Stochastic Approach to Kick Tolerance Determination in Risk Based Designs

Dedenuola, Adetola D. (Shell Petroleum Development Company of Nigeria) | Iyamu, Isaac E. (Shell Petroleum Development Company of Nigeria) | Adeleye, Olayinka A. (Shell Petroleum Development Company of Nigeria)


Abstract Drilling experience in the Niger Delta has shown that inappropriate selection of kick tolerance could constitute a serious setback to the implementation of cost effective and innovative solutions in well designs. While a conservative selection of kick tolerance could make a well design uneconomical, attracting severe cost penalties by fostering use of expensive / extreme pressure-ratings BOPs and extra casing strings sometimes, undue tightening of the tolerance on the other hand could make a well undrillable from safety point of view. This consideration poses serious dilemma to the well engineer who must balance the demands of conflicting alternatives in his management of design risks. To resolve this, the approach of stochastic analysis procedures to the prediction of appropriate design kick tolerance in any given geological setting and operational environment is being presented in this paper. Since this approach depends largely on the use of historical kick records, certain concerns regarding the reliability and acceptability of the use of historical data for the purpose of kick tolerance modeling and selection were equally addressed. Finally, a well example shows how this method has been used to integrate kick tolerance considerations in the management of design risks associated with the deployment of new technology application, expandable sand screen, in a high rate horizontal oil producer. Introduction In well planning process, kick tolerance is one of the key parameters in setting depth determination and burst designs of non-production casing strings. Also, various practices and techniques are generally used in the selection of appropriate design value for this parameter, the commonest being the use of gas flow deterministic models combined with computer simulations (also commonly based on Monte Carlo simulation engine) to investigate the likelihood of the selection. While the popularity of this method is rooted in arguments against presumed limitations inherent in the use of historical drilling data for the modeling kick tolerance distribution, it is mind boggling that input variables with similar historical origin (i.e. permeability, porosity, crew reaction time, etc) which could even be less reliable are often used in these simulations. This development suggests that the problem really is not with the use of historical well control data for kick tolerance modeling, but rather with what approach has been employed to transform these data into usable forms. Needless to say, a general consensus throughout the drilling industry is the desirability of an acceptable and reliable method of kick tolerance determination from historical kick data. The fact that historical records usually encompass a wide range of events, environment, and data quality poses a great challenge to the development of historic probability distribution with any great confidence. However, stochastic analysis procedures combined with various data mining techniques provide a veritable means of transforming historical well control data into usable probabilistic models for the reliable prediction of kick tolerance within a given geological settings. The outstanding feature of this approach is that any probabilistic model developed using historical records must be validated by sound engineering judgment based upon historical experience of drilling and operational practices particular to the environment. This represents a classic case of the combination of both qualitative and quantitative risk assessment approaches to the management of design risks.

Duplicate Docs Excel Report

None found

Similar Docs  Excel Report  more

None found