Understanding RAM analysis and typical availability values for LNG plants Page 1 of 4

SIGN UP Gas Processing is Produced by Gulf Publishing Company

Your source for concise technology and business information for the gas processing industry.

HOME NEWS SUPPLEMENT PROJECTS EVENTS

Understanding RAM analysis and typical

availability values for LNG plants
F. J. Malaret, Technip, Paris, France

The overall production availability of an LNG plant is a major criterion for economic evaluation. It is,
therefore, important to have an estimation of the overall availability of an LNG plant, especially at the
early stage of development.

A simple way to perform this estimation that does not require any calculations is to use availability
data from a reference project, or to use typical mean production availability (PA) values for LNG
projects. To perform meaningful comparisons, the reliability, availability and maintainability (RAM)
model and the reliability data used must be fully understood.

PA for an LNG plant is discussed, encompassing the major considerations of RAM analyses in LNG
plants. Typical availability values are given to perform economic evaluations during the early phase of
a project (i.e., conceptual or feasibility studies) when detailed information to perform rigorous
calculations is not available.

Estimating production efficiency. The overall production efficiency of an LNG project encompasses
all of the steps in the LNG production chain (Fig. 1). These steps include upstream units (i.e.,
offshore platform, pipelines and subsea system), the LNG plant itself, the storage and loading
system, the LNG carrier fleet, and sometimes the regasification terminal that will receive the LNG. Market data is delayed and provided by Barchart.

FEATURED COLUMNS
Fig. 1. LNG chain flow diagram.
Boxscore Construction Analysis
-Adrienne Blume
The production loss due to LNG storage and loading facilities is addressed on a high-level basis. The Large quantities of shale gas, tight gas and
availability of LNG storage is generally proportional to the size of the export fleet. Detailed studies can coalbed methane are being discovered and
be performed to determine the optimum fleet size required to distribute free-on-board (FOB) LNG extracted in the vast Western Canadian
product to customers. These studies can also help optimize the LNG tank storage capacity to Sedimentary Basin (WCSB), which covers
minimize production losses due to “tank tops,” which occur when tanks are full and no storage is southwestern Manitoba, southern
available for additional LNG production. These studies must take into account the reliability of the Saskatchewan, Alberta (AB), northeastern
loading arms, the access to the loading berth(s) and the influence of inclement weather. British Columbia (BC) and the southwest
corner of the Northwest Territories.
The estimated availability of an LNG plant can impact profit loss and dictate the size and scope of the Editorial comment
distribution fleet, the upstream facilities and the LNG regasification terminal(s) served. -Adrienne Blume
Falling crude oil prices have impacted profit
Calculating production availability. Generally speaking, the availability is the probability of an item forecasts for oil and gas producers, as well as
(in this case, an LNG plant) to be in a state to perform a required function at a given point of time or for petrochemical makers and gas processors.
over a given time interval. PA is often used when describing the production efficiency of a system,2 as
this metric can be easily used for economic evaluations. Executive Q&A viewpoint
-Jérôme Ferrier
The oil and gas industry has undergone
Here, the PA is defined as the percentage of the predicted achievable production (PAP) over the massive changes during the past three years
potential production (PP), as shown in Eq. 1: and continues to be in flux.
Gas seperation
-Thomas Perschke
The Ichthys LNG project is a JV between
(1) INPEX, Total and the Australian subsidiaries of
Tokyo Gas, Osaka Gas, Chubu Electric Power
The PAP is the total production over the field life as predicted by a model (for example, a RAM and Toho Gas.
analysis), taking into account the production losses due to scheduled downtime required for
maintenance and inspections, and unscheduled downtime as a result of the random failure of
components or equipment.

Unscheduled events are random in nature. Downtime due to equipment failures takes into account
all repair and logistics time (i.e., active repair time; mobilization delays; diagnosis of the problem;
unavailability of spare parts; production losses due to startup, restart and ramp-up to full production,
etc.).

Reliability data1, 5 for unscheduled events is based on statistical data collected on the availability of
similar pieces of equipment. For pieces of equipment dedicated to the LNG industry, such as
cryogenic heat exchangers, LNG turboexpanders, gas turbines used as mechanical drivers for
refrigeration compressors and other equipment, reliability data is generally taken directly from

http://www.gasprocessingnews.com/features/201504/understanding-ram-analysis-and-ty... 4/4/2015
Understanding RAM analysis and typical availability values for LNG plants Page 2 of 4

vendors.

PP2, 3 (Fig. 2) is the production over the field life as determined by the deliverability profile for the
system—i.e., the maximum LNG production, assuming that the LNG plant will run continuously,
without any interruption. For chemical engineers, this profile corresponds to the value indicated on
heat and material balances.

Fig. 2. Potential LNG production. New! US Gas Processing

Plant Directory Hydrocarbon
Processing and Gas
Fig. 3 shows a hypothetical LNG production profile over the life of an LNG plant. Some production Processing have published
loss is seen during the operating years due to unscheduled events, which typically vary only slightly in their first US Gas Processing
years without planned maintenance. Plant Directory, a
comprehensive list of all
natural gas processing plants
in the US. This includes
natural gas processing,
cryogenic and fractionation.
The directory is available in a
fully searchable, electronic
format. The directory provides
detailed information for more
than 500 gas processing
plants. Gain the market
knowledge to grow your
business and inform your
decisions in the booming
Fig. 3. Annual trend in LNG PA over the project life. United States gas market.
Order your Copy
It can also be seen that LNG production is reduced in years where planned maintenance and/or
Today!
inspections take place, since a total LNG plant shutdown is required. The predicted average LNG PA
of the facility is shown as a purple line. Note: The PA shown in the chart is a through-life average,
where the actual value will vary from year to year.

For example, if the maximum LNG production of a facility is 3 MMtpy and the predicted average
annual production from a RAM model is 2.7 MMtpy, then the average PA would be 90%. In other
words, on average, 90% of the time the LNG plant will produce at its maximum capacity of 3 MMtpy,
and 10% of the time it will not produce LNG due to scheduled and unscheduled events. This figure
can also be interpreted on a 365-day basis as an average of 328.5 days producing at maximum
capacity and 36.5 days of no production.

From this point on, the term “availability” will be used to refer to the yearly mean production average
throughout the life of the project.

Understanding RAM analysis of LNG plants. RAM analysis predicts production losses for
estimating the real throughput of an LNG plant. RAM techniques can be applied at any stage of a
project, from engineering design to the operational stage.

The principal objectives of a RAM study are to:

 Quantify the expected PA for the LNG, based on the configuration of the LNG plant and
reliability data
 Evaluate the PA of other products, such as LPG, ethane, propane, butane or condensates
 Provide a detailed analysis of the LNG production loss (by system or equipment) to identify
the key contributors
 Quantify the impact on production of various changes to the system configuration, operating
philosophy and maintenance strategy, as well as the impact on the achieved system
performance, through sensitivity analysis and availability improvement measures
 Provide recommendations on how to improve the PA of the facilities.

Battery limits for RAM studies in LNG plants. In general, the battery limits extend from the
isolation valves upstream of the turret (for FLNG facilities) or upstream of the inlet facilities (for
onshore plants), up to the isolation valves upstream of the LNG storage tanks. However, RAM
analysis sometimes includes the LNG storage, offloading facilities and shipping operations.

The scope of RAM studies for LNG plants generally excludes the following:

 Early life and wear-out period

 Upstream system (offshore platform, pipelines, subsea system, etc.)
 Shipping (LNG carrier) operations
 Utility systems that are noncritical to operations
 LNG storage and offloading facilities
 Meteo-marine conditions impacting mooring, offloading and departure of LNG carriers
 Extreme climatic conditions impacting production
 Support hull operations with no impact on production (only applicable for FLNG)
 Meteo-marine conditions impacting LNG production/hull motion (only applicable for FLNG)
 Catastrophic events.

As catastrophic events are very rare, the expected contribution from such events is not representative
and is normally a small quantity. These events are not normally included in RAM analyses, but rather
addressed in risk analyses. If the catastrophic event occurs, the actual loss would be large, which
could mean a dramatic reduction in PA.

According to the International Organization for Standardization’s ISO 20815, “Some serious,
infrequent events will cause long-term shutdown of production.”4 These so-called catastrophic events
include:

 Earthquakes
 Fires and explosions
 Blowouts
 Sabotage
 Structural collapse

http://www.gasprocessingnews.com/features/201504/understanding-ram-analysis-and-ty... 4/4/2015
Understanding RAM analysis and typical availability values for LNG plants Page 3 of 4

 Major problem with casing or wellheads

 Riser or export pipeline ruptures
 Falling loads with large damage potential
 Other events, or combinations of events, with large damage potential.

Typical availability values of LNG plants. Availability information from several LNG projects has
been collected and analyzed to generate statistical data. The projects selected are considered to be a
good representative sample for LNG plants, as they cover a wide range of situations:

1. Environmental conditions: from mild climates to extreme conditions, such as the Arctic
2. Onshore and offshore projects (with different sea motions)
3. Process schemes, capacities, flaring management, technologies, drivers, sparing
philosophies, etc.
4. Different availability and reliability data
5. Various maintenance requirements and strategies
6. Country regulations, and company safety rules and specifications
7. Crew availability, spare parts management, etc.

The results in this section are expressed with one-decimal figures. The intent is to show the tendency,
rather than an exact value. Mean values are reported with ± and their standard deviation.

So, what is the typical PA for an LNG plant? The typical LNG plant PA, including both scheduled and
unscheduled events, ranges from 89.4%–95.5%, with a mean value of 92.6% ± 2.2%.

Another important question is: What is the typical average production loss due to scheduled events in
an LNG plant? The production loss can be accurately evaluated if the following items are known:

 Interval for maintenance and inspections

 Duration of each total shutdown
 Amounts of time needed to stop and to ramp up production.

Typically, the unavailability of LNG plants is driven by the maintenance pattern of the drivers used for
the refrigeration-cycle compressors, but other factors must be taken into account. The information
needed to perform detailed calculations may not be available during the conceptual phase, so an
average value for losses of 2% can be considered.

The exact values calculated for the typical production losses due to scheduled maintenance ranges
from 1.4%–3%, with a mean value of 2.1% ± 0.6%.

Production loss corresponds to unavailability. Eq. 2 can be used to calculate the associated
availability:

Availability = 100% – Unavailability (2)

Note: The downtime due to scheduled events could be reduced by optimizing maintenance patterns.

A third essential question is: What is the typical average production loss due to unscheduled events
in an LNG plant? The production loss due to unscheduled events is a consequence of the random
failure of equipment. Reliability block diagrams (RBDs) and failure modes and effect analyses
(FMEAs) must be performed to accurately predict the production loss. Nonetheless, the information
needed to perform such detailed calculations for each process unit may not be available during the
conceptual phase.

The calculated value for yearly average production losses due to unscheduled events ranges from
2.5%–9.5%, with a mean value of 5.2% ± 2.6%.

Takeaway. When comparing the PA value for different LNG projects, it is important to understand the
basis of each study, as the results will be significantly impacted by differences in the battery limits,
exclusions, reliability data set and other factors.

Due to the random and statistical nature of the data used to estimate the PA, the results obtained
from a RAM analysis for a given project can vary significantly if other bases for the study are
considered. Each LNG project is unique, and it is difficult to compare one to another from a reliability
point of view. The values provided here are not intended to substitute for detailed RAM studies to
predict the actual output of an LNG plant. Rather, they are intended to offer an understanding of the
order of magnitude of the loss of LNG production due to planned and unplanned events.

Table 1 gives a summary of the typical PA values for LNG plants obtained from RAM studies
performed at the FEED stage, and Fig. 4 shows typical LNG production loss in LNG plants. The
average PA can be used to estimate the design capacity of an LNG plant, as shown in Eq. 3:

(3)

Average PA can be adjusted depending on the context of a new LNG plant. For example, for a project
that will be constructed in a place with extreme climatic conditions, the availability of the LNG plant
will be lower than average; therefore, the average PA can be corrected before evaluating the design
capacity. GP

Note

Any opinions, findings and conclusions or recommendations expressed in this material are those of
the author and do not necessarily reflect the views of Technip.

Acknowledgments

The author thanks Nathalie Millot and Vincent Tirilly, of the process and technology division of the
LNG department at Technip, for their technical assistance.

http://www.gasprocessingnews.com/features/201504/understanding-ram-analysis-and-ty... 4/4/2015
Understanding RAM analysis and typical availability values for LNG plants Page 4 of 4

Literature cited

1Collins, S., et al., “Improving LNG importation reliability,” proceedings of LNG XI, Birmingham, UK,
1995.

2Mokhatab, S., “Plant reliability and availability,” Hydrocarbon Processing, December 2007.

3International Organization for Standardization, NF EN ISO 14224, “Petroleum, petrochemical and

natural gas industries: Collection and exchange of reliability and maintenance data for equipment,”
June 1, 2012.

4International Organization for Standardization, NF EN ISO 20815, “Petroleum, petrochemical and

natural gas industries: Production assurance and reliability management,” June 1, 2010.

5DNV, OREDA, Offshore Reliability Data Handbook, 5th Ed., 2009.

Francisco Malaret is a senior process engineer with more than eight years of
experience in the oil and gas industry, specifically in onshore and offshore LNG
projects. His experience covers conceptual to detailed studies, including basic
engineering; front-end engineering design; and engineering, procurement and
construction. Mr. Malaret earned a BScEng degree in chemical engineering and a BSc
degree in chemistry, graduating cum laude from Universidad Simon Bolivar in
Venezuela. He also holds an MS degree in oil refining from the IFP School in France.

| Contact Us | About Us | Site Map | GulfPub.com | WorldOil.com | HydrocarbonProcessing.com | ConstructionBoxScore.com | Webcasts | White Papers | Events |
Gas Processing Magazine | Newsletter Sign-Up | Advertise | US Directory |

Please read our Term and Conditions, Cookies Policy, and Privacy Policy before using the site. All material subject to strictly enforced copyright laws.
© 2013 Gulf Publishing Company.

http://www.gasprocessingnews.com/features/201504/understanding-ram-analysis-and-ty... 4/4/2015