During the production phase of the PW1 well, an unintentional operation depressurized the annulus A below its design limit, causing a progressive casing collapse from the surface casing to the tubing string, resulting in a well failure and a subsequent an abandonment operation. This event was detailed in a previous publication (Campos et al., 2019). After a detailed investigation of the events that drove the well failure and using casing design software to simulate the well behavior with the historical time series data, it was possible to identify the exact moment of the casings collapse through mathematical modelling. Based on that, considering innovation as critical success factor (Rocha, 2012), PETROBRAS started an R&D project to evaluate, in real time, critical integrity events during production phase. The approach proposed consists on the integration of a transient fluid flow model (Bonizzi, 2003) and (Bonizzi, Andreussi and Banerjee, 2009) with a structural model (AraÚjo et al., 2019) of the well and loads being calculúated in real time as a function of the production data (PDG’s, X-TREE and Topside). The process described was named as Well Integrity Surveillance (WIS) and its being designed to work 24/7 in a collaborative environment, integrating well handover information, inspections and intervention reports, production data and flow modeling. The objective of this paper is to present the well integrity case study of a deep-water field in the Brazilian coast, through a real time digital twin approach, highlighting the main technological solutions, developed to ensure the assets operational envelope, during its lifetime. With the well integrity surveillance digital twin, PETROBRAS successfully implemented a pilot project on three offshore production units in Brazil, increasing the knowledge of the well structural behavior and consequentially the assets operational safety during its lifetime.

Summary: This paper explores the application of large language models (LLMs) in the oil and gas (O&G) sector, specifically within well construction and maintenance tasks. The study evaluates the performances of a single agent and a multiagent LLM-based architecture in processing different tasks, offering a comparative perspective on their accuracy and the cost implications of their implementation. The results indicate that multiagent systems offer improved performance in question and answer (Q&A) tasks, with a truthfulness measure 28% higher than single-agent systems but at a higher financial cost. Specifically, the multiagent architecture incurs costs that are, on average, 3.7 times higher than those of the single-agent setup due to the increased number of tokens processed. Conversely, single-agent systems excel in Text-to-SQL (structured query language) tasks, particularly when using the Generative Pre-Trained Transformer 4 (GPT-4), achieving a 15% higher score compared to multiagent configurations, suggesting that simpler architectures can sometimes outpace complexity. The novelty of this work lies in its original examination of the specific challenges presented by the complex, technical, unstructured data inherent in well construction operations, contributing to strategic planning for adopting generative artificial intelligence (AI) (Gen-AI) applications and providing a basis for optimizing solutions against economic and technological parameters.

Today's society is increasingly susceptible to global economic crisis. The energy is still extremely dependent on oil, whose extraction scenarios usually have great complexity. The technology in the oil industry has undergone a vertiginous development in recent times. Advances in information, data and telecommunications have been reflected in the connectivity of systems and people with tremendous impacts on improvements of operational and business decisions. In this context, came to light an important movement of Integrated Operations (IO), which has systematized the implementation of this concept in the oil and gas operators and service providers. Each company has created its program, with peculiarities and challenges. This paper, including an intensive literature search, analyzed the objectives, concepts and architectures of Integrated Operations deployments in five major operators. Through a reference framework, the differences and similarities of these models were discussed. There was also an analysis in recent bibliography in order to better understand the value for Integrated Operations.

Summary: The demand for ultradeepwater scenarios invoked the frequent application of managed pressure drilling (MPD) in the last few years. In an ultradeepwater scenario, oil companies face issues such as narrow pressure windows and severe loss zones. Many wells are considered undrillable without the aid of MPD technology. MPD operations need to be correctly evaluated with consideration given to increased time and cost/benefit analysis. In this paper, we propose a probabilistic model to evaluate MPD demand by estimating the optimal number of rigs equipped with MPD and a rotating control device (RCD), and we analyze which intervention strategy is the most cost– and time–effective. Reducing uncertainty is an important factor when making decisions about drilling. We adopted a Monte Carlo simulation using loss–zone estimation, probability of prediction error, the number of rigs equipped with MPD, and several strategies. Better MPD strategies were determined on the basis of available data and the optimal number of rigs equipped with an MPD system and RCD equipment, reducing subjectivity in the decision–making process. The originality of our paper lies in the new quantitative approach to dealing with uncertainty in the prediction of fluid losses and the cost and duration of different MPD strategies, numerically simulating the possible scenarios.

In the oil industry, undesirable events in oil wells can cause economic losses, environmental accidents, and human casualties. In 2019, recognizing the importance and the lack of public datasets related to undesirable events in oil wells, Petrobras developed and publicly released the first version of the 3W Dataset, which is essentially a set of Multivariate Time Series labeled by experts. Since then, the 3W Dataset has been developed collaboratively and has become a foundational reference for numerous works in the field. This data article describes the current publicly available version of the 3W Dataset, which contains additional instances, more variables, and a new label. Furthermore, a new data structure has been developed to make data access more robust and efficient. The detailed description we provide encourages and supports the 3W community and new 3W users to improve previous published results and to develop new robust methodologies, digital products and services capable of detecting undesirable events in oil wells with enough anticipation to enable corrective or mitigating actions.