Summary
The acquisition of gas in mud data while drilling for geological surveillance and safety is an almost universal practice. This source of data is only rarely used for formation evaluation because of the widely accepted presumption that it is unreliable and unrepresentative. Recent developments in the mud-logging industry to improve gas data acquisition and analysis have led to the availability of better quality data.
Within a joint Elf/Eni-Agip Div. research program, a new interpretation method has been developed following the comprehensive analysis and interpretation of gas data from a wide range of wells covering different types of geological, petroleum, and drilling environments.
The results, validated by correlation and comparison with other data such as logs, well tests, and pressure/volume temperature (PVT) data, enable us to characterize lithological changes; porosity variations and permeability barriers; seal depth, thickness, and efficiency; gas diffusion or leakage; gas/oil and hydrocarbon/water contacts; vertical changes in fluid over a thick monolayer pay zone; vertical fluid differentiation in multilayer intervals; and biodegradation.
The comparison of surface gas, PVT, and geochemistry data clearly confirms the consistency between the drilling gas data (gas shows) and the corresponding reservoir fluid composition.
The near real-time availability, at no extra acquisition cost, of such data has led to:The optimization of future well operations (such as logging and testing).A better integration of while-drilling data to the well evaluation process.A significant improvement in both early formation evaluation and reservoir studies, especially for the following applications, in which traditional log analysis often remains inconclusive:Very-low-porosity reservoirs.Thin beds.Dynamic barriers and seal efficiency.Low-resistivity pay.Light hydrocarbons.
Examples show gas while drilling (GWD) wellsite quicklook interpretations with simple lithological and fluid interpretations, as well as more complex reservoir and fluid characterization applications in varied geographical and geological contexts; both demonstrate how GWD data are integrated with more standard data sets.
Introduction
The measurement of gas shows is standard practice during the drilling of exploration and development wells.
Continuous gas monitoring sometimes enables us to indicate, in general terms, the presence of hydrocarbon-bearing intervals, but it rarely allows us to define the fluid types (oil, condensate and/or gas, and water).
Gas data are at present largely underused because they are considered unreliable and not fully representative of the formation fluids.
There are many reasons for this. On one hand, poorly established correlations exist between reservoir fluids and shows at surface; on the other hand, numerous drilling parameters strongly influence the recorded gas data, such as formation pressure, mud weight and type, gas-trap position in the shaker ditch, and mud-out temperatures. One reason may be the very low cost of such data, often equated with low value.
Until a few years ago, the analysis performed on gas shows was generally restricted to the use of Pixler and/or Geoservices diagrams (or equivalent), wetness, balance, character, and gas normalization.1–4
Recent improvements in gas-acquisition technology and the new GWD methodology allow us to perform reservoir interpretation in near real time for fluid identification and contacts [oil/water contact (OWC), gas/oil contact (GOC), etc.], lithological changes, and barrier efficiency, thus allowing operations optimization (e.g., coring, wireline recording and sampling, and testing operations). It is also possible to integrate the GWD interpretation in reservoir, geochemical, PVT analysis, and comprehensive studies.
Method
Data Acquisition.
The measurement of gas shows in the circulating drilling mud was introduced in the early days of mud logging (ML) with two objectives: first, as a safety device to indicate well behavior to drillers, and second, as an indicator of hydrocarbon-bearing zones. Today, gas-shows measurement is systematically acquired in the petroleum industry for the same reason, but it is seldom used to its full potential, mainly because of an ongoing prejudice that the data are not representative of the formation fluids and/or that the recording of these data is strongly influenced by varying drilling parameters.
The ML gas system is composed of three parts:A "gas trap" to extract gas from the mud stream situated somewhere between the bell nipple and the shaker box (often in the latter).Lines, pumps, and filters enabling the transport of a dry-gas sample to the ML unit.A detection system in the ML unit.
Recent efforts in the mud-logging industry to improve gas-data acquisition and analysis have led to the availability of better quality data, which has provided reliable lithological and fluid information since the 1990s.
In the 1980s, most of the ML companies introduced the flame ionization detectors (FID) to replace previous total gas (TG) and chromatograph measurements. The TG measurement gives the total amount of hydrocarbon components extracted from the mud and burned in the detector. The TG could now be correlated with the C1-C5 readings from the new breed of chromatographs.5
Finally, over the past few years, several ML companies have introduced fast-gas chromatographs with improved resolution (C1-C5 in less than 1 minute), improved C1/C2 separation, and, above all, improved reliability and repeatability. High-speed chromatographs using a thermal-conductivity detector have also appeared on the market, but they were not tested within this project.
Work carried out by Texaco in the early 1990s led to a significant improvement in basic trap design with the introduction of the quantitative gas measurement (QGM) trap, which was a major step in reducing the effect of environmental changes.6 An alternative proposition from Geoservices was to replace the trap, generally situated in the shaker box, with a pumping system supplying the trap with a constant volume of mud sucked from a probe situated close in the flowline to the bell nipple.7
Detection of Abnormalities in Real-Time Computer Network Traffic Empowered by Machine Learning
