Hub Digital - COPPE - UFRJ

Petrophysical characterization of carbonate reservoirs remains a challenge. The widespread use of X-ray computed microtomography (micro-CT) images in the classical reservoir characterization workflow allows the use of recent artificial intelligence algorithms to improve the process. This work presents an end-to-end workflow for permeability prediction using deep learning models and micro-CT images. A dataset of 37,600 slices from 376 plug samples from Brazilian presalt carbonate rock, along with the laboratory determined absolute permeability of each sample, were used for model training. Three models were tested: two convolutional neural network models (CNN and CNNSPP) and an ImageNet pretrained model (Densenet161). The models were trained using MSE or the Huber loss and with/without data augmentation.All experiments were performed using 10-fold cross-validation, and the models performance were evaluated by the average prediction of all slices for each sample. In this study, the Densenet161 model achieved the best results. The comparison with other models shows that pretrained models have less influence of data augmentation and almost no difference with respect to the loss function. This shows the effect of transfer learning, even if micro-CT images are very different from ImageNet. The results show that the proposed workflow can automate and speed up the characterization of Brazilian presalt carbonate samples by processing micro-CT slices thereby allowing accurate estimations of absolute permeability within a few seconds.

The wake steering control in wind farms has gained significant attention in the last years. This control strategy has shown promise to reduce energy losses due to wake effects and increase the energy production in a wind farm. However, wind conditions are variable in wind farms, and the measurements are uncertain what should be considered in the design of wake steering control strategies. This paper proposes using the probabilistic learning on manifold (PLoM), which can be viewed as a supervised machine learning method, to enable the wake steering optimization under uncertainty. The expected power generation is estimated considering uncertainties in wind speed and direction with good accuracy and reduced computational cost for two wind farm layouts, which expand the application of machine learning models in wake steering. Furthermore, the analysis shows the potential gain with the application of wake steering control.

Digital twins seek to replicate a physical structure in a digital domain. For a digital twin to have close correspondence to its physical twin, data are required. However, it is not always possible, or cost-effective, to collect a complete set of data for a structure in all configurations of interest. It is nonetheless useful to repurpose data to help validate predictions for different configurations and scenarios. This statement is true in drilling applications, where, for example, the length of the drill string is altered throughout operation. This paper demonstrates how transfer learning, in the form of three domain-adaptation methods, — transfer component analysis (TCA), maximum independence domain adaptation (MIDA) and geodesic flow kernel (GFK) — can be used to construct a digital twin for localising torsional friction in deviated wells under structural changes (e.g., when the drill column gets longer). The method uses a physics-based torsional model to train a machine-learning classifier that can localise torsional friction for a given drill string length and diameter, where friction localisation labels are known (source). As the length or diameter of the drill string are altered in the field, transfer learning is utilised to map the classifier from the labelled (source) scenario onto these unlabelled (target) scenarios. As a result, transfer learning improves the performance of the classifier when applied to the target data, and increases the domain of validity for the classifier. The performance of the classifier, and therefore its suitability to new drill-string configurations, is estimated by utilising two different distance metrics between the source and a proposed target dataset.

Black oil delumping, also known as a stream conversion method, converts a black oil wellstream into a compositional wellstream. This procedure ensures consistent flowrate allocations and monitoring of well’s performance. This method requires volumetric oil and gas flowrates given in well-test reports, an equation of state model, and additional black oil information reported in the Well Test, PVT Analysis, and Gas Chromatographic Analysis. This work proposes an improvement on the method to convert black oil data into compositional wellstream. The method’s performance was tested using data of three wells of a platform from an offshore oil field. This improvement significantly increased the accuracy of the method by decreasing the maximum percentage relative error from 16.50% to 4.44% when comparing the calculated and measured oil and gas properties for Well 1, for example. The method also preserves the gas and oil ratio reported in the well tests.

An appropriate techno-economic assessment of biorefineries is essential for consolidating a bio-based economy. Herein, the effects of scale and seasonality on the design of a sugarcane to ethanol biorefinery were assessed using an extensive process simulation and economic modeling methodology. Four sub-processes were simulated for eight different scales: ethanol from sugarcane, ethanol dehydration to ethylene, ethylene to mono ethylene glycol (MEG), and electricity and steam generation from bagasse. Models for capital expenditure and minimum selling price versus product capacity were developed for ethanol, bio-ethylene, and MEG. The integrated biorefinery was analyzed for 200 kt/yr and 440 kt/yr bio-ethylene capacities, considering three regions of Brazil: Northeast, South-central, and São Paulo/Paraná. Seasonality effects were mitigated by establishing a storage park alongside the biorefinery. The results indicated that the selling price of bio-MEG must be at least 20–30% higher than the fossil-based MEG to ensure competition. The 440 kt/yr biorefinery located in the South-central region had the lowest bio-MEG selling price, however, it required an outsized distillery. Moreover, scale and seasonality had a significant impact on the competitiveness of the biorefinery. The bio-MEG selling price might vary by up to 22%, depending on the specific combination of these factors.

In addition to the ongoing development, pre-salt carbonate reservoir characterization remains a challenge, primarily due to inherent geological particularities. These challenges stimulate the use of well-established technologies, such as artificial intelligence algorithms, for image classification tasks. Therefore, this work intends to present an application of deep learning techniques to identify lithological patterns in Brazilian pre-salt carbonate rocks using microtomographic images. Four convolutional neural network models were proposed. The first model includes three convolutional layers, followed by a fully connected layer. This model is used as a base model for the following proposals. In the next two models, we replace the max pooling layer with a spatial pyramid pooling and a global average pooling layer. The last model uses a combination of spatial pyramid pooling followed by global average pooling in place of the final pooling layer. All models are compared using original images, when possible, as well as resized images. The dataset consists of 6,000 images from three different classes. The model performances were evaluated by each image individually, as well as by the most frequently predicted class for each sample. According to accuracy, Model 2 trained on resized images achieved the best results, reaching an average of 75.54% for the first evaluation approach and an average of 81.33% for the second. We developed a workflow to automate and accelerate the lithology classification of Brazilian pre-salt carbonate samples by categorizing microtomographic images using deep learning algorithms in a non-destructive way.

Over the last decades, oil and gas companies have been facing a continuous increase of data collected in unstructured textual format. New disruptive technologies, such as natural language processing and machine learning, present an unprecedented opportunity to extract a wealth of valuable information within these documents. Word embedding models are one of the most fundamental units of natural language processing, enabling machine learning algorithms to achieve great generalization capabilities by providing meaningful representations of words, being able to capture syntactic and semantic features based on their context. However, the oil and gas domain-specific vocabulary represents a challenge to those algorithms, in which words may assume a completely different meaning from a common understanding. The Brazilian pre-salt is an important exploratory frontier for the oil and gas industry, with increasing attractiveness for international investments in exploration and production projects, and most of its documentation is in Portuguese. Moreover, Portuguese is one of the largest languages in terms of number of native speakers. Nonetheless, despite the importance of the petroleum sector of Portuguese speaking countries, specialized public corpora in this domain are scarce. This work proposes PetroVec, a representative set of word embedding models for the specific domain of oil and gas in Portuguese. We gathered an extensive collection of domain-related documents from leading institutions to build a large specialized oil and gas corpus in Portuguese, comprising more than 85 million tokens. To provide an intrinsic evaluation, assessing how well the models can encode domain semantics from the text, we created a semantic relatedness test set, comprising 1,500 word pairs labeled by selected experts in geoscience and petroleum engineering from both academia and industry. In addition, we performed an extrinsic quantitative evaluation on a downstream task of named entity recognition in geoscience, plus a set of qualitative analyses, and conducted a comparative evaluation against a public general-domain embedding model. The obtained results suggest that our domain-specific models outperformed the general model on their ability to represent specialized terminology. To the best of our knowledge, this is the first attempt to generate and evaluate word embedding models for the oil and gas domain in Portuguese. Finally, all the resources developed by this work are made available for public use, including the pre-trained specialized models, corpora, and validation datasets.