A typical example of reservoir rock from the study area (in thin section) under plane polarized light (top) and crossed nicols (bottom). The sample is mostly unidentifiable crystals and volcanic glass. (Images courtesy of Schlumberger)

Finding hydrocarbons in the region known as the ring of fire is difficult enough, but when the traditional rules of formation evaluation are bent out of shape by the presence of volcanic rocks, an innovative approach is required.

In northeastern China, the YingCheng Group reservoirs in the SongLiao Basin are characterized by a complex heterogeneity consisting of crystalline volcanic and volcaniclastic rocks. Traditional lava flows punctuated by violent eruptions of volcanic ash and magma as well as steam explosions that leveled entire mountainsides took millions of years of overlying sediments and mixed them into a vast geological puzzle.

Natural gas was discovered in volcanic rocks of the YingCheng Group near the Daqing oil field in the early 1970s. The difficult drilling and production environment and the challenging reservoir geology precluded development.

PetroChina knew there were hydrocarbons, but the company did not know the quantity or if the reservoir could be profitably produced.
New formation evaluation technology emerged since the discovery of natural gas reservoirs near Daqing. In 2004, a joint PetroChina/Schlumberger study team began to investigate how new measurements could be combined in complementary workflows to solve the challenges posed by volcanic reservoirs. The approach involved drilling, coring and logging nine appraisal wells, then developing an integrated reservoir evaluation program using a suite of complementary logging measurements benchmarked by a comprehensive core analysis dataset.

Volcanics defy traditional methodology

Traditional well log interpretation is based on relationships that have worked reasonably well for decades in conventional clastic or carbonate reservoirs. But these relationships provide sustainable results only under a limited set of mineralogical and lithological conditions. The relationship relating formation density measurements to porosity, for example, depends on accurate knowledge of matrix density. In sandstones or carbonates, this is fairly easy to determine — not so in volcanic rocks.
Electrical resistivity is the traditional measurement for determining hydrocarbon saturation. However, the relationship between saturation and resistivity becomes increasingly complex as pore geometry becomes more complicated or if conductive minerals such as zeolites and clays are present. Complex pore geometry and mineralogy found in volcanic reservoirs make using the standard Archie relationship impractical.

Traditional reservoir characterization techniques have depended on classification of reservoirs into either clastics or carbonates, each with its specialized interpretation approach. Because volcanic reservoirs consist of a rock matrix very similar to complex clastics, but with pore network geometry more common to carbonates, a non-traditional approach to formation evaluation and reservoir characterization was indicated.

Complementary calculations

The team determined that an appropriate suite of complementary logging measurements would stand the best chance of accurately characterizing the reservoir. The question was how logging measurements could be linked to reservoir properties. The solution was to perform exhaustive analyses of the nine whole cores and to use this information to calibrate the responses of the logging suite tools.

The problems were considered from each angle:
• Lithology – Although the fine crystalline nature of volcanic rocks complicates the use of conventional techniques to determine mineralogy, geochemistry provides a way to identify lithology.
• Porosity – Traditional approaches to porosity estimation give less than satisfactory results when there is an unknown matrix density and gas is present in the formation. The solution is to integrate measurements with particular sensitivity to the rock matrix with those having particular sensitivity to the pore fluids, addressing the complexities of both the rock matrix and contained fluids.
• Pore geometry and permeability – The volcanic rocks in the study could be classified as fractured crystalline volcanics, autoclastic breccias, and microporous tuffs. Each type has a characteristic pore geometry and permeability response, and so contributes differently to the overall storage capacity and flow capacity of the reservoir.
• Saturation – Due to the very complicated relationship between resistivity and fluid saturations in these volcanic reservoir rocks, an innovative nuclear magnetic resonance (NMR) to capillary-pressure relationship, calibrated by core measurements, was used to compute saturation by balancing the capillary pressure with the buoyancy pressure in the reservoir.

An “elementary” approach

Evaluating mineralogy in volcanic rocks is often impossible due to the very small crystals and large amounts of volcanic glass that form during rapid cooling of magma (Figure 1). Therefore, a common approach to understanding volcanic lithologies is to use geochemical data obtained from rock samples.

Leveraging a capability to obtain chemical information from logging measurements, concentrations of elements comprising the formation ware determined using a neutron capture spectroscopy tool, the ECS Elemental Capture Spectroscopy sonde. Often the presence of high concentrations of chlorine from high salinity drilling mud or barite from mud weighting materials can complicate calculations, but in the nine appraisal wells, care had been taken to maintain very low borehole salinity and drill without barite mud. This permitted the team to discriminate the maximum number of elements from neutron capture spectroscopy, namely silicon, aluminum, iron, titanium, calcium, sodium, potassium, sulfur and gadolinium. Using these data, a geochemistry-based classification scheme devised by the International Union of Geological Sciences — the Total Alkali’s vs. Silica scheme — provided a continuous description of the volcanic lithologies (Figure 2).

Porosity was determined by combining the results of density logging with those of NMR logs from the CMR-Plus logging tool with high-logging-speed capability. The first is sensitive to both the rock matrix and fluids, the second to fluids alone. Using these measurement together allowed fluid effects to be overcome, but complex variations in rock matrix density still needed to be addressed (Figure 3). Core measurements of chemistry and matrix density were used to build a series of transforms to compute matrix nuclear properties from the geochemistry data available from neutron capture spectroscopy. When the derived matrix density was input to the bulk density equation and combined with NMR data, a useful byproduct of these calculations was invaded zone gas saturation, an indicator of the presence of hydrocarbons.

The true shape of things

Pore geometry and permeability were estimated by applying a technique first developed for use in carbonates whereby pore network geometry is characterized into three main pore classes: microporosity, mesoporosity and macroporosity. First, the total pore volume is subdivided into fractions associated with these classes with assistance from NMR measurements, specifically the T2 distribution. From this measurement, a short cutoff value can be established to isolate the microporosity. Similarly, a long T2 cutoff is used to isolate the macroporosity. What’s left is mesoporosity.

Borehole microresistivity images such as those provided by the FMI Fullbore Formation MicroImager tool are used in conjunction with T2 to further quantify the macroporosity. A comparison of these results to core values of pore throat radius distributions derived from mercury injection capillary pressure (MICP) measurements was very encouraging, and the approach provided reasonable permeability estimates for this complex environment (Figure 4).

Determining saturation normally involves using formation resistivity measurements in conjunction with the Archie relationship. However, a different method was needed because of the complexities of the pore geometry, resulting in extreme variability in values for the cementation factor, m, and the saturation exponent, n, coupled with the presence of highly conductive minerals such as zeolites and clays.

The approach implemented a resistivity-independent method of finding saturation. Pseudo-capillary pressure curves derived from NMR T2 distributions yielded the best result. In an iterative approach, four fitting parameters of a transform equation were calibrated by comparing pseudo-capillary pressures from measured T2 distributions with corresponding MICP curves to achieve the best match. A balance exists between capillary pressure and buoyancy pressure in a reservoir at equilibrium. Saturation can be deduced by determining the buoyancy pressure.

The team constructed a reservoir model that considered the entire volcanic interval as a single pressure system. A single free water-level and buoyancy pressure distribution were hypothesized from this model. Saturation was computed by applying the buoyancy pressure distribution to the pseudo-capillary pressure curves. The final solution was achieved in an iterative process aiming to satisfy all observations and measurements of fluid distribution.

Implementing these complementary measurements and methodologies in an integrated workflow enabled a comprehensive evaluation of reservoir properties in this extremely challenging environment (Figure 5). The techniques pioneered in China open the door to resolution of petrophysical analyses of volcanic reservoir rocks elsewhere in the world.