Data mining/analysis

Real-Time Analytics Improves Process Safety in a Drilling-Contractor Operations Center

This paper explains how an ultradeepwater drilling contractor is applying real-time analytics and machine learning to leverage its real-time operations center to improve process safety and performance.

jpt-2019-05-28824hero.jpg

Digitally enabled collaboration between operators and service companies is driving significant improvements in process safety in the upstream oil and gas industry. However, drilling contractors generally have been considered only the owners of the systems that generate part of the surface drilling data and, as such, have not been engaged in the development of real-time analytics tools. The role of a real-time operations center (RTOC) with a drilling contractor is itself a new factor. This paper explains how an ultradeepwater drilling contractor is applying real-time analytics and machine learning to leverage its RTOC to improve process safety and performance.

Introduction

The increasing availability of digital drilling data and the emergence of reliable communication between offshore rigs and well operators’ offices enabled a proliferation of RTOCs in the 2000s, spawning a collaborative environment between well operators and service companies that enabled enhanced support and optimized use of expert resources to improve process safety and operational efficiency. Drilling contractors generally were not engaged in developing the real-time analytics tools for the RTOCs. One consequence is that the majority of RTOC software solutions are only appropriate for wellsite information transfer standard markup language (WITSML) data visualization, or they feature embedded tools for well operators and service companies focused on drilling efficiencies but the data and models they use are not accessible to the rig operators.

Essential aspects of process safety and operational performance, such as procedural discipline and critical equipment health, can be addressed by drilling contractors in real time if appropriate resources are in place. The contractor’s RTOC has enhanced process safety and surveillance after moving to an approach in which drilling engineers, assisted by software developers with robust experience in real-time data analytics, began writing automated operation-identification and problem-detection algorithms. The cooperative work allowed the creation of tools adapted to the company’s needs and provided a streamlined process to implement, customize, and repair these tools over time. During the planning phase, it was agreed that initial developments should be grouped into six important dimensions that strongly influence the overall process safety and operational performance of the rig.

The contractor applied real-time analytics and machine learning to identify and alert RTOC engineers of abnormal situations automatically. The detection of the abnormal situations by RTOC engineers and the comparison of the standard by which various drilling crews react to them feeds a lessons-learned database, which, in turn, sustains the procedural advancement initiatives, ensuring continuous improvement.

The complete paper introduces a methodology by which the described RTOC platform can integrate and synchronize data coming from different sources, including third-party companies and the well operator. Current technologies have supported the rise of integrated dashboards with data from different sources, such as drilling packages, blowout-preventer- (BOP) control systems, downhole drilling data, and mud logging incorporated into the same real-time platform, enhancing the overall ­decision-making process considerably.

Data analysis is supported by a ­data-collection mechanism communicating with a real-time analytics Internet-of-things platform. It allows the integration of data from drilling packages, the BOP, and the unit-drilling data server, which may provide basic environmental information.

Operation modes are detected automatically in real time and used as primary information. Algorithms developed by the RTOC team search for, and alert personnel to, any abnormal situation. These tools are organized into six main categories, all related to process safety and linked with the company’s operational track record.

A flexible dashboard platform enables the RTOC engineers to develop algorithms and graphics that are customized to improve monitoring accuracy and fulfill operational needs. The dashboards target critical operations and allow much more data to be taken into consideration, resulting in a more-structured decision-making process.

Project Timeline

After gathering drilling and BOP data from the systems installed on the drilling rigs, the first stage of development for the RTOC platform was to implement a routine in which RTOC engineers were responsible for operational surveillance. Procedures were developed to rank critical operations, documenting the way each should be approached.

The system was intended mainly for data visualization and permitted setting up threshold alarms. Detection was based primarily on training and personnel awareness.

The initial methodology worked relatively well and prevented some situations from escalating, allowing a considerable number of lessons to be learned and adding to existing procedures. However, sometimes problems occurred in operations deemed as less critical at the beginning, leading to a lack of RTOC support when it was required.

After ineffective attempts to enhance the way operations should be classified according to their process-safety risk and increase operational coverage without a proportional need for personnel, a change emerged in the way the problem was addressed. All solutions, in fact, would involve some level of automation in operations identification, problem detection, and inclusion of awareness alarms when an abnormal situation was detected.

A few WITSML analytic tools available on the market were field tested, and the conclusion was that they were not adequate to deal with the specific demands of this new approach. Additionally, in most cases, the tools were oversized to the type and extent of accessible data. Customizing the existing products and using more than one solution were not economically viable.

The contractor decided instead to collaborate with a software company to develop a platform on which real-time routines could be developed. Drilling engineers who broadly knew the well and rig-equipment issues that needed to be overcome worked together with software developers who knew how to address the issues computationally. The primary drivers were the possibility of adapting the software to the company’s requirements rather than vice versa and having the drilling engineers develop most of the logics and dashboards without the need for in-depth programming expertise. The adjustments identified to improve detection performance are performed in house upon demand and tested immediately with real field data.

Automatic identification allows one engineer to cover more operations in a system that is less susceptible to errors. Once an alarm is activated, the process that generates it starts to deserve particular attention and, thereby, may require additional support.

Operational Routine

According to the authors, despite the significant volume of data gathered and recorded on today’s drilling rigs, ordinary drilling crews continue to make decisions on the basis of the same underlying data they used 15 years ago. The critical factor in the development of the RTOC platform in question was the way the data were consolidated and addressed to cope with typical well and rig problems. The complete paper explains the operational routine for the RTOC (Fig. 1).

jpt-2019-05-28824f1.jpg
Fig. 1—Routine operation for drilling-contractor RTOC.

Six-Dimensions Concept

To organize the development initiatives better and balance the resources correctly, the RTOC platform categorizes work into six dimensions, all related to process safety and linked with the company’s operational track record. They are

  • Well control
  • Well-barrier assessment
  • Equipment safety operational window
  • Drilling-problem prediction
  • Process-sensor integrity
  • BOP monitoring

The categories have been approached through routines prioritized according to their immediate effect on process safety, available resources, complexity, and development time. Each of these dimensions is discussed in detail in the complete paper, and the types of routines are exemplified.
This process turns data into useful information and detects abnormal situations faster, taking into account much more data than an overall drilling crew could. Alarms are associated with the scenarios and alert RTOC engineers of any abnormality. After verification, the RTOC communicates with the rig. Additional actions depend on the type and ­severity of the recognized situation.

Conclusions

  • Process safety and surveillance have improved considerably using real-time analytics tools developed in the contractor’s RTOC. A software platform affords flexibility to allow the creation of a fit-for-purpose solution that covers primary operational objectives.
  • Drilling engineers are in charge of monitoring processes and development of most detection algorithms, engaging the support of software experts when needed. The result is an optimized solution reached in a short period of development.
  • Better results are achieved using a hybrid method with detection algorithms based on the automatically detected operation mode running in the background, combined with procedures in which the RTOC engineers should carry out specific measures.
  • Proper planning during initial phases of the well ensures that the most-affected well subprojects are prioritized and the efforts for its conclusion are well-defined. This approach fosters an environment in which RTOC engineers are engaged in the project and follow up on all phases, from the planning to the delivery of relies the service.
  • Successful implementation also relies on a well-structured communication protocol, which should define at least what type of alarms or deviations will be communicated, to whom the information should go, how it should flow, and when it is to be triggered.
  • Procedural discipline, a critical aspect of safe and efficient operation, must be addressed and reported; time-surveillance algorithms detect deviations in critical operations. Programming is less complicated when the procedure is standardized. To improve procedures on the basis of what was learned from the real-time monitoring will make operations performance simultaneously more consistent and safer.
  • RTOC outcomes depend directly on data availability. High costs could be involved in mapping additional channels and upgrading drilling data servers if data availability has not been taken into account during the system’s design and acquisition phase.

This article, written by JPT Technology Editor Judy Feder, contains highlights of paper OTC 28824, “Operational and Safety Improvements of Applying Real-Time Analytics in a Drilling-Contractor RTOC,” by A.L.F. Madaleno, S.L.S. Neto, L.A. dos Santos, and C.A.L. de Oliveira, QGOG Constellation, prepared for the 2018 Offshore Technology Conference, Houston, 30 April–3 May. The paper has not been peer reviewed. Copyright 2018 Offshore Technology Conference. Reproduced by permission.