Data driven approach to risk management and decision support for dynamic positioning systems

https://doi.org/10.1016/j.ress.2020.106964Get rights and content

Highlights

  • A risk management framework for dynamic positioning operation is proposed.

  • Model output is risk status and pre-warnings of possible deviations in the system.

  • The framework assists operators in decision-making process.

  • Dynamic positioning drilling operations are considered as a case study.

Abstract

Offshore oil and gas operations are inherently associated with risk and may have catastrophic consequences to life, property, and environment. Risk management is thus performed during the design, planning, and operation phases to control risk. Operational risk models are only periodically updated and do not always reflect the available real-time data. This is also the case for dynamic positioning (DP) operation. Monitoring the risk levels of the system during the operational phase could reduce the accident risk by providing additional decision support information for operators.

In this paper, a framework for the risk management of DP operation is proposed to assist operators in decision-making. The risk management output will provide operators a real time risk status and pre-warnings of possible deviations in the system. This framework is developed to support the decision-making process of the operators with providing failure probability of alternative decision scenarios, and it can be applied to any other engineering system and operation.

In order to validate the effectiveness of the framework, DP drilling operations are considered as a case study. The results demonstrate the value and effectiveness of the framework, which reduces the risk level of operations by contributing to the risk-informed decision-making of operators.

Introduction

Over the past few decades, online risk-informed decision-making has performed an increasingly important function in aerospace, nuclear, and marine technologies [1,2]. The frequency of dynamic positioning (DP) system failure is an ever-increasing problem, as reported in [3]. It is therefore necessary to further identify fundamental techniques for improving the DP system to reduce its failure frequency.

The safety improvements resulting from online risk management applications lead to advanced improvements in system operation [4,5]. For this reason, online risk monitoring and risk management are necessary to reflect system changes and enhance the understanding of the current safety state of a system.

The online risk management framework updates the failure probability of the system as necessary to account for the changes in system design and operation, thus improving system comprehension [6]. In contrast to conventional risk management methods, online risk management is presented as a dynamic development process. Generally, online risk management is employed to reflect the real-time risk of the system, thus indicating the actual status of (sub) systems and operational/environmental conditions.

Moreover, during critical situations, the online decision support and alarm system may prevent critical unwanted events or provide earlier situation awareness and increased response time to allow for early manual intervention [7]. The online risk level of the system can be used as an input to a decision support tool in order to aid operators in making better decisions more expeditiously and efficiently. As shown in Fig. 1, the action result returns to the risk management model, and the updated risk value is calculated in real time; this is a continuous iterative process.

Risk-informed decision-making has been applied for many years in different fields, and the role of risk insights in safety-related decision-making has received considerable attention. For instance, in [8], an overall methodology for risk-informed decision-making is proposed. In [9], the necessity of structural repair of aging naval ships is investigated based on risk-informed decision-making. In [10], a value-risk graph that visualizes the risk level of alternative decisions in a manufacturing process is proposed.

According to a literature review, however, research works related to risk-informed decision-making in dynamic positioning systems are limited [11,12]. In [7], the importance of decision support in reducing the shuttle tanker collision risk in floating production storage and offloading is investigated. In their study, the hazard, barriers, and risk reduction potentials are assessed, and the necessity of considering advanced estimation and data assimilation in risk management is discussed.

In this study, a novel framework that facilitates decision-making in DP systems is proposed. The main part of the risk-informed decision-making framework is the risk management model. The risk analysis of the DP system has been studied in detail at different complexity levels. In [13], a conceptual model for risk analysis is proposed. In most research works, the power structure of DP systems is investigated in detail [14,15]. Studies related to the risk and reliability analyses of overall DP systems are limited. In [16], a general fault tree for DP classes 1, 2, and 3 is presented. This study indicates that fault trees can be employed for reliability-based design and maintenance scheduling of multi-megawatt capacity DP systems. The human and organizational factors, some of which are important in incident occurrence, however, are ignored [17]. In this study, these factors are considered in the risk management model. As a result, the decision-making process, as well as the human and organizational factors, is facilitated.

The proposed risk management model quantifies the failure probabilities of different operating scenarios of DP systems and can be used as a basis for developing a decision support tool. One of the main parts in system failure quantification includes component failure frequencies. The sources and research on DP system failure mode quantification are also limited. In this study, failure modes are identified, and the rates of their occurrence are quantified based on the International Marine Contractors Association (IMCA) annual incident reports on DP systems (2004–2015). The data gathered from these reports are filtered (missing and inaccurate data are removed), and the failure frequencies of a generic DP drilling system are presented. Although incomplete, these data provide a comprehensive insight into the failure modes of DP systems. As a result, apart from the failure mode frequencies, the detailed fault trees and risk management model are proposed for DP systems. It should be noted that this dataset is used as a basis for the risk level calculation, and the failure frequencies are updated based on the information gathered over time, as presented in Fig. 1.

The online risk management model should satisfy two basic requirements to be applicable in a decision support tool [18]. First, risk level updating should reflect the information on real system configuration; second, a rapid solution to support the real-time application of risk management and decision-making is necessary. The feasibility of online risk management and decision support tool therefore considerably depends on the conversion and calculation times of the solution methodology. It is thus important to develop a highly efficient calculation engine to enable a rapid solution of the online risk management model.

Probabilistic risk assessment (PRA) is a systematic and comprehensive methodology for evaluating risks associated with a complex system while considering the uncertainties of operational and environmental conditions [19]. In the present study, an efficient solution approach based on the PRA method is proposed. Any significant changes in the system risk level can therefore be perceived by the operator, providing a basis for risk-informed decision-making. The full system description and boundaries utilized in this study are provided in Section 3.1. The main contributions of this study are summarized as follows.

  • A comprehensive risk management framework for DP operations is developed.

  • Human and organizational factors are considered in the risk management model.

  • Failure frequencies are calculated according to IMCA reports from 2004 to 2015.

The paper is organized as follows. A brief overview of the developed framework is introduced in Section 2. The details of the framework are presented in Section 3. In Section 4, a DP drilling unit is presented as a case study and results are presented. In Section 5, the results are analyzed, and the usefulness and drawbacks of the proposed framework are discussed. Finally, conclusions and contributions are presented in Section 6.

Section snippets

General concept of risk-informed decision making

A conceptual framework of risk-informed decision-making is presented in Fig. 2.

As presented in Fig. 2, the first step is data collection. Some data are non-observable; hence, these data cannot be considered in the modeling. These data are the main source of model uncertainty, and this limitation is further presented in the Discussion section (Section 5.4- Model uncertainties). Operators’ beliefs and desires in controlling the system are examples of non-observable data.

Observable data can be

Risk management model

An overview of the proposed risk management methodology is presented in this section. Fig. 3 illustrates the flow diagram of sub-models of the proposed methodology. In the first step, the boundary of the system should be defined. Based on the system boundary, the preferred end states can be determined. Thereafter, initial events that can lead to the realization of the target end state are determined. In the next step, event sequence diagrams (ESD), which can aid in deriving event trees, are

Risk management model applied to a MODU drift-off incident at Skarv field

An incident investigation report of a MODU is used to study the applicability of the developed framework. Mobile offshore drilling units are employed in the exploratory offshore drilling of new oil and gas wells. They rest on columns and pontoons and can be moored with anchors. For deep water drilling operations, however, these units (such as the studied MODU in the presented incident investigation report) rely on the DP system to maintain position. The DP system of MODUs follows the same rules

Applicability of proposed risk management model

In order to evaluate the effectiveness of the model, a case study with two operating scenarios is employed. The comparison between the results derived from the two scenarios indicates the sensitivity of the model to the PSFs of an operator. According to the results, when the stress level is high and the procedures are inadequate, the system failure probability is high (as expected). This demonstrates the usefulness of the model. Moreover, two operational modes, manual and automatic, in each

Conclusion

This paper proposes a risk management framework for a dynamic positioning system to assist operators in decision-making. The output of the risk management framework provides operators with a real-time risk status that can aid in making better decisions within a limited time. The framework presented in this paper also takes a more holistic approach to the risk modeling of DP operations by including human error scenarios both as initiating events and potentially escalating events.

The paper

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

CRediT authorship contribution statement

Tarannom Parhizkar: Data curation, Writing - original draft. Sandra Hogenboom: Data curation, Investigation, Writing - review & editing. Jan Erik Vinnem: Conceptualization, Methodology, Writing - review & editing, Supervision. Ingrid Bouwer Utne: Conceptualization, Methodology, Writing - review & editing, Supervision.

References (33)

  • S. Hogenboom et al.

    Human reliability and the impact of control function allocation in the design of dynamic positioning systems

    Reliab Eng Syst Saf

    (2020)
  • T. Parhizkar et al.

    Efficient health monitoring of buildings using failure modes and effects analysis case study: air handling unit system

    J Build Eng

    (2020)
  • J. Rasmussen

    Human errors. A taxonomy for describing human malfunction in industrial installations

    J Occup Accid

    (1982)
  • K. Groth et al.

    Hybrid causal methodology and software platform for probabilistic risk assessment and safety monitoring of socio-technical systems

    Reliab Eng Syst Saf

    (2010)
  • A.F. Mehr et al.

    Risk-based decision-making for managing resources during the design of complex space

    J Mech De.

    (2006)
  • Dynamic positioning station keeping review

    Incidents and events reported for

    (2016)
  • Cited by (0)

    View full text