Ensemble Learning Regression for Estimating Unconfined Compressive Strength of Cemented Paste Backfill

Lü, Xiang; Zhou, Wei; Ding, Xiaohua; Shi, Xuyang; Luan, Boyu; Li, Ming

doi:10.1109/access.2019.2918177

Cited by 64 publications

(26 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…At the same time, an intelligent modelling framework for the mechanical properties prediction using machine learning (ML) algorithms and genetic algorithm (GA) was proposed [15][16][17]. Lu et al [18] improved a method to estimate the unconfined compressive strength of CPB. e strength of paste is usually determined mainly on the demand of the adopted mining technology.…”

Section: Introductionmentioning

confidence: 99%

A Novel Prediction Model of Strength of Paste Backfill Prepared from Waste-Unclassified Tailings

Zhang

2019

Advances in Materials Science and Engineering

View full text Add to dashboard Cite

Paste backfilling is an important support for the development of green mines and deep mining. It can effectively reduce a series of risks of underground goaf and surface tailings ponds. Reasonable strength of backfill is an effective guarantee for controlling ground stress and realizing safe mining function. Under the combination of complex materials and local conditions, ensuring the optimal design and effective proportion for paste backfill strength is the bottleneck problem that restricts the safety, economy, and efficiency of filling mining. The strength developing trend of paste backfilling prepared from waste rock and unclassified tailings has been studied. Different levels of cement contents, tailings-waste ratios, and slurry concentrations were investigated through orthogonal design to obtain the relationship between the UCS and the multi-influential factors. Combined with the experimental results and the previous strength prediction models, the waste rock-unclassified tailings paste strength prediction model was proposed. Introducing the water-cement ratio, the cement-tailings ratio, the amount of cement, and the packing density that characterizing the overall gradation of unclassified tailings and waste rock, as well as the curing time, a strength prediction model of multifactors was developed. Moreover, the microscopic structure of the paste prepared from waste-unclassified tailings was analyzed with an Environment Scanning Electron Microscope (ESEM), and the influence mechanism was ascertained. The weight coefficient of strength development is carded in this paper, and the strength model of unclassified tailings-waste paste considering five factors is obtained, which is of great significance to guide the mining engineering.

show abstract

Section: Introductionmentioning

confidence: 99%

A Novel Prediction Model of Strength of Paste Backfill Prepared from Waste-Unclassified Tailings

Zhang

2019

Advances in Materials Science and Engineering

View full text Add to dashboard Cite

show abstract

“…Support Vector Regression (SVR) is a nonparametric ML model which can deal with nonlinear regression problems using kernel function [24]. Extreme Gradient Boosting (XGBoost) is a boosting algorithm based on decision tree [25], and therefore it enables users to analyze the relative importance of features in prediction response. Artificial Neural Network (ANN) is a biologically-inspired method to tackle complex ML tasks with a large number of simple processing elements called neurons [26].…”

Section: Introductionmentioning

confidence: 99%

On the Use of Machine Learning Models for Prediction of Compressive Strength of Concrete: Influence of Dimensionality Reduction on the Model Performance

Wan

Šavija

2021

Materials

View full text Add to dashboard Cite

Compressive strength is the most significant metric to evaluate the mechanical properties of concrete. Machine learning (ML) methods have shown promising results for predicting compressive strength of concrete. However, at present, no in-depth studies have been devoted to the influence of dimensionality reduction on the performance of different ML models for this application. In this work, four representative ML models, i.e., Linear Regression (LR), Support Vector Regression (SVR), Extreme Gradient Boosting (XGBoost), and Artificial Neural Network (ANN), are trained and used to predict the compressive strength of concrete based on its mixture composition and curing age. For each ML model, three kinds of features are used as input: the eight original features, six Principal Component Analysis (PCA)-selected features, and six manually selected features. The performance as well as the training speed of those four ML models with three different kinds of features is assessed and compared. Based on the obtained results, it is possible to make a relatively accurate prediction of concrete compressive strength using SVR, XGBoost, and ANN with an R-square of over 0.9. When using different features, the highest R-square of the test set occurs in the XGBoost model with manually selected features as inputs (R-square = 0.9339). The prediction accuracy of the SVR model with manually selected features (R-square = 0.9080) or PCA-selected features (R-square = 0.9134) is better than the model with original features (R-square = 0.9003) without dramatic running time change, indicating that dimensionality reduction has a positive influence on SVR model. For XGBoost, the model with PCA-selected features shows poorer performance (R-square = 0.8787) than XGBoost model with original features or manually selected features. A possible reason for this is that the PCA-selected features are not as distinguishable as the manually selected features in this study. In addition, the running time of XGBoost model with PCA-selected features is longer than XGBoost model with original features or manually selected features. In other words, dimensionality reduction by PCA seems to have an adverse effect both on the performance and the running time of XGBoost model. Dimensionality reduction has an adverse effect on the performance of LR model and ANN model because the R-squares on test set of those two models with manually selected features or PCA-selected features are lower than models with original features. Although the running time of ANN is much longer than the other three ML models (less than 1s) in three scenarios, dimensionality reduction has an obviously positive influence on running time without losing much prediction accuracy for ANN model.

show abstract

“…The surface subsidence risk after mining can be reduced by pumping the PTT underground as CPB materials [5][6][7]. Tailings are generally used as fine aggregates in CPB materials to support the goaf surrounding rock.…”

Section: Introductionmentioning

confidence: 99%

Pore Connectivity and Dewatering Mechanism of Tailings Bed in Raking Deep-Cone Thickener Process

et al. 2020

View full text Add to dashboard Cite

Paste and thickened tailings (PTT) technology can improve the utilization and management of tailings from processing plants. The pore size distribution (PSD) and microstructure evolution affected by the rake shear in thickening tailings beds are essential to produce a high-density tailings underflow. Continuous thickening and computed tomography (CT) scanning tests were conducted to study the PSD with and without shear. The pore morphology was studied to reveal the shearing-dewatering performance of the tailings bed. The results show that at a flocculant solution concentration of 0.01 wt % and a feed slurry concentration of 10 wt%, the underflow concentration with and without shear can reach 58.5 wt %and 55.8 wt %, respectively. The CT image reconstruction models demonstrated that the porosity of the sheared tailings bed increased with the bed height. When the bed height increased from 2.5 to 10 cm, the porosity increased from 35.1% to 41.9%, the pore fractal dimension increased from the range 1.8–1.95 to the range 2.1–2.15, and the pore quantity decreased by 21.39%. The average pore volume increased with increasing height by 13.93%, 16.57% and 12.07%. The pore structure became more complex with the bed height, and the connectivity between pores increased to form water-flow channels, which were beneficial to the drainage of sealed water.

show abstract

Ensemble Learning Regression for Estimating Unconfined Compressive Strength of Cemented Paste Backfill

Cited by 64 publications

References 58 publications

A Novel Prediction Model of Strength of Paste Backfill Prepared from Waste-Unclassified Tailings

A Novel Prediction Model of Strength of Paste Backfill Prepared from Waste-Unclassified Tailings

On the Use of Machine Learning Models for Prediction of Compressive Strength of Concrete: Influence of Dimensionality Reduction on the Model Performance

Pore Connectivity and Dewatering Mechanism of Tailings Bed in Raking Deep-Cone Thickener Process

Contact Info

Product

Resources

About