METHOD OF TEMPERATURE PREDICTION AT DEPTH BELOW WELL BOTTOM Russian patent published in 2020 - IPC G01V3/08 E21B47/07 

FIELD: geophysics.

SUBSTANCE: invention relates to geophysics and can be used for evaluation of temperature to depth below bottomhole of drilled wells. On the surface of the Earth in the vicinity of the drilled wells, for which electric logging data are known, horizontal components of the natural magnetotelluric field are measured in the frequency range sufficient for the field penetration to the depth to which the temperature forecast is to be made. Components of the impedance tensor Ẑ(ω) are determined from two field polarizations in the primary field. Then one performs a one-dimensional inversion from the values of Ẑ(ω), as a result of which a vertical profile of specific electric resistance is plotted at the field measurement point. Resistance values are determined for points of temperature measurement in well. By built in the vicinity of the well of the resistivity profile, as well as according to electric logging in the well itself, the profile of pseudo-electric logging is built to the specified depth using the pre-trained first artificial neuronet according to the profile of resistivity and logging data built in the vicinity of the well in the well itself. Then, using the trained neural network, the pseudo-electric logging is predicted at the depths below the bottomhole of the well as per data of specific electric resistance in its vicinity at these depths. Second neuron network is trained to comply with the data of electric and thermal logging in the well itself; after that, using its forecast temperature curve below the bottomhole of the pseudo-electric logging device built at the first stage.

EFFECT: high accuracy of predicting temperature at depths below well bottom, which in turn increases reliability of estimating potential of geothermal resources, as well as enables to detect deposits of hydrocarbons based on temperature anomalies created in their vicinity.

3 cl, 1 tbl, 4 dwg