Evaluating and predicting blastinduced ground vibration in opencast mine using ANN: a case study in Vietnam
 109 Downloads
Abstract
Blasting is one of the cheapest and effective methods for breaking rock mass in openpit mines. However, its side effects are not small such as ground vibration (PPV), air overpressure, fly rock, back break, dust, and toxic. Of these side effects, blastinduced PPV is the most dangerous for the human and surrounding environment. Therefore, evaluating and accurately forecasting blastinduced PPV is one of the most challenging issues facing openpit mines today. In this paper, a series of artificial neural network models were applied to predict blastinduced PPV in an openpit coal mine of Vietnam; 68 blasting events were used in this study for development of the ANN models. Of the whole dataset, 80% (approximately 56 observations) were used for the training process, and the rest of 20% (12 observations) were used for the testing process. Five ANN models were developed in this study with the difference in the number of hidden layers. The ANN 251; ANN 2861; ANN 2531; ANN 28641; and ANN 210851 models were considered in this study. An empirical technique was also conducted to estimate blastinduced PPV and compared to the constructed ANN models. For evaluating the performance of the models, rootmeansquared error (RMSE) and determination coefficient (R^{2}) were used. The results indicated that the ANN 210851 model (10 neurons in the first hidden layer, 8 neurons in the second hidden layer, and 5 neurons for the third hidden layer) yielded a superior performance over the other models with an RMSE of 0.738 and R^{2} of 0.964. In contrast, the empirical performed poorest performance with an RMSE of 2.670 and R^{2} of 0.768. This study is a new approach to predict blastinduced PPV in opencast mines aim to minimize the adverse effects of blasting operations on the surrounding environment.
Keywords
ANN Machine learning Blasting Ground vibration Openpit mine1 Introduction
Blasting is one of the cheapest and most effective methods for hardrock fragmentation in openpit mines and civil engineering. However, many previous researchers have concluded that only 25–30% of the explosive energy was directly involved in breaking rock, the rest produces undesirable side effects such as ground vibration (PPV), air overpressure, fly rock, and back break [1, 2, 3, 4, 5, 6, 7]. Of these side effects, PPV is the most dangerous side effects for humans and the environment. Therefore, precise prediction of blastinduced PPV is an essential requirement in openpit mines.
To assess and predict blastinduced PPV, many scientists have access to empirical techniques based on mathematical statistics [8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18]. Empirical techniques were mostly based on mathematical statistical methods and used two input variables including explosive charge per delay/maximum explosive charge capacity (W) and monitoring distance (R). They were considered the two most influential factors for blastinduced PPV [19]. However, their performance was not high in some cases [20, 21, 22, 23, 24, 25].
Some studies of PPV prediction using soft computational models
References  Technique  No. of datasets  Performance 

Monjezi et al. [37]  MLPNN  269  R^{2} = 0.954; RMSE = 0.03 
Monjezi et al. [38]  ANN  182  R^{2} = 0.949 
Armaghani et al. [39]  PSOANN  44  R^{2} = 0.930; MSE = 10.71 
Saadat et al. [21]  ANN  69  R^{2} = 0.957; MSE = 0.000722 
Hasanipanah et al. [6]  SVM  80  R^{2} = 0.957; RMSE = 0.340 
Amiri et al. [40]  ANNKNN  75  R^{2} = 0.880; RMSE = 0.540 
Hasanipanah et al. [41]  PSO  80  R^{2} = 0.938; RMSE = 0.240 
Faradonbeh and Monjezi [42]  GEP  115  R^{2} = 0.874; RMSE = 6.732 
Taheri et al. [23]  ABCANN  89  R^{2} = 0.920; RMSE = 0.220 
Armaghani et al. [43]  ICA  73  R^{2} = 0.940; RMSE = 0.370 
Behzadafshar et al. [44]  ICA  76  R^{2} = 0.939; RMSE = 0.320 
Abbaszadeh Shahri and Asheghi [45]  ANN  37  R^{2} = 0.954; RMSE = 0.157 
Mokfi et al. [46]  GMDH  102  R^{2} = 0.911; RMSE = 0.889 
Torres et al. [47]  MLREmpirical  178  R^{2} = 0.898 
Although the studies of PPV predictions in openpit mines using artificial intelligence (AI) have been approached, no approach or model was optimal for every area. In addition, AI is an approach that needs to be continually evolving and diverse. Thus, several ANN models for predicting blastinduced PPV at an openpit mine in Vietnam were developed in this study. The USBM empirical technique was also developed in this study to predict and compare with the ANN technique.
The composition of the article includes four parts: Part 1 reviews a number of published studies and the reasons for this study; Sect. 2 summarizes the data used and the methodology used; the results of this study and discussion are presented in Sect. 3; Finally, the conclusions and recommendations are given in the last section.
2 Data used and methodologies
2.1 Summary of the data used
Characteristic of the datasets used
W  R  PPV 

Min.: 136.0  Min.: 68.0  Min.: 17.80 
1st Qu.: 219.8  1st Qu.: 112.0  1st Qu.: 24.92 
Median: 305.5  Median: 134.5  Median: 27.30 
Mean: 316.5  Mean: 131.7  Mean: 27.47 
3rd Qu.: 405.0  3rd Qu.: 159.5  3rd Qu.: 29.61 
Max.: 670.0  Max.: 204.0  Max.: 35.63 
Before constructing the forecasting models, a data splitting procedure was performed for this study. Of the 68 observations, 80% of the whole data (56 observations) were used as the training dataset; the remaining 20% (12 observations) were used as the testing dataset. For the development of the PPV predictive models, the training dataset was used. For evaluating the performance of the models, the testing dataset was used as unseen data based on the developed models.
2.2 Empirical technique
2.3 Overview of ANN
ANN model works in the following manner: At the input layer, neurons receive input signals with weights. Then, they are processed and sent to the neurons of the first hidden layer via the transfer function. Here, the neurons will receive the results from the input and processing parameter classes, calculate the weights and send them to the second hidden layer via the transfer function. The process continues until the results are passed to the output layer and give the final output [53].
The results of the ANN model depend heavily on the learning process of the network, also known as the training process. The learning process of ANN includes two types of learning: supervised learning and unsupervised learning [54]. Input data for PPV prediction are numerical data and using regression algorithms, so most uses supervised learning based on input data and output requirements.
In this study, five ANN models with one, two, and three hidden layers were considered and developed to predict blastinduced PPV.
2.4 Establish the predictive models
Site factors and empirical equations for predicting PPV in this site study
No.  Site factors  Equation  

k  p  
1  9.520  − 0.359  \({\text{PPV}} = 9.520\left( {\frac{R}{{\sqrt[3]{W}}}} \right)^{0.359}\) 
2  6.640  − 0.480  \({\text{PPV}} = 6.640\left( {\frac{R}{{\sqrt[3]{W}}}} \right)^{0.480}\) 
3  9.427  − 0.362  \({\text{PPV}} = 9.427\left( {\frac{R}{{\sqrt[3]{W}}}} \right)^{0.362}\) 
4  6.602  − 0.482  \({\text{PPV}} = 6.602\left( {\frac{R}{{\sqrt[3]{W}}}} \right)^{0.482}\) 
5  9.378  − 0.364  \({\text{PPV}} = 9.378\left( {\frac{R}{{\sqrt[3]{W}}}} \right)^{0.364}\) 
In Fig. 3, I1 and I2 are input parameters corresponding to W and R; H1 to H10 is the number of neurons in each hidden layer; O1 is the output predicted value (PPV); B1 to B4 are biased layers that apply constant values to the nodes. The black line represents for positive weights, and the gray line represents for negative weights. Line thickness is in proportion to the magnitude of the weight relative to all others.
3 Results and discussion
3.1 Empirical models
Performance indices of the empirical models
No.  Training datasets  Testing datasets  

RMSE  R ^{2}  RMSE  R ^{2}  
Empirical 1  2.480  0.603  2.670  0.768 
Empirical 2  2.329  0.616  2.814  0.677 
Empirical 3  2.468  0.599  2.705  0.599 
Empirical 4  2.266  0.618  3.042  0.663 
Empirical 5  2.631  0.573  1.823  0.761 
Performance indices of the empirical models and their ranking
Method  Model  RMSE  R ^{2}  Rank for RMSE  Rank for R^{2}  Total rank 

Empirical  Training 1  2.48  0.603  2  3  5 
Training 2  2.329  0.616  4  4  8  
Training 3  2.468  0.599  3  2  5  
Training 4  2.266  0.618  5  5  10  
Training 5  2.631  0.573  1  1  2  
Testing 1  2.67  0.768  4  5  9  
Testing 2  2.814  0.677  2  3  5  
Testing 3  2.705  0.599  3  1  4  
Testing 4  3.042  0.663  1  2  3  
Testing 5  1.823  0.761  5  4  9 
Total rank of the empirical models
Method  Model no.  Total rank 

Empirical  1  14 
2  13  
3  9  
4  13  
5  11 
3.2 The ANN models
Performance indices of the ANN models
Model  Training datasets  Testing datasets  

RMSE  R ^{2}  RMSE  R ^{2}  
ANN 251  1.007  0.933  1.035  0.936 
ANN 2531  0.903  0.946  0.831  0.958 
ANN 2861  0.842  0.953  0.806  0.96 
ANN 28641  0.856  0.952  0.648  0.973 
ANN 210851  0.821  0.956  0.738  0.964 
Performance indices of the ANN models and their ranking
Method  Model  RMSE  R ^{2}  Rank for RMSE  Rank for R^{2}  Total rank 

ANN  Training 1  1.007  0.933  1  1  2 
Training 2  0.903  0.946  2  2  4  
Training 3  0.842  0.953  4  4  8  
Training 4  0.856  0.952  3  3  6  
Training 5  0.821  0.956  5  5  10  
Testing 1  1.035  0.936  1  1  2  
Testing 2  0.831  0.958  2  2  4  
Testing 3  0.806  0.96  3  3  6  
Testing 4  0.648  0.973  5  5  10  
Testing 5  0.738  0.964  4  4  8 
Total rank of the ANN models
Method  Model no.  Total rank 

ANN  1  4 
2  8  
3  14  
4  16  
5  18 
3.3 Performance evaluation between the ANN model and empirical model
Comparisons of performance between the empirical and ANN selected
Method  Training datasets  Testing datasets  

RMSE  R ^{2}  RMSE  R ^{2}  
Empirical  2.480  0.603  2.670  0.768 
ANN  0.821  0.956  0.738  0.964 
4 Conclusion and remarks

ANN is an advanced and robust technique that should be used to predict blastinduced PPV in openpit mines. This study has developed a robust ANN model with high accuracy (RMSE = 0.738, R^{2} = 0.964). It should be applied in practical engineering to control the undesirable effects on the surrounding environment. However, the development of the ANN models in openpit mine is often complicated, requiring the user to have an understanding of mathematics and programming.

The ANN model with a hidden layer should not be used to predict PPVs because it does not reflect all the characteristics of the data that lead to the inaccuracy of the forecasting model. The ANN model with three hidden layers has been proposed for predicting blastinduced PPV in practical engineering based on the results of this study. ANN models with many hidden layers should be considered in the future for predicting blastinduced PPV.

The empirical technique is a rapid and straightforward method of estimating blastinduced PPV, but further research is needed to improve the accuracy of empirical models.

The other influence parameters should be considered and supplemented to improve the accuracy of PPV predictive models, especially for empirical methods.
The results of this study are the basis for the development of blastinduced PPV predictive models for other openpit mines with similar conditions. At the same time, it is useful for managers, engineers, and blasters in optimizing blasting efficiency and minimizing the negative impacts caused by blasting operations in openpit mines.
Notes
Acknowledgements
We thank Hanoi University of Mining and Geology (HUMG), and the Center for Mining, ElectroMechanical research of HUMG.
Compliance with ethical standards
Conflict of interest
The author declares that there is no conflict of interests regarding the publication of this paper.
References
 1.Khandelwal M, Singh T (2005) Prediction of blast induced air overpressure in opencast mine. Noise Vib Worldw 36(2):7–16CrossRefGoogle Scholar
 2.Khandelwal M, Singh T (2009) Prediction of blastinduced ground vibration using artificial neural network. Int J Rock Mech Min Sci 46(7):1214–1222CrossRefGoogle Scholar
 3.Verma A, Singh T (2011) Intelligent systems for ground vibration measurement: a comparative study. Eng Comput 27(3):225–233CrossRefGoogle Scholar
 4.Monjezi M, Khoshalan HA, Varjani AY (2012) Prediction of flyrock and backbreak in open pit blasting operation: a neurogenetic approach. Arab J Geosci 5(3):441–448CrossRefGoogle Scholar
 5.Monjezi M, Hasanipanah M, Khandelwal M (2013) Evaluation and prediction of blastinduced ground vibration at Shur River Dam, Iran, by artificial neural network. Neural Comput Appl 22(7–8):1637–1643CrossRefGoogle Scholar
 6.Hasanipanah M, Monjezi M, Shahnazar A, Armaghani DJ, Farazmand A (2015) Feasibility of indirect determination of blast induced ground vibration based on support vector machine. Measurement 75:289–297CrossRefGoogle Scholar
 7.Hasanipanah M, Shahnazar A, Amnieh HB, Armaghani DJ (2017) Prediction of airoverpressure caused by mine blasting using a new hybrid PSO–SVR model. Eng Comput 33(1):23–31CrossRefGoogle Scholar
 8.Duvall WI, Petkof B (1958) Spherical propagation of explosiongenerated strain pulses in rock. Bureau of Mines. Report of Investigations 5483. US Department of the Interior, Bureau of Mines, Washington, DCGoogle Scholar
 9.Langefors U, Kihlstrom B (1963) The modern techniques of rock blasting. Wiley, New YorkGoogle Scholar
 10.Ambraseys NN, Hendron AJ (1968). In: Stagg KG, Zienkiewicz OC (eds) Rock mechanics in engineering practice. Wiley, LondonGoogle Scholar
 11.Standard I (1973) Criteria for safety and design of structures subjected to underground blast. ISI, IS6922Google Scholar
 12.Ghosh A, Daemen JJ (1983) A simple new blast vibration predictor (based on wave propagation laws). In: The 24th US symposium on rock mechanics (USRMS). American Rock Mechanics AssociationGoogle Scholar
 13.Roy PP (1991) Vibration control in an opencast mine based on improved blast vibration predictors. Min Sci Technol 12(2):157–165MathSciNetCrossRefGoogle Scholar
 14.Murmu S, Maheshwari P, Verma HK (2018) Empirical and probabilistic analysis of blastinduced ground vibrations. Int J Rock Mech Min Sci 103:267–274CrossRefGoogle Scholar
 15.Rai R, Singh T (2004) A new predictor for ground vibration prediction and its comparison with other predictors. Indian J Eng Mater Sci 11:178–184Google Scholar
 16.Ak H, Konuk A (2008) The effect of discontinuity frequency on ground vibrations produced from bench blasting: a case study. Soil Dyn Earthq Eng 28(9):686–694CrossRefGoogle Scholar
 17.Simangunsong GM, Wahyudi S (2015) Effect of bedding plane on prediction blastinduced ground vibration in open pit coal mines. Int J Rock Mech Min Sci 79:1–8CrossRefGoogle Scholar
 18.Kumar R, Choudhury D, Bhargava K (2016) Determination of blastinduced ground vibration equations for rocks using mechanical and geological properties. J Rock Mech Geotechn Eng 8(3):341–349CrossRefGoogle Scholar
 19.Siskind DE, Stachura VJ, Stagg MS, Kopp JW (1980) Structure response and damage produced by airblast from surface mining. Bureau of Mines. Report of Investigations 8485. US Department of the Interior, Bureau of Mines, Washington, DCGoogle Scholar
 20.Monjezi Ghafurikalajahi M, Bahrami A (2011) Prediction of blastinduced ground vibration using artificial neural networks. Tunn Undergr Space Technol 26(1):46–50CrossRefGoogle Scholar
 21.Saadat M, Khandelwal M, Monjezi M (2014) An ANNbased approach to predict blastinduced ground vibration of GolEGohar iron ore mine, Iran. J Rock Mech Geotechn Eng 6(1):67–76CrossRefGoogle Scholar
 22.Sheykhi H, Bagherpour R, Ghasemi E, Kalhori H (2017) Forecasting ground vibration due to rock blasting: a hybrid intelligent approach using support vector regression and fuzzy Cmeans clustering. Eng Comput 34(2):357–365. https://doi.org/10.1007/s0036601705466 CrossRefGoogle Scholar
 23.Taheri K, Hasanipanah M, Golzar SB, Majid MZA (2017) A hybrid artificial bee colony algorithmartificial neural network for forecasting the blastproduced ground vibration. Eng Comput 33(3):689–700CrossRefGoogle Scholar
 24.Nguyen H, Bui XN (2018) A comparison of artificial neural network and empirical technique for predicting blastinduced ground vibration in openpit mine. In: Mining sciences and technology—XXVI, Mong Cai, Hanoi, Vietnam. Industry and Trade of the Socialist Republic of Vietnam, pp 177–182Google Scholar
 25.Nguyen H, Bui XN, Thanh NT, Anh NT (2018) Effectiveness comparison between linear regression and nonlinear regression algorithms in predicting blastinduced ground vibration in Nui Beo openpit coal mine—Quang Ninh. J Min Ind 02:24–27Google Scholar
 26.Longjun D, Xibing L, Ming X, Qiyue L (2011) Comparisons of random forest and support vector machine for predicting blasting vibration characteristic parameters. Procedia Eng 26:1772–1781CrossRefGoogle Scholar
 27.Hasanipanah M, Faradonbeh RS, Amnieh HB, Armaghani DJ, Monjezi M (2017) Forecasting blastinduced ground vibration developing a CART model. Eng Comput 33(2):307–316CrossRefGoogle Scholar
 28.Chandar KR, Sastry V, Hegde C (2017) A critical comparison of regression models and artificial neural networks to predict ground vibrations. Geotech Geol Eng 35(2):573–583CrossRefGoogle Scholar
 29.Faradonbeh RS, Monjezi M (2017) Prediction and minimization of blastinduced ground vibration using two robust metaheuristic algorithms. Eng Comput 33(4):835–851. https://doi.org/10.1007/s0036601705016 CrossRefGoogle Scholar
 30.Faradonbeh RS, Armaghani DJ, Majid MA, Tahir MM, Murlidhar BR, Monjezi M, Wong H (2016) Prediction of ground vibration due to quarry blasting based on gene expression programming: a new model for peak particle velocity prediction. Int J Environ Sci Technol 13(6):1453–1464CrossRefGoogle Scholar
 31.Hasanipanah M, Golzar SB, Larki IA, Maryaki MY, Ghahremanians T (2017) Estimation of blastinduced ground vibration through a soft computing framework. Eng Comput 33(4):951–959. https://doi.org/10.1007/s003660170508z CrossRefGoogle Scholar
 32.Hajihassani M, Armaghani DJ, Marto A, Mohamad ET (2015) Ground vibration prediction in quarry blasting through an artificial neural network optimized by imperialist competitive algorithm. Bull Eng Geol Env 74(3):873–886CrossRefGoogle Scholar
 33.Armaghani DJ, Momeni E, Abad SVANK, Khandelwal M (2015) Feasibility of ANFIS model for prediction of ground vibrations resulting from quarry blasting. Environ Earth Sci 74(4):2845–2860CrossRefGoogle Scholar
 34.Monjezi M, Baghestani M, Faradonbeh RS, Saghand MP, Armaghani DJ (2016) Modification and prediction of blastinduced ground vibrations based on both empirical and computational techniques. Eng Comput 32(4):717–728CrossRefGoogle Scholar
 35.Ghoraba S, Monjezi M, Talebi N, Armaghani DJ, Moghaddam M (2016) Estimation of ground vibration produced by blasting operations through intelligent and empirical models. Environ Earth Sci 75(15):1137CrossRefGoogle Scholar
 36.Mohamad ET, Noorani SA, Armaghani DJ, Saad R (2012) Simulation of blasting induced ground vibration by using artificial neural network. Electron J Geotech Eng 17:2571–2584Google Scholar
 37.Monjezi M, Ahmadi M, Sheikhan M, Bahrami A, Salimi A (2010) Predicting blastinduced ground vibration using various types of neural networks. Soil Dyn Earthq Eng 30(11):1233–1236CrossRefGoogle Scholar
 38.Monjezi M, Ghafurikalajahi M, Bahrami A (2011) Prediction of blastinduced ground vibration using artificial neural networks. Tunn Undergr Space Technol 26(1):46–50CrossRefGoogle Scholar
 39.Armaghani DJ, Hajihassani M, Mohamad ET, Marto A, Noorani S (2014) Blastinginduced flyrock and ground vibration prediction through an expert artificial neural network based on particle swarm optimization. Arab J Geosci 7(12):5383–5396CrossRefGoogle Scholar
 40.Amiri M, Amnieh HB, Hasanipanah M, Khanli LM (2016) A new combination of artificial neural network and Knearest neighbors models to predict blastinduced ground vibration and airoverpressure. Eng Comput 32(4):631–644CrossRefGoogle Scholar
 41.Hasanipanah M, Naderi R, Kashir J, Noorani SA, Qaleh AZA (2017) Prediction of blastproduced ground vibration using particle swarm optimization. Eng Comput 33(2):173–179CrossRefGoogle Scholar
 42.Faradonbeh RS, Monjezi M (2017) Prediction and minimization of blastinduced ground vibration using two robust metaheuristic algorithms. Eng Comput 33(4):835–851CrossRefGoogle Scholar
 43.Armaghani DJ, Hasanipanah M, Amnieh HB, Mohamad ET (2018) Feasibility of ICA in approximating ground vibration resulting from mine blasting. Neural Comput Appl 29(9):457–465CrossRefGoogle Scholar
 44.Behzadafshar K, Mohebbi F, Soltani Tehrani M, Hasanipanah M, Tabrizi O (2018) Predicting the ground vibration induced by mine blasting using imperialist competitive algorithm. Eng Comput 35(4):1774–1787CrossRefGoogle Scholar
 45.Abbaszadeh Shahri A, Asheghi R (2018) Optimized developed artificial neural networkbased models to predict the blastinduced ground vibration. Innov Infrastruct Solut 3:1–10CrossRefGoogle Scholar
 46.Mokfi T, Shahnazar A, Bakhshayeshi I, Derakhsh AM, Tabrizi O (2018) Proposing of a new soft computingbased model to predict peak particle velocity induced by blasting. Eng Comput 34(4):881–888. https://doi.org/10.1007/s0036601805786 CrossRefGoogle Scholar
 47.Torres VN, Silveira LG, Lopes PF, de Lima HM (2018) Assessing and controlling of bench blastinginduced vibrations to minimize impacts to a neighboring community. J Clean Prod 187:514–524CrossRefGoogle Scholar
 48.Bach NV, Thanh NV (1996) The effects of blasting operations in the surrounding structures. J Min Ind 4:13–14 (in Vietnamese) Google Scholar
 49.Bach NV, Quyen LV, Nam BX, An ND, Phuc NV (2006) Reducing the undesirable effects of blastinduced ground vibration in Nui Beo openpit coal mine. J Sci Technol Hanoi Univ Min Geol 14:58–62 (in Vietnamese) Google Scholar
 50.Bach NV, Nam BX, An ND, Hung TK (2012) Determination of blastinduced ground vibration for nonelectric delay blasting. J Sci Technol Hanoi Univ Min Geol 38:25–28 (in Vietnamese) Google Scholar
 51.Schalkoff RJ (1997) Artificial neural networks, vol 1. McGrawHill, New YorkzbMATHGoogle Scholar
 52.Nguyen H, Bui XN, Bui HB, Mai NL (2018) A comparative study of artificial neural networks in predicting blastinduced airblast overpressure at Deo Nai openpit coal mine, Vietnam. Neural Comput Appl. https://doi.org/10.1007/s0052101837175 CrossRefGoogle Scholar
 53.Zerguine A, Shafi A, Bettayeb M (2001) Multilayer perceptronbased DFE with lattice structure. IEEE Trans Neural Netw 12(3):532–545CrossRefGoogle Scholar
 54.Perez LG, Flechsig AJ, Meador JL, Obradovic Z (1994) Training an artificial neural network to discriminate between magnetizing inrush and internal faults. IEEE Trans Power Deliv 9(1):434–441CrossRefGoogle Scholar
 55.Norusis M (2008) SPSS 16.0 advanced statistical procedures companion. Prentice Hall Press, Upper Saddle RiverGoogle Scholar