Hydrocarbon production from mudrocks (“shales”) is vital to global economic growth and smooth transition to a clean energy infrastructure. The commercial development prospect of a shale play depends on its evolution history over millions of years. Economic hydrocarbon production from shale starts after hydraulic fracturing, that creates a multiscale fracture network leading to an increased overall permeability. The properties of the stimulated rock can be assessed via parameters at different scales (nano-, micro- and macro-scale). Better understanding of these parameters is the key to predicting well productivity and profitability. This work aims to deepen the understanding of the multiscale parameters that define effective hydraulic fracturing. To investigate permeability increase in shales, we start with a model of micro-capillary in contact with nanopores . We show that the nanopores that discharge gas into a fracture network in the source rock significantly increase and extend gas flow into the hydrofractured horizontal wells. We then use a fractal stimulated reservoir volume model to match production histories of 45 Barnett gas wells and to quantify connectivity between the nanopores and the fracture network. This model relies on a source term, ${s}$, and fracture permeability $k_f$ . Our analysis shows that the different degrees of coupling between ${s}$ and $k_f$ create distinctly different types of fracture networks after rock stimulation and impact the well production profiles. We then couple the fractal SRV model with universal scaling $τ − M$ model to simulate production history of 1000 wells each in the Barnett, Marcellus, Haynesville and Eagle Ford shale plays. The analysis shows the coupled effect of stimulated surface area $A$, fracture half-distance, $d$, and the fractal dimension ,$D$, on production and economics of gas production. These parameters define the key differences between different shale plays in the US. Finally, we simulate microfracturing associated with hydrocarbon expulsion in the Tuwaiq Mountain source rock, Saudi Arabia, and propose the pore/microchannel blocking by bitumen/pyrobitumen as a viable mechanism of sustaining the high pore pressure in the source rock for millions of years.

[pt] Prever precisamente a produção de óleo é essencial para o planejamento e administração de um reservatório. Entretanto, prever a produção de óleo é um problema complexo e não linear, devido a todas as propriedades geofísicas que com pequenas variações podem resultar em differentes cenários. Além disso, todas as decisões tomadas durante a exploração do projeto devem considerar diferentes algoritmos para simular dados, fornecer cenários e conduzir a boas deduções. Para reduzir as incertezas nas simulações, estudos recentes propuseram o uso de algoritmos de aprendizado de maquina para solução de problemas da engenharia de reservatórios, devido a capacidade desses modelos de extrair o maxiomo de informações de um conjunto de dados. Essa tese propôe o uso ed duas tecnicas de machine learning para prever a produção diaria de óleo de um reservatório. Inicialmente, a produção diária de óleo é considerada uma série temporal, é pré-processada e reestruturada como um problema de aprendizado supervisionado. O modelo Random Forest, uma extensão das arvores de decisão muito utilizado em problemas de regressão e classificação, é utilizado para predizer um passo de tempo a frente. Entretanto, as restrições dessa abordagem nos conduziram a um modelo mais robusto, as redes neurais recorrentes LSTM, que são utilizadas em varios estudos como uma ferramenta dee aprendizado profundo adequada para modelagem de séries temporais. Várias configurações de redes LSTM foram construidas para implementar a previsão de um passo de tempo e de multiplos passos de tempo, a pressão do fundo de poço foi incorporada aos dados de entrada. Para testar a eficacia dos modelos propostos, foram usados quatro conjunto de dados diferentes, três gerados sintéticamente e um conjunto de dados reais do campo de produção VOlve, como casos de estudo para conduzir os experimentos. Os resultados indicam que o Random Forest é suficiente para previsões de um passo de tempo da produção de óleo e o LSTM é capaz de lidar com mais dados de entrada e estimar multiplos passos de tempo da produção de óleo. / [en] Precisely forecasting oil field performance is essential in oil reservoir planning and management. Nevertheless, forecasting oil production is a complex nonlinear problem due to all geophysical and petrophysical properties that may result in different effects with a bit of change. Thus, all decisions to be made during an exploitation project must consider different efficient algorithms to simulate data, providing robust scenarios to lead to the best deductions. To reduce the uncertainty in the simulation process, recent studies have efficiently introduced machine learning algorithms for solving reservoir engineering problems since they can extract the maximum information from the dataset. This thesis proposes using two machine learning techniques to predict the daily oil production of an offshore reservoir. Initially, the oil rate production is considered a time series and is pre-processed and restructured to fit a supervised learning problem. The Random Forest model is used to forecast a one-time step, which is an extension of decision tree learning, widely used in regression and classification problems for supervised machine learning. Regardless, the restrictions of this approach lead us to a more robust model, the LSTM RNN s, which are proposed by several studies as a suitable deep learning technique for time series modeling. Various configurations of LSTM RNN s were constructed to implement single-step and multi-step oil rate forecasting and down-hole pressure was incorporated to the inputs. For testing the robustness of the proposed models, we use four different datasets, three of them synthetically generated and one from a public real dataset, the Volve oil field, as a case study to conduct the experiments. The results indicate that the Random Forest model could sufficiently estimate the one-time step of the oil field production, and LSTM could handle more inputs and adequately estimate multiple-time steps of oil production.

[pt] Alguns projetos típicos de explotação de petróleo em águas profundas no Brasil exigem que se adotem alternativas de menor custo de desenvolvimento para que possam ser viabilizados. Nesses casos, deve-se também buscar reduzir o tempo para o desenvolvimento da produção, incluindo a perfuração de poços, o sistema de coleta submarino e a instalação das unidades de produção. Além disso, esses projetos podem ser significativamente afetados pelo regime fiscal vigente. É importante salientar que projetos de desenvolvimento offshore exercem um forte impacto sócio-econômico no país, não só pela geração de receita fiscal oriunda da produção petrolífera como também pela geração de emprego e renda no suprimento de bens e serviços para o desenvolvimento da produção. O presente trabalho tem por objetivo analisar uma nova proposta de um algoritmo de previsão de produção e da viabilidade econômica dos campos offshore baseado no regime fiscal vigente. / [en] Some typical deep water offshore reservoirs in Brazil requires the use some less expensive alternatives in order to make them economically attractive. In these cases we also need to reduce the development time including the perforation of the fields, the under water collecting system and the installation of the production units. These projects can also very affected according with the standing tax structure. It is also important to enforce that that offshore projects has a large impact on the social-economics of the country, not only for the generation of revenue form the tax income but also because of the job generation and for the needs of goods and services from the surrounding region. The present work has the objective to analyze a new numerical production algorithmic for offshore fields and also economic viability of the offshore fields based on the current tax structure.

This paper proposes an innovative way to address real cases of production prediction. This approach consists in the decomposition of original time series into time sub-series according to a group of factors in order to generate a predictive model from the partial predictive models of the sub-series. The adjustment of the models is carried out by means of a set of statistic techniques and Automatic Learning. This method was compared to an intuitive method consisting of a direct prediction of time series. The results show that this approach achieves better predictive performance than the direct way, so applying a decomposition method is more appropriate for this problem than non-decomposition.

Submitted by Gabriel Goulart Ferreira (gabrielggfsc@gmail.com) on 2017-08-24T21:23:43Z No. of bitstreams: 1 FGV Dissert - Final.pdf: 1134356 bytes, checksum: bae808529697b0480d6465ae8534ccf7 (MD5) / Approved for entry into archive by GILSON ROCHA MIRANDA (gilson.miranda@fgv.br) on 2017-08-25T13:40:06Z (GMT) No. of bitstreams: 1 FGV Dissert - Final.pdf: 1134356 bytes, checksum: bae808529697b0480d6465ae8534ccf7 (MD5) / Made available in DSpace on 2017-08-31T12:51:18Z (GMT). No. of bitstreams: 1 FGV Dissert - Final.pdf: 1134356 bytes, checksum: bae808529697b0480d6465ae8534ccf7 (MD5) Previous issue date: 2009-05-25 / This paper presents the analysis about the importance of the Consumer Confidence Indexes in the United States and the potential utilization of the Brazil’s similar index, the ICC FGV, to predict and track performance local economy. As practical result, this paper proposes a new alternative model to predict, in the short term, Monthly Industrial Production (PIM), a nationwide survey of industrial activity. / Esta dissertação apresenta análise sobre a importância dos indicadores de Confiança do Consumidor nos EUA e o potencial de utilização do indicador paralelo nacional, o ICC FGV, para previsão e acompanhamento do desempenho da economia brasileira. Como resultado prático, faz-se a proposição de novo modelo alternativo para previsão de curto prazo da PIM, Pesquisa Mensal Industrial do IBGE.