Using random forest for the risk assessment of coal-floor water inrush in Panjiayao Coal Mine, northern China

  • Dekang Zhao
  • Qiang Wu
  • Fangpeng Cui
  • Hua Xu
  • Yifan Zeng
  • Yufei Cao
  • Yuanze Du
Paper
  • 25 Downloads

Abstract

Coal-floor water-inrush incidents account for a large proportion of coal mine disasters in northern China, and accurate risk assessment is crucial for safe coal production. A novel and promising assessment model for water inrush is proposed based on random forest (RF), which is a powerful intelligent machine-learning algorithm. RF has considerable advantages, including high classification accuracy and the capability to evaluate the importance of variables; in particularly, it is robust in dealing with the complicated and non-linear problems inherent in risk assessment. In this study, the proposed model is applied to Panjiayao Coal Mine, northern China. Eight factors were selected as evaluation indices according to systematic analysis of the geological conditions and a field survey of the study area. Risk assessment maps were generated based on RF, and the probabilistic neural network (PNN) model was also used for risk assessment as a comparison. The results demonstrate that the two methods are consistent in the risk assessment of water inrush at the mine, and RF shows a better performance compared to PNN with an overall accuracy higher by 6.67%. It is concluded that RF is more practicable to assess the water-inrush risk than PNN. The presented method will be helpful in avoiding water inrush and also can be extended to various engineering applications.

Keywords

Water inrush Risk assessment Mining Random forest China 

Utilisation d’une forêt aléatoire pour l’évaluation des risques liés à l’irruption de l’eau dans le charbon de la mine de charbon de Panjiayao, dans le nord de la Chine

Résumé

Les incidents liés à l’irruption de l’eau dans le charbon représentent une grande partie des catastrophes causées par les mines de charbon dans le nord de la Chine, et une évaluation précise des risques est cruciale pour la production de charbon en toute sécurité. Un modèle d’évaluation novateur et prometteur pour l’appel d’eau est proposé basé sur la forêt aléatoire (RF), qui est un puissant algorithme intelligent d’apprentissage automatique. Les RF présentent des avantages considérables, notamment une précision de classification élevée et la capacité d’évaluer l’importance des variables; en particulier, il résiste aux problèmes compliqués et non linéaires inhérents à l’évaluation des risques. Dans cette étude, le modèle proposé est appliqué à la mine de charbon de Panjiayao, dans le nord de la Chine. Huit facteurs ont été sélectionnés comme indices d’évaluation en fonction de l’analyze systématique des conditions géologiques et d’une étude de terrain de la zone d’étude. Des cartes d’évaluation des risques ont été générées sur la base de RF, et le modèle de réseau neuronal probabiliste (PNN) a également été utilisé pour l’évaluation des risques en tant que comparaison. Les résultats démontrent que les deux méthodes sont cohérentes dans l’évaluation du risque d’irruption de l’eau à la mine, et que la RF montre une meilleure performance par rapport au PNN avec une précision globale supérieure de 6.67%. Il est conclu que la RF est plus pratique pour évaluer le risque d’irruption d’eau que PNN. La méthode présentée sera utile pour éviter l’appel d’eau et peut également être étendue à diverses applications d’ingénierie.

Usando Bosque Aleatorio para la evaluación de riesgo de afluencias de agua en mina Panjiayao, una mina de carbón en el Norte de China

Resumen

Incidentes de filtraciones de agua en suelos de carbón representan una gran proporción de los desastres en minas de carbón en el norte de China, y una evaluación precisa de los riesgos es crucial para la producción de carbón. Un modelo novedoso y prometedor para simular los flujos de agua ha sido propuesto basado en el bosque aleatorio (RF por sus siglas en inglés), el cual es un potente algoritmo de aprendizaje de máquinas inteligentes. RF tiene ventajas considerables, incluida su alta precisión de clasificación y la capacidad para evaluar la importancia de las variables; en particular, es robusto en lidiar con los problemas complejos y no lineales inherentes en la evaluación de riesgos. En este estudio, el modelo propuesto se aplica a la mina de carbón Panjiayao, en el norte de China. Ocho factores fueron seleccionados como los índices de evaluación de acuerdo a un análisis sistemático de las condiciones geológicas y un estudio de campo de la zona de estudio. Mapas de asesorías de riesgos fueron generados sobre la base de RF, y el modelo de red neuronal probabilística (PNN por sus siglas en inglés) también fue utilizado para la evaluación de riesgos como punto de comparación. Los resultados demuestran que los dos métodos son consecuentes con evaluaciones de riesgo de filtraciones de agua en la mina, y RF muestra un mejor rendimiento en comparación con PNN con una precisión superior por 6.67%. Se concluye que RF es más eficiente para evaluar el riesgo de filtraciones de agua que PNN. El presente método será útil para evitar filtraciones de agua y también puede extenderse a diferentes aplicaciónes de ingeniería.

基于随机森林的中国北方潘家窑煤矿煤层底板突水危险性评价

摘要

煤矿底板突水事故在中国北方煤矿灾害中占很大比例,准确的风险评估是煤矿安全生产的关键。本文提出了一种基于随机森林的突水评价模型,该模型是一种功能强大的智能机器学习算法。随机算法具有很高的分类精度和评估变量重要性的能力,特别是在处理复杂和非线性的问题时具有很强的鲁棒性。本研究将该模型应用于我国北方潘家窑煤矿。根据对地质条件的系统分析和对研究区的实地调查,选取了8个主控因素作为评价指标。基于随机森林模型生成了危险性评价图,并采用概率神经网络模型进行对比。结果表明,这两种方法在矿井突水危险性评价中是基本一致的,随机森林模型与概率神经网络模型相比具有更好的性能,总体精度提高了6.67%。在评价突水危险中,随机森林模型比概率神经网络模型更有效。本文提出的方法既有利于矿井突水的防治,也可以推广到各种工程应用中。

Uso de floresta aleatória para a avaliação do risco de inrush da água do piso de carvão na mina de carvão de Panjiayao, norte da China

Resumo

Incidentes de entrada de água em piso de carvão são responsáveis por uma grande proporção de desastres em minas de carvão no norte da China, e a avaliação precisa dos riscos é crucial para a produção segura de carvão. Um modelo de avaliação inovador e promissor para o inrush da água é proposto com base na floresta aleatória (RF), que é um poderoso algoritmo inteligente de aprendizado de máquina. RF tem vantagens consideráveis, incluindo alta precisão de classificação e capacidade de avaliar a importância das variáveis; em particular, é robusto ao lidar com os problemas complicados e não lineares inerentes à avaliação de risco. Neste estudo, o modelo proposto é aplicado à Mina de Carvão Panjiayao, norte da China. Oito fatores foram selecionados como índices de avaliação de acordo com a análise sistemática das condições geológicas e um levantamento de campo da área de estudo. Mapas de avaliação de risco foram gerados com base em RF, e o modelo de rede neural probabilística (PNN) também foi utilizado para a avaliação de risco como comparação. Os resultados demonstram que os dois métodos são consistentes na avaliação de risco de inrush de água na mina, e RF mostra um melhor desempenho comparado ao PNN com uma precisão geral maior de 6.67%. Conclui-se que a RF é mais praticável para avaliar o risco de entrada de água do que o PNN. O método apresentado será útil para evitar o inrush da água e também pode ser estendido para várias aplicações de engenharia.

Notes

Acknowledgements

The authors would like to thank the editor and reviewers for their constructive suggestions.

References

  1. Adam E, Mutanga O, Odindi J, Abdelrahman EM (2014) Land-use/cover classification in a heterogeneous coastal landscape using RapidEye imagery: evaluating the performance of random forest and support vector machines classifiers. Int J Remote Sens 35(10):3440–3458CrossRefGoogle Scholar
  2. Bonissone P, Cadenas JM, Garrido MC, Díaz-Valladares RA (2010) A fuzzy random forest. Int J Approx Reason 51(7):729–747CrossRefGoogle Scholar
  3. Boulesteix A, Janitza S, Kruppa J, König IR (2012) Overview of random forest methodology and practical guidance with emphasis on computational biology and bioinformatics. WIREs Data Mining Knowl Discov 2(6):493–507CrossRefGoogle Scholar
  4. Breiman L (2001) Random forest. Mach Learn 45:5–32CrossRefGoogle Scholar
  5. Breiman L, Friedman JH, Olsen RA, Stone CJ (1984) Classification and regression trees. Wadsworth, Belmont, CAGoogle Scholar
  6. Chawla NV, Bowyer KW, Hall HO, Kegelmeyer WP (2002) SMOTE: synthetic minority over-sampling technique. J Artif Intell Res 16(1):321–357Google Scholar
  7. Coalfield Geological Central Bureau of China (2000) Coalfield hydrogeology of China. Coal Industry Publishing House of China, Beijing, pp 15–59Google Scholar
  8. Death G, Fabricius KE (2000) Classification and regression trees: a powerful yet simple technique for the analysis of complex ecological data. Ecology 81(11):3178–3192CrossRefGoogle Scholar
  9. Eisavi V, Homayouni S, Yazdi AM, Alimohammadi A (2015) Land cover mapping based on random forest classification of multitemporal spectral and thermal images. Environ Monit Assess 187(5):291CrossRefGoogle Scholar
  10. Goldscheider N (2005) Karst groundwater vulnerability mapping: application of a new method in the Swabian Alb, Germany. Hydrogeol J 13(4):555–564CrossRefGoogle Scholar
  11. Harris T (2013) Quantitative credit risk assessment using support vector machines: broad versus narrow default definitions. Expert Syst Appl 40(11):4404–4413CrossRefGoogle Scholar
  12. Hong HY, Pourghasemi HR, Pourtaghi ZS (2016) Landslide susceptibility assessment in Lianhua County (China): a comparison between a random forest data mining technique and bivariate and multivariate statistical models. Geomorphology 259:105–118CrossRefGoogle Scholar
  13. Iounousse J, Er-Raki S, El Motassadeq A, Chehouani H (2015) Using an unsupervised approach of Probabilistic Neural Network (PNN) for land use classification from multitemporal satellite images. Appl Soft Comput 30(C):1–13Google Scholar
  14. Krkač M, Špoljarić D, Bernat S, Arbanas SM (2016) Method for prediction of landslide movements based on random forests. Landslides 14(3):947–960CrossRefGoogle Scholar
  15. Kubal C, Haase D, Meyer V, Scheuer S (2009) Integrated urban flood risk assessment: adapting a multicriteria approach to a city. Nat Hazards Earth Syst Sci 9(6):1881–1895CrossRefGoogle Scholar
  16. Li LP, Zhou ZQ, Li SC, Xue YG, Xu ZH (2015) An attribute synthetic evaluation system for risk assessment of floor water inrush in coal mines. Mine Water Environ 34(3):288–294CrossRefGoogle Scholar
  17. Li T, Mei TT, Sun XH, Lv YG, Sheng JQ, Cai M (2013) A study on a water-inrush incident at Laohutai coalmine. Int J Rock Mech Mining Sci 59(5):151–115CrossRefGoogle Scholar
  18. Loh WY (2011) Classification and regression trees. WIREs Data Mining Knowl Discov 1(1):14–23CrossRefGoogle Scholar
  19. Meng ZP, Li G, Xie X (2012) A geological assessment method of floor water inrush risk and its application. Eng Geol 143–144:51–60CrossRefGoogle Scholar
  20. Menze BH, Kelm MB, Masuch R, Himmelreich U, Bachert P, Petrich W, Hamprech FA (2009) A comparison of random forest and its Gini importance with standard chemometric methods for the feature selection and classification of spectral data. BMC Bioinform 10(1):1–16CrossRefGoogle Scholar
  21. Naghibi SA, Dashtpagerdi MM (2017) Evaluation of four supervised learning methods for groundwater spring potential mapping in Khalkhal region (Iran) using GIS-based features. Hydrogeol J 25(1):169–189CrossRefGoogle Scholar
  22. Pal M (2005) Random forest classifier for remote sensing classification. Int J Remote Sens 26(1):217–222CrossRefGoogle Scholar
  23. Polishchuk PG, Muratov EN, Artemenko AG, Kolumbin OG, Muratov NN, Kuz’Min VE (2009) Application of random forest approach to QSAR prediction of aquatic toxicity. J Chem Inform Model 49(11):2481–2488CrossRefGoogle Scholar
  24. Pradhan B (2013) A comparative study on the predictive ability of the decision tree, support vector machine and neuro-fuzzy models in landslide susceptibility mapping using GIS. Comput Geosci 51(2):350–365CrossRefGoogle Scholar
  25. Qiu M, Shi LQ, Teng C, Zhou Y (2016) Assessment of water inrush risk using the fuzzy delphi analytic hierarchy process and grey relational analysis in the Liangzhuang Coal Mine, China. Mine Water Environ.  https://doi.org/10.1007/s10230-016-0391-7
  26. Rodriguez-Galiano VF, Ghimire B, Rogan J, Chica-Olmo M, Rigol-Sanchez JP (2012) An assessment of the effectiveness of a random forest classifier for land-cover classification. Isprs J Photogram Remote Sens 67(1):93–104CrossRefGoogle Scholar
  27. Samuel OW, Asogbon GM, Sangaiah AK, Peng F, Li G (2017) An integrated decision support system based on ANN and fuzzy-AHP for heart failure risk prediction. Expert Syst Appl 68:163–172CrossRefGoogle Scholar
  28. Shi LQ, Qiu M, Wei WX, Xu DJ, Han J (2014) Water inrush evaluation of coal seam floor by integrating the water inrush coefficient and the information of water abundance. Int J Min Sci Technol 24(5):677–681CrossRefGoogle Scholar
  29. Singh KP, Gupta S, Rai P (2013) Predicting carcinogenicity of diverse chemicals using probabilistic neural network modeling approaches. Toxicol Appl Pharmacol 272(2):465–475CrossRefGoogle Scholar
  30. Specht DF (1990) Probabilistic neural networks. Neural Netw 3(1):109–118CrossRefGoogle Scholar
  31. Specht DF (2002) Probabilistic neural networks and the polynomial Adaline as complementary techniques for classification. IEEE Trans Neural Netw 1(1):111–121CrossRefGoogle Scholar
  32. Sun WJ, Wu Q, Dong DL, Jiao J (2012) Avoiding coal–water conflicts during the development of China’s large coal-producing regions. Mine Water Environ 31(1):74–78CrossRefGoogle Scholar
  33. Tatsumi K, Yamashiki Y, Torres MAC, Taipe CLR (2015) Crop classification of upland fields using random forest of time-series Landsat 7 ETM+ data. Comput Electron Agric 115:171–179CrossRefGoogle Scholar
  34. Tabatabaei S (2016) A probabilistic neural network based approach for predicting the output power of wind turbines. JETAI 29(2).  https://doi.org/10.1080/0952813X.2015.1132272
  35. Ulrike G (2009) Variable importance assessment in regression: linear regression versus random forest. Am Stat 63(4):308–319CrossRefGoogle Scholar
  36. Wang Y, Yang WF, Li M, Liu X (2012) Risk assessment of floor water inrush in coal mines based on secondary fuzzy comprehensive evaluation. Int J Rock Mech Mining Sci 52(6):50–55CrossRefGoogle Scholar
  37. Wang ZL, Lai CG, Chen XH, Yang B, Zhao SW, Bai XY (2015) Flood hazard risk assessment model based on random forest. J Hydrol 527:1130–1141CrossRefGoogle Scholar
  38. Wu Q, Zhou WF (2008) Prediction of groundwater inrush into coal mines from aquifers underlying the coal seams in China: vulnerability index method and its construction. Environ Geol 56(2):245–254CrossRefGoogle Scholar
  39. Wu Q, Zhou WF, Wang JH, Xie SH (2009) Prediction of groundwater inrush into coal mines from aquifers underlying the coal seams in China: application of vulnerability index method to Zhangcun Coal Mine, China. Environ Geol 57(5):1187–1195CrossRefGoogle Scholar
  40. Wu Q, Liu YZ, Liu DH, Zhou WF (2011) Prediction of floor water inrush: the application of GIS-based AHP vulnerable index method to Donghuantuo Coal Mine, China. Rock Mech Rock Eng 44(5):591–600CrossRefGoogle Scholar
  41. Wu Q, Fan SK, Zhou WF, Liu SQ (2013) Application of the analytic hierarchy process to assessment of water inrush: a case study for the no. 17 coal seam in the Sanhejian Coal Mine, China. Mine Water Environ 32(3):229–238CrossRefGoogle Scholar
  42. Wu Q, Fan ZL, Zhang ZW, Zhou WF (2014) Evaluation and zoning of groundwater hazards in Pingshuo no. 1 underground coal mine, Shanxi Province. Chin Hydrogeol J 22(7):1693–1705CrossRefGoogle Scholar
  43. Wu Q, Liu YZ, Wu XL, Liu SQ, Sun WJ, Zeng YF (2016) Assessment of groundwater inrush from underlying aquifers in Tunbai Coal Mine, Shanxi Province, China. Environ Earth Sci 75(9):1–13Google Scholar
  44. Wu Q, Zhao DK, Wang Y, Shen JJ, Mu WP, Liu HL (2017) Method for assessing coal-floor water-inrush risk based on the variable-weight model and unascertained measure theory. Hydrogeol J.  https://doi.org/10.1007/s10040-017-1614-0
  45. Zeng YF, Wu Q, Liu SQ, Zhai YL, Zhang W, Liu YZ (2016) Vulnerability assessment of water bursting from Ordovician limestone into coal mines of China. Environ Earth Sci 75(22):1431CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  • Dekang Zhao
    • 1
    • 2
  • Qiang Wu
    • 1
    • 2
  • Fangpeng Cui
    • 1
    • 2
  • Hua Xu
    • 3
  • Yifan Zeng
    • 1
    • 2
  • Yufei Cao
    • 4
  • Yuanze Du
    • 1
    • 2
  1. 1.China University of Mining & Technology (Beijing)BeijingChina
  2. 2.National Engineering Research Center of Coal Mine Water Hazard ControllingBeijingChina
  3. 3.Information Engineering CollegeBeijing Institute of Petrochemical TechnologyBeijingChina
  4. 4.Beijing Urban Construction Exploration and Surveying Design Research Institute Co. Ltd.BeijingChina

Personalised recommendations