Predictive model of overburden deformation: based on machine learning and distributed optical fiber sensing technology

Abstract

Purpose The purpose of this paper is to establish a strain prediction model of mining overburden deformation, to predict the strain in the subsequent mining stage. In this way, the mining area can be divided into zones with different degrees of risk, and the prevention measures can be taken for the areas predicted to have large deformation. Design/methodology/approach A similar-material model was built by geological and mining conditions of Zhangzhuang Coal Mine. The evolution characteristics of overburden strain were studied by using the distributed optical fiber sensing (DOFS) technology and the predictive model about overburden deformation was established by applying machine learning. The modeling method of the predictive model based on the similar-material model test was summarized. Finally, this method was applied to engineering. Findings The strain value predicted by the proposed model was compared with the actual measured value and the accuracy is as high as 97%, which proves that it is feasible to combine DOFS technology with machine learning and introduce it into overburden deformation prediction. When this method was applied to engineering, it also showed good performance. Originality/value This paper helps to promote the application of machine learning in the geosciences and mining engineering. It provides a new way to solve similar problems.

Study on deep learning methods for coal burst risk prediction based on mining-induced seismicity quantification

“…The backpropagation (BP) method is the core part of the feed-forward neural network, solving non-linear optimization problems through the input and output of a set of samples. Any non-linear function can be approximated with arbitrary precision by adjusting the BP connection weights and network size (including n, m, and the number of hidden layer neurons) (Liu et al 2021). The non-linear relationship between the coal burst risk being inverted and a set of neural networks describes the impact factors (n, h 1 , … , h p , m) as follows:…”

Analysis of Overburden Deformation and Migration under Huge Thick Unconsolidated Layers Based on Multiple Approaches

“…), and drilling are implemented as an in-site field measurement method [17][18][19][20]. Moreover, distributed fiber optic sensing (DOFS) technology with the virtue capacity of portable installation, low environmental requirements, and real-time online monitoring becomes an effective strategy to measure the deformation and strain distribution of overburdened strata [21][22][23]. Approximately reflecting the actual geological condition of a targeted area, similar material modeling embedded in DOFS can analyze the deformation characteristic of overburdened strata from the small-scale perspective in the laboratory [24][25][26][27].…”