Operation well data is a foundation of effective analysis. Unfortunately, some parts of data are not always correct, so it is difficult to perform meaningful analysis. The following problems are often encountered: the lack of information for some time intervals, measurements do not always correspond to a physical model or do not coordinate with each other. Similar problems can be associated with both malfunction of gauge devises and human factor.

Results of production solutions directly depend on the quality of initial geological and production data. Solutions based on inaccurate information can lead to negative consequences, therefore it is necessary to verify data operatively for accuracy and consistency. Currently, the amount of incoming information increases and requirements for data quality rise; therefore it is impossible to verify data manually. As a result it becomes an urgent task to develop algorithms for automated analysis of field data.

This paper presents an overview of hierarchy of approaches to data management in production and development. The most promising areas for development are mentioned in the paper. Three approaches to verify data are considered in details. Also a practical example is given for each approach. The first is based on conventional statistical methods, the second relies on physical models and the last one uses algorithms for processing large amount of information - machine learning methods. Considering constant growth of incoming information quantity, it seems most promising to develop the area of Data Science (in collaboration with analytical models), which focuses on integrated approach for analyzing large amounts of information using machine learning methods.

References

1. Mirzadzhanzade A.Kh., Khasanov M.M., Bakhtizin R.N., Modelirovanie protsessov neftegazodobychi. Nelineynost', neravnovesnost', neopredelennost' (Modelling of oil and gas production processes. Nonlinearity, disequilibrium, uncertainty), Moscow-Izhevsk: Publ. of Institute of Computer Science, 2004, 368 p.

2. Shulenin V.P., Matematicheskaya statistika (Math statistics), Part 2: Neparametricheskaya statistika (Nonparametric statistics), Tomsk: NTL Publ., 2012, 388 p.

3. El-Khatib N.A., Waterflooding performance of communicating stratified reservoirs with log-normal permeability distribution, SPE 37696-MS, 1997.

4. Brooks R.H., Corey A.T., Hydraulic properties of porous media, Colorado State Univ. Hydrol. Paper, 1964, no. 3, 27 p.

5. Breunig M.M., Kriegel H.-P., Ng R.T., Sander J., Identifying density-based local outliers, Proceedings of the 2000 ACM SIGMOD International Conference on Management of Data, SIGMOD, 2000, pp. 93–104.Operation well data is a foundation of effective analysis. Unfortunately, some parts of data are not always correct, so it is difficult to perform meaningful analysis. The following problems are often encountered: the lack of information for some time intervals, measurements do not always correspond to a physical model or do not coordinate with each other. Similar problems can be associated with both malfunction of gauge devises and human factor.

Results of production solutions directly depend on the quality of initial geological and production data. Solutions based on inaccurate information can lead to negative consequences, therefore it is necessary to verify data operatively for accuracy and consistency. Currently, the amount of incoming information increases and requirements for data quality rise; therefore it is impossible to verify data manually. As a result it becomes an urgent task to develop algorithms for automated analysis of field data.

This paper presents an overview of hierarchy of approaches to data management in production and development. The most promising areas for development are mentioned in the paper. Three approaches to verify data are considered in details. Also a practical example is given for each approach. The first is based on conventional statistical methods, the second relies on physical models and the last one uses algorithms for processing large amount of information - machine learning methods. Considering constant growth of incoming information quantity, it seems most promising to develop the area of Data Science (in collaboration with analytical models), which focuses on integrated approach for analyzing large amounts of information using machine learning methods.

References

1. Mirzadzhanzade A.Kh., Khasanov M.M., Bakhtizin R.N., Modelirovanie protsessov neftegazodobychi. Nelineynost', neravnovesnost', neopredelennost' (Modelling of oil and gas production processes. Nonlinearity, disequilibrium, uncertainty), Moscow-Izhevsk: Publ. of Institute of Computer Science, 2004, 368 p.

2. Shulenin V.P., Matematicheskaya statistika (Math statistics), Part 2: Neparametricheskaya statistika (Nonparametric statistics), Tomsk: NTL Publ., 2012, 388 p.

3. El-Khatib N.A., Waterflooding performance of communicating stratified reservoirs with log-normal permeability distribution, SPE 37696-MS, 1997.

4. Brooks R.H., Corey A.T., Hydraulic properties of porous media, Colorado State Univ. Hydrol. Paper, 1964, no. 3, 27 p.

5. Breunig M.M., Kriegel H.-P., Ng R.T., Sander J., Identifying density-based local outliers, Proceedings of the 2000 ACM SIGMOD International Conference on Management of Data, SIGMOD, 2000, pp. 93–104.