Study On Assessment Methodology For Offshore Drilling Dynamic Risk And Safety Barriers Reliability

Offshore drilling operations are likely to be affected by a variety of uncertainties,especially in areas where there are the limestone formation,or narrowed drilling fluid density windows.Drilling lost circulation and well kick are regard as some typically hazardous events that may occur during drilling phases and it also may develop into a blowout if they are not well handled.The dynamic risk prediction and diagnosis of offshore drilling operations plays an important role in drilling safety.Based on the development and practice of some oil and gas fields,four major themes have been carried out including dynamic quantitative risk analysis,real-time risk early warning,reliability assessment of safety barriers and maintenance optimization.The main contents are as follows:Firstly,with combination of the characteristics of the managed pressure drilling system and process technology and the bottom hole pressure control mechanism,a comprehensive evolution model of drilling incidents under the multi-operating condition is established according to the drilling engineering system theory and the safety barrier theory.This model describes the development tendency of well lost circulation and kicks,and can provide theoretical guidance for the prevention of catastrophic accidents such as offshore drilling blowout occurrence.Secondly,the main weakness of the existing risk assessment approaches is their inability to capture dynamic effects of a drilling operation,the state probability of events,such as drilling fluid density changing,is difficult to be updated.And what also seems to be missing is the possibility to incorporate the effects of uncertainties on quantitative results.A new approach is therefore presented to handle the above-mentioned issues by allowing uncertainties analysis with DBN theory.This approach can systematically perform incidents evolution prediction and root cause reasoning for risk assessment using the predictive,diagnostic and sensitivity analysis technology.Thirdly,the existing early warning for drilling cannot treat the problems such as the nonlinear variation of the drilling parameters and the delayed or unapparent change trend of data.The judgment of conventional indicators can’t meet the requirements of early incidents monitoring.The real-time early warning model based on the multi-indicators fusion is established.Such model is based on the dynamic safety threshold judgement and Bayes estimation of drilling parameters to predict whether the drilling is abnomal.The application of the model provides the basis for real-time early warning.Fourthly,many simplified formulas for PFD_avg for safety barriers assume the failures are exponentially distributed,and the failure rates are generally assumed to be constant,which mean that all channels restored components after a proof test are in an as-good-as-new state.A new approach is therefore proposed for the reliability assessment on safety barriers with time dependent failure rates.Approximation formulas for evaluation of average probability of failure on demand(PFD_avg)are developed to handle the degradation effects in a serious subsequent proof testing intervals.The change trend of the PFD_avg and the evaluation of the safety integrity level of the different testing intervals are analyzed for decision making.Fifthly,challenges from subsea context are not well handled with,for instance,the non-ignorable time to repair even for a revealed failure.Repairs for subsea facilities are always postponed due to some potential well accidents.A newly multi-state-based approach is developed for preventive testing and maintenance by incorporating repair waiting state into multiphase Markov process.Multi-phase Markov Models for different configurations are established.The effects of testing errors,postponed repairs and non-periodic characteristics on the tendency of unavailability are checked in the proposed models for unavailability analysis.