Operational Aspects and Risk Management of an Autonomous Well Service Pump for Offshore Installations

Abstract

The paper discusses operational aspects and establishes the basics of risk management for autonomous technology applied to the topside well service pump for offshore installations. In the analysis, a specific machine is investigated, an electrically driven pump with a power of 725 horsepower equipped with a computer-controlled system. The system enables performing basic tests, control checks, and maintenance, for example, from the operator’s office onshore. The research is divided into two parts. The first describes machine learning applications and determines autonomy levels for offshore well service pumps. The computer-controlled system in this paper is considered the first stage of autonomy. The level of autonomy (LoA) was gradually increased by applying machine learning, implementing predictive maintenance techniques, and creating the digital twin. The highest autonomy stage enables the machine to make critical decisions. That feature brings many profits, for instance, reducing the number of people in dangerous places or prevents from making bad decisions. The risk assessment analysis was performed in the second part of the paper. The risk description for every LoA was provided, and the hazard events were specified and described providing the causes and solutions. Lastly, the experienced team assessed the risk, presenting the results in the risk matrix. The analysis shows that the most hazardous events are related to the connection and environmental conditions with the unit. The research shows that there is a potential for the application of machine learning in machinery systems for the offshore industry. Therefore, there is a need for more research in that field.

Received: 31 March 2023 | Revised: 6 June 2023 | Accepted: 18 June 2023

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.