Mark Fuglem

Implementation of ISO 19906 for probabilistic assessment of global sea ice loads on offshore structures encountering first-year sea ice

Ice failure during ice-structure interactions is a complex process and the development and improvement of ice load models is challenging, in large part because of difficulties obtaining full-scale data and scaling issues when extrapolating from model test data. The ISO 19906 standard provides guidance for the calculation of design ice loads on offshore structures in Arctic and subarctic regions. This paper examines issues in implementing available formulae for probabilistically determining design ice loads from first-year ice on sloping and vertically-faced offshore structures. Methodologies suggested in ISO 19906 are applied to simulate structure interactions with ice floes and embedded ice ridges in subarctic regions, such as the Northern Caspian Sea and Baltic Sea where no multi-year ice is anticipated to contribute to design loads. In these regions design ice loads are determined by first-year ice, which is less severe than the ice loads that result from multi-year ice features in Arctic regions. When compared multi-year ice features, lower ice strength coefficients are appropriate for both first-year level ice and first-year ridges. Structure interactions with first-year ridges are modeled differently than with multi-year ridges. The keels of first-year ridges are not yet consolidated, so that a different model approach is required for keel failure than for the failure of fully consolidated multi-year ridges. Challenges in defining the required input data for the appropriate ISO 19906 formulae are discussed in this paper. Sensitivity analyses performed using the Sea Ice Loads Software (SILS), a Monte-Carlo type simulator developed by C-CORE for determining first and multi-year sea ice loads using probabilistic methods are also presented and discussed. Suggested ISO 19906 models of potential failure modes are implemented in SILS for first-year level ice and ridge interactions with different structure types. The sensitivity analyses compare design ice loads from ice impacts with different structures and for different ice loading scenarios, illustrating the influence of different model assumptions on calculated design loads.

Risk Analysis and Hazards of Ice Islands

When considering man-made structures for offshore Arctic regions where ice islands may transit, the probability of encountering different sizes of ice islands and ice island fragments needs to be estimated. This can be difficult to determine in near-shore regions where the occurrence of ice islands is infrequent, yet if there is an occurrence and the ice island grounds in shallow water, it could break into a number of smaller ice island fragments. Probabilistic design methods can be used to determine if consideration of ice island impact in structural design criteria is required and to choose appropriate levels of ice strengthening. This chapter includes a brief description of the design issues involved, available approaches and areas where additional information would be useful.

Methodology to Evaluate Sea Ice Loads for Seasonal Operations

Offshore structures designed for operation in regions where sea ice is present will include a sea ice load component in their environmental loading assessment. Typically ice loads of interest are for 10−2, 10−3 or 10−4 annual probability of exceedance (APE) levels, with appropriate factoring to the required safety level.The ISO 19906 standard recommends methods to determine global sea ice loads on vertical structures, where crushing is the predominant failure mode. Fitted coefficients are proposed for both Arctic and Sub-Arctic (e.g. Baltic) conditions. With the extreme ice thickness expected at the site of interest, an annual global sea ice load can be derived deterministically. Although the simplicity of the proposed relation provides quick design load estimates, it lacks accuracy because the only dependencies are structure width, ice thickness and provided coefficients; no consideration is given to site-specific sea ice conditions and the corresponding exposure. Additionally, no term is provided for including ice management in the design load basis.This paper presents a probabilistic methodology to modify the deterministic ISO 19906 relations for determining global and local first-year sea ice loads on vertical structures. The presented methodology is based on the same ice pressure data as presented in ISO 19906, but accounts better for the influence of ice exposure, ice management and site-specific sea ice data. This is especially beneficial for ice load analyses of seasonal operations where exposure to sea ice is limited, and only thinner ice is encountered. Sea ice chart data can provide site-specific model inputs such as ice thickness estimates and partial concentrations, from which corresponding global load exceedance curves are generated. Example scenarios show dependencies of design loads on season length, structural geometry and sea ice conditions. Example results are also provided, showing dependency of design loads on the number of operation days after freeze-up, providing useful information for extending the drilling season of MODUs after freeze-up occurs.Copyright

Estimating Iceberg-Wave Companion Loads Using Probabilistic Methods

Many challenges arise when designing offshore structures for iceberg loads in arctic and subarctic regions. To help the designer, the ISO 19906:2010 standard provides guidance for the calculation of design ice loads using both deterministic and probabilistic approaches. In determining design loads for different environmental factors, both principal and companion actions must be taken into account; an example is iceberg actions and companion wave actions. ISO 19906 allows the designer to calculate the companion wave action as a specified fraction (combination factor) of the extreme level (EL) design wave load. Alternatively, the designer can calculate appropriate companion wave loads explicitly.A methodology has been developed at C-CORE in which representative iceberg actions are determined using a software package, the Iceberg Load Software (ILS). This is a probabilistic tool which uses Monte Carlo simulation to obtain a distribution of global impact forces based on the expected range of iceberg and environmental conditions that a structure would likely encounter. The software provides a reasonably accurate representation of the iceberg loading situation, following the provisions of ISO 19906:2010, without introducing unnecessary conservatism in the design load. In the software, the influence of waves on the iceberg actions are considered, but companion wave loads must be calculated and added externally to the software, The software accounts for the probability of different sea state conditions and the influence of the sea state on the probability and severity of iceberg impact, given the correlations between the sea state, iceberg management effectiveness and iceberg drift and wave-induced velocity.The additional hydrodynamic pressure due to the wave during the period of the impact; is not considered. This wave loading will be complicated by the influence that the presence the iceberg and structure have on the local sea state. In this paper, brief descriptions are provided of background studies on companion wave loading and the application of companion load factors in ISO 19906. The companion load factors allow the designer to apply the design wave load, which is calculated for situations with no iceberg present, to the case of iceberg impacts. In this study, a methodology is presented for determining companion wave loads based on the distribution of sea states expected during an iceberg impact. These sea states are significantly less severe than that associated with the design wave load as iceberg impacts are rare events. The companion wave loads are determined without accounting for the influence of the iceberg; this is thought to be quite conservative. An example application of the methodology is presented for a hypothetical platform located on the Grand Banks, off the east coast of Newfoundland. Iceberg actions, wave actions and combined iceberg-wave actions are estimated using the described methodology. Comparisons are provided for the resulting companion loads and those based on ISO 19906:2010 companion load factors applied to the extreme level wave load.Copyright

Iceberg Drift Forecast Requirements for Offshore Platforms Utilizing Facility Side-Tracking to Avoid Impacts

A number of solutions have been successfully implemented for producing oil and gas off Canada’s east coast, where impacts by icebergs are a possibility. In future, as operators move further offshore to deeper water and further north where the numbers and sizes of icebergs may increase, new solutions for avoiding impacts will be required to limit ice strengthening requirements and ice related downtime costs. A potential solution is the use of facility side-tracking, where a floating system is designed to move laterally to avoid approaching icebergs. This paper discusses the issues involved including the need for improved short-term iceberg drift forecasting.Copyright

Definition for technical success of iceberg towing operations

Offshore platforms proposed for operation in regions with icebergs must be designed to withstand ice loads associated with impacts for cases where icebergs cannot be managed or avoided. Both fixed and floating systems have been successfully used on the Grand Banks off Canadas east coast without incident to date. The Hibernia platform is designed to withstand impact loads; these loads were estimated without consideration of the efficiencies of ice management. The Terra Nova and White Rose FPSOs are designed to disconnect and move off site if threatening icebergs cannot be managed. These systems have ice strengthening, with the benefits of ice management considered in determining the design ice loads. Ice management efficiencies were based in large part on overall towing success rates determined from ice management records. Towing success was based on a number of criteria, including whether the iceberg was obviously deflected from its course and whether downtime resulted. It is expected that in the near future there will be pressures to reduce platform costs in order to make more marginal fields viable, platforms will be installed further north where iceberg and sea ice conditions are more severe and as more platforms come on stream, there will be restrictions regarding where icebergs can be safely towed so as not to increase risks to neighboring installations. As assessment of ice management becomes more critical, it will be desirable to develop more accurate measures of efficiency for assessing both downtime and potential for impacts. In this paper, a number of alternative methods for evaluating ice management efficiency with respect to impact avoidance and the potential influence on ice design loads are explored and criteria for further development of these discussed.

Dealing With Knowledge Uncertainties in Pipeline Reliability Based Design and Assessment

Knowledge uncertainties result from limitations of the data and other information required to define parameters that are used in estimating reliability with respect to a given failure threat. The parameters affected typically represent distribution parameters of input random variables used in the calculation; for example, the mean corrosion growth rate for a given pipeline segment. Knowledge uncertainties are distinct from randomness, which is typically manifested in variations in the basic input parameters affecting a given limit state; for example, variations in the excavator force applied to the pipeline in different impact events. Randomness is reflected in the probability distributions used to model the input variables affected and is automatically built into the reliability estimate. However, the reliability estimate is conditional on the values used for parameters affected by knowledge uncertainty. Since these parameters can take a range of values with different probabilities, knowledge uncertainty is best represented as a distribution or confidence interval on the calculated failure probability. Two approaches are proposed to deal with knowledge uncertainties in Reliability Based Design and Assessment (RBDA) applications in which design and operational choices are accepted if they meet a specified reliability target. The first is a formal approach in which reliability targets must be met with a specified level of confidence (e.g. meet the reliability targets with 90% confidence). The second approach is an informal one in which a single conservative value is used for each parameter affected by knowledge uncertainties. Although this approach relies on the judgment of the user, it has the advantage of being simple. In the context of standardizing RBDA, it is recommended that epistemic uncertainty be identified as an important issue that must be considered in demonstrating compliance. It is also recommended that both formal and informal approaches be permitted as viable means of accounting for epistemic uncertainty. The informal approach should be included as a minimum requirement, whereas the formal approach should be presented as an option. This recommended strategy addresses epistemic uncertainty without creating a significant obstacle to the application of RBDA.Copyright