Analysis of Windows of Opportunity for Weather Sensitive Operations

Abstract

The paper presents an approach and real life results for the statistical analysis of operational windows for weather sensitive marine activities of limited durations (hours to days), based on the application of two statistical distributions: the four parameters Johnson distribution and the K distribution. In the offshore industry, operations like heavy lift, pipeline tie-in, topside float-over can be carried out only when strict multi parameter limit conditions in term of wave height, wave period, wave direction - and, occasionally, wind and current speeds - are met over periods ranging from a few hours to a few days. Reliable statistical analysis of occurrence and duration of such conditions - windows of opportunity - is an important planning tool to identify the most suitable season for operation, the vessel characteristics to minimize downtime and even the most appropriate operative approach - e.g. active vs. passive ballasting in float-over. This allows the minimization of overall cost and duration of the operation. Present availability of long time series of metocean data, mostly from numerical models, allows the identification of windows of opportunity subject to complex sets of operative conditions, such as the concomitant occurrence wave, current and wind impacting on the vessel. However, the subsequent evaluation of the statistical properties (average duration, percentiles etc.) of operational windows from sample data is usually quite noisy, especially when the analysis is required on seasonal/monthly basis. The robustness of the assessment can be improved with the use of suitable statistical distributions to fit sample data. In this context, the four parameters Johnson distributions and the K-distribution have the flexibility to fit windows of opportunity data and be used to evaluate their statistical properties. The paper reviews the characteristics of both distributions, describes methods to fit them on sample data (estimators) and provides some examples of applications and preliminary comparison of performance with data and conditions derived from real life float-over operations.