The concept of temperature seems simple enough. Everyone knows that a thermometer can be used to measure the temperature of a substance such as air and that by knowing the forecast, people can adequately prepare for the day’s outdoor activities. While it might be marginally problematic to be underdressed or overdressed for the weather conditions, there are no serious repercussions if the forecast is wrong. For Arctic oil and gas operations, however, poorly developed predictions can have catastrophic consequences if the equipment does not function reliably when the temperature drops.
While some work has been done in this field, there is much more to be done to actually define the temperature that should be used for design, equipment procurement specification, and operational risk management for Arctic operations.
The most recognized definition/ selection of a design service temperature comes from the International Association of Classification Societies (IACS) Unified Requirement (UR) S6 for structural steel intended for service at lower temperatures. IACS UR S6 and the ABS Low Temperature Environment (LTE) Guide require that the design service temperature (DST) be selected as the lowest mean daily average (LMDAT) for the operational window and geographical location. The LMDAT, which has been in use for many years and is widely recognized in the industry, is easy to calculate using temperature data for the operational area.
Prediction and problems
Inaccurate temperature prediction has the potential to impact many operational elements. If an engineer designing a new support vessel or offshore platform to operate in a specific location has the wrong DST, the designs are wrong. If an equipment manufacturer designing and building equipment for cold environment applications uses the wrong DST, the equipment may not be reliable. And if a regulatory body developing guidance for Arctic operations applies the wrong numbers, the regulatory guidance will not facilitate safe operations.
Before engineering a unit for Arctic operations, a temperature specification has to be established based on measured data from the work site or from a site as close to that location as possible. Then, all of the systems on the unit, including the equipment to be supplied from vendors, have to be designed and manufactured to operate at that temperature.
The problem lies in answering the question, “What is the temperature for the location?” Temperature varies minute to minute, hour by hour, day by day, and even decade by decade. Often, temperature is expressed as an average of all the temperatures ever recorded in a particular area, the historic average temperature for the day on which operations are expected to take place.
To further complicate calculations, materials vary in their sensitivity to temperatures and temperature fluctuation. A sudden drop in temperature for a few hours will have significantly less effect on the hull material (with its thermal bulk inertia) than it will have on an exposed, 100mm diameter freshwater pipe.
While a unit will be required to have a certified design service temperature, there has to be cognizance that every unit is a combination of all its systems. So the question of how those systems will be tested to verify functionality at the design temperature must be answered. And consideration has to be given to whether an additional temperature safety factor should be required for testing and certification.
Research efforts
The DST is applied to the unit’s structural steel; however, a second temperature is required for machinery, namely the minimum anticipated temperature (MAT), which currently can be defined by the owner, operator, shipyard, or designer—or taken as 20°C colder than the DST. There is very little guidance offered to define the MAT with greater accuracy.
Certain systems that are exposed to a sudden, but temporary, low temperature, such as when polar lows move, require a design temperature value that incorporates the probability and duration of such an occurrence. Probability of occurrence is a risk concept that the marine and offshore industry understands well, but duration of occurrence is new.
A simple analogy helps to illustrate the concept. A man living where the outside temperature is cold but who works indoors probably would find it unnecessary to don an expensive, extreme harsh environment jacket if he only needs to run a short distance to a nearby building. On the other hand, for someone who works outdoors all day in that same cold environment, it might make very good sense to invest in the expensive jacket.
This same philosophy can be applied to an installation. There is no need to winterize the entire unit for a temperature that may occur for only a few minutes. Winterization can be applied selectively on the basis of the probability of occurrence and duration of occurrence. Further, it would be possible to delay some operations if there were an understanding that the duration of the cold occurrence would be short and that the delay would not reduce safety levels.
A temperature analysis concept, resulting from collaborative research by ABS Harsh Environment Technology Center and Memorial University, in St. John’s, NL, Canada, is offering a new rational statistical method for defining a minimum anticipated temperature.
Analysis
A temperature data set, spanning 1999 to 2011, for Barrow, Alaska, demonstrates the concept (Fig. 1). The mean daily average temperature line would be used per IACS UR S6 to determine the design service temperature for a unit operating in Barrow. The DST would be selected as the LMDAT, equal to -29.1°C. The blue line, the lowest recorded daily temperatures, would be used to set the minimum anticipated temperature (MAT), here equal to -48.3°C. Note that this data set justifies the practice of setting the MAT 20°C colder than the DST.
Consider a crane on an OSV as an example for the use of the TDF plot (Fig. 2). If the crane was designed for a temperature of -45°C, the unit should not be used at times when the ambient temperature is below -45oC. The statistical temperature data indicate that there is about a 2% probability of having colder than -45oC lasting seven hours and a 1% probability of having it last for 20 hours.
The way forward
A statistical representation of temperature feeds directly into a risk-based approach to winterization. Knowing the probabilities of a temperature occurring and the temperature at which operational degradation begins to occur allows engineers to determine a probability of failure. A risk level can be defined based on the criticality of the equipment under consideration. A design can be modified to achieve an acceptable risk level if the probability of temperature occurrence is known and an acceptable risk level for the system has been defined.
As experience and more data are brought to bear, progress will take place more rapidly. Recognizing that further research is required, ABS is working through the Harsh Environment Technology Center in St. John’s on collaborative efforts that will help move the industry forward. OE
Based in Houston, James Bond is Director, Shared Technology in the American Bureau of Shipping Corporate Technology group. He is responsible for guiding ABS research, ABS rule development and industry guidance. Bond has worked in the marine and offshore industries for more than 25 years.
Dan Oldford, P.Eng, worked as an ABS surveyor in Canada before joining the ABS Harsh Environment Technology Center in St. John’s, NL, where he is involved in R&D efforts targeting Arctic issues, including winterization. He is a graduate of the Ocean and Naval Architectural Engineering program Memorial University in St. John’s.
Caption Image (top): Fig. 1: Temperature data for Barrow, Alaska, from 1999 to 2011.
Caption Image (bottom): Fig. 2: Comparing Figs. 1 and 2, one first notices that the lowest point on the 1% return line from Fig. 2 does not go as low as the record low in Fig. 1, but Fig. 2 is for the month of January and Fig. 1 is for an entire year. The record lows are observed in February. Fig. 1 provides a graphical representation of temperatures recorded in the past, while Fig. 2 takes those temperatures and projects them forward to give a probability of a temperature return and the time over which that temperature can be expected to last.
Subscribe
