Find in Library

The relevance. While drilling, the drill bits reinforced with PDC plates do not fully work out their life, as part of the cutting PDC elements fails due to their chipping, breakage or loss, which largely affects the final technical and economic indicators of well drilling. The aim of the research is to develop and to propose a neural network model to solve the problem of determining the percentage of breakage of cutting elements of drill bits reinforced with PDC, based on the available field data of drilling wells on rocks of VI–VIII category of drilling capacity. Objects: the reasons causing failure of PDC plates on drill bits in well drilling in rocks of VI–VIII category of drilling capacity depending on the operating parameters of drilling. Methods. The load on the bit, its speed and drilling speed were used as the determining factors for defining the percentage of breakdowns at constant values of the parameters of the washing liquid and bottom-hole assembly. Regression and neural network models of different configurations were used for data analysis. Result. The analysis of regression models showed their unsuitability due to the nonlinear nature of the drilling speed at low pressures and high rpm. The authors proposed a two-stage neural network model in which the first neural network is used to determine the drilling speed, and the second – to predict the percentage of PDC plate breakdowns. Errors of the neural network ensemble on the test data do not exceed the values of the suitability of nonlinear models – 4,5 % for the relative error and 12–15 % of its maximum value. The proposed neural network model can be used in the development of technological regulations of working bits reinforced with PDC cutters.