentally friendly, or “green,” technologies with a focus on reducing emissions and finding more efficient sources of energy. Traditional marine diesel propulsion relies on the combustion of finite sources of energy and is ultimately an inefficient generator of electrical power and also creates harmful emissions. We have designed an alternate propulsion plant that uses fuel cells instead of engines, and makes excellent use of the boil off gas that naturally occurs aboard liquefied natural gas (LNG) carriers.
The primary objective of our thesis is to design and to evaluate a conceptual propulsion plant for an LNG carrier that uses fuel cells as the propulsive energy source. This includes the selection of an appropriate type of fuel cell and its associated fuel, and the design and optimization of the propulsion system and any related systems. Additionally, we have performed a comparison against a modern dual-fuel diesel-electric LNG carrier to evaluate the economic, environmental, and operational feasibility of our design.
There are many reasons to use an LNG carrier for the investigation of the fuel cell propulsion plant. The LNG trade is a profitable venture with expected longevity that is actively investigating alternate propulsion methods to reduce shipping costs while reducing negative environmental impacts. Additionally, modern technology has made natural gas acceptable for most fuel cell types because of advances in reformation technology that allow the hydrogen to be extracted immediately before the chemical process. LNG tankers possess a plentiful supply of fuel, and are already equipped with time-tested systems capable of dealing with this explosive gas. Additionally, the boil-off gas (BOG) that is naturally generated onboard an LNG tanker can be used as the main fuel source for the fuel cell system. Loading times in port would be reduced if the ship uses the BOG from the cargo as a main source of fuel, and does not need to spend time loading liquid fuel at a separate location. We have referenced a 2008 Webb Institute thesis by Adam Van Doren extensively in our work. He created a methodology for comparing LNG carrier propulsion systems in order to find the most economically efficient case. In his work, he concluded that a dual-fuel diesel-electric (DFDE) system was the most efficient. Thus, we have chosen to compare our fuel cell results to his best-case scenario.
We have used a concept solid-oxide fuel cell system produced by Fuel Cell Energy Corporation, the DFC-3000, to model our 23 MW propulsion system. A schematic of the internal process in one fuel cell is shown below.
In the simplest case, pure hydrogen gas is provided to the anode, where it is combined with the oxygen ions flowing through the electrolyte. This combination yields steam and electrons. The water is ejected from the cell while the electrons migrate through the load circuit to reach the cathode, where they combine with oxygen to create ions capable of passing through the electrolyte and continuing the conversion process. The work done by the movement of the electrons through the load circuit provides the electric power output.
However, it is difficult to obtain pure hydrogen and pure oxygen. Instead, standard air is used to meet the oxygen requirements, and a process known as internal reformation is used to extract hydrogen gas from hydrocarbon fuels. This procedure consists of mixing a hydrocarbon gas with steam in a high-temperature environment while in the presence of a catalyst.
We propose to use BOG as the main fuel for propelling the ship. The rate of production of BOG is unsteady, and may have to be stimulated at times. Because of this, the gas will be stored in holding tanks as it boils off from the cargo tanks. The gas can then be used in the fuel cells. Should the contents of the supply tank become low, and BOG is not being produced in sufficient quantities in the cargo tanks, a supply will be created by stimulating boil off of the cargo. This involves pumping some LNG out of the tank and passing it through a seawater heat exchanger to boil it.
With the added power produced from the waste heat recovery system, we predict that we can produce 408 kW per fuel cell stack. Designing to our goal of 23 MW gives us a fuel cell system that includes 56 DFC-3000 fuel cell stacks. The combined weight of the 56 fuel cell stacks is 680 tonnes. Estimating the weight of the auxiliary systems for the fuel cell system, including the fuel, water, reformation, and electrical conditioning equipment, we get a total combined weight of 800 tonnes for the fuel cell system. So, in terms of weight, it is feasible to use fuel cells for shipboard power.
The weight of the fuel cell system is comparable to that of other power production plants for the same power output. Four medium-speed, marine diesel engines with the required power output weigh about 576 tonnes. However, this is the dry weight of the engines alone, and does not include the weight of supporting systems, oils, and fuels. Because of this, we estimate that our fuel cell system will not weigh much more than the total diesel engine system, and may even weigh less. With the rapidly improving technology in this field that leads to more power-dense cells, it is increasingly likely that the fuel cell system will see further reductions in weight before the first commercial application.
For a laden voyage, we predict that 117% of our total required fuel supply is produced by natural boil off gas. This means that the natural BOG will be more than sufficient to power the voyage. The additional BOG can be stored for the return voyage in the holding tanks, burned in the boiler to increase efficiency, or incinerated. All of this BOG is accounted for in our economic analysis.
For the ballasted return voyage, we predict that 72% of our total fuel supply will come from natural boil off gas. This means that additional BOG will have to be stimulated. On the return voyage, a “heel”, or small amount of cargo, is left in the tanks of an LNG carrier to maintain their temperature so as to avoid thermal cracking and minimize boil-off during the next cargo loading operation. Our design requires that a slightly larger amount remain in the tanks, with the extra being allocated for power generation during the return voyage. In addition to stimulated BOG on the return voyage, extra BOG may be left over from the laden voyage and stored in the holding tanks. This could reduce the amount of retained LNG needed for the ballast voyage, thereby increasing the amount of cargo delivered.
On a round trip voyage, we have estimated that approximately 3% of the total cargo loaded is ultimately used as fuel. This can effectively be treated as a 3% reduction in the cargo carrying capacity of the vessel, but the cost is offset by the fact that no money is spent on traditional fuel for the round-trip voyage. The ship in Van Doren’s thesis seems to act similarly, as he makes the assumption that no HFO will be burned in the engines on the round trip voyage and power will be supplied primarily from BOG. For our purposes we are assuming similar amounts of cargo delivery between the two cases.
Van Doren’s thesis estimated that a DFDE system would spend approximately $765,000 on a roundtrip voyage, not including maneuvering and port operations. Our system will spend approximately $655,000 on a similar voyage assuming 3% consumption of LNG, which results in a 14% savings of approximately $110,000 per voyage. Over the course of the year, a ship with a similar operating profile would save about $1,250,000. 
Our cost analysis confirms our expectation that money will be saved during transit by using a fuel cell system instead of a DFDE system. Capital costs for the fuel cell system will likely be much higher than those for a DFDE system, even when the technology is ready for its initial application. However, with such large savings in fuel costs, it may be possible to recoup that cost and make the system more economically competitive. If its lifetime cost can be within the range of that of a diesel, other advantages like environmental cleanliness could make it a very desirable option. Eventually, as further investment is made in this technology, engineering advances and the benefits of mass production should radically decrease the costs from their current values.
We would have liked to give an estimate of the cost of this system, so as to draw a more thorough analysis between it and existing ships. However, the technology is only just being applied commercially and many corporations closely guard data on costs. We have had difficulty obtaining any reasonable estimate for a system of this size, or for a comparable system that could be scaled for this application. However, we have estimated the net present value of the fuel savings over a 20 year life to be between $10.7 and $15.6 million. This is a significant amount of savings, a portion of which could be applied to the capital cost of the fuel cell system to offset any price difference between it and a DFDE system.
There are other advantages to the fuel cell system that make it more desirable for shipboard application. The reduced emissions make this a clear leader in environmental cleanliness. With increasingly stringent legal and financial implications for ships that produce harmful emissions, it is in the operator’s best interests to ensure that his ship is as “clean” as possible. The fuel cell system not only fulfills all current and pending legal restrictions on CO2, NOX, and SOX emissions, but it far surpasses any other propulsion choice
After comparing available experimental emissions data, we determined that our system would emit only about 45% of the carbon dioxide, 0.002% of the SOX, and 0.033% of the NOX of a comparable DFDE system. Please see Figures 2 -4 below. These values assume that some MDO is being burned in the DFDE case, and that the LNG fuel and fuel cell exhaust contain a
We were unable to find any thorough documentation of the byproducts resulting from fuel cell creation. These products could influence the fuel cell’s absolute environmental impact. However, that is not within the scope of this analysis. Ship operators generally do not account for the lifetime “carbon footprint” of their propulsion systems, although in the future it could be a consideration in the event that proposed “carbon taxes” are implemented. This could result in higher capital costs because of the carbon taxes paid during the fuel cell production process, the burden of which would be shifted to the buyer. However, because we have not found any concrete research into the byproducts of fuel cell production versus diesel engine production, it is impossible for us to draw any real conclusion on this topic, other than to note that it should be investigated while implementing the system. The real concern for a business is the relative environmental impact of a propulsion system choice after purchase, in which case a fuel cell system is the obvious decision.
In conclusion, we believe that fuel cell technology is readily adaptable for shipboard applications, especially as a main propulsion plant for an LNG carrier. It carries with it a promise for a brighter future of reduced emissions and cost savings, while allowing for increased efficiency and flexibility in ship design. We hope that our research inspires others to investigate this technology.