Логин:
Пароль:
Регистрация
Забыли свой пароль?

The application of machine learning methods for predicting porosity of rocks based on data of x-ray fluorescence analysis and gamma-ray spectrometry

UDK: 553.98.001
DOI: 10.24887/0028-2448-2018-7-64-69
Key words: forecast, porosity, x-ray fluorescence analysis, gamma-spectrometry, machine learning
Authors: S.V. Shadrina (Tyumen Branch of SurgutNIPIneft, RF, Tyumen), A.A. Shadrin (University of Oslo, Norway, Oslo)
Developing of the non-traditional hydrocarbon reservoirs and underground water deposits allotted a number of tasks to specialists the solution of which goes beyond the traditional methodologies. The most significant problem among them is unable to predict petrophysical parameters (in particular porosity coefficient) for lithology undifferentiated section. There was examined the dependence of porosity coefficient of rocks upon the chemical composition of weakly differentiable volcanic rocks. It takes much time and much more petrophysical parameters to decide this problem by traditional mathematical-statistical methods. These factors make the topical application of modern mathematical methods to obtain evaluation of porosity based on characteristics of the rock and does not depend on the condition of the core. On the other hand, such approach allows taking into account the collectors and fluid traps geochemical mechanism formation in non-traditional objects to predict its spatial distribution. Comparison of the various modern techniques of machine learning to predict porosity is showed in the article. The best predictive capability showed methods based on decision trees-forest – Random forest and Extra trees, providing an average coefficient of determination for test samples equal to 0,52, Pearson product moment correlation coefficient is 0,722 and 0,701 respectively. In addition, these methods allow to rank factors used for predicting porosity of rocks, on the degree of their influence on the accuracy of the prediction. The received results permit to use of the methods of machine learning as a promising approach for predicting porosity through description of mechanisms of secondary mineral formation

References

1. Strontsiy i bariy v endogennykh protsessakh (Strontium and barium in endogenous processes): edited by Pozharitskaya L.K., Moscow: Nauka Publ., 1973, 215 p.

2. Portnov A.M., Kandinov M.N., Carbon dioxide as an ore deposition controller (In Russ.), Priroda, 1992, no. 11, pp. 64–69.

3. Arbuzov S.I., Rikhvanov L.P., Geokhimiya radioaktivnykh elementov (Geochemistry of radioactive elements), Tomsk: Publ. of TPU, 2010, 300 p.

4. Stolbov Yu.M., Fomin Yu.A., Stolbova N.F., Vozmozhnost' primeneniya prikladnoy geokhimii urana pri issledovanii protsessov nalozhennogo epigeneza terrigennykh otlozheniy Zapadnoy Sibiri (The possibility of applying applied geochemistry of uranium in the study of the processes of superimposed epigenesis of terrigenous deposits of Western Siberia), Proceedings of II International Conference “Geokhimicheskoe modelirovanie i materinskie porody neftegazonosnykh basseynov Rossii i stran SNG” (Geochemical modeling and parent rocks of oil and gas bearing basins of Russia and CIS countries), St. Petersburg: Publ. of VNIGRI, 2000, pp. 160–171.

5. Shaldybin M.V., Geokhimicheskie kriterii otsenki vliyaniya protsessov nalozhennogo epigeneza na fil'tratsionno-emkostnye svoystva oblomochnykh porod-kollektorov (na primere neftyanykh mestorozhdeniy Tomskoy oblasti) (Geochemical criteria for assessing the impact of superimposed epigenesis on the reservoir properties of clastic rock reservoirs (for example, oil deposits in the Tomsk region)): thesis of candidate of geological andВ В  mineralogical science, Tomsk, 2005.

6. Bocharov E.I., Stolbov Yu.M., Evaluation of postsedimentational process influence on filtration-capacitor properties of Chalky deposits of the Western Siberia north (In Russ.), Izvestiya Tomskogo politekhnicheskogo universiteta = Bulletin of the Tomsk Polytechnic University, 2007, V. 311, no. 1, pp. 64–66.

7. Gavshin V.M., Shcherbov B.L., Sukhorukov F.V. et al., Zakonomernosti raspredeleniya mikroelementov v profile vyvetrivaniya Barlakskogo granitnogo massiva (Regularities of distribution of microelements in the weathering profile of the Barlak granite massif), In: “Geokhimiya rudnykh elementov v protsessakh vyvetrivaniya, osadkonakopleniya i katageneza” (Geochemistry of ore elements in the processes of weathering, sedimentation and catagenesis), Novosibirsk: Nauka Publ., 1979, pp. 3–19.

8. Zhmodik S.M., Geokhimiya radioaktivnykh elementov v protsesse vyvetrivaniya karbonatitov, kislykh i shchelochnykh porod (Geochemistry of radioactive elements in weathering of carbonatites, acidic and alkaline rocks), Novosibirsk: Nauka Publ., 1984, 170 p.В  В В 
Developing of the non-traditional hydrocarbon reservoirs and underground water deposits allotted a number of tasks to specialists the solution of which goes beyond the traditional methodologies. The most significant problem among them is unable to predict petrophysical parameters (in particular porosity coefficient) for lithology undifferentiated section. There was examined the dependence of porosity coefficient of rocks upon the chemical composition of weakly differentiable volcanic rocks. It takes much time and much more petrophysical parameters to decide this problem by traditional mathematical-statistical methods. These factors make the topical application of modern mathematical methods to obtain evaluation of porosity based on characteristics of the rock and does not depend on the condition of the core. On the other hand, such approach allows taking into account the collectors and fluid traps geochemical mechanism formation in non-traditional objects to predict its spatial distribution. Comparison of the various modern techniques of machine learning to predict porosity is showed in the article. The best predictive capability showed methods based on decision trees-forest – Random forest and Extra trees, providing an average coefficient of determination for test samples equal to 0,52, Pearson product moment correlation coefficient is 0,722 and 0,701 respectively. In addition, these methods allow to rank factors used for predicting porosity of rocks, on the degree of their influence on the accuracy of the prediction. The received results permit to use of the methods of machine learning as a promising approach for predicting porosity through description of mechanisms of secondary mineral formation

References

1. Strontsiy i bariy v endogennykh protsessakh (Strontium and barium in endogenous processes): edited by Pozharitskaya L.K., Moscow: Nauka Publ., 1973, 215 p.

2. Portnov A.M., Kandinov M.N., Carbon dioxide as an ore deposition controller (In Russ.), Priroda, 1992, no. 11, pp. 64–69.

3. Arbuzov S.I., Rikhvanov L.P., Geokhimiya radioaktivnykh elementov (Geochemistry of radioactive elements), Tomsk: Publ. of TPU, 2010, 300 p.

4. Stolbov Yu.M., Fomin Yu.A., Stolbova N.F., Vozmozhnost' primeneniya prikladnoy geokhimii urana pri issledovanii protsessov nalozhennogo epigeneza terrigennykh otlozheniy Zapadnoy Sibiri (The possibility of applying applied geochemistry of uranium in the study of the processes of superimposed epigenesis of terrigenous deposits of Western Siberia), Proceedings of II International Conference “Geokhimicheskoe modelirovanie i materinskie porody neftegazonosnykh basseynov Rossii i stran SNG” (Geochemical modeling and parent rocks of oil and gas bearing basins of Russia and CIS countries), St. Petersburg: Publ. of VNIGRI, 2000, pp. 160–171.

5. Shaldybin M.V., Geokhimicheskie kriterii otsenki vliyaniya protsessov nalozhennogo epigeneza na fil'tratsionno-emkostnye svoystva oblomochnykh porod-kollektorov (na primere neftyanykh mestorozhdeniy Tomskoy oblasti) (Geochemical criteria for assessing the impact of superimposed epigenesis on the reservoir properties of clastic rock reservoirs (for example, oil deposits in the Tomsk region)): thesis of candidate of geological andВ В  mineralogical science, Tomsk, 2005.

6. Bocharov E.I., Stolbov Yu.M., Evaluation of postsedimentational process influence on filtration-capacitor properties of Chalky deposits of the Western Siberia north (In Russ.), Izvestiya Tomskogo politekhnicheskogo universiteta = Bulletin of the Tomsk Polytechnic University, 2007, V. 311, no. 1, pp. 64–66.

7. Gavshin V.M., Shcherbov B.L., Sukhorukov F.V. et al., Zakonomernosti raspredeleniya mikroelementov v profile vyvetrivaniya Barlakskogo granitnogo massiva (Regularities of distribution of microelements in the weathering profile of the Barlak granite massif), In: “Geokhimiya rudnykh elementov v protsessakh vyvetrivaniya, osadkonakopleniya i katageneza” (Geochemistry of ore elements in the processes of weathering, sedimentation and catagenesis), Novosibirsk: Nauka Publ., 1979, pp. 3–19.

8. Zhmodik S.M., Geokhimiya radioaktivnykh elementov v protsesse vyvetrivaniya karbonatitov, kislykh i shchelochnykh porod (Geochemistry of radioactive elements in weathering of carbonatites, acidic and alkaline rocks), Novosibirsk: Nauka Publ., 1984, 170 p.В  В В 


Attention!
To buy the complete text of article (a format - PDF) or to read the material which is in open access only the authorized visitors of the website can. .

Mobile applications

Read our magazine on mobile devices

Press Releases

16.07.2019
08.07.2019
04.07.2019
SPE 2019
ТАТАРСТАНСКИЙ НЕФТЕГАЗОХИМИЧЕСКИЙ ФОРУМ