## Mathematical software complex

EM-Data processor (EMDP) program complex is the base tool for interpretation of data (1D and 3D).

EM-Data Processor is designed to perform the most adequate images of geoelectric sections along the linear profiles and of spatial geoelectrical voxel models of areas of surveys based on TEM, M-TEM soundings, time-domain EM sounding from fixed source, airborne TEM, MT and EM scanning in geological engineering surveys.

EM-DataProcessor combines modern graphical interface, the set of comfort user tools providing efficiency of work. EMDP utilizes performance capabilities of modern hardware (Fig.1). The EMDP demo-version can be downloaded for appreciation of program capabilities.

Fig.1. The distribution of the measured signal with the section of resistivity along one of the profiles

Main capabilities:

- Interactive processing of picket, profile and areal data;
- Signal filtering by time/profile/area;
- Apparent media parameters calculation, drawing of cuts and sections (RDI, CDI) considering shape of the sounding impulse, transmitter-receiver geometry and flying height;
- Picket and batch 1D-inversion, including induction induced polarization (IP) effect correction based on the Cole-Cole formula;
- Automated induction and polarization signals separation while processing sounding data from polarizable media;
- Combined picket and batch 1D-inversion of the data obtained from transmitter-receiver installations with different geometries and also of MT data;
- Automated estimation of the medium dimension in cases of adaptive 3D electromagnetic survey;
- 3D-inversion of ground surveys;
- 3D-inversion of airborne TEM data;
- 2D- and 3D- visualization (including voxel view) of interpretation results (transformations/1D-inversion/3D-inversion);
- Export of processing and interpretation results to Surfer, Oasis, Montaj, etc;
- Support of different formats of input data obtained from various equipment.

### Impulse electromagnetic survey data interpretation

EM DataProcessor capabilities for impulse electromagnetic survey data processing and interpretation include necessary data preprocessing (interactive data editing, time/space filtering , etc.), quantitative data interpretation algorithms (batch 1D-inversion, linear and nonlinear 3D-inversions) and also 2D- and 3D-interpretation results visualization.

During the software complex development the main attention was given to a speed of 1D- and mainly 3D-inversions, intuitiveness of user interface for process and interpret data, and also to informativity of obtained interpretation images of the medium (interpretation results) using implemented visualization tools.

### 1D inversion

One of the basic procedures in EM-DataProcessor is 1D quantitative data interpretation that fits many experimental geological and geophysical tasks. User is able to reconstruct geoelectrical parameters of a medium (including parameters of IP described by the Cole-Cole formula) in quasi-layered model mode both in one picket or along the whole profile, considering relative locations of sounding points (Fig.2). This approach enables to accelerate the process of inversion and obtain more stable and «smooth» distribution of medium parameters along the profile. 1D-inversion is realized for all types of data to process: multi-offset TEM, MT, airborne TEM.

The feature of inversion of data complicated by induction IP is considerable growth of the number of parameters to be defined: from two to five parameneters for each layer (thikness h, resistivity ρ and Cole-Cole parameters: polarizability, polarization time constant, the degree factor). The solution of inverse task in such a case may therefore be problematic or even impossible. To solve this problem the two approaches in the EM-DataProcessor is implemented to interpret such data. The first one is based on the technology of separation of induction and polarization fields and consists in subsequent interpretation of separated signals: on the base of induction signal the medium parameters without IP-effect are inversed, after that on its base IP parameters using extracted IP signal are calculated. The second approach is based on combined inversion of the data obtained from different transmitter-receiver installations (or from different receivers, located at different heights from the transmitter loop). It allows significantly decrease an equivalence in a solution.

Fig.2. The section of electrical resistivity along the profile. 1D-inversion with correction for IP-effect.

### 3D-inversion

Fast non-linear 3D-inversion (accounting non-linearity of testing objects signal response from its anomalous conductivity) based on dividing of studied volume into the mesh of objects, in each of which its conductivity is calculated is implemented in the EM-DataProcessor. The 3D-inversion procedure is provided on a personal computer (notebook) in a short time (from several minutes).

The results of two types of 3D-inversions in comparison with 1D quantitative interpretation of the same data are presented on fig. 3.

a) The model with two objects

b) 3D-inversion results (non-linear – at the top, linear – in the middle) in comparison with 1D-inversion results (at the bottom)

Fig.3. 1D- and 3D- inversion results in comparison with the model

The figures show that the most adequate result was obtained when using non-linear 3D-inversion. However linear 3D-inversion also gave the result not distorted by occurrence of the false edge objects that can quite be used for express analysis of 3D-distribution of conductivity in geological medium.

Let us note, that during 3D-inversion the studied volume has been divided into 120 objects and the time of one iteration was 1 minute (Intel Core i7 processor at 3,6 GHz clock). It took 5 iterations to decrease rms deviation 7 times.

One more capability of the EM-DataProcessor is 3D-inversion of airborne TEM survey data. The feature of such data interpretation is its large volume and large exploration areas. For inversion of such data the approach of “moving footprint” considering sounding data to calculate the response from selected test object is used. Moreover an acceleration is also achieved by capability of the representation of a source as a magnetic dipole uplifted to the flight height.

### Complex M-TEM and MT data interpretation

The technology of complex interpretation of M-TEM and MT data in the EM-DataProcessor program is based on a staged parameterization of a cross-section.

The first stage is recovery of 3D model of upper part of section based on M-TEM data. Then a normalization (a removal of shift-effect) in MT data using 3D model obtained according to M-TEM data is performed. This normalization allows more adequate interpretion of the deep part of observed medium in comparison with methods of normalization that use only MT data.

At the second stage the 3D interpretation of the normalized MT data with fixed 3D model of the upper part of section obtained according to the M-TEM data is performed. The calculation is performed only for the deep part of the section.

**Total conductivity graph at depth interval from 3000 to 5000 m**

a) section based on 3D interpretation

b) section based on initial model

Fig.4. 3D interpretaion result in comparison with test model

EM-DataProcessor is registered in the public register of computer programs. Certificate of state registration #2011611248.

Version | 1D | 3D express | 3D TEM | 3D MT |

Capabilities | 1D inversion | 3D inversion | Accurate 3D interpretation of TEM data | Complex 3D interpretation of TEM+MT data |

Signal to apparent characteristics of medium transformation | √ | √ | √ | √ |

3D-visualization of field data and interpretation results | √ | √ | √ | √ |

Export of processing and interpretation results | √ | √ | √ | √ |

1D inversion | √ | √ | √ | √ |

Fast 1D-inversion of airborne TEM and EM scanning data | √ | √ | √ | √ |

Dimension of the medium estimation (for M-TEM data) | √ | √ | √ | √ |

Linear 3D-inversion | — | √ | √ | √ |

Non-inear 3D-inversion | — | √ | √ | √ |

Accurate 3D interpretation of TEM data | — | — | √ | √ |

Complex 3D interpretation of MT data | — | — | — | √ |