In many oil and gas enterprises there is a problem of qualitative selection of geological and engineering operations for stabilization of oil production. The paper suggests a methodical approach to solving the problem of preliminary assessment of the effectiveness of planned geological and engineering operations using decision trees. The standard method for constructing decision trees involves a recursive algorithm for selecting the most influential factor and searching for a better separation of the original sample of data into two new ones, depending on the value of this factor. The problem of the technique is the deterioration of the possibility of assessing the impact on the efficiency of geological and engineering operations for each of the factors, in addition, the complexity of algorithms for building decision trees requires the availability of special software or programming skills, as well as a certain level of engineer training. Due to the specific nature of the initial data, it is possible to simplify the algorithms for constructing decision trees to analyze the influence of factors on the efficiency of the geological and engineering operations. The proposed simplified methodology has a transparent mechanism of operation, the implementation of which does not cause much difficulty in common spreadsheet editors. As a result of effective analysis using a simplified technique for constructing trees quantitative selection criteria for processing wells are appeared. As a demonstration of this statement, an analysis of the effectiveness of acidic hydraulic fracturing of the formation was carried out, quantitative criteria were determined for further selection of wells for this technology. A method for ranking wells based on the predicted processing efficiency is proposed depending on the values of the influencing factors, whereas the calculated well rank is a sufficient factor for selection of the well, because it contains the values of the other factors. An example of a map of favorable and unfavorable zones for carrying out geological and engineering operations is shown.
1. Pichugin O.N., Prokof'eva Yu.Z., Aleksandrov D.M., Application of decision trees as an efficient method of analysis and prediction (In Russ.), Neftepromyslovoe delo, 2013, no. 11, pp. 69 –75.
2. Pichugin O.N., Solyanoy P.N., Fatikhova Yu.Z., From “mistakes corrected” to effective treatment prediction (In Russ.), Neft'. Gaz. Novatsii, 2012, no. 3, pp. 28–31.
3. URL: http://statsoft.ru/home/ textbook/modules/stclatre.html.
4. Kaftannikov I.L., Parasich A.V., Decision tree’s features of application in classification problems (In Russ.), Vestnik YuUrGU. Seriya “Komp'yuternye tekhnologii, upravlenie, radioelektronika” = Bulletin of the South Ural State University. Ser. Computer Technologies, Automatic Control, Radio Electronics, 2015, V. 15, no. 3, pp. 26 –32.
5. Mel'nikov G.A., Gubarev V.V., Method for regression tree induction based on the ant algorithms (In Russ.), Doklady TUSUR, 2014, no. 4(34), pp. 72–78.вЃ