Learn some strategies to handle missing or incomplete log data for formation evaluation and uncertainty analysis in oil and gas, such as using existing log data, core data, seismic data, machine learning, or uncertainty analysis.

Kern məlumatları ilə karotaj diaqramlarındakı məlumat itkilərini tamamlamaq o qədər də real olmur adətən. Çünki, kern məlumatları daha çox vəsait tələb edən bir tədqiqatıdır və kern nümunələri yalnız müəyyən intervallardan götürülə bilir. Kern nümunələri adətən mövcud karotaj məlumatlarının dəqiqləşdirilməsində istifadə olunur. Yəni, karotaj məlumatları ilə kern məlumatları kompleks şəkildə tətbiq edilərək, quyu lüləsini təşkil edən süxurlar haqqında müfəssəl informasiya əldə edilir. Bununla yanaşı, əgər müəyyən intervalda karotaj məlumatları yoxdursa və eyni intervaldan kern nümunəsi götürülübsə, çatışmayan karotaj məlumatları tamamlana bilər.

Data imputation is an infamous problem when related to filling out missing data from a given data log file (LAS, DLIS).Many methods exist(mathmatical and statistical) that can perform this task with ease, however the decision to use which is subject to engineering judgment. Here is a short list of possible imputation methodology: 1- Carry forward. 2- Backward fill. 3- Mean Value fill 4-Constant value fill. 5-Negative value fill (-9999 is common) 6-Data synthesis using ML and offset well logs 7-Trend analysis and Linear Regression. Not all software packages support all features listed above.

Last updated on Sep 10, 2024

What strategies can you use to handle missing or incomplete log data?

Missing or incomplete log data can pose significant challenges for formation evaluation and uncertainty analysis in oil and gas exploration and production. Log data provide essential information about the physical properties and characteristics of the subsurface layers, such as porosity, permeability, fluid saturation, lithology, and mineralogy. However, due to various reasons, such as operational constraints, technical limitations, data quality issues, or data loss, some log data may be unavailable or unreliable for certain intervals or wells. In this article, we will discuss some strategies that you can use to handle missing or incomplete log data and improve your formation evaluation and uncertainty analysis results.

1 Use existing log data

One of the simplest and most common strategies to handle missing or incomplete log data is to use the existing log data from the same or nearby wells to fill in the gaps or replace the unreliable data. This can be done by using correlation techniques, such as crossplots, histograms, or regression models, to establish relationships between different log measurements and estimate the missing or incomplete values based on the available values. For example, if you have missing density log data, you can use a crossplot of density versus neutron porosity to estimate the density values from the neutron porosity values. However, this strategy assumes that the subsurface layers are laterally homogeneous and consistent, which may not always be the case.

Add your perspective

Mohit Srivastava

Building Innovative Products | Development Strategy | Consulting | Digital | Energy | Application Delivery & Support | NHH | IIMB (Director's Merit List)
Report contribution
Data transfer over any protocol and standard is always prone to data loss due to multiple factors. Many of these factors, like connection between servers, are beyond our scope to keep up always. Even the best connections would be 99.99% availability, and hence, one should always have a process to monitor real-time data and any update in previous data as it happens. This process is best done through scripts instead of manual comparison. Any loss in data can be then identified almost immediately and repaired with the raw channel from the sensor source from the rig.

Like
Muhammad Said Abdallah 🍉🌲

Measurement While Drilling (MWD) Field Engineer at TAQA | Property Officer | Property Management | LWD Field Specialist - RPS/DDx | Research & Development | MSc. Environmental Sciences | BSc. Petroleum Engineering
Report contribution
The use of well-log data from nearby offset wells, combined with seismic data, can help estimate missing well-log data in a target well. By mapping similarities in stratigraphy across multiple wells, geoscientists can fill gaps in the data more accurately. Statistical averaging methods or AI can be applied to improve estimations, with the effectiveness of the approach depending on the quantity and quality of available data. This integrated approach enhances understanding of subsurface geology and supports informed decision-making in oil and gas exploration and production.

Like
İlqar Cəfərov

Geoloq | Azərbaycan dilini geologiyada yaşadıram
Report contribution
Karotaj diaqramlarında hansısa intervallarda məlumat əksikliyi olduqda həmin məlumatları əvəz etməyin ilk üsulu ətraf quyularda müvafiq intervalların məlumatlarından istifadə etməkdir. Məsələn, biz qamma karotaj ilə 11-750 metr intervalında ölçü işi aparmışıq, amma 550-562 metr intervalında məlumat əksikliyi var. Bu zaman ətraf quyuların karotaj məlumatları ilə quyu kəsilişi korrelyasiya (müqayisə) edilməlidir. Məlumat əksikliyi olan intervalın yuxarı və aşağı hissələri ilə kəsiliş dəqiqləşdirilməli və çatışmayan hissələr qonşu quyuların məlumatlarına müvafiq olaraq tamamlanmalıdır. Onsuz da bir-birinə yaxın yerləşə quyuların kəsilişləri də təqribi eyni olur. Bu səbəbdən də belə yanaşma ilə diaqramı tamamlamaq düzgün hesab oluna bilər.

Translated

Like

2 Use core data

Another strategy to handle missing or incomplete log data is to use core data, which are physical samples of the rock formations obtained from drilling. Core data can provide more accurate and detailed information about the rock properties and fluid content than log data, especially for complex or unconventional reservoirs. However, core data are usually sparse and expensive to obtain, so they may not cover the entire interval or well of interest. Therefore, you can use core data to calibrate or validate your log data and extrapolate the core data to fill in the missing or incomplete log data using interpolation or extrapolation techniques. For example, if you have missing porosity log data, you can use core porosity measurements to calibrate your neutron-density porosity log and interpolate the porosity values for the missing intervals.

Add your perspective

İlqar Cəfərov

Geoloq | Azərbaycan dilini geologiyada yaşadıram
Report contribution
Kern məlumatları ilə karotaj diaqramlarındakı məlumat itkilərini tamamlamaq o qədər də real olmur adətən. Çünki, kern məlumatları daha çox vəsait tələb edən bir tədqiqatıdır və kern nümunələri yalnız müəyyən intervallardan götürülə bilir. Kern nümunələri adətən mövcud karotaj məlumatlarının dəqiqləşdirilməsində istifadə olunur. Yəni, karotaj məlumatları ilə kern məlumatları kompleks şəkildə tətbiq edilərək, quyu lüləsini təşkil edən süxurlar haqqında müfəssəl informasiya əldə edilir. Bununla yanaşı, əgər müəyyən intervalda karotaj məlumatları yoxdursa və eyni intervaldan kern nümunəsi götürülübsə, çatışmayan karotaj məlumatları tamamlana bilər.

Translated

Like

3 Use seismic data

A third strategy to handle missing or incomplete log data is to use seismic data, which are acoustic waves that reflect and refract from the subsurface layers and are recorded at the surface. Seismic data can provide information about the structure and geometry of the subsurface layers, as well as some indications of their acoustic properties, such as velocity and impedance. However, seismic data have lower resolution and higher uncertainty than log data, so they may not capture the fine-scale variations and features of the rock formations. Therefore, you can use seismic data to complement or constrain your log data and estimate the missing or incomplete log data using inversion or integration techniques. For example, if you have missing acoustic log data, you can use seismic inversion to estimate the acoustic impedance values from the seismic amplitude values.

Add your perspective

4 Use machine learning

A fourth strategy to handle missing or incomplete log data is to use machine learning, which is a branch of artificial intelligence that uses algorithms and models to learn from data and make predictions or decisions. Machine learning can help you handle missing or incomplete log data by using various methods, such as imputation, classification, regression, or clustering, to infer the missing or incomplete values from the available data or to identify patterns or anomalies in the data. For example, if you have missing lithology log data, you can use a machine learning classifier to predict the lithology classes from the other log measurements. However, machine learning requires sufficient and reliable data to train and validate the models, as well as careful selection and evaluation of the methods and parameters.

Add your perspective

Ir. Muhammad R.
Report contribution
This is one of my favorite methods, but it requires enough data from the surrounding wells and formations. The quality of these data must also be controlled to ensure their validity. After all, as with any model, garbage in will create garbage out. Domain knowledge plays a significant portion in the success of this method.

Like

5 Use uncertainty analysis

A fifth strategy to handle missing or incomplete log data is to use uncertainty analysis, which is a process of quantifying and assessing the effects of uncertainties on your formation evaluation and decision making. Uncertainty analysis can help you handle missing or incomplete log data by using various techniques, such as sensitivity analysis, error propagation, or probabilistic modeling, to estimate the range and distribution of the possible values for the missing or incomplete data and to evaluate the impact of the uncertainty on your formation evaluation results and objectives. For example, if you have incomplete saturation log data, you can use a probabilistic model to generate multiple scenarios of saturation values based on the uncertainty of the log measurements and the reservoir parameters and to compare the scenarios in terms of their economic and technical feasibility.

Add your perspective

6 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

Add your perspective

Nashat Jumaah Omar

Production F.A Engineer & Instructor | Trained +600 Petroleum Engineers on Python |
(edited)
Report contribution
Data imputation is an infamous problem when related to filling out missing data from a given data log file (LAS, DLIS).Many methods exist(mathmatical and statistical) that can perform this task with ease, however the decision to use which is subject to engineering judgment. Here is a short list of possible imputation methodology: 1- Carry forward. 2- Backward fill. 3- Mean Value fill 4-Constant value fill. 5-Negative value fill (-9999 is common) 6-Data synthesis using ML and offset well logs 7-Trend analysis and Linear Regression. Not all software packages support all features listed above.

Like

What strategies can you use to handle missing or incomplete log data?

1

2

3

4

5

6

1 Use existing log data

2 Use core data

3 Use seismic data

4 Use machine learning

5 Use uncertainty analysis

6 Here’s what else to consider

Oil & Gas

Rate this article

Thanks for your feedback

More articles on Oil & Gas

More relevant reading