- Title
- Prediction of blast-induced ground vibration at a limestone quarry : an artificial intelligence approach
- Creator
- Arthur, Clement; Bhatawdekar, Ramesh; Mohamad, Edy; Sabri, Mohanad; Bohra, Manish; Khandelwal, Manoj; Kwon, Sangki
- Date
- 2022
- Type
- Text; Journal article
- Identifier
- http://researchonline.federation.edu.au/vital/access/HandleResolver/1959.17/190017
- Identifier
- vital:17513
- Identifier
-
https://doi.org/10.3390/app12189189
- Identifier
- ISSN:2076-3417 (ISSN)
- Abstract
- Ground vibration is one of the most unfavourable environmental effects of blasting activities, which can cause serious damage to neighboring homes and structures. As a result, effective forecasting of their severity is critical to controlling and reducing their recurrence. There are several conventional vibration predictor equations available proposed by different researchers but most of them are based on only two parameters, i.e., explosive charge used per delay and distance between blast face to the monitoring point. It is a well-known fact that blasting results are influenced by a number of blast design parameters, such as burden, spacing, powder factor, etc. but these are not being considered in any of the available conventional predictors and due to that they show a high error in predicting blast vibrations. Nowadays, artificial intelligence has been widely used in blast engineering. Thus, three artificial intelligence approaches, namely Gaussian process regression (GPR), extreme learning machine (ELM) and backpropagation neural network (BPNN) were used in this study to estimate ground vibration caused by blasting in Shree Cement Ras Limestone Mine in India. To achieve that aim, 101 blasting datasets with powder factor, average depth, distance, spacing, burden, charge weight, and stemming length as input parameters were collected from the mine site. For comparison purposes, a simple multivariate regression analysis (MVRA) model as well as, a nonparametric regression-based technique known as multivariate adaptive regression splines (MARS) was also constructed using the same datasets. This study serves as a foundational study for the comparison of GPR, BPNN, ELM, MARS and MVRA to ascertain their respective predictive performances. Eighty-one (81) datasets representing 80% of the total blasting datasets were used to construct and train the various predictive models while 20 data samples (20%) were utilized for evaluating the predictive capabilities of the developed predictive models. Using the testing datasets, major indicators of performance, namely mean squared error (MSE), variance accounted for (VAF), correlation coefficient (R) and coefficient of determination (R2) were compared as statistical evaluators of model performance. This study revealed that the GPR model exhibited superior predictive capability in comparison to the MARS, BPNN, ELM and MVRA. The GPR model showed the highest VAF, R and R2 values of 99.1728%, 0.9985 and 0.9971 respectively and the lowest MSE of 0.0903. As a result, the blast engineer can employ GPR as an effective and appropriate method for forecasting blast-induced ground vibration. © 2022 by the authors.
- Publisher
- MDPI
- Relation
- Applied Sciences (Switzerland) Vol. 12, no. 18 (2022), p.
- Rights
- All metadata describing materials held in, or linked to, the repository is freely available under a CC0 licence
- Rights
- https://creativecommons.org/licenses/by/4.0/
- Rights
- Copyright © 2022 by the authors.
- Rights
- Open Access
- Subject
- MD Multidisciplinary; Artificial intelligence; Backpropagation neural network; Blast-induced ground vibration; Gaussian process regression
- Full Text
- Reviewed
- Funder
- The research is partially funded by the Ministry of Science and Higher Education of the Russian Federation under the strategic academic leadership program ‘Priority 2030′ (Agreement 075-15-2021-1333 dated 30 September 2021).
- Hits: 516
- Visitors: 553
- Downloads: 54
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | SOURCE1 | Published version | 2 MB | Adobe Acrobat PDF | View Details Download |