The problem of timely and accurate evaluation of well integrity is becoming increasingly relevant in the context of mature field development, high wellstream water cut, and a growing number of old wells. For production casing diagnostics, geophysical methods are typically used to identify damage and determine its interval. However, high workload of field personnel hinders prompt deployment of wireline crews to survey the integrity of wells. This results in lost oil production, increased water cut, environmental risks, increased non-productive injected volumes, and reduced key economic indices. To address these challenges, a novel approach to evaluation of casing string integrity based on machine learning models has been proposed. The paper presents a procedure for application of interpretable machine learning to detect production casing leakage and provides a comparison of this approach with the ROC-AUC statistical analysis method. The novel approach integrates the LightGBM machine learning algorithm and SHAP analysis to evaluate contribution of key features to well integrity prediction and determine their threshold values. The model training was based on data from 14,318 well surveys conducted between 2000 and 2022. The results indicate that the most important features are sulfate content, solution supersaturation ratio, and water cut. The study confirms the efficiency of interpretable machine learning methods for diagnosing complex technical systems. These results show the potential for application of such models in well integrity monitoring and well workover planning. This approach can also be used in other oil and gas applications, such as prediction of various problems and optimization of well operation conditions.

It is known that much of the technology aimed at intensifying fluid inflow by means of hydraulic fracturing involves the use of proppant. In order to transport and position grains in the fracture, a uniform supply of proppant with a given concentration into the fracturing fluid is ensured. The aim of the operation is to eliminate the occurrence of distortions in the injection program of proppant HF. A mathematically accurate linear increase of concentration under given conditions is possible only if the transient concentration is correctly defined. The proposed approach allows to correctly form a proppant HF work program for both linear and non-linear increase in proppant concentration. The scientific novelty of the work lies in application of a new mathematical model for direct calculation of injection program parameters, previously determined by trial and error method. A mathematical model of linear and non-linear increase of proppant concentration during HF was developed. For the first time, an analytical solution is presented that allows direct calculation of parameters of the main HF stages, including transient concentrations for given masses of the various types of proppant. The application of the mathematical model in formation of a treatment plan allows maintaining correct proppant mass distribution by fractions, which facilitates implementation of information and analytical systems, data transfer directly from a work program into databases. It is suggested to improve spreadsheet forms used in production, which would allow applying mathematical model of work program formation at each HF process without additional labour costs. The obtained mathematical model can be used to improve the software applied in the design, modelling and engineering support of HF processes.