Log interpretation workflow

Interpretation answers four questions: what the section is made of, where the reservoirs are, what is in them, and how much there is. We answer them step by step, and the dependencies set the order of the steps. The effective porosity of a shaly bed needs the shale volume. The water resistivity from a water-bearing bed needs the porosity of that bed. The water saturation needs the porosity, the water resistivity and the true resistivity of the formation. So an error at any step flows down the chain, and we place checks between the steps, not only at the end.

We can interpret in two ways. In bed-by-bed interpretation, we divide the section into beds, take the characteristic readings opposite each bed, and calculate the properties of the bed. This is how we do a quick-look. It is also how the traditional report on the results of well logging is organised. For each bed it gives the interval, the thickness, the properties and the fluid type: oil, gas, water or “unclear”. In sample-by-sample interpretation, a program calculates the properties at every recording depth and then sums the intervals that pass the cut-offs. The sample-by-sample approach shows changes within a bed better. The bed-by-bed approach is clearer and easier to check. The steps are the same in both, and in this lesson the quick-look goes bed by bed.

Before calculating, we check the data: the heading, the calibrations and the repeat section, the logging speed, the quality indicators of the methods, and the depth matching of the runs (checklist). Then we check the borehole: the caliper log against the bit, the type of mud, and the expected invasion profile (first checks). We mark unreliable intervals at once. In washouts the density and the neutron porosity are distorted, so we evaluate such a bed by other methods or exclude it.

Then we divide the section into beds. This step is qualitative (it uses no calculations), but everything that follows depends on it: each bed gets its own readings and its own parameters.

• Lithology. We recognise shales by a high gamma ray (GR) and by the spontaneous potential (SP) at the shale baseline, the line that the SP follows opposite shales (Gamma ray log and natural gamma ray spectrometry, Spontaneous potential (SP) log). We recognise sandstones, limestones, dolomites and evaporites (such as salt and anhydrite) by density, the photoelectric factor and the crossplots of the porosity methods (Lithology and porosity from crossplots, Lithology from the photoelectric factor and spectral gamma ray). We place bed boundaries at the middle of the transition of the curves.
• Reservoirs. The direct signs of permeability are mudcake on the caliper log (a diameter smaller than the bit), separation of resistivity curves of different depths of investigation, and positive microlog separation on the microresistivity tools (more). An indirect sign is the deflection of the SP from the shale baseline. High density with a low GR and coincident resistivity curves indicate a tight streak (a thin bed with very little pore space).
• Fluids. We compare beds of the same lithology and porosity by their deep resistivity: hydrocarbons raise it many times over (Resistivity logs: electrical and induction tools). With fresh mud, the ratio of the flushed-zone resistivity to the true resistivity, $R_{xo}/R_t$, is close to $R_{mf}/R_w$ in a water-bearing bed. In an oil-bearing bed it is much lower (resistivity profile). The crossover of the neutron and density curves shows gas (more). The oil–water contact (OWC) shows up as a drop in resistivity within one clean bed. The drop is abrupt, or gradual if there is a capillary transition zone above the contact (Capillary pressure and the transition zone).

From here on, we carry out all the steps on one assumed vertical well. The lesson Log acquisition, processing and presentation covers its heading and runs. The formation temperature is about 80 °C, the mud is fresh, and the filtrate resistivity at formation temperature is $R_{mf}$ = 0.191 Ω·m. The density run is tied to the resistivity run by the GR, and the neutron curve is recorded in sandstone units.

We determine shale volume before porosity. We need it to correct the porosity of shaly beds and to choose the saturation model. The main method is the gamma ray index, which shows how far the GR reading lies between the clean-formation baseline and the shale baseline of the same sequence. We use it with a relationship checked against core (Determining shale volume). We check the result with other indicators: the SP and the difference between neutron and density porosity (more). We draw the shale baseline through thick, uniform shales, not through the peak of a highly radioactive streak.

In a sand–shale (siliciclastic) section we most often take porosity from the density log, using the matrix density (the density of the rock grains) from the lithology and the filtrate density (Porosity from the density log). In a shaly bed we subtract the contribution of the shale. This gives the effective porosity $\phi_e$ (shale correction). The neutron and sonic logs serve as checks (Porosity from the neutron log, Porosity from the sonic log). In mixed lithology we find porosity and composition from crossplots (Lithology and porosity from crossplots). In gas-bearing beds we use the density–neutron pair with a gas correction (Gas, shale, carbonate and pressure effects on porosity logs). Nuclear magnetic resonance (NMR) logging divides the pore space into bound and free water (Nuclear magnetic resonance (NMR) logging).

$R_w$ is a single number for the whole interval that contains one water, and all the saturations depend on it (more). We take it at formation temperature and check it in several ways. The ways are water samples taken in tests, the SP, the apparent water resistivity $R_{wa}$ (the water resistivity calculated from the resistivity and porosity of a clean water-bearing bed), and the Pickett plot (a log–log plot of resistivity against porosity) (Determining formation water resistivity, from the SP, Resistivity–porosity crossplots). $R_{wa}$ needs the porosity of the water-bearing bed, which is why this step comes after porosity. The bed itself must be clean and thick, because in shaly beds $R_{wa}$ comes out too low.

We calculate saturation from the true formation resistivity $R_t$, the porosity and $R_w$. $R_t$ is the deep tool reading after corrections for the borehole, the enclosing rocks (the beds above and below) and invasion (how to find it). In clean beds Archie's equation works, with coefficients from core (Archie's equation and saturation of clean reservoirs). In shaly beds the clays conduct current, and Archie's equation overestimates water saturation. There we need a shaly-sand model. We choose it according to the form of the shale distribution (how the shale is spread in the rock), and we check it on water-bearing shaly beds (Water saturation in shaly sands). We choose the model before the calculation, not by which result we like better.

The flushed-zone resistivity gives a second saturation, $S_{xo}$, and with it a check of oil moveability. If $S_{xo}$ is noticeably larger than $S_w$, the filtrate has displaced some of the oil, which means that the oil is moveable (more).

Cut-offs (in Russian-language reserves estimation also called “conditional limits”, from the Russian konditsionnye predely) divide the section. Shale volume and porosity cut-offs separate reservoir from non-reservoir rock. A water saturation cut-off separates productive reservoir from water-bearing reservoir (more). We apply the cut-offs in sequence, and each level gives its own thickness. The gross thickness is the whole bed from top to base. The net thickness is the sum of the reservoir intervals. The net pay is the reservoir with hydrocarbons above the contact. The cut-offs themselves are a decision, not a property of the rock. The advanced section covers where they come from.

Then we average the parameters over the same intervals for which we determined the thickness: porosity over thickness, and water saturation over pore volume. The result for the estimate of hydrocarbons in place is the hydrocarbon pore thickness $H = \sum h\,\phi_e\,(1 - S_w)$. This is the thickness of the layer of hydrocarbons that we would get if we gathered them out of the pores (more). We present the result as a table by bed and a summary for the unit. In the logging report we give each bed a fluid type and recommendations: which intervals to test and perforate.

Each step brings its own uncertainty. It comes from the parameters (the Archie coefficients $a$, $m$, $n$, and $R_w$, $R_{sh}$, the matrix and shale densities), the choice of model and relationship, the GR baselines, the cut-offs, and the data themselves (depth matching, corrections, thin beds). The influence is of two kinds. Continuous parameters shift the result smoothly: an error of 10% in $R_w$ changes $S_w$ by 5% of its value at $n = 2$ (more). “Yes or no” decisions can include or exclude a marginal bed entirely, and so change the net thickness in a jump. These decisions are the saturation model, the shale-volume relationship and the cut-offs.