All over the world, shear strengthening of reinforced concrete elements using external fiber-reinforced polymer jackets could be used to improve building sustainability. However, reports issued by ...the American Concrete Institute called for heavy scrutiny before actual field implementation. The very limited number of proposed shear equations lacks reliability and accuracy. Thus, further investigation in this area is needed. In addition, machine-learning techniques are being implemented successfully to develop strength models for complex problems including shear, flexure, and torsion. This study aims to provide a reliable machine-learning model for reinforced concrete beams strengthened in shear using externally reinforced fiber polymer sheets. The proposed model was developed and validated against the experimental database and the very limited models in existing literature. The model showed better agreement with the experimentally measured strength compared to the previous models, which accounted for the effect of various parameters including but not limited to: the element geometry, strengthening details, and configurations. The model could guide the further developments of design codes and mechanical models.
•Extensively compared 10 machine learning algorithms for predicting rate of penetration (ROP).•The significance of feature selection, noise removal, data scaling, and managing features' ...dimensionality was emphasized.•The importance of understanding and considering prediction uncertainty in model evaluation was quantified.•Extreme tree (ET), light gradient boost (LGBoost), random forest (RF), and gradient boosting (GBR) algorithms demonstrated superior performance for ROP prediction.
Enhanced Geothermal Systems (EGS) present a compelling means to unlock the considerable yet largely untapped thermal energy within the earth's crust globally. The EGS drilling cost constitutes a substantial portion, up to 60 %, of the overall expenses. Consequently, streamlining and optimizing the drilling processes for these systems hold immense economic significance, with the potential to substantially lower costs, advance the utilization of geothermal energy, and contribute significantly to the reduction of carbon emissions. This research aims to enhance EGS drilling operations by seeking ways of reducing drilling cost.
In this research, we harnessed the predictive power of 10 state-of-the-art machine learning (ML) algorithms to anticipate a crucial drilling parameter: ROP. Using the FORGE dataset, we developed a code tailored for the intensive preprocessing of drilling data. This code offers many options, including various noise-removing techniques and scaling approaches. However, the primary focus of our work extends to quantifying uncertainties intrinsic to the predictions of the 10 algorithms employed. To achieve this, a comprehensive approach involved subjecting each algorithm to 40 runs, utilizing the best model from the tuning process.
The results show that unaddressed uncertainties may lead to unstable model behavior, where small changes in the input data or random initialization result in significant prediction variations. Thus, models that do not account for uncertainty may be overfit to the noise in the training data, leading to poor generalization to new, unseen data. Remarkably, the results highlight the superiority of the Extreme Tree, Light Gradient Boost, Random Forest, and Gradient Boosting algorithms, showcasing mean absolute error (MAE) and R2 values within the ranges of 3.5–4.5 and 0.9–0.95, respectively. Conversely, artificial neural networks and support vector machine algorithms demonstrate comparatively lower performance, with MAE ranging from 4.7 to 6.3 and R2 from 0.85 to 0.91.
Although the results presented in this work are only based on the drilling parameters from one well of the FORGE site and do not include rock properties or geological parameters, however, the proposed nuanced understanding of algorithmic performance is valuable for refining predictive models in geothermal drilling applications, ensuring robustness and reliability in the face of diverse operational scenarios. Navigating a successful workflow, adeptly processing data, and meticulously quantifying uncertainties linked to each predictive algorithm emerge as formidable yet indispensable tasks. Our research not only sheds light on these complexities but also paves the way for a strategic optimization approach based on data-driven ROP models. Therefore, the potential for substantial savings in real-time drilling operations becomes apparent, unlocking a new realm of possibilities for the geothermal energy sector.
Mud filtrate invasion is a vital parameter that should be optimized during drilling for oil and gas to reduce formation damage. Nanoparticles (NPs) have shown promising filtrate loss mitigation when ...used as drilling fluid (mud) additives in numerous recent studies. Modeling the influence of NPs can fasten the process of selecting their optimum type, size, concentration, etc. to meet the drilling conditions. In this study, a model was developed, using artificial neural network (ANN), to predict the filtrate invasion of nano-based mud under wide range of pressures and temperatures up to 500 psi and 350 °F, respectively. A total of 2,863 data points were used in the development of the model (806 data points were collected form conducted experiments and the rest were collected form the literature). Seven different types of NPs with size and concentration ranges from 15 to 50 nm and 0 to 2.5 wt%, respectively, had been included in the model to ensure universality. The dataset was divided into 70 % for training and 30 % for validation. A total of 6,750 different combinations for the model’s hyperparameters were evaluated to determine the optimum combination. The N-encoded method was used to convert the categorical data into numerical values. The model was evaluated through calculating the statistical parameters. The developed ANN-model proofed to be efficient in predicting the filtrate invasion at different pressures and temperatures with an average absolute relative error (AARE) of less than 0.5 % and a coefficient of determination (R2) of more than 0.99 for the overall data. The ANN-model covers wide range of pressures, temperatures as well as various NPs’ types, concentrations, and sizes, which confirms its useability and coverability.
HIGHLIGHTS
Artificial neural network (ANN)-model was developed to predict the volume of filtrate of water-based mud (WBM) modified with nanoparticles (NPs)
A total of 2,863 data points were collected to build the ANN-model from both experimental work and literature considering 3 types of WBM modified with 7 types of NPs (SiO2, TiO2, Al2O3, CuO, MgO, ZnO, Fe2O3) with size and concentration ranges from 15 to 50 nm and 0 to 2.5 wt%, respectively, under wide range of pressures and temperatures up to 500 psi and 350 °F
A total of 6,750 different combinations for the model’s hyperparameters were evaluated to determine the optimum combination and the N-encoded method was used to convert the categorical data into numerical values
The ANN-model proofed to be efficient with an average absolute relative error (AARE) of less than 0.5 % and a coefficient of determination (R2) of more than 0.99 for the overall data
GRAPHICAL ABSTRACT
