Utilizando segmentação de perfis de poços e o algoritmo DTW para o ajuste automático de perfis LWD e cabo

• Ano de defesa: 2024
• Autor(a) principal: Rozendo, Luiz Antonio
• Orientador(a): Silva, Diego Furtado
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos; Câmpus São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação - PPGCC
• Palavras-chave em Português: Segmentação de séries temporais; Segmentação de perfis de poços; Alinhamento LWD e cabo; Ajuste perfil LWD
• Palavras-chave em Inglês: Time series segmentation; Well log segmentation; LWD and wireline log alignment; LWD log adjustment
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/20194

The well logging phase in oil drilling aims to obtain rock characteristics or properties from indirect measurements acquired by sensors. Typically, these data are characterized by sequences of numerical values representing one or more soil properties. This phase can be executed through various approaches, with two of the most widely recognized and employed being: Logging While Drilling (LWD), which occurs during drilling, and Wireline Logging, which takes place after the well has been drilled. Due to vibrations and other phenomena occurring during drilling, there may be inaccuracies in the depth of LWD logs. To correct this, manual alignment of LWD logs is performed using Wireline logs as a reference. In this context, this work proposes a workflow that performs automatic alignment of LWD and Wireline logs using the Dynamic Time Warping (DTW) algorithm in a segmented manner. The proposed workflow is divided into three parts: finding the common region between the two series, segmenting the LWD and Wireline logs so that the segments are similar, and performing segmented alignment using the DTW algorithm. The results obtained during the experiments in this work demonstrated a significant improvement in the alignment performed by the DTW algorithm when conducted in a segmented manner, resulting in an LWD curve more faithful to the original.