Petroleum Engineering: Formation Evaluation
1. Since the formation evaluation techniques cover over 10 orders of magnitude (gross structure using satellite imagery [10^6 m] to micro-pore structure using SEM [10^-6 m]), while the frequencies used by the relevant measuring processes cover nearly 20 orders of magnitude {10^9 Hz for Dielectric to 10^-9 Hz for Material Balance}, how exactly ML/AI would be able to efficiently make use of the involved data with such a wide range of measurement techniques over such a wide range of physical dimensions?
2. What is the fraction of ‘actually’ gathered data from formation evaluation (such as by seismic records; by coring; by mud logging; by formation testing; and by conventional logging) - as against the interpreted data - that is required for both economic analysis and production planning?
3. While deducing time-depth relationship towards calibrating conventional seismic and VSP surveys, whether a geophysicist would be able to distinguish between a single-continuum and a multi-continuum hydrocarbon reservoir?
4. How easy for a geologist to deduce the reservoir boundaries along with sources and skins - by delineating the stratigraphy of the formations, the structural and sedimentary features, and the mineralogy of the formations through which the well remains drilled?
5. Whether a reservoir engineer would be able deduce the reservoir geometry, reservoir dimension as well as the required coordinates (radial/Cartesian) by knowing the vertical and lateral extent of the reservoir, its porosity/permeability, fluid content, and recoverability?
6. Whether a production engineer will be able to have a sound control over reservoir pressure (abnormal/sub-normal pressures) – towards pressure maintenance and/or water-flooding - by knowing reservoir rock properties; sand management; completion problems; formation injectivity and residual water saturation?
7. Whether a reservoir manager will be able to deduce an economic study precisely - with the aid of ML/AI – by knowing the amount of hydrocarbons in place; their recoverability; the cost of development; and the profitability of producing the reservoir?
8. While log measurements can provide either a direct measurement or a good indication of lithology; sedimentary environment; formation dip and structure; travel times of elastic waves; type of pore fluids; water saturation; primary/secondary porosity; (along with the details of pressure, residual oil saturation, flow rates and fluid types associated with the logging in producing wells); - what exactly, we mean by ‘permeability’ from log measurements?
For that matter, even, if we integrate logging along with coring, core-analysis and formation testing; whether can we still succeed in capturing the ‘dynamic reservoir behavior’ (and not the ‘dynamic well behavior’)?
These are a set of complex questions involving multiple aspects of petroleum engineering. To help, I'll focus on your last question about 'permeability' from log measurements.
Permeability measures how easily fluids can flow through a porous medium (in this case, the reservoir rock). It's usually quantified in Darcies (D) or milliDarcies (mD), and it's a crucial parameter for predicting the productivity of a well or reservoir.
In the context of log measurements, permeability is often not directly measured but inferred from other properties. These may include:
- Resistivity Logs: Higher permeability zones often correlate with higher resistivity because they contain more hydrocarbons and less conductive formation water.
- Porosity Logs: Permeability often correlates with porosity because higher porosity usually means more pore spaces for fluid to flow through.
- Nuclear Magnetic Resonance (NMR) Logs: These logs can provide a more direct measure of permeability by measuring the response of hydrogen nuclei in the pore fluids to a magnetic field. NMR logs can provide information on pore size distribution, which can be used to infer permeability.
- Formation pressure tester (FPT) or Formation Tester: These tools can provide direct permeability measurements. They establish a pressure drawdown on a small formation section and measure the pressure response.
However, it's important to note that logs tend to measure properties at a smaller scale, often in the near-wellbore area. They might not fully capture larger-scale heterogeneities in permeability that can significantly affect production. This is why integrated studies that combine logs with core data, well tests, and possibly seismic data are often necessary to understand the reservoir fully.
As for capturing 'dynamic reservoir behaviour,' this is typically the role of reservoir simulation, which integrates all available data (logs, cores, tests, production data, etc.) to build a reservoir model that can be used to predict future performance under different production scenarios. These models attempt to capture the dynamic behaviour of the reservoir, although they inevitably involve some simplifications and assumptions.
For the other queries, each involves an extensive discussion that would be better addressed individually. Most of them revolve around how different disciplines in petroleum engineering, from geology to reservoir management, can integrate data, possibly using Machine Learning/AI, to understand better and predict the behaviour of hydrocarbon reservoirs. The use of ML/AI in this context is a rapidly evolving field, and its effectiveness would depend on the specific application and the quality and quantity of the data available.
Agree with Joshua's post..
Estimation of permeability from logs is most often done using semi-empirical formulas based on a capillary bundle model (Karman-Kozeny). These include estimated porosity and estimated bound water saturation as key parameters. In most cases these are provided by the log data indirectly. Consequently, the often strong anisotropy of e.g. permeability cannot be estimated very well. The description of the friction at the boundary between flowing water and the inner surface of the rock is of course rather understated.
(an exception to this is the formation tester, which gives direction-dependent results)
It is not clear for me whether you want to get definition or to know the mathematical formula of permability logs
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Atmosphere
- Air