Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Comprehensive analyses of transport phenomena and thermodynamics of complex multiphysics systems are laborious and computationally intensive. Yet, such analyses are often required during the design of thermal and process equipment. As a remedy, this paper puts forward a novel approach to the prediction of transport behaviours of multiphysics systems, offering significant reductions in the computational time and cost. This is based on machine learning techniques that utilize the data generated by computational fluid dynamics for training purposes. The physical system under investigation includes a stagnation-point flow of a hybrid nanofluid (Cu−Al2O3/Water) over a blunt object embedded in porous media. The problem further involves mixed convection, entropy generation, local thermal non-equilibrium and non-linear thermal radiation within the porous medium. The SVR (Support Machine Vector) model is employed to approximate velocity, temperature, Nusselt number and shear-stress as well as entropy generation and Bejan number functions. Further, PSO meta-heuristic algorithm is applied to propose correlations for Nusselt number and shear stress. The effects of Nusselt number, temperature fields and shear stress on the surface of the blunt-body as well as thermal and frictional entropy generation are analysed over a wide range of parameters. Further, it is shown that the generated correlations allow a quantitative evaluation of the contribution of a large number of variables to Nusselt number and shear stress. This makes the combined computational and artificial intelligence (AI) approach most suitable for design purposes.

A machine learning approach to the prediction of transport and thermodynamic processes in multiphysics systems - heat transfer in a hybrid nanofluid flow in porous media

Modeling the SOFC by BP neural network algorithm

ABSTRACT Machine learning (ML) is widely used in many industries, resulting in recent interests to explore ML in geotechnical engineering. Past review papers focus mainly on ML algorithms while this paper advocates an agenda to put data at the core, to develop novel algorithms that are effective for geotechnical data (existing and new), to address the needs of current practice, to exploit new opportunities from emerging technologies or to meet new needs from digital transformation, and to take advantage of current knowledge and accumulated experience. This agenda is called data-centric geotechnics and it contains three core elements: data centricity, fit for (and transform) practice, and geotechnical context. The future of machine learning in geotechnics should be envisioned with this “data first practice central” agenda in mind. Data-driven site characterization (DDSC) is an active research topic in this agenda because an understanding of the ground is crucial in all projects. Examples of DDSC challenges are ugly data and explainable site recognition. Additional challenges include making ML indispensable (ML supremacy), learning how to learn (meta-learning), and becoming smart (digital twin).

Future of machine learning in geotechnics

Key factors of marine shale conductivity in southern China—Part II: The influence of pore system and the development direction of shale gas saturation models

This paper presents a detailed numerical trend analysis performed on a mechanistic compositional model using a multiple fractured, single well huff-n-puff technique. A wide range of formation and fluid properties, hydraulic fracture, and operational design considerations with several injectant types using single and multiple huff-n-puff cycles are investigated with different operational alignments. The objective of this study is very well achieved, which is to determine a fluid flow performance trend under different circumstances such that to come up with an optimal EOR and hydraulic fracture design to develop a UHR.

WITHDRAWN: Unconventional EOR Applications in Unconventional Hydrocarbon Reservoirs – Numerical Trend Analysis

This paper presents an assessment of infill drilling opportunities in a complex multi-layered heterogeneous carbonate formation located in Abu Dhabi offshore. The subject field was developed last century and is currently undergoing further development with line drive horizontal wells. The field is being redeveloped for improved oil recovery at higher production rates with long horizontal wells. An infill assessment process is defined using sector model reservoir simulations for the specific reservoir. Reservoir simulations are performed on pattern sector models to establish the optimum grid for infill evaluation. Also, the models with an appropriate grid size are used to optimize infill placement (vertical and lateral well placement) and infill drilling timing for a couple of geologically similar areas. In the second step, the sector model results are applied to test the full-field infill development plan. The relatively homogeneous geological area shows very uniform displacement with 1-km spaced wells and gives no considerable benefit of incremental recovery through infill drilling. However, in a comparatively heterogeneous geological area, considerable incremental oil recovery is quantified. In brief, this paper presents a detailed infill drilling and well placement assessment process workflow for the re-development of a multi-layered heterogeneous reservoir.

/pdf/infill-drilling-and-well-placement-assessment-for-a-multi-5dvoqipj94.pdf

Infill drilling and well placement assessment for a multi-layered heterogeneous reservoir

Shale oil reservoir is one of the promising unconventional oil reservoirs with large deposits worldwide. Due to its complex nature and development strategy planning, understanding the behavior and devising a development strategy through numerical simulation is inexpensive and time-consuming. Hence, this study proposes to implement an Artificial neural network to understand the mechanism involved in the shale oil matrix and predict its performance. This study involves, at first, generating a mechanistic model in order to develop a dataset featuring the responses of fracture and rock properties on shale oil recovery. Later, an artificial neural network is trained and tested to predict the oil recovery under the influence of a defined set of fracture and rock properties. The neural network is validated by k-fold cross-validation. Moreover, an optimizer is used to minimize the loss in the trained proxy model. The results show that the generated proxy model efficiently predicts the oil recovery on different fracture and rock properties. Moreover, the model results are found to be consistent with the trend analysis showing the robust nature of the model in predicting oil recovery.

Amirmasoud K. Dahaghi

Papers

WITHDRAWN: Unconventional EOR Applications in Unconventional Hydrocarbon Reservoirs – Numerical Trend Analysis

Infill drilling and well placement assessment for a multi-layered heterogeneous reservoir

Subsurface Physics Inspired Neural Network to Predict Shale Oil Recovery under the Influence of Rock and Fracture Properties

Low-Rank Tensors Applications for Dimensionality Reduction of Complex Hydrocarbon Reservoirs

Low-rank tensors applications for dimensionality reduction of complex hydrocarbon reservoirs