Data Hierarchy

Data Hierarchy – The Correct Basis for Petrophysical Data Integration

The ever increasing diversity and abundance of petrophysical data presents choices as to how to best use numerical data and qualitative information in a quantitative petrophysical workflow.  The decision as to which data to use, where and how is usually a subjective process left to the expert.  He/she rarely discusses options with other disciplines and often remains unaware of possibilities. This process of which data to use, where and how should be further examined and formalised with guidelines as to why data is or is not useful.  This thinking does not yet seem to have evolved in companies but is the necessary first step to robust integration.  Much useful data is ignored by petrophysicists.  We must remain eternally vigilant as to the relevance and possible function of new data items in our existing workflows.  Workflows must evolve with new data, that is our task, to use everything at our disposal in the most efficient manner.  Most established workflows today need a complete re-think to accomodate and exploit all the data available for that field, Reservoir Type and Reservoir Rock Type (RRT)

The basis for how to rank, characterise and fully exploit new data is the objective of Data Hierarchy. This topic is covered formally in IPRC, IPCFR and other PETROPHYSICS Pty Ltd training.

A Passive “Comparison with” or an Active “Calibrate to”?  What does the experienced and competent petrophysicist intuitively use as criteria for reservoir Data Hierarchy, that is ranking data for its usefulness? This petrophysicist suggests Directness, Accuracy and Spatial Definition are the root criteria we intuitively look for. How do we navigate through competing data sets and methods? Experienced petrophysicists, geologists and engineers: What do you use?  Analyse That!


Why Log Free Fluid Volume?  Think Data Hierarchy

Do we really need FFV for Permeability ? The PETROPHYSICS method does not require a log of Free Fluid Volume (FFV) or Log Mean T2 to solve for permeability. This is advantageous during logging operations because it avoids the need to log the far more problematic FFV with an NMR tool – for example FFV cannot easily be obtained in gas zones and logging is often slowed dramatically to an expensive crawl in the belief that the full T2 spectrum must be logged to achieve useful NMR results. Not so. Bound Fluid Volume (BFV) is the easiest and fastest log to obtain with an NMR tool. It is usually 100% water, rapidly polarizes and is rapidly counted in early T2 time. With this, a robust and accurate permeability is achieved by the PPL approach, in clastics and non-vuggy carbonates. Whilst service companies may be aware of how this can be achieved it usually is not in their interests to advance such an approach. Service companies job is to run logs and create data not to evaluate your reservoir. Indeed, many wells can be accurately evaluated for permeability without even acquiring NMR data, set your “Interested Petrophysicist” to work on that one.


Elan, Multimin, Mineral Solver. Probabilistic or Deterministic Petrophysics?

“It should be noted that whilst over-determined, error-minimization (probabilistic) log analyses appear attractive they must demonstrate that the discontinuous core data has in fact been correctly extrapolated beyond core, rather than simply matched over core. If they fail to demonstrate mechanism(s) by which the cored answers are locked into the log analysis model applied beyond core they have failed to exploit the expensive and more accurate core data, and revert back to ‘log analysis’ – defeating the petrophysical objective of coring. The rigid extension of core data into uncored intervals is more difficult to demonstrate with probabilistic petrophysics than with deterministic petrophysics, and is a primary reason why simple deterministic methods persist in the face of what might appear as superior log analysis techniques. Again, it is the intention of this author that xx Field petrophysics is core based, not log based.”


“The interested petrophysicist may wish to consider that the use a multi-well, multiple linear regression of these same curves (portmlr) provided a better fit to core but a poorer overall porosity. This is because the fixed point zero porosity is lost in the MLR process and the scaling of pordn2sc (density-neutron shale corrected porosity) is forced to a non-feasible value (*0.374), because core porosity is of limited range. We know the pordn2sc factor needs to be ~ *1.00! It is worth noting that the deterministic fixed point technique (Juhasz 1988, Deakin 2002) overcomes the all too common limited range of core data by providing a zero porosity data point, and is the reason why the often ignored core grain density should be integrated into the log analysis. Also, the ‘interested petrophysicist’ may wish to ask him/herself how an Elan/Multimin approach integrates this hard data?”