Coal mine production disaster risk prediction method and system based on large model

By constructing a coal mine production disaster risk prediction method based on a large model, collecting gas concentration and fault information, generating an equivalent topological bias matrix, and combining it with scaling cue vectors to calculate the risk probability distribution, the problem of neglecting the influence of geological faults in traditional methods is solved, and more accurate risk warning is achieved.

CN122066048BActive Publication Date: 2026-07-07GUIZHOU ZHILIE TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
GUIZHOU ZHILIE TECH CO LTD
Filing Date
2026-04-15
Publication Date
2026-07-07

AI Technical Summary

Technical Problem

Traditional methods for predicting disaster risks in coal mines cannot effectively capture the impact of complex geological fault distribution underground, and it is difficult to adaptively adjust assessment standards, resulting in lag and error in the judgment of precursor indicators and a lack of predictive sensitivity.

Method used

By collecting gas concentrations on the upwind and downwind sides, calculating gas diffusion, constructing a spatial topological concentration term sequence, calculating the penalty amount by combining the fault crossing state of the probe and the coal mining machine connection, generating an equivalent topological bias matrix, deeply fusing geological structural features using scaling cue vectors, injecting them into a large model to calculate the risk probability distribution, and finally generating a coal mine emergency risk warning instruction.

Benefits of technology

It enables accurate risk assessment of multidimensional nonlinear environments, improving the accuracy and adaptive adjustment capabilities of early warning of sudden dangerous situations in coal mines.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122066048B_ABST
    Figure CN122066048B_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of risk prediction, in particular to a coal mine production disaster risk prediction method and system based on a large model, comprising the following steps: comparing a gas diffusion amount with a reference amount to construct a concentration word sequence, analyzing a probe connection line through a fault state to calculate a penalty amount to construct a bias matrix, analyzing a gas concentration and an alarm upper limit difference to construct a scaling prompt vector, using the prompt vector to obtain fusion features and injecting the bias matrix to calculate a disaster probability distribution amount, and comparing the distribution amount with the reference amount to construct an early warning instruction. In the present application, the equivalent bias matrix is constructed by deeply fusing the spatial geological physical obstruction factors and the gas dynamic diffusion characteristics, the large model reasoning adjustment mechanism is established by extracting the variable approximation rate, the feature is spliced by using the prompt vector and injected into the model node, the limitations of static numerical judgment are broken, and the multi-dimensional nonlinear environmental data is accurately mapped, thereby effectively improving the accuracy of the coal mine complex danger situation early warning.
Need to check novelty before this filing date? Find Prior Art