Production of oil and gas, at the mature stage of field development, is accompanied by extraction of ground water to the surface, which leads to an increase in the intensity of complications, including scaling. Oil companies use various methods to deal with complications, the effectiveness of which, in most cases, depends on the chemical composition of the water. The justified choice of technology for combating complications requires laboratory and pilot studies to determine the chemical composition of water, as well as the effectiveness of the technology in relation to current operating conditions. Reduction of the required number of studies is of undoubted practical interest.
One way to reduce the number of studies is to generalize the objects of research within the groups on the basis of their proximity to the totality of characteristics, the solution of such a problem is to solve the clustering problem. In the article, the solution of the clustering problem is realized by the joint use of methods to reduce the dimension of the source space of features describing the object of research and establishing dominant relationships within the data with the subsequent application of clustering algorithms to the obtained representation of the initial data.
The proposed calculation method was applied at the facilities of Bashneft Company. Depending on the chemical composition of the reservoir waters, samples are pooled into groups. The description of both qualitative and quantitative properties of samples of formation water was carried out on the basis of calculating the intensity of scaling, for each of the clusters obtained, according to the Oddo – Thomson method. The main practical value of the proposed methodology is the possibility of its application to the solution of the problem of optimizing the number of laboratory and pilot-field tests of oil production reagents at facilities allocated to clusters.
Further development of the proposed methodology assumes an increase in the dimension of the initial data and a search for such a representation space that provides a scientific and technical justification for testing oilfield reagents at existing and new oil production facilities. Among promising methods that require an assessment of the applicability to the problems being solved, the following can be distinguished: the problem of clustering-autocoders, Markov chains; the task of describing the properties of formation water is the Pitzer method.
References
1. Bouhlel J., Bouveresse D.J.-R. et al., Comparison of common components analysis with principal components analysis and independent components analysis: Application to SPME–GC–MS volatolomic signatures, Talanta, 2018, V. 178, no. 1, February, pp. 854–863.
2. Marghade D., Malpe D.B., Subba Rao N., Identification of controlling processes of groundwater quality in a developing urban area using principal component analysis, Environmental Earth Sciences, 2015, V. 74(7), pp. 5919–5933.
3. Agarwal A., Maheswaran R., Kurths J., Khosa R., Wavelet Spectrum and self–organizing maps–based approach for hydrologic regionalization – a case study in the western United States, Water Resources Management, 2016, V. 30(12), pp. 4399–4413.
4. Ley R., Casper M.C., Hellebrand H., Merz R., Catchment classification by runoff behaviour with self–organizing maps (SOM), Hydrology and Earth System Sciences, 2011, V. 15(9), pp. 2947–2962.
5. Chang F.J., Chang L.C., Huang C.W., Kao I.F., Prediction of monthly regional groundwater levels through hybrid soft–computing techniques, Journal of Hydrology, 2016, V. 541, pp. 965–976.
6. Haga J., Siekkinen J., Sundvik D., Initial stage clustering when estimating accounting quality measures with self–organizing maps, Expert Systems with Applications, 2015, V. 42, no. 21, pp. 8327–8336.
7. García H.L., González I.M., Self–organizing map and clustering for wastewater treatment monitoring, Engineering Applications of Artificial Intelligence, 2004, V. 17(3), pp. 215–225.
8. Hentati A., Kawamura A., Amaguchi H., Iseri Y., Evaluation of sedimentation vulnerability at small hillside reservoirs in the semi-arid region of Tunisia using the self–organizing map, Geomorphology, 2010, V. 122(1–2), pp. 56–64.
9. Kriegel H.-P., Schubert E., Zimek A., The (black) art of runtime evaluation: Are we comparing algorithms or implementations, Knowledge and Information Systems, 2016, V. 52, pp. 341–378.
10. Halim Z., Waqas M., Efficient clustering of large uncertain graphs using neighborhood information, International Journal of Approximate Reasoning, 2017, V. 90, November, pp. 274–291.
11. De Amorim R.C., Feature relevance in Ward’s hierarchical clustering using the Lp norm, Journal of Classification, 2015, V. 32(1), pp. 46–62.
12. Domokos E.-K., Bálint C., Definition of user groups applying Ward’s method, Transportation Research Procedia, 2017, V. 22, pp. 25–34.
13. Moosavi V., Pre-specific modeling: Diss., Eidgenössische Technische Hochschule ETH Zürich, no. 22683, 2015.
14. Oddo J.E., Tomson M.B., Why scale forms and how to predict it, SPE 21710-PA, 1994.
15. Zahedzadeh M., Karambeigi M.S. et al., Comprehensive management of mineral scale deposition in carbonate oil fields – A case study, Chemical Engineering Research and Design, 2014, V. 92, no. 11, pp. 2264–2272.вЃ
