Mine water inrush source identification method and identification system based on big data and deep learning

Abstract

The invention discloses a mine water inrush source identification system based on big data and deep learning. The mine water inrush source identification system comprises a ground workstation, an acquisition device and a computer detection module. The mine water inrush source identification method based on big data and deep learning comprises the following steps of: S1, performing classification according to known chemical component characteristics of a regional water source, and establishing a sample library; S2, when water inrush occurs in the mine, pumping the inrush water into a containerthrough a liquid pump for storage; S3, detecting and analyzing each chemical component in the inrush water through an electrode detector array module, and determining the content of each chemical component; and S4, inputting the chemical components into a data processing module for processing. According to the invention, the internal multi-layer characteristics of a water sample are autonomously obtained through a deep feedforward network model based on deep learning, repeated training, error prediction and model parameter adjustment are carried out, the accuracy is improved along with increase of water sample information, and water inrush source identification is still accurate under the condition of big data.

Description

technical field [0001] The invention relates to the technical field of water inrush source identification, in particular to a mine water inrush source identification method and an identification system based on big data and deep learning. Background technique [0002] Coal is the main energy source in our country, and the safe mining of coal resources is related to the economic development of the country. The frequent occurrence of water inrush has caused huge casualties and economic losses. Therefore, quickly and accurately identifying the source of water inrush is the key to water inrush control, and it also provides a theoretical basis and correct guidance for the follow-up water prevention and control work. [0003] The methods for identification of mine water inrush sources mainly include hydrochemical analysis method, Fisher discriminant analysis theory, gray relational degree theory, fuzzy variable set method and so on. These include identifying the water source of t...

AI Technical Summary

Problems solved by technology

[0004] Most of the traditional water source identification methods mentioned above are difficult to identify water inrush sources due to data redundancy and complex operations in the case of large data.

The early neural network and deep learning network have a similar hierarchical structure. The difference between the two is that the early neural network generally has a two- to three-layer network, and the computing unit and learning ability are limited, which makes the judgment of the water inrush source inaccurate.

Images