A Combination of Monte-Carlo and Damped Least Square Inversion Method For Determining Radon Source in Geothermal Case

Radon measurement on the surface can represent the subsurface condition. the measured radon in geothermal field is caused by the source, which is usually a geothermal reservoir. this study did the inversion process for determining the depth and value of radon source. another fact, nonuniqueness of the solution can produce a result with different model parameter combinations. hence, it can confuse the interpreter to determine the correct model. based on this case, we proposed an inversion scheme that can minimize the nonuniqueness effect in the radon data inversion. the scheme is started by montecarlo inversion and finished by damped leastsquare. monte-carlo inversion, as one of the global optimizations, produce an appropriate starting model for the damped least squares. the damped least square method will finish the scheme fast. in order to be sure with the result, the whole scheme is repeated 19 times. the relative rms error for the synthetic data is 0.07% to 0.32%  to  a depth difference of 7% from the synthetic model. with this synthetic data inversion test, the inversion scheme was applied to the real data from the rajabasa geothermal field. with this scheme, the section aa’ gives an error of 0.51% to 0.88% with a depth of 712 m and section bb’ gives an error of 5.79% to 5.27% with a depth of 728 m. this result is coherent with the magnetotelluric data in this area.

A Combination of Monte-Carlo and Damped Least Square Inversion Method for Determining Radon Source in Geothermal Case

Radon measurement on the surface can represent the subsurface condition. The measured Radon in geothermal field is caused by the source, which is usually a geothermal reservoir. This study did the inversion process for determining the depth and value of Radon Source. Another fact, nonuniqueness of the solution can produce a result with different model parameter combinations. Hence, it can confuse the interpreter to determine the correct model. Based on this case, we proposed an inversion scheme that can minimize the nonuniqueness effect in the Radon data inversion. The scheme is started by MonteCarlo inversion and finished by damped leastsquare. MonteCarlo inversion, as one of the global optimizations, produce an appropriate starting model for the damped least squares. The damped least square method will finish the scheme fast. In order to be sure with the result, the whole scheme is repeated 19 times. The relative RMS error for the synthetic data is 0.07% to 0.32%  to  a depth difference of 7% from the synthetic model. With this synthetic data inversion test, the inversion scheme was applied to the real data from the Rajabasa Geothermal field. With this scheme, the section AA’ gives an error of 0.51% to 0.88% with a depth of 712 m and section BB’ gives an error of 5.79% to 5.27% with a depth of 728 m. This result is coherent with the magnetotelluric data in this area.