No abstract

Change points are abrupt alterations in the distribution of sequential data. A change-point detection (CPD) model aims at quick detection of such changes. Classic approaches perform poorly for semi-structured sequential data because of the absence of adequate data representation learning. To deal with it, we introduce a principled differentiable loss function that considers the specificity of the CPD task. The theoretical results suggest that this function approximates well classic rigorous solutions. For such loss function, we propose an end-to-end method for the training of deep representation learning CPD models. Our experiments provide evidence that the proposed approach improves baseline results of change point detection for various data types, including real-world videos and image sequences, and improve representations for them. * This work is supported by the Russian Science Foundation (project 20-71-10135) Preprint. Under review.

During the directional drilling, a bit may sometimes go to a nonproductive rock layer due to the gap about 20m between the bit and high-fidelity rock type sensors. The only way to detect the lithotype changes in time is the usage of Measurements While Drilling (MWD) data. However, there are no general mathematical modeling approaches that both well reconstruct the rock type based on MWD data and correspond to specifics of the oil and gas industry. In this article, we present a data-driven procedure that utilizes MWD data for quick detection of changes in rock type. We propose the approach that combines traditional machine learning based on the solution of the rock type classification problem with change detection procedures rarely used before in Oil&Gas industry. The data come from a newly developed oilfield in the north of western Siberia. The results suggest that we can detect a significant part of changes in rock type reducing the change detection delay from 20 to 1.8 meters and the number of false positive alarms from 43 to 6 per well.

Majority of the accidents while drilling have a number of premonitory symptoms notable during continuous drilling support. Experts can usually recognize such symptoms, however, we are not aware of any system that can do this job automatically. We have developed a Machine learning algorithm which allows detecting anomalies using the drilling support data (drilling telemetry). The algorithm automatically extracts patterns of premonitory symptoms and then recognizes them during drilling. The machine learning model is based on Gradient Boosting decision trees. The model analyzes real time drilling parameters within a sliding 4-hour window. For each measurement, the model calculates the probability of an accident and warns about anomaly of particular type, if the probability exceeds the selected threshold. Our training sample comes from 20+ oilfields and consists of sections related to 80+ accidents of the following types: stuck pipe, mud loss, gas-oil-water show, washout of pipe string, failure of drilling tool, packing formation, that occurred while drilling, trip-in, trip-out, reaming. We have designed the prediction model to work during drilling new wells and to distinguish the normal drilling process from the faulty one. One can configure the anomaly threshold to balance amount of false alarms and the number of missed accidents. To evaluate quality of the model we measure such data science metrics as ROC AUC score and confusion matrices. While testing model can identify 24 accident from 30 with high confidence, whereas for the others there is still a room for improvement. Our findings suggest that including more accidents of underrepresented types will improve quality. Other data science metrics also support aptitude of the model. Finally, having data from multiple heterogeneous oilfields, we expect that the model will generalize well to new ones. This paper presents a good practice of development and implementation of a data-driven model for automatic supervision of continuous drilling. In particular, the model described in the paper will assist specialists with drilling accidents prediction, optimize their work with data and reduce the nonproductive time associated with the accidents by up to 20%.

No abstract