A.R. Dudaev (NSU, IPGG SB RAS), A.Y. Sobolev (IPGG SB RAS, NSU, NSTU), V.N. Glinskikh* (IPGG SB RAS, NSU, NSTU) SUMMARYWe have developed a prototype of cloud computing software for logging-while-drilling and geosteering. For the first time ever, during the development of software for addressing the problems of petroleum geophysics, we apply software solutions based on state-of-the-art IT-technologies in the field of cloud computing, namely, cross-platform scalable distributed computing. We have created a client-server application written in JavaScript, which stores all user data in MongoDB. The application allows a user to send input data through message queues for processing by means of computing applications that run in a virtual environment. Processing results are displayed on a web page. We have tested the software under discussion on real practical data from the interval of a subhorizontal well from one of the Latitude Priob oil fields. When drilling highly deviated boreholes, an effective well targeting in the productive part of a reservoir is essential. Successful Russian experience of studying oil and gas wells is demonstrated by using the first Russian telemetry system for logging-while-drilling, intended for drilling wells with horizontal completions. At the present time, it is necessary to develop a multifunctional automated computer system for processing, visualization and interpretation of the data obtained with the complex under consideration.

No abstract

The paper presents an algorithm for estimating the error of the 1D geomechanical model taking into account the measurement errors of the instruments and the correlation dependencies used. The sensitivity analysis of the resulting error to the quality of the source information was also performed. A confidence parameter has been introduced to assess the contribution of a measurement to the calculations. An approach is proposed for the transition from the convergence of correlation dependencies to the error of the correlation used. The application of this methodology allows at the initial stage to draw a conclusion about the quality of simulation results and take measures to increase the reliability of calculations.

Geomechanical modeling is an integral part of the oil and gas industry and is used in all life cycles of the field - monitoring and improving the efficiency of well construction, choosing a completion system, modeling hydraulic fracturing processes, modeling development processes taking into account changes in the stress state of the reservoir, taking into account the fault, salt tectonics, control over the development of the reservoir, control of subsidence of the earth's surface. The success of geomechanical modeling directly depends on the quantity and quality of input data. In contrast to the geological and hydrodynamic models, in geomechanics there is still no unified approach and algorithm for quantifying the model error. The quality of the geomechanical model is defined as "satisfactory" / "not satisfactory" and "confirmed by actual data" / "not confirmed by actual data". In a series of articles on "Metrological support of a geomechanical model", the authors show an algorithm for a quantitative assessment of the error of a geomechanical model. The proposed algorithm takes into account the measurement error (in the well and in the laboratory), the quality of logging data, direct measurements or reconstructed measurements, the tightness of correlations (both for the results of core studies and for the reconstruction of missing logging data), the calculation of the uncertainty taking into account the calibration information. This paper describes a generalized algorithm for quantifying the error of a geomechanical model, presented in previous articles, and provides a method for quantifying calculate the uncertainty, taking into account calibration information, such as measurements of horizontal stresses, core studies in laboratory conditions.