Research On CSAMT Inversion For Collapse Column Detection Based On Clustering Constraint

As a big energy country,coal occupies the main position of China’s energy strategy compared with oil and natural gas,and it may not change greatly in a long period of time.The coal mining process will undoubtedly result in a variety of unexpected safety accidents,posing a severe danger to people’s lives and property safety.Safety accidents caused by water inrush account for a significant percentage of them,and the coalfield collapse column created by complicated hydrogeological conditions is one of the primary sources of water inrush accidents.Furthermore,the production safety issues created by water inrush accidents tangentially raise mining costs,limiting resource exploitation and consumption by a certain extent.Therefore,increasing the precision of coalfield collapse column exploration is crucial for the following safe and effective production as well as for the reasonable and effective avoidance of mine water inrush catastrophes in premining.A manufactured source frequency domain hearing technique called Controlled Source Audio-frequency Magnetotellurics(CSAMT)was created on the foundation of Magnetotelluric(MT)and Audio-frequency Magnetotellurics(AMT).This method’s fake field source has the benefits of a broad frequency range and a substantial depth of investigation.Geothermal resource discovery,finding non-ferrous metal deposits,finding concealed catastrophe sources like coalfield fall columns and goafs,and hydrocarbon research and development all make extensive use of it.Unfortunately,this method’s detection accuracy and resolution are constrained by the reaction of the volume effect,and there are still some issues with the precise detection of collapse columns.This research suggests an enhanced inversion technique based on fuzzy c-means clustering(FCM)that seeks to realize accurate inversion of controlled source audio magnetotelluric data and enhance detection accuracy of collapse column in coalfield.The numerical simulation of a three-dimensional electromagnetic response is first carried out in three dimensions,and then it is converted into solving algebraic equations.A basic field and a secondary field are created from the entire field.In order to achieve the numerical simulation of a three-dimensional electromagnetic field of a controlled source,the calculation technique of a long wire field source is developed from Maxwell’s equations,and the solution method of a secondary field is generated using the finite element approach.On the basis of confirming the program’s validity,several collapse column models were created for numerical simulation,and the features of the controlled source audio magnetotelluric response under various diameters,buried depths,and overburden thicknesses were examined.The analysis’ s findings indicate that the buried depth,diameter,and overburden thickness have a considerable impact on the controlled source audio magnetotelluric technique.As a consequence,the detection effect of deep burials and small-scale collapse columns will be drastically diminished.On the basis of forward modeling,inversion-related research was subsequently conducted.The one-dimensional inversion is enhanced and the objective function of the controlled source audio magnetotelluric is generated by including a lateral constraint term in the objective function of one-dimensional inversion.The controlled source audio magnetotelluric inversion technique is then created by including the fuzzy clustering constraint into the inversion process.The inversion outcomes of various inversion methodologies are compared and studied using the collapse column model as an example.The findings demonstrate that controlled source audio magnetotelluric inversion based on clustering restrictions yields results that are clearly superior to those of quasi-twodimensional inversion and one-dimensional inversion,hence increasing the effectiveness of collapse column identification.The findings of one-dimensional inversion,pseudo-two-dimensional inversion,and cluster restricted inversion of controlled source audio magnetotelluric inversion are then compared,along with the detection example of suspected collapse column at Huangyuchuan Coal mine.The drilling verification demonstrates that the controllable source audio magnetotellus inversion method based on cluster constraints can effectively improve the interpretation accuracy of the subsidence column by delineating the shape of the subsidence column well,inferring the boundary of the subsidence column accurately,and delineating the subsidence column well.It proves the applicability of the controlled source audio magnetotellus inversion technique based on cluster limitations presented in this study.