A Review of Engineering and Safety Considerations for ...€¦ · 68 Oil nd as acilities • June 2015 PB Oil and Gas Facilities • June 2015 June 2015 • Oil and Gas Facilities

68 Oil and Gas Facilities • June 2015PB Oil and Gas Facilities • June 2015 June 2015 • Oil and Gas Facilities 1

A Review of Engineering and Safety Considerations for Hybrid-Power

(Lithium-Ion) Systems in Offshore Applications

D. M. Hill, A. Agarwal, and B. Gully, DNV GL

topside and subsea systems enjoy the same benefits (increased ef-ficiency and reduced costs), yet for different reasons. Oil and gas stakeholders must determine when and where hybrid-power sys-tems provide the most value for operations, how they should be implemented, what technologies are acceptable, what safety con-siderations there may be, system suitability for extreme environ-ments, and how these technologies can improve the bottom line. There is a wealth of information on Li-ion batteries, though it is not all consistent—cost data are unclear, lifetime and energy density considerations vary under different conditions, and ruggedness and application to harsh environments constitute a large uncertainty. In the following sections, we will address these issues to help provide clarification for the oil and gas operator.

Industry Precedents and Economic BenefitsFrom prior experience in the automotive sector and more recently in marine applications, hybridization of power systems is known to increase energy efficiency by means of reduced fuel consump-tion (Chen et al. 2012). Increased energy efficiency of operations can reduce cost and improve energy security of oil and gas oper-ations by reducing the energy cost of exploration and production. Currently, remote power generation on platforms runs at less than optimum efficiency without hybridization. The subsea power market is to grow 7 to 8% per year and is already valued at USD 3.9 bil-lion (De Beaupuy 2013). The emissions controls cited previously, in tandem with growth in the subsea market and the greater costs that will accompany that growth, imply that solutions are needed that will increase energy efficiency, reduce emissions, and reduce capital and operational costs. Captive generation is the largest energy-generation mode across platforms, though there is a shift toward the exploration of incorporating renewables. There are power-system manufacturers developing new grid solutions. As captive generation becomes more complex, there will be a need to incorporate energy storage and hy-brid solutions to optimize generation and load together. Hybridiza-tion can reduce both emissions and costs for these power systems across the board. There are all-electric subsea-device designs that are passing extensive qualification requirements (SAFT 2013).

It should be noted that subsea battery concepts have been ad-dressed in the last decade through a number of research projects. The most notable and recent subsea effort examined optimum cost for subsea power systems, and that study was conducted from 2008 to 2010 (Wolf et al. 2010). The finding of that work was that so-dium sulfur (NaS) batteries, coupled to portable nuclear-power sys-tems similar to those in naval submarines and ships, would provide the optimum system for low cost over long life. Since this study started, tremendous growth in the lithium-ion (Li-ion) -battery market presents new opportunities from battery technologies emerging as stable solutions for marine, stationary, and mobile power. NaS batteries are demonstrated in large-scale systems for onshore electric grids (> 50 kW), but the market remains nascent and cost remains an issue (Nichols 2005). With relation to power generation, the Macondo well incident, as well as the Fukushima

Copyright © 2015 Society of Petroleum Engineers

This paper (SPE 174091) was revised for publication from paper OTC 25129, first presented at the Offshore Technology Conference, Houston, 5–8 May 2014. Original manuscript received for review 30 January 2014. Revised manuscript received for review 29 January 2015. Paper peer approved 2 February 2015.

SummaryFrom prior experience in the automotive sector, and now the mari-time sector, hybridization of power systems is known to increase en-ergy efficiency and reduce emissions, with lower fuel consumption. With impending emissions-control areas in the US continental shelf, and nitrogen oxide enforcement mechanisms in the North Sea, emis-sions reduction in oil and gas exploration-and-production operations is increasingly relevant. Hybrid-power systems can address some of these issues with batteries to offset peak loads, thereby reducing size requirements for the total system. The challenge that the oil and gas industry faces is to decide when and where hybrid-power systems pro-vide the most value for operations, how they should be implemented, what technologies are acceptable, what safety considerations there may be, and how these technologies can improve the bottom line. There is a wealth of information on lithium-ion batteries, though it is not all consistent—cost data are unclear, lifetime and energy density considerations vary under different conditions, and ruggedness and application to harsh environments constitute a large uncertainty. A re-view of these technologies is provided to serve as a selection guide.

IntroductionThe offshore oil and gas industry faces two challenges in the near future: increased regulation on emissions and increasing cost of op-eration. The following discussion outlines the way in which hybrid-power systems—primarily those enabled with lithium-ion (Li-ion) battery technologies—can address both of these challenges, though not without some safety and technology qualification needs. US emissions-control-area regulations will require nitrogen oxide (NOx) reductions by 80% by 2020 (DNV 2012b). In preparation for this, engines use urea injection to reduce emissions, which adds cost to the power system. These systems reduce particulates, NOx, and sulfur oxides, yet hybrid systems can be used to level out transient loads, which will also lead to more-efficient urea management. In addition, the push for deepwater exploration and production implies increased costs in transmission of power to the seafloor, and greater distribution of subsea equipment in deep water. Also, because of stringent safety requriements, subsea equipment will require in-creased reliability by means of backup power and self-sustaining power systems. When connected, this equipment will require trans-mission lines that must be sized for maximum capacity, even if that capacity is met infrequently, unless hybridization can reduce the transmission requirement. Battery-based energy storage on the sea-floor can be used to discharge during peak loads, thereby reducing capacity requirements for the overall power system. The smaller-diameter cables or reduced capacity of umbilicals translates to re-duced capital cost and less losses in the transmission system. Thus,

June 2015 • Oil and Gas Facilities 692 Oil and Gas Facilities • June 2015 June 2015 • Oil and Gas Facilities 3

a high priority. As power systems become more complex, and gen-eration and distribution of power is moved subsea, there will be an increased need for safe, redundant, efficient, and aggregated power sources in the offshore sector to ensure multiple safety gates for backup and redistribution of power in the event of a failure. Energy storage and batteries will be the crux of this solution.

Battery-System PropertiesNew systems-integration efforts are bringing together battery tech-nology with generators, power electronics, inverters, and battery-man-agement systems (Maritime Journal 2010; Vardtal and Chryssakis 2011; Chen et al. 2012; Gully 2012). Some of these systems are rugge-dized for marine environments, and others are building systems for on-shore environments for energy-storage applications (Hill et al. 2014a). Since 2008, the lithium-ion (Li-ion) -battery market has grown dramat-ically. To those not familiar with the battery market, the term “lithium ion” is used generically for many products, but the details of the chem-istry imply very different performance characteristics for batteries. It is said in automotive racing that “you can have speed, reliability, and low cost: pick two. Similarly, for batteries and hybrid-power systems, there are engineering and cost tradeoffs to consider that will affect cost, life-time, power capability, energy (duration), and reliability.

DNV GL collected databases on battery cycle life, safety per-formance, cost metrics, and failure modes since 2008. These data-bases are collected through testing data and industry surveys. The parameters in the database are used to determine what conditions affect the ability of a battery to perform in its required environment and what batteries are suitable for a given application. The applica-tion requirement can be as simple as a power and duration metric, or it may involve more-advanced considerations such as tempera-ture considerations, pressure, containment, or additional safety re-quirements. As battery technologies mature, more tolerance will be afforded to their risk of use, and they will be implemented in more applications that push the outer limits of the battery-performance envelope. This is witnessed already in the automotive sector, where the previous 80% capacity threshold was defined as end of life, but now it is more common to see 70%. In addition, some batteries are deployed without thermal management (LaMonica 2012). These lessons learned are bound to extend to other industries.

DNV GL conducted its own survey comprising 16 different Li-ion battery manufacturers and compiled factors such as cycle life, temperature stability, power density, energy density, and cumula-tive energy throughput that were all examined in close detail. The purpose of this study was to begin a classification of battery types and explore their suitability for different applications. The findings are consistent with other industry standards on batteries; however, the need became apparent to use the survey results to educate po-tential end users in the oil and gas industry about what is available, what capabilities exist, and what other factors to consider before implementation. Here are some general findings:

1. Power density (W/kg) and energy density (W-hr/kg) are gen-erally inversely proportional.

2. Cycle lifetime under conventional testing conditions does not always imply high lifetime energy throughput.

3. Cost and weight are directly proportional.4. Cost and cycle life are directly proportional, and with Item

2, it is therefore true that cycle life and weight are directly proportional.

5. Iron phosphate chemistries, as well as chemistries that imple-ment titanate anodes, tend to be more stable at higher power rates, measured as C-rate*. Iron phosphate chemistries can achieve very high C-rates (10 to 100).

6. Titanate-anode chemistries advertise long cycle lifetimes.

power-plant disaster, highlights the need for safety and environ-mental protection, and have delayed the implementation of subsea nuclear systems in the near future. In short, the landscape changed significantly in the last 5 years, and it is time to revise available power-generation and energy-storage solutions.

Operations in offshore exploration-and-production activity can benefit from hybridization with Li-ion batteries in the following cited cases. Systems as basic as elevators and cranes benefit from hybridization, and one study shows that replacement of a redun-dant generator with a battery pack for crane operations pays back in less than 1 year, with continued fuel savings afterward (Ovruum 2013). Particularly for transient power loads, such as drilling or position stabilization, hybridization offers the same benefits as it does in the automotive sector—the battery pack absorbs transient loads and permits the combustion engine to be downsized and op-timized for peak efficiency. With a smaller, more-efficient engine, the emission of particulates, nitrogen oxides (NOx), and sulfer ox-ides is reduced (Gully 2012). If the aim is to reduce NOx emis-sions, however, hybridization is a competing technology with urea injection. Urea-injection systems require capital investment and/or retrofit for existing systems. NOx emissions can be reduced by 90% with selective catalytic reduction or urea injection (Caterpillar 2002). The efficiency of these processes, however, is reduced when the load is transient. Hybridized powertrains, on the other hand, re-duce NOx emissions by 29 to 51% (Vardtal and Chryssakis 2011). In marine propulsion and dynamic positioning, hybrid-power ap-plications reduce fuel costs, some by 10% or more (Chen et al. 2012). In other hybrid applications, fuel savings can be reduced by as much as 20 to 29%, lowering operational costs and providing a value proposition for lowered overall cost per kilowatt-hour (Vartdal and Chryssakis 2011; DNV 2012b; Ovruum 2013). There may be a place for both hybridization and downsized urea injection on the same platform, with lower overall operational costs than urea injec-tion on conventional powertrains.

Regulatory NeedBecause the oil and gas industry is global, the regulatory impact is also global. Emissions-reductions regulations for the Gulf of Mexico, in particular, are targeted for the US under International Convention for the Prevention of Pollution from Ships (MARPOL) agreements, and for Mexico under the emission-control area (ECA). In addition, on 14 January 2013, the US Environmental Protection Agency (EPA) finalized amendments to the National Emissions Standards for Hazardous Air Pollutants for stationary reciprocating-internal-combustion engines, which apply to many areas, but include the outer continental shelf, where offshore opera-tions occur. The North American ECA, under MARPOL, went into effect from 1 August 2012, bringing in stricter controls on emis-sions of sulfur oxides, nitrogen oxides (NOx), and particulate matter for ships trading off the coasts of Canada, the US, and the French overseas collectivity of Saint-Pierre and Miquelon. Both the US EPA and the US Coast Guard will enforce these regulations, imple-menting fines for violators (EPA 2012). These regulations are per-haps less stringent than efforts in Scandanavia and Europe, where the European Union (EU) Emissions Trading System (ETS) is the first multinational cap-and-trade program to limit global-warming pollution. Since its 2005 inception, the EU ETS is credited with a 13% decrease in emissions, with increasing stringency until 2020 (Brown et al. 2009). In Scandanavia, which proves to be a strong market for electric and hybrid ships, the Norwegian NOx Fund im-poses a EUR 0.5/kg NOx tax on shipping and a EUR 1.5/kg tax on oil and gas operations, which funds an overall program to support emissions controls and NOx-reduction measures (Johnsen 2013). The program in turn funds liquefied-natural-gas conversions, low-NOx engine modifications, and selective-catalytic-reduction mea-sures. Hybridization will also fall under this program (Sandvik 2014). Therefore, strong regulatory trends intend to curb emissions from offshore sources. In addition, safety in offshore operations is

*A 1C discharge or charge rate is one full discharge/charge in an hour, respectively. A 2C discharge rate is a full discharge in one-half hour. By extension, a discharge or charge rate of xC is a full discharge/charge in 1/x hours.

70 Oil and Gas Facilities • June 20152 Oil and Gas Facilities • June 2015 June 2015 • Oil and Gas Facilities 3

7. There is much innovation in the nickel/cobalt/manganese (NCM) chemistry variants to acquire a balance of multiple properties, such as good power, good energy density, long cycle life, and reasonable temperature stability.

8. Many cells labeled as lithium cobalt oxide (LiCoO2) adver-tise high power capability, but lower cycle lifetimes. Other LiCoO2 chemistries use undisclosed chemistry additives to improve thermal stability and cycle life.

Not all Li-ion batteries were designed for large-format systems. The portable-electronics sector led the initial growth of the Li-ion market (tens of watts to hundreds), and the automotive markets led large-scale battery-pack developments (kilowatt scale) in the 2010 era. Currently, kilowatt- to megawatt-scale batteries are deployed in stationary grid storage, heavy industries, and marine and off-shore sectors. Some battery types will emerge and dominate these more-extreme applications. Therefore, it should be said that a bat-tery designed for portable electronics is not necessarily scalable to large-format power systems. The cost requirements will also differ from electronics and automotive markets. It is important for the buyer of such systems to understand that the quality of the battery will determine the quality of the total system.

Weight may be a prime consideration. Battery construction in-volves an anode, cathode, separator, and electrolyte. These layers can be rolled into a cylindrical form factor, or sandwiched in a pouch-type or prismatic cell. Most Li-ion batteries use a graphite carbon anode on a copper collector. Some batteries use a lithium ti-tanate anode, which is heavier, but mechanically more stable. The weight of this battery is evident in its energy density, which ranges from 50 to 75 W-hr/kg, which is roughly 2X lead acid (lead acid can be assumed to have an energy density of 30 to 50 W-hr/kg). The cathode is the electrode that determines the common name for a battery. Lithium/iron/phosphate cathodes are sometimes called LFP or LiFePO4. LFP batteries exhibit lower to moderate energy density compared with other chemistries (≈90 W-hr/kg), but pos-sess very high power capability and somewhat moderate cycle life-times. It is generally true that more weight implies more cost, and therefore an increased cycle life will imply a greater capital in-vestment. Many variations of the NCM batteries exist, sometimes called NMC. These batteries have medium to long cycle lifetimes, reasonable power capability, and the potential for higher energy density (>130 W-hr/kg). Some NCM cells are tolerant to high dis-charge rates for power applications. The NCM chemistry is versa-tile in that it is tuned for multiple applications while maintaining thermal stability. The lithium-manganese-oxide chemistries, some-times called LMO or “spinel” for their microstructure, demonstrate an energy density greater than 200 W-hr/kg with long cycle life-times. The nickel-cobalt-aluminum (NCA) chemistry may vary its performance with additions of manganese, and can also have higher energy density (≈200 W-hr/kg). While typical NCA batteries claim lower energy density, some manufacturers make successful additions to the NCA chemistry to achieve high energy density and high power. Some technology outliers cause significant spread in the energy density for each chemistry, and this is a testament to the evolving market.

Performance CharacteristicsA common battery-performance metric is cycle life. This metric, while useful, is not alone a sufficient characterization of battery performance. Factors that affect performance include tempera-ture, depth of discharge (DOD), rest times between cycles, and the charge and discharge rate. If charge and discharge rates and tem-perature are included, the rest times are rarely mentioned. In partic-ular for high cycle-life claims, one should be careful to determine whether the depth of discharge was less than 100%. Higher tem-peratures are strong abuse factors (i.e., high temperatures degrade battery capacity more than moderate temperatures). The most-common cycle-life chart for a battery is performed at 20 to 25°C, with 1C charge and discharge rates and 100% DOD at each cycle.

However, most batteries are deployed in systems with very dif-ferent characteristics. They will likely not experience 100% DOD, their temperature may be less controlled, and they may be expected to endure high C-rates to meet the power load.

The survey notably finds that the lithium/titanate/oxide (LTO) chemistries claim long cycle lifetimes. Most nickel/cobalt/manga-nese (NCM) batteries for large-format systems have a 2,000- to 3,000-cycle life and as much as 6,000 cycles under less-abusive conditions, such as a partial DOD. Some LTO batteries claim 6,000 to as much as 16,000 cycles at 100% DOD. There is another standout in the survey—some lithium-manganese-oxide (LMO) batteries have long cycle lifetimes, but only when cycled at 50% DOD/cycle. This again highlights the need to identify the specific conditions under which cycle lifetimes are reported.

For design purposes, the battery system may be needed for high power or high energy. For battery systems with a uniform chem-istry, these may be mutually exclusive properties. The survey rep-resents a significant cross section of the present-day industry and determines that with energy densities greater than 150 W-hr/kg, it is unlikely that the battery is capable of power densities greater than 2000 W/kg. However, for lower energy densities, some chemistries can offer up to 12 000 W/kg. The cell design and its chemistry will have an impact on its ability to deliver high power. Power density and energy density are inversely proportional across surveyed tech-nologies. Power density in some cases is as high as 10 000 W/kg or more, but the power densities of most lithium-ion (Li-ion) bat-teries range from 500 to 2000 W/kg. Batteries with energy densi-ties near 150 to 225 W-hr/kg are typically the batteries with the lower power density. Power density is a specialized case for battery types that can be “tuned” with subtle chemistry changes and elec-trolyte modifications. Li-ion battery manufacturers do this while balancing thermal stability and safety to reduce self-heating during high power discharge.

It is also true from the survey results that heavier batteries tend to have longer cycle lifetimes. Higher cycle lifetimes are often as-sociated with lower energy density, or conversely, more weight per megawatt-hour. On the basis of the energy density of LTO, for example, a cycle life of 10,000 to 16,000 cycles would require a system weight of nearly 25 tons (just batteries) vs. a system with a cycle life of 2,000 cycles weighing just 2.5 to 5 tons.

An oil and gas operation requires marinizatoin of equipment to endure harsh conditions. An operator may wish to know if a bat-tery system can endure the temperatures that it may see in the field, or if it will have the lifetime needed to complete its service. Most battery-specification sheets emphasize cycle life, energy density, and recommended temperature range. Beyond these specifications, the batteries may respond to other environmental stresses (such as shock, mechanical deformation, and vibration), which are covered in the Safety Considerations for Battery Systems section. In Fig. 1, three example paradigms are shown for battery management. In the far left of the figure, the typical and most-conservative ap-proach is shown. In these cases, the batteries are never expected to endure charge or discharge rates higher than 1C, and great care is taken to maintain a temperature at or near 25°C and an average state of charge (SOC) well within the limits established in Fig. 1. This is the typical approach taken in the automotive sector. In the energy-storage sector, battery-management system protocols are similar even though there may not be capability for active cooling. For distributed energy-storage systems, however, there is an as-sumption that the air-cooled system will operate in an underground vault, which should have consistent temperatures at or near 50°F (CES 2009). Oil and gas operations may be less conservative than the automotive and onshore utility industries in some applications. There may be a need for high power applications such as drilling or lifting, cranes or elevators, dynamic positioning, or actuation of valves and chokes. In addition, the environment will be less con-trolled, entailing large swings in temperature, or perhaps cold tem-peratures (in subsea or Arctic applications) or high temperatures


(in Gulf Coast or high-power applications). In this case, the C-rate needs to be balanced against temperature and average SOC. This is the state of the art in battery safety: average SOC is maintained con-servatively, C-rates are kept at 1C or less (unless the chemistry and application are specifically designed otherwise), and temperature is maintained at or near 25°C. This conservative battery management adds cost burden because some energy and power potential is left unused in exchange for enhanced safety margins. For power appli-cations (middle of Fig. 1), the C-rate comes into play and affects the capacity-loss severity near the end of life (EOL). The longevity of the system may matter less in this case than the ability for the system to deliver power reliably during critical periods. In the far right case of Fig. 1, extreme conditions, such as high temperature, may dominate the performance of the system.

A major battery-system cost factor is the overdesign to offset capacity degradation and maximum expected capacity. Just as a conventional generator is sized for its maximum load (even if that maximum load will be encountered infrequently), a battery system must be similarly designed. Once this maximum capacity is deter-mined, the battery system will need an additional factor to account for degradation over its operating life. A carry-over rule of thumb from the automotive sector is that EOL for the battery system oc-curs when the battery capacity is 80% of its as-new state. For ex-ample, in Fig. 2, the actual electrochemical capacity of a battery pack may be 1.3 kW-hr, but conservative limits on the upper and lower voltage are placed such that the system SOC never exceeds 90% or never drops below 20%. These restrictions reduce the us-able capacity of the battery pack to 1 kW-hr. However, the cost of the system is then inflated because the battery cost may be USD 500/kW-hr (for example), but the additional sizing considerations inflate the cost to USD 650/kW-hr. An additional size factor may also then be instituted to account for capacity at EOL, which will inflate the cost further.

Most Li-ion batteries have maximum temperature ratings be-tween 0 and 60°C. The high temperature limit may be lowered during charging. Charging is more abusive to batteries than dis-charging, and for this reason, many batteries have higher allowable discharge rates than charge rates. The C-rate affects battery temper-ature. Higher C-rates induce greater cell heating, and this heating af-

fects capacity loss (DNV 2013). From 2008 to 2012, battery-system costs averaged approximately USD 1,000/kW-hr. However, the bal-ance of system cost will change this factor, and variations in battery technology are driving some disparity in these numbers. In addi-tion, volume will affect the cost. In some cases, an assumption of USD 500/kW-hr for the batteries alone may be more appropriate, but the titanate chemistries may cost more than USD 1,000/kW-hr. The capital cost is less important than the system requirements for safety, expected lifetime, and total energy throughput. For example, the lifetime-cost-per-throughput kW-hr may be more relevant than the upfront cost per kW-hr of capacity. The best way to measure this is to extract the battery-capacity curve from test data, integrate the energy throughput, and compare that with the upfront cost.

The cumulative energy throughput (normalized by weight) for each battery chemistry varies significantly. Some batteries with low energy density, such as the titanates, can have long cycle lifetimes and therefore deliver a significant energy throughput over their long lifetime. Cumulative throughput can be measured in W-hr/kg to normalize their performance by energy density across battery chemistries. Cumulative throughput is proportional to the square root of the percent capacity lost over time (i.e., there are dimin-ishing returns past a certain threshold).

However, high throughput energy is not excluded to just one chemistry. Certain manufacturers have obtained outlier life-time performance with NCM and lithium/iron/phosphate and are learning to balance lifetime and total throughput with evident suc-cess. It may be a question of whether the added cost for higher cycle life is a function of diminishing returns. For example, one might expect the titanate batteries with their long cycle lifetimes to provide a higher cumulative throughput than competitors. But Fig. 2 shows that some LMO and NCM batteries can approach the performance of the titanates when measured in this manner. The lifetime throughput energy can be verified with limited short-term-testing campaigns or life-prediction tools.

Qualifying Technologies for ServiceA major risk in the deployment of the battery system is that it will last long enough to provide a payback on its investment. Indepen-dent testing supported the use of databases and lifetime-prediction tools for battery types. The tool assesses whether a battery system will meet the required lifetime given a duty cycle, system size, and average operational temperature. The purpose of these tools are several:

1. To understand capacity fade and its impact on system per-formance for the long term, as well as initial capital cost and payback

2. To identify risks that may impact system performance in its intended application

3. To identify safety factors for battery systemsThe purpose of battery-life-prediction models is to accelerate

the testing cycle for new battery chemistries. An example of a com-

Actualelectrochemical

capacity (1.3 kW-hr)

Deliveredcapacity (1 kW-hr)

90% SOC

130% × USD500/kW-hr =USD650/kW-hr

20% SOC

Fig. 2—Overdesign is the simplest way to ensure that a battery is never stressed in its application, though this practice adds cost.

Temperature

Conservative Management: 25°C, 1C/1C Chargeand Discharge Rates

Power Applications,2C Rates or Higher

Extreme Conditions(High Temperature)

AverageSOC

AverageSOC

AverageSOC

Temperature

C-rateC-rate

Temperature

Fig. 1—Balancing aging factors for different battery applications.


parative analysis is shown in Fig. 3, in which pack size and temper-ature impact pack longevity directly. This analysis can also be used to estimate battery-size requirements, which leads into economic trade-off and cost/benefit analysis.

In the 25°C condition and with conservative state-of-charge (SOC) management, the pack can maintain up to 75% capacity through its service life. However, with the same duty cycle, if the pack size is reduced from 1,000 to 900 kW-hr (which reduces cap-ital cost), if the battery is permitted to use more available SOC (up to 90% as opposed to 80%), and if the cooling requirements are relaxed, allowing it to reach a higher average temperature of 30°C, one can see that it loses capacity faster, reaching 40% by Year 10. In this case, the cycling component dominates the degra-dation. Under these conditions, the pack would fail its functionality requirements; therefore, the life-prediction tools permit sizing ex-ercises to balance cost and performance.

Because the lithium-ion battery industry is growing quickly, it is not often practical to implement long-term cycling tests to verify performance because a new and better chemistry may already be on the verge of commercialization. Modeling techniques were de-

veloped with available data and independent testing to interpolate results with semiempirical equations and generate a capacity-fade curve. These are the techniqes that generated Fig. 3. Modeling can then be used as it is often intended elsewhere—to consolidate in-dustry knowledge and downsize the amount of testing required, identify areas of high consequence, and focus testing efforts on the most-relevant cases for model verification.

Battery-life prediction is useful for a number of applications. Hy-brid and fully electric vessels are implementing large-format bat-teries that may experience a number of unique conditions, such as cold temperatures, fast charging, high-rate discharging, or long periods of dormancy (Vardtal and Chryssakis 2011). Battery-life-prediction models verify battery-size requirements and test battery compatibility virtually in environments with contrasting temper-atures (DNV 2014). Separating out the effect of these factors on battery capacity and designing the battery to ensure the same per-formance at Year 10 as at Year 1 are of great importance to the end user. Battery-life prediction estimates the degradation curve and can be used to make recommendations for optimum system size to min-imize cost and maximize performance, even after years of service

25°C ambienttemperature

80% SOC upper limit

1,000-kW-hr pack

75% capacity by Year 10

30°C ambienttemperature

90% SOC upper limit

900-kW-hr pack

40% capacity by Year 10

Capacity Prediction Results

Capacity Prediction Results

35

Bat

tery

Tem

pera

ture

(°C

)B

atte

ry T

empe

ratu

re (

°C)

Cap

acity

Los

s (%

)C

apac

ity L

oss

(%)

Bat

tery

SO

C (

%)

Bat

tery

SO

C (

%)

30

25

100

50

0

100

50

0

40 100

80

60

40

20

0

38

36

34

32

300

0 500 1,000 1,500 2,000Time (days)

2,500 3,000 3,500 4,000

0.5 1 1.5 2.52 3

0 500 1,000 1,500 2,000Time (days)

Time (days)

2,500 3,000 3,500 4,000

0 0.5 1 1.5 2 2.5 3Time (days)

–100

Total lossCalendar lossCycling loss

Total lossCalendar lossCycling loss

0

100

Fig. 3—Prediction algorithms are the combined effect of temperature and average state of charge (SOC) on calendar and cycling factors, which drive capacity loss.


life. The tool also helps inform techno-economic analysis that may include tradeoff analysis, such as

• Are there field conditions that may threaten the life or perfor-mance of the battery?

• Is it better to use a cooling system or oversize the pack?• Is shelf life a consideration for a battery that may see periods

of dormancy?

Safety Considerations for Battery SystemsAdvanced batteries are designed and engineered for high energy density and long life cycles. However, a battery is a forced close-proximity encasement of materials, which, when permitted to com-bine without constraint, exhibit exothermic recombination. The very electrochemical features that provide utility are the same features that limit its lifetime, and in rare cases, in a catastrophic fashion. The hazard is the same for fuel because high temperature, fire, spark, crush, or puncture may lead to fuel fires. One must take the same precautions to avoid such hazards with battery systems. There are myriad safety-testing standards to examine the volatility or stability of battery cells and modules, such as UN 38.3 (United Nations 2011), IEC 62281:2012(E), NAVSEA S9310-AQ-SAF-010 (2010), UL 2580 (2013), and UL 1642 (2012), though most of these standards identify similar testing methods such as nail penetration, external shot, shock, vibration, and crush conditions as pass/fail criteria. The fail criterion is if the battery fails catastrophically, with flame and deflagration.

Ideally, battery failure is slow, noncatastrophic, and benign. Bat-tery-performance degradation is analogous to other material-deg-radation mechanisms, such as metal fatigue. Standards outlining the methodology for assessing metal-fatigue life have been devel-oped with great consistency and agreement among engineering disciplines to help accelerate materials selection and design pro-cesses for pipelines and other structures. Modern battery systems are moving toward this kind of standardization, but the knowledge is not distributed widely because battery developers hold this in-formation as trade secrets. Currently, battery-system integrators have developed their own testing programs and battery-manage-ment protocols to find maximum-allowable operational parame-ters to minimize failure risk while maximizing the utility of their product. In most cases, their safety margins involve setting usage limits below the maximum recommended voltage and above the minimum recommended voltage, with additional pack oversizing to reduce the cell C-rate. This conservative overdesign increases cost and weight. An example of this practice is shown in Fig. 1.

While standards committees are increasing activity around bat-tery systems (such as the myriad Society of Automotive Engineers, Institute of Electrical and Electronics Engineers, and International Electrotechnical Commission committees on battery-management-system design, charging devices, and other circuitry features), there does not yet exist a standard lifetime database for batteries for the public sector that is analogous to fatigue curves or strength data available for materials. In addition, the growing, but still maturing industry is still undergoing turnover, and therefore, some manufac-turers and chemistries that were highly regarded 2 to 3 years ago are no longer supported. Therefore, there still exists a research par-adigm in battery management that is using multiple monitoring and control methods to monitor and estimate the state of health of the battery system.

Known Failure MechanismsIn practicality, the electrochemical elegance of the latest battery chemistry becomes less relevant as the inelegance of actual op-erating conditions imposes unexpected consequences on battery health. Battery manufacturers assert that a cell will not experi-ence off-gas release unless it is already on the verge of failure. This assertion likely comes from the expectations of a perfect bat-tery. There are no perfect batteries, and battery cells (as an inte-grated system in materials packaging and assembly) are capable of leaking during operation without triggering alarms in conventional voltage- or temperature-monitoring systems. Battery failures are not always immediate and sudden. As recent events would indicate, the failure is slowly impending and began with prior conditions that were beyond the sensing and control of the battery-manage-ment system (BMS)—sometimes deliberately and sometimes ac-cidentally. A few safety incidents are summarized here to establish the hazards that battery systems may face in a real-world applica-tion. In 2012, a hybrid tugboat experienced a fire because of BMS failure and overcharging (Tyler 2012). In 2013, an airliner experi-enced a short within the battery pack that caused a fire and smoke (NTSB 2013). In 2011, a hybrid vehicle experienced a latent fire because of shorting of the battery pack with liquid coolant (Smith 2012), and in 2012, a pack caused an explosion as a result of bat-tery electrolyte off-gassing at elevated temperature (Blanco 2012). Finally, an electric vehicle in 2013 experienced a fire because of an impact (Shepardson 2014). Hence, a brief overview of the failure modes are BMS failure, external short, high temperature, and ex-ternal impact. In nearly all cases, the stimulating event was ex-ternal. In consideration of the external factors that are known to cause failures, testing against standards such as UN 38.3 (United Nations 2011) is often considered sufficient because it captures shock, external short, and puncture consequences of the system. However, the combined factors resulting from increased-temper-ature operation and/or the subtle off-gassing modes are not cov-ered by this standard. In all the preceding scenarios, the battery system responded unpredictably to unforeseen stimuli (i.e., an over-charge error, an arc or electrical short, mechanical impact, or high-temperature use). However, in real-world scenarios, these incidents are not always predictable or preventable. Not unlike the hazards associated with liquid fuels, battery hazards must be managed in-telligently with training, procedures, and recommended practices.

Categorizing Safety RisksThree levels of deflagration-like hazards are described in the fol-lowing subsections, each of increasing risk such that separate but integrated safety systems may need to be considered to manage each stage of potential risk. These stages are shown in Fig. 4. The stages shown in the figure are time correlated according to the stages through which a battery moves as it approaches thermal runaway.

Stage 1: Off-Gas and Ventilation. In most cases, any abuse event to the battery will first result in a compromise to its packaging, which then leads to the first hazard, off-gas (Kumai et al. 1999). An ongoing project with the US Department of Energy’s Advanced Research Projects Agency—Energy (ARPA-E) shows under mul-tiple circumstances that off-gassing is a repeatable precursor to catastrophic battery failure. In some cases, off-gas detection be-fore failure precludes action to arrest the failure before it occurs (Energy Storage News 2012; Hill et al. 2014b). Battery off-gases are generated from the battery electrolyte solvents, and for most lithium-ion chemistries, these solvents are in the ethylene car-bonate family. The boiling point is near 120°C, but the vapor pres-sure is low such that the odor is evident at room temperature. The electrolyte solvents tend to condense to a gel-like liquid on all sur-faces once evaporated, which leaves an enduring odor and evidence of a battery leak. When the gases are at elevated temperature, the vapor pressure increases, and, in contained spaces, it is possible for these gases to build in concentration and reach the flammability

Hazard Management

Stage 2

Stage 1

Stage 3

Extinguishing

Increasing Risk

Ventilation

ThermalManagement

Fig. 4—Stages of battery hazard.


heat or damage to nearby cells, there is risk that additional batteries will undergo thermal runaway such that the entire module will be consumed. At this stage, fire is likely, and most batteries undergoing thermal runaway do so in a deflagration event that not only consumes the battery electrochemical materials, but its packaging, wiring, and any other nearby flammable materials. The smoke and fumes emitted from the fire present two hazards: The first is uncombusted off-gas, which poses an explosion risk. The second is toxic fumes similar to those of a polyvinyl chloride fire. For both hazards, ventilation (from Stage 1) is paramount, and the BMS may isolate the failing cell from the rest of the module, as it is designed to do. If this does not occur, the second immediate need is to quench the fire. Unfortunately, bat-tery fires are unique in that the consumption of the battery cathode results in direct oxygen generation. This generation of oxygen im-plies that fire-extinguishing methods designed to suffocate fires will be ineffective. Therefore, most fire-fighting techniques have less rel-evance to a battery fire, and the fire-management technique will be designed to manage the fire until it burns itself out.

A note on explosion risk: Upon failing catastrophically, the battery cell itself does not usually release its energy explosively, but it may in-stead experience a deflagration failure mode when a short occurs be-tween the anode and cathode (the most-likely thermal-runaway cause). This failure mode is more common with pouch cells because the seal around their perimeter allows for pressure release. In some cases, ex-plosion risk is possible. This explosion may be a result of ignited off-gases, or it may be a sudden chemical energy release. Few incidents of this nature are reported. The risk of such an event is more likely when the battery is overcharged beyond its maximum recommended voltage. Studies show that mild overcharging (by ≈0.1 V over the max-imum recommended voltage) can degrade capacity quickly without a catastrophic failure (Hill et al. 2013). Deliberate overcharge tests will involve charging to 150% or more of the upper voltage limit, at which point either the mechanical deformation (swelling) of the cell or the self-generated heat cause a separator failure or packaging breach, which leads to a deflagration failure mode.

Stage 3: Total Fire. With this knowledge, it is understood that the fire will likely continue to burn until all battery materials are con-sumed, and so this risk must be accepted and the fire-management goal shall be to manage the heat and prevent these thermal events from causing peripheral damage. The last stage is a total fire, which poses the risk of consuming the entire battery system. This risk may come from an external fire hazard, large electrical disturbances or shorts, or physical impact. This is a high-risk category and represents the most-extreme courses of action in which all common fire-mit-igation techniques may pose additional risks. In this case, the path of least risk should be considered, and a priority should be placed on heat management. In Fig. 6, the typically recommended fire- extinguishing techniques for each class of fire are shown (US code). Class A fires are solid materials, Class B fires involve flammable liq-uids and gases, Class C fires involve electrical elements, and Class

limit, which is at or near 3.6% by volume (similar to ethylene gas). Testing shows (Fig. 5) that early off-gas release is detectable and in some cases reversible, and that batteries may function despite a breached package. The gases released are carbonate solvents for the electrolyte. This hazard poses a danger because the building off-gas hazard may not give prior warning before ignition. There-fore, adequate ventilation and air turnover in these enclosed spaces are the relevant courses of action to mitigate the first risk stage. As cited previously, if off-gases reach the flammability limit in the en-closed area, an explosion risk may be present.

The appropriate ventilation parameters for enclosed spaces will need to account for the following:

1. Volume of the enclosed area2. Flammability limit of the entrapped gases3. Potential toxicity limits4. Spark hazard

The turnover rate (if any) or air changes per hour will depend on the flammability limit, expected release rate, and toxicity hazards. In the case of the ethylene carbonate off-gases, the air changes per hour would be based on the maximum estimated release rate (L/min), the volume of the enclosed space (L), and the volume rating of the venti-lation system (L/min). The turnover rate would need to ensure that the flammability limit does not reach 3.6% by volume. The toxicity limits for ethylene carbonate vapor, in this case, are higher than the flam-mability limit, so the air-turnover rate would mitigate both hazards.

Stage 2: Thermal Runaway. The second stage of risk occurs if one or more cells within a battery module reach the irreversible stages of thermal runaway. In this case, the battery is experiencing exothermic and destructive reactions, which will inevitably lead to fire caused by an internal short between anode and cathode. Typically, the bat-tery will begin this self-consuming process at temperatures greater than 100 to 120°C. Most battery-management systems (BMSs) are designed to isolate single cells that are undergoing such transforma-tions to isolate them from the rest of the module or pack. Monitoring is accomplished by means of temperature, voltage and/or current, and off-gas. In the unlikely event that the single cell is able to transmit

GenerallyAccepted

Class orFunction

ThermalManagement

A

B

C

D

Water Foam CO2 CFC DryChemical

PotentiallyUnsafe

Fig. 6—Recommendation of fire-extinghishing media for each class of fire (CFC = chlorofluorocarbon.)

Battery cell temperature (°C)250

200

150

100

Cel

l Tem

pera

ture

(°C

)

Sen

sor

Res

pons

e (%

)

50

0100

10000

1000

100

10

1

0.1100 120 140 160 180

Time (minutes)200 220 240

120 140 160 180Time (minutes)

MEC (ppm)

ME

C C

onte

nt (

ppm

)

200 220

Sensor response leadstemperature spike

240

Sensor response (%)

0102030405060708090100

Fig. 5—This thermally abused cell eventually went into thermal runaway, but off-gased before the thermal-runaway event, providing some early warning. The lower chart is gas composition confirmed by post-test gas-chromatography analysis. (MEC = methyl ethylene carbonate.)


D fires involve metals. Unfortunately, a battery fire may involve all of these materials and classes simultaneously, causing a significant challenge in extinguishing the fire. As mentioned in the preceding, the battery fire may generate its own oxygen and be self-feeding, hence most oxygen-suppression methods may not be effective.

A few conditions of battery failure are shown in the following. Previous tests confirm that a cylindrical cell continues to cycle while off-gas monitoring detects it is indeed leaking. Subtle, non-catastrophic off-gassing occurs as a function of voltage, current, or temperature-abuse vectors (Hill et al. 2014b). This cell did not fail catastrophically, but instead lost capacity gradually over time. Voltage and current measurements would provide no indication of this issue until capacity began to decrease significantly. When a cell is punctured (Fig. 7), heated, crushed, or overcharged, the internal separator between anode and cathode may fail, at which point the battery shorts, the voltage drops instantaneously, and gas-sing occurs immediately, leading to eventual fire. Temperature rises quickly after the puncture, as shown in Fig. 7.

Updated class rules for marine and offshore systems include ten-tative rules for battery power and updated safety recommendations for large-format battery systems, including ventilation and fire management (DNV 2012a). A number of techniques and method-ologies are currently under way to assess the risks of deployment of battery systems.

Bowtie analysis permits the analysis of external hazard factors (Fig. 8). The model visualizes all possible threats that may cause a top event, such as battery failure. Putting barriers in place to prevent such events may increase safety of the system overall. The diagram

illustrates a generic battery-failure model and shows that threats (left side of the diagram) are prevented from leading to the top event by barriers such as active monitoring and proactive controls. An example shown is mechanical damage by the red arrows progressing from the left of the diagram to the right. In this example, there may be moni-toring methods in place that did not react quickly enough to identify and prevent consequences of mechanical damage, and other barriers (such as physical barriers) may fail. If these barriers are breached and the top event occurs, then a possible consequence is thermal runaway. Preventive controls, such as a rapid battery discharge, to minimize the chemical potential energy in the system may be executed. There may also be reactive controls such as fire alarms, automatic module dis-connects, or emergency cooling systems to draw heat from the battery before the thermal-runaway threshold is reached. Either side of the bowtie model may be expanded into multiple threat or consequence layers, depending on the detail of the model.

The bowtie model is the highest level of analysis and comple-ments failure-mode effects and criticality analysis. Technology qualification procedures follow this methodology and will involve stakeholders across the value chain, including technology vendors and end users (Recommended Practice DNV-RP-A203 2011). Mod-eling tools with greater detail may evaluate the design criteria against the goals of the application. These detailed steps are relevant when a technology is unproved in a specific application or when it is a unique application with little prior history or experience on which to rely. Fig. 9 illustrates that bowtie models influence fault-tree models,

5

4

3

2

Vol

tage

(V

)

1

0

–18 10 12 14 16 18 20

Time (minutes)22 24 26 28 30

0

10

20

Sen

sor

Res

pons

e (1

,000

Ω)

30

40

50

60

Fig. 7—It was found that nail puncture and failure of a cell are seen instantly by voltage monitoring, followed closely by off-gas monitoring.

System Understanding

BowtieAnalysis

Fault-Tree orInfluenceDiagrams

ProbabilisticModels Risk

Recommendand Reassess

Detailed Understanding Action

Fig. 9—Model-based risk-assessment process for estimating safety risk in battery systems.

TemperatureExcursion

Threats Consequences

Mon

itorin

g

Pro

activ

eM

easu

res

Pre

vent

ive

Con

trol

s

Rea

ctiv

e C

ontr

ols

C-RateExcursions

MechanicalDamage

Overvoltage

Undervoltage

Off-gasFlammability

TopEvent:Loss ofBatteryControl

Loss

ReducedOperational Life

SystemFunctionality

Failure

ThermalRunaway

Fig. 8—Bowtie analysis permits the visualization of threats to a top event, such as loss of battery control, and ties these threats to consequences.


which influence probabilistic models, which lead to sensitivity fac-tors and recommendations for design revision or additional controls.

Unique Considerations for Offshore EnvironmentsSubsea batteries will encounter unique environmental conditions that onshore battery systems are inadequately designed to tolerate. Cold temperatures and high pressure, corrosion risk, humidity, and physical damage (crush, puncture, collision, and shock) are all hazards that have higher probabilities in a subsea environ-ment. Longevity and calendar life may be relevant if the system is expected to last the life of the well. Distributed-energy storage-system-design lifetimes are greater than 5 years in a mild environ-ment (CES 2009) compared with the extreme subsea environment in which an oil well may have an expected life of 20 to 30 years. Some systems, if designed to actuate valves or chokes, may have long periods of dormancy interrupted by intense periods of cy-cling. These unique duty cycles will impact cycle fade and may lead to less-predictable capacity loss over time. The effect of these conditions on battery life should be evaluated, especially for crit-ical operations.

Topside systems will also need to endure upgrades for protec-tion against the harsh environments encountered in the sea envi-ronment. Ingress protection, fire safety, and fire training may all be required to adopt large-format battery systems in topside or subsea service. Batteries do not have the same energy density of fuels, yet they pose a fire risk. Under the worst cases, the batteries cannot be extinguished by conventional means because the consumption of the cathode materials feeds oxygen into the fire, thereby making oxygen suppression difficult or impossible (Fig. 6). Therefore, there are different requirements for extinguishing, depending on the stage of the fire. In addition, there may be a greater need to re-move heat from the fire to prevent possible cascading of the fire to neighboring modules or packs. For this reason, a battery fire does not fall within the typical class requirement; there may be addi-tional site-specific factors that would make heat management more relevant than fire extinguishing in case of fire.

Conflicts in class rules exist for offshore vessels, which place technical challenges in battery-system safety management. For ex-ample, if a fire triggers an alarm on a ship, one reactive procedure may be to shut down the ventilation systems to prevent an external oxygen source to fuel the fire. However, if a battery system is off-gassing, there will be a need to vent the area and remove the pos-sibility of trapped and uncombusted flammable gases. There will be overlapping needs to design safety systems to work in tandem without interfering with one another.

ConclusionsThe following conclusions arise from the preceding discussion:

• Cost savings and emissions reductions are relevant to up-coming emissions regulations in controlled areas.

• Given the use cases cited, near-term battery hybrid systems will deploy lithium-ion batteries with different power-density, energy-density, lifetime, and cost characteristics.

• Some battery types will be more suitable than others, de-pending on environmental conditions and engineering re-quirements.

• Safety considerations for batteries will depend on whether they are used in enclosed spaces, the potential hazards that may exist for off-gassing, thermal abuse, mechanical damage, or other factors.

• The intended environment for the application will influence whether the battery technology is qualified for use.

• Testing for such systems can be performed on a cell or module level, and the results are scaled to system-level deployments.

• Modeling tools are used to reduce the time and cost of qual-ification, and testing is minimized to the most-critical tech-nical factors.

Recommendations.• Qualification of battery systems should consider whether the

use is topside or subsea.• Safety considerations should involve ventilation of off-gases,

protection from external forces, temperature control, and bat-tery-management-system backup or redundancy.

• The battery-system cost/benefit analysis should involve both operational and capital considerations, and this will vary be-tween topside and subsea systems.

• Qualification will require review against class rules, but also comparison to known International Electrotechnical Commis-sion, Institute of Electrical and Electronics Engineers, and Underwriters Laboratories standards that may be cross refer-enced by the class rules.

• Beyond UN 38.3 (United Nations 2011), there may need to be testing to examine temperature-induced failure modes and off-gassing behavior to better understand ventilation require-ments.

ReferencesBlanco, S. 2012. GM Battery Lab Explosion Cost Could Reach $5m. Au-

toBlog (18 April 2012), http://www.autoblog.com/2012/04/18/gm-bat-tery-lab-explosion-cost-could-reach-5m/ (accessed 3 February 2015).

Brown, L. M., Hanafia, A., and Petsonk, A. 2009. The EU Emissions Trading System: Results and Lessons Learned. Environmental De-fense Fund, Washington, D.C. http://www.edf.org/sites/default/files/EU_ETS_Lessons_Learned_Report_EDF.pdf.

Caterpillar. 2002. Selective Catalytic Reduction. Caterpillar. http://www.al torfer.com/power/powersystems/GENSET/PDF/LEHE2611.PDF (ac-cessed 3 February 2015).

CES. 2009. Functional Specification for Community Energy Storage (CES) Unit Revision 2.2. AEP Dolan Technology Center. http://www.dolan techcenter.com/focus/distributedenergy/docs/cesunitspecifications_rev2_2.pdf (accessed 3 February 2015).

Chen, J. W., Lindtjørn, J. O., and Wendt, F. 2012. Hybrid Marine Electric Propulsion System with Super-Capacitors Energy Storage. ABB, Inc. http://www05.abb.com/global/scot/scot293.nsf/veritydisplay/b48197614a6b304cc1257a8a003c3c43/$file/ABB%20Generations_23%20Hybrid%20marine%20electric%20propulsion.pdf (accessed 3 Feb-ruary 2015).

De Beaupuy, F. 2013. Nexans Mulls U.S. Subsea Cable Plant for $3.9 Billion Market. Bloomberg (16 July 2013, Business): http://www.bloomberg.com/news/articles/2013-07-16/nexans-mulls-u-s-subma rine-cable-plant-for-3-9-billion-market.

DNV. 2012a. Rules of Classification of Ships/High Speed, Light Craft and Naval Surface Craft. DNV. https://exchange.dnv.com/publishing/ruleshslc/2012-01/ts628.pdf (accessed 3 February 2015).

DNV. 2012b. Shipping 2020. Report, Det Norske Veritas, Oslo, Norway (November 2012). http://www.dnv.nl/binaries/shipping%202020%20-%20final%20report_tcm141-530559.pdf

DNV. 2013. Cell Level Risk Based Evaluation of Batteries for Maritime, Energy Storage, and Other Applications: Rev. 3.1. DNV Bærum, Norway (November 2013). http://www.dnv.com/binaries/battery%20performance%20rp%20rev%203%20november%202013_tcm4-584726.pdf.

DNV. 2014. DNV GL Guideline for Large Maritime Battery Systems: Joint Project between ZEM, Grenland Energy and DNV-GL. DNV, Bærum, Norway http://www.zemenergy.com/content/download/1181/6231/file/DNV_GL_Guideline_large_mar_battery_systems_tcm4-597233.pdf.

Energy Storage News. 2012. ARPA-E Funds DNV KEMA Battery Limits Testing. Energy Storage News (30 September 2012).

EPA. 2012. US Emission Control Areas: Briefing for the Mobile Source Technical Review Subcommittee. US Environmental Protection Agency, Washington, D.C. (13 December 2012).

Gully, B.H. 2012. Hybrid Powertrain Performance Analysis for Naval and Commercial Ocean-Going Vessels. PhD dissertation, The University of Texas at Austin, Austin, Texas, USA (August 2012).

June 2015 • Oil and Gas Facilities 7710 Oil and Gas Facilities • June 2015 June 2015 • Oil and Gas Facilities PB

Hill, D., Gully, B., Agarwal, A. et al. 2013. Detection of Off Gassing from Li-Ion Batteries. Proc., Energytech, 2013 IEEE, 21–23 May, 1–7. http://dx.doi.org/10.1109/EnergyTech.2013.6645307.

Hill, D. M., Agarwal, A., and Gully, B. 2014a. Li-Ion Batteries for Energy Storage and Hybrid Power: Safety and Suitability Considerations for Offshore Applications. Presented at the Offshore Technology Conference, Houston, 5–8 May. OTC-25129-MS. http://dx.doi.org/10.4043/25129-MS.

Hill, D., Agarwal, A., Gully, B. et al. 2014b. Sensor Enhanced and Model Validated Life Extension of Lithium Batteries for Energy Storage. ARPA-E AMPED Program, Fifth Quarterly Report, January 2014.

IEC 62281:2012(E): Safety of primary and secondary lithium cells and batteries during transport. 2012. Geneva, Switzerland: International Electrotechnical Commission (IEC).

Johnsen, T. 2013. The Norwegian NOx Fund – How Does It Work and Re-sults So Far. DLFE-1547, NOx-fond, Helsinki http://www.ndptl.org/c/document_library/get_file?folderId=19620&name=DLFE-1547.pdf.

Kumai, K., Miyashiro, H., Kobayashi, Y. et al. 1999. Gas Generation Mech-anism Due to Electrolyte Decomposition in Commercial Lithium-Ion Cell. J Power Sources 81–82 (September 1999): 715–719. http://dx.doi.org/10.1016/S0378-7753(98)00234-1.

LaMonica, M. 2012. Are Air Cooled Batteries Hurting Nissan Leaf Range? MIT Technology Review (19 September 2012). http://www.technolo gyreview.com/view/429282/are-air-cooled-batteries-hurting-nissan-leaf-range/ (accessed 3 February 2015).

Maritime Journal. 2010. Diesel Hybrid Drive Available for Smaller Boats. Maritime Journal (3 October 2010). http://www.maritimejournal.com/news101/power-and-propulsion/diesel-hybrid-drive-available-for-smaller-boats (accessed 3 February 2015).

NAVSEA S9310-AQ-SAF-010: Technical Manual for Navy Lithium Battery Safety Program Responsibilities and Procedures. 2010. Washington, D.C.: Naval Sea Systems Command (NAVSEA).

Nichols, D. 2005. AEP Sodium-Sulfur (NaS) Battery Demonstration: Final Report. 1012049, Electric Power Research Institute, Palo Alto, Cal-ifornia, USA (14 June 2005). http://www.epri.com/abstracts/Pages/ProductAbstract.aspx?ProductId=000000000001012049.

NTSB. 2013. Interim Factual Report. NTSB Case Number: DCA13IA037, National Transportation Safety Board: Office of Aviation Safety, Wash-ington, D.C. (7 March 2013). http://www.ntsb.gov/investigations/Ac cidentReports/Reports/DCA13IA037-interim-factual-report.pdf.

Ovruum, E. 2013. DNV and Grieg Star Take Batteries on Board. DNV GL (10 April 2013). http://www.dnv.com/press_area/press_releases/2013/dnv_and_grieg_star_take_batteries_on_board.asp (accessed 3 Feb-ruary 2015).

Recommended Practice DNV-RP-A203: Qualification of New Technology. 2011. Oslo, Norway: Det Norske Veritas (DNV).

SAFT. 2013. Clean Battery Power for Cameron’s Electric Subsea Chokes. SAFT Groupe S.A. (February 2013). http://www.saftbatteries.com/force_download/21852_2_0213_CS_Cameron_Protected.pdf (ac-cessed 3 February 2015).

Sandvik, E.T. 2014. Procurement of Environment- and Energy Efficient Ferries. Norwegian Public Roads Administration (21 May 2014). http://www.testsitesweden.com/sites/default/files/content/PDF/elec trified_public_transport_may_22_edvard_sandvik.pdf (accessed 3 February 2015).

Shepardson, D. 2014. Tesla Recalls 29k Vehicle Chargers That Could Spark Fire. The Detroit News (14 January 2014).

Smith, B. 2012. Chevrolet Volt Battery Incident Overview Report. DOT HS 811 573, NHTSA: Office of Vehicle Safety Compliance, Washington, D.C. (January 2012). www.nhtsa.gov/staticfiles/nvs/pdf/Final_Re ports.pdf.

Tyler, D. A. 2012. Battery-Related Fire Damages Famed Hybrid Tug, Puts It out of Service. Professional Mariner (8 November 2012). http://www.professionalmariner.com/December-January-2013/Battery-related-fire-damages-famed-hybrid-tug-puts-it-out-of-service/ (ac-cessed 3 February 2013).

UL 1642: Scope for Lithium Batteries. 2012. Northbrook, Illinois: Under-writers Laboratories (UL).

UL 2580: Batteries for Use in Electric Vehicles. 2013. Northbrook, Illinois: Underwriters Laboratories (UL).

United Nations. 2011. Section 38.3: Lithium metal and lithium ion bat-teries. In Recommendations on the Transport of Dangerous Goods: Manual of Tests and Criteria, fifth revised edition, Amendment 1. United Nations, New York and Geneva.

Vartdal, B. J. and Chryssakis, C. 2011. Potential Benefits of Hybrid Pow-ertrain Systems for Various Ship Types. Presented at the RHEVE 2011: International Scientific Conference on Hybrid and Electric Ve-hicles, Rueil-Malmaison, France, 6–7 December.

Wolf, K., Farmer, J., Karpinski, A. et al. 2010. Techinical Memorandum: Conclusions and Recommendations Regarding the Deep Sea Hy-brid Power Systems Initial Study. TM-2413-OCN, Naval Facil-ities Engineering Command: Engineering Service Center, Port Hueneme, California, USA (June 2010). www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA522631.

Davion Hill is group leader and principal engineer for DNV GL’s Energy and Materials group. Before his current position, he led DNV’s internal research programs on battery testing and new battery materials. Hill has also worked for CC Technologies, Honda Research Institute, and Aeroflex Lintek Corporation. His areas of expertise include energy storage, batteries, and renewable energy, and he has authored or coauthored more than 30 publications in these areas. Hill holds a PhD degree in physics from Ohio State University. He is the 2015 president of NAATBatt International.

Arun S. Agarwal is a senior research engineer in DNV GL’s Strategic Research and Innovation department. His research focus areas in-clude development of technologies to convert carbon dioxide to useful products and strategies for safe usage of batteries for diverse applica-tions. Agarwal has 15 years of experience working on multidisciplinary problems in electrochemistry, chemical engineering, and materials sci-ence. He has coauthored more than 20 publications in the fields of carbon dioxide recycling, fuel cells, corrosion modeling, batteries, and semiconductors. Agarwal holds a PhD degree in chemical engineering from Case Western Reserve University.

Ben Gully is a project manager and engineer in DNV GL’s Research and Innovation department. His background is in power-system modeling and controls. Particular areas of expertise include control-algorithm de-velopment, naval and maritime powertrain analysis, and energy-storage-system integration and optimization. At DNV GL, Gully has focused on experimental and analytical evaluation of lithium-ion battery systems, particularly load-profile analysis and the correlation to degradation be-havior. He holds a PhD degree in mechanical engineering from the Uni-versity of Texas at Austin.

Documents

A Review of Engineering and Safety Considerations for ...€¦ · 68 Oil nd as acilities • June 2015 PB Oil and Gas Facilities • June 2015 June 2015 • Oil and Gas Facilities