REPRINTED FROM JUL/AUG 2015 LNGINDUSTRY

The rising cost of LNG facility construction and the future volatility of natural gas prices are motivating stakeholders to pay even more attention to optimising value while mitigating investment risks.

Concerns dictate a second or even third look at optimising lifecycle risk and performance – steps beyond lifecycle cost. This article will consider the role of the insulation system in these efforts, since it has a material impact not only on lifecycle cost (material/labour capital cost, as well as long-term operation/maintenance costs), but also on lifecycle performance and risk.

Ted Berglund

and Joe Hughes, Dyplast Products, USA,

examine the role of the insulation system in optimising value in LNG facilities.

Mission critical

critical

LNGINDUSTRY REPRINTED FROM JUL/AUG 2015

Some key risks include: � Engineering, procurement and construction (EPC)

schedule delays.

� Operational energy/process inefficiencies.

� Outage/curtailments.

� Poor system resilience (e.g. from incidents, extreme weather events).

� Low margins of error.

� Inflexibility (e.g. in process modifications, more cycling and expansion).

While the CAPEX of the insulation system is relatively small compared to the rest of the facility, it is substantive; and the additional impacts an insulant can have on the aforementioned parameters make the insulation system ‘mission critical’. For example, credible yet subjective analysis indicates that approximately half of delays to LNG facility completion dates are attributable to construction schedules that do not allocate sufficient time for a quality installation of the insulation system. Moreover, on average, the cost of an insulation failure results in ~US$100 000 for each occurrence, and sometimes more impact on operating costs (curtailment of production, and repair material/labour).

Insulant selection

Basic impact on the insulation systemAll LNG facilities need insulation on piping, tanks, and equipment. A typical large LNG facility may have 30 000 − 50 000 or more linear ft of piping with diameters ranging from 1 − 24 in., sometimes larger. Consider a sample of 100 linear ft of 20 in. dia. LNG pipe. Depending on the insulant selected, the following parameters apply:

� The insulant can range from <5000 to >8000 board ft (a 60% differential).

� Capital cost of the insulants (in required thickness, and ignoring amounts of vapour barriers, jackets, etc.) can vary by >200%.

� Installed cost of the insulation system can vary by ~300%.

� Weight of the insulants can vary by 600%. With the heavier insulants, this is approximately 70% of the weight per ft of a 20 in. pipe (0.365 in. wall), resulting in increased structural support cost.

NoteThere is virtually no correlation between cost per board ft and thermal resistance, and in some cases, an inverse correlation exists. Similarly, a higher weight insulant may or may not have better thermal resistance, yet can be assumed to add to the cost of the installation.

Stakeholders are increasingly insisting on objective answers as to why more money is often paid for poorer thermal performance simply because an insulant ‘is specified.’

Due diligenceWhen insulation design engineers select an LNG insulant on behalf of their client, their selection fundamentally dictates the design/cost/risk/labour profile associated with the entire insulation system. Each insulant requires different approaches to vapour barriers, expansion joints, accommodations for the overall weight of the system with stress on pipe hangers, as well as installation procedures, long-term maintenance, etc. Additionally, there are new approaches to LNG system design, including modularisation and pre-insulation of pipe before shipment. Some designers have not yet considered the impacts on the insulation systems necessary to support new LNG design approaches. Examples include shear key systems for module interconnection, pre-insulation of modules prior to shipping to the site, and factory-applied vapour retarders when contractor installation guides only address field-applied.

Insulation system designers generally start with two initial choices: select the insulant that was previously qualified by the organisation or consider requalification of the old and/or qualification of newcomers in light of what is known today.

The periodic requalification is important due to a number of factors, including changes in the following:

� Specification standards (e.g. ASTM, CINI).

� Chemical formulation (such as the chemical’s molecular components, or changes in additives such as fire retardants, catalysts, etc.).

� Manufacturing methodologies.

� Manufacturing locations.Figure 2. Dyplast polyiso with vapour stop at Elba Island.

Figure 1. Elba Island LNG facility.

REPRINTED FROM JUL/AUG 2015 LNGINDUSTRY

In the latter case, for instance, an insulant manufacturing facility may relocate the insulation manufacturing to the LNG equipment assembly site in order to be more responsive/cost effective. However, most product certification entities, including Factory Mutual (FM) and Underwriters Laboratories (UL), require physical properties from each location of manufacture to be tested in order to remain certified. This ensures the product delivered is the same as the product specified. In all cases, due diligence is paramount. The process of due diligence and selection of insulants is multifaceted; this article will examine a few of the components of due diligence that sometimes escape decision makers/influencers.

Lifecycle performance/riskIt is not possible to fully discuss lifecycle performance/risk in detail in this article. However, the insulant selection stage is a critical step in the process. Being better informed means mitigating risks and potentially improving performance. To be better informed, the following must be understood:

� What information is pertinent?

� What information is factual?

� Is there full disclosure of all pertinent facts?

In other words, ‘better informed’ does not simply mean more information.

What information is pertinent?When considering what is pertinent there is the ‘must have’ of thermal conductivity (K-factor or lambda) − the lower the better. At the top of the list are:

� The temperature at which K-factor is measured.

� Initial vs aged (clearly defined by ASTM and other standards). For instance, ‘cured’ or ‘fresh’ have no meaning within any credible standard.

� Water vapour transmission (WVT) and water absorption (WA).

� Other, depending on circumstance.

Temperature vs K-factorThe testing standards/protocols appear to be increasingly focused on K-factors across a variety of temperatures, whereas the traditional approach had been to compare insulants at approximately 70 − 75°F. The newest ASTM C591 standard (the overall standard to which polyisocyanurate [polyiso or PIR] must comply) now requires K-factors to be measured across multiple temperatures from -200°F to +200°F. Taking into account the fact that the K-factor of most insulants improves at lower temperatures, thickness calculation programmes, such as 3E-Plus, now incorporate a range of K-factors vs temperature for each insulant.

Thermal ageingThermal ageing represents the process by which insulants using blowing agents such as hydrocarbons (not air) lose some of their thermal resistance over time, as some blowing agents diffuse out of the cells. ASTM testing protocols involve ageing these insulants for six months in a controlled environment, intending to anticipate the actual thermal conductivity over the life of the insulant. The aged K-factors of some prevalent insulants are

still better than the initial K-factors of insulants that purportedly do not age. Moreover, thermal ageing slows down at lower temperatures, and at cryogenic temperatures ageing can be assumed as nil or at least considerably reduced.

WVT/WAThe other major pertinent property of insulants is WVT and WA. It is important to prevent water vapour reaching the piping. Additionally, any water absorbed into an insulant worsens its K-factor (reducing thermal resistance). If a vapour barrier (meaning zero-permeability) sheet or mastic is applied over the insulation, the WVT and the WA of the system should each be zero. However, in a conservatively designed insulation system (which all LNG installations should be), the designer must consider both errors in application of vapour barriers and/or breach of the barrier, which can be caused by several factors. Thus, a second line of

Figure 3. Dyplast ISO-C1 on the Elba Island LNG insulation systems.

Figure 4. Dyplast polyiso bunstock production.

LNGINDUSTRY REPRINTED FROM JUL/AUG 2015

defense is an advantage, i.e. the insulant itself, exclusive of any vapour barrier, should have low permeability.

OtherBeyond these primary physical properties, others can be important. Higher compressive strength, for example, is important when mechanical abuse is likely, and in circumstances where insulants carry load within pipe hangers. Flame/smoke ratings can be important, however some engineers conclude a metal jacket virtually eliminates flame spread caused by insulation; and in case of a fire at an LNG plant, most engineers agree that smoking insulation is the last thing to worry about. Dimensional stability, weight, and flexural strength also have an impact on insulation system design and must be appropriately integrated by a qualified engineer.

Noting CINI as a dominant standard in many LNG plants, there is also the CINI Cryogenic Thermal Stress Resistance (CTSR) factor.

What information is factual?At the top of the list for consideration by evaluating engineers is third party testing of physical properties. This is followed closely by listed approvals supported by a periodic third party audit by a reputable entity, such as FM, UL, International Code Council (ICC), or other. Too often, manufacturers and suppliers quote physical properties measured within their own laboratories, using sampling and testing protocols that may not comply with industry standards (e.g. ASTM and CINI, etc.). These protocols include the following:

� Sample curing/ageing.

� Parallel vs perpendicular cut samples.

� Sample selection (middle of sample, edge of sample, skins removed, an average, etc.).

� Equipment calibration.

� Date of last test, and whether formulations or regulations have changed since the last test.

Regarding the first three protocols, it is possible to bias a test by testing samples that may not be representative of the product being sold. The chemical formulation of an insulant, for example, can be tweaked specifically for a test sample to yield improved thermal conductivity or strength measurements, while compromising other properties such as flame/smoke. This is why third party audit adds additional assurances. In the absence of third party audit, the engineer/end user should, at a minimum, ask relevant questions and insist on contractual representations from the seller of the insulant’s properties.

With respect to the final protocol, the date of the test is critical. Testing reports must reflect the current products being offered. Credible authorities require re-testing when insulant formulations or methods of manufacturing change. Even minor adjustments to blowing agents, flame retardants, catalysts, etc. will alter physical and fire performance properties. A good example is the recent withdrawal from the market of a common polyol used in polyiso insulants, requiring manufacturers to switch to alternative polyol molecules. Proper selection of the replacement polyol can improve performance; poor due diligence in selection can degrade it. Decision makers should enquire about any changes to inputs or manufacturing procedures that should justify product re-testing.

A final consideration is that ASTM, CINI, and other standards/protocols often make ‘like to like’ comparisons challenging, and requisite codes may not reflect rapidly changing regulatory and technology environments. This can result in either far too lenient standards, or at the other extreme, unachievable standards. A case in point is the variation in WA tests defined by ASTM for different insulants, which vary from 24 − 96 hr. Other tests sometimes differ in their specifying percentage based on weight, while others specify percentage based on volume. The specifying engineer should clearly understand the essence of each testing protocol and should verify the pertinence and the facts. Using the WA example, assuming ≤1% WA is the specification, why is it pertinent without specifying per weight or per volume, and what if a 1% moisture gain in one insulant may materially impact K-factor while it may not in another?

Is there full disclosure of all relevant facts?With respect to full disclosure, insulation suppliers and manufacturers often do not expose key physical properties that could be perceived as negative. As mentioned earlier, this could be the aged K-factor, the parallel strength vs the perpendicular, the dimensional stability (volume vs length) at low temperature, the actual WA or WVT of the insulant itself without the skin or the laminate, flexibility at low temperatures, cost per board ft, delivery times/flexibilities, prior successes and failures (plus root causes), etc. Stakeholders should insist upon full disclosure, or at least disclosure comparable to competing suppliers. Insistence on conformance with ASTM or CINI standards for cryogenic applications and requisite compliance mitigates these issues.

Case study: Elba IslandElba Island, an LNG facility off the coast of South Carolina, US, selected polyisocyanurate manufactured by Dyplast Products as the core of its insulation system. ISO-C1® was selected after multi-value assessments by the turnkey EPC contractor and the client. Key factors in the decision included the following:

� Physical properties audited and validated by an independent laboratory.

� Physical properties of ISO-C1 met or exceeded requirements set by ASTM C591.

� ISO-C1 exhibited the low aged K-factor – this allowed a total insulant thickness of only 7.5 in. on a 20 in. pipe, compared to the 10 in. required if cellular glass was used.

� ISO-C1’s K-factor improved significantly as temperatures dropped.

� The density of ISO-C1 at 2 lb/ft3 made handling and shipping easier than 7.5 lb/ft3 cellular glass. Structural engineers were able to minimise the number of pipe supports, reducing cost.

� Customised bunstock sizing provided efficient shipping logistics and scrap minimisation during fabrication.

� Availability of higher density polyiso (up to 6 lb/ft3) provided the higher compressive strengths to support pipe hanger applications.

� Ability to fabricate blocks to close tolerances allowed for tight seams and joints.

� Flexibility and responsiveness in deliveries and technical advice enabled reduced costs, improved schedules, and enhanced relationships.

� Easy to handle and work in the field, with minimal breakage.

� Quick product turnaround and delivery (e.g. 2 − 3 days).

Dyplast’s scope of work included polyiso insulation for 31 895 linear ft of piping that connects the ship offloading facilities with storage, regasification and transportation facilities. The insulation system consisted of double-layer insulation for piping with outside diameters varying in size from 3.5 − 41.25 in., covered with a combination of vapour barrier sheeting and mastic, enveloped in aluminium colour-coded jacketing. Specifications required stringent adherence to validated physical properties for insulation system components, as well as specific standards for shop fabrication of shaped insulation segments, such as dimensional tolerances for hemicylindrical sections, pipe ells for small elbows, mitered sections for large elbows, and tees.

The ability to customise polyiso bunstock dimensions was an advantage, since bun sizes could be matched to minimise waste as Dyplast cut the bunstock into blocks (‘pipe chunks’), which were, in turn, sized for minimising waste during shape fabrication. Optimally sized pipe chunks also allowed for efficient packing in transportation containers. The company

shipped over 1.25 million board ft of ISO-C1 in 43 semi-trailers to support the construction schedule.

The project schedule could be met only if parallel construction work was effectively executed and coordinated so that schedule savings could be realised at every opportunity. Dyplast’s flexible delivery schedule and ability to execute just-in-time deliveries were advantageous. Communication feedback from the company ensured that installation of the insulation system proceeded efficiently and in sync with other contractors.

ConclusionNatural gas markets remain unpredictable. Markets have increasingly regional dynamics, and LNG construction costs rise. Selection of the optimal insulant supported by its appropriate insulation system becomes ‘mission critical’. Engineers, specifiers, owners, and stakeholders should be increasingly vigilant regarding the facts, the relevance of the facts, and full disclosure as validated by independent parties. Specifications, standards, regulations, and technologies change. Insulant manufacturers either adapt to comply with the latest standards in order to support their client’s objectives, or they choose to be parochial in advancement of short-term interests. Ultimately, it is up to the owner or the investor to ask the tougher questions, or to ensure that their engineers are.