1. Introduction
Absolute liquid permeability value is an important parameter for conducting dynamic reservoir simulation for any oil and gas reservoir. Industrial practice is to use either Klinkenberg-corrected permeability or water permeability as an absolute liquid permeability value [
1,
2]. Theoretically, Klinkenberg-corrected permeability and water permeability of a sample should be similar; however, experimental work shows clear differences between Klinkenberg-corrected permeability and water permeability values.
Measured gas permeability is subjected to gas slippage effect, also known as Klinkenberg effect, which can result in overestimation of gas permeability value. This theory was first established by L. Klinkenberg et al. [
3]. Later, M. Muskat et al. [
4] observed significant differences in permeability values between air and water [
5]. Thus, L. Klinkenberg [
3] defined that Klinkenberg effect occurs when the mean free path of gas molecules in any porous media approaches the pore dimension. This phenomenon will lead to more frequent collision between gas molecules and the pore wall than the collisions between gas molecules, which reduces viscous drag, thus enhancing gas slip flow and increasing the gas permeability values [
2,
6,
7,
8]. As a result, gas permeability must be corrected to Klinkenberg-corrected permeability after applying infinite differential pore pressure [
7]. Since at infinite differential pore pressure, gas flows as a liquid-like fluid, theoretically Klinkenberg-corrected permeability of a core sample must be similar to its water permeability value. This observation is supported by [
5,
7,
9,
10], which state that the permeability of a sample should be independent of its pore fluid. However, experimental results have shown that Klinkenberg-corrected permeability and its respective water permeability values show significant differences. These observations are in close agreement with previous studies [
1,
7,
9,
10,
11,
12,
13,
14,
15,
16,
17,
18]. Due to these discrepancies, significant amounts of time and cost have been spent on measuring both permeability values to obtain an absolute liquid permeability value. Thus, by developing a correlation between Klinkenberg-corrected permeability and water permeability for specific regions, the amount of time and cost spent on obtaining absolute liquid permeability values will be reduced by using the developed correlation.
Based on previous research, there are a wide range of factors influencing the water permeability value of core samples such as clay swelling, rehydration of unreacted minerals, dissolution/precipitation of matrix, fines migration and also water adhesion to the smallest pores of the matrix [
7,
9,
14,
19,
20,
21]. B. Kanimozhi et al. [
22] reported that fines migration is becoming a serious problem in carbonate reservoirs and disturbing the well productivity of affected reservoirs. Although fine particles are usually observed in sandstone reservoirs, there are specific carbonate reservoirs that have reported the presence of fine particles. Thus, carbonate rocks are also susceptible to water permeability reduction caused by fine migration and plugging, reactive flows and geochemical reactions and alterations [
23].
Similar studies [
10,
14,
16,
18] have argued that the inaccessibility of water to flow through micro-pores and micro-cracks compared to inert gases caused the difference between Klinkenberg-corrected and water permeability. The ability of inert gases to flow through the double network system (matrix + naturally induced fractures) will cause the gas to cover more area of the pore system to have a higher permeability value [
10,
16]. Water, in turn, will chemically react with clay particles in the core samples, which will form layers of clay-bound water film on the pore throat and micro-crack surfaces that will reduce the effective radius of these pore systems and cause inaccessibility of water to flow through the micro-pores and micro-cracks [
10,
18]. However, other researchers [
14,
16] have argued that the values of water permeability were still lower than Klinkenberg-corrected permeability of samples that contains less than 0.1% clay particles. M. Chen et al. [
14] also stated that the difference between Klinkenberg-corrected and water permeability values of their samples were not due to clay–water reaction but rather it was due to the steady-state flow that formed thin water films around the grain matrix through fluid storage and mechanical coupling. Moreover, the presence of water films at high curvature contact points, inter-grain separations, and the grain surfaces were also proved. This observation is also supported by M. Heap et al. [
16], which stated that the tortuous, kinked nature of the samples and the rough nature of the micro-cracks will allow adsorption of water, which will reduce water permeability values. It was also demonstrated that the adsorbed water molecules do not have to block the entire length of a micro-crack, but only a small part, for instance, a rough-walled section or at a tight bend that will hinder the flow of water and result in a water permeability reduction.
A number of studies have been conducted previously [
1,
7,
9,
10,
11,
12,
13,
14,
15,
16,
17,
18] regarding the relationship between Klinkenberg-corrected and water permeability values, however most of the studies were based on sandstone formations and also very few selective carbonate formations such as Shiuaba carbonate formation in Oman [
1], Zelatowa dolomite formation in Poland [
10] and selected unnamed carbonates [
15]. The correlation established based on sandstone samples cannot be applied on carbonate formations due to the complexity and uncertainty of pore geometry that ranges from big and interconnected pore systems to micro pore structures that only contain intra granular pores or even irregular pore systems such as vugs and pore spaces created from grain dissolution [
1,
10,
24].
Field Geological Background
PETRONS gas field is located offshore in Sarawak, Malaysia and contains prolific quantities of gas reserves. Based on the report of on the PETRONS-1 well [
25], the depositional environment was interpreted to be predominantly shallow platform, back reef to reef flat below 1635.5 m (Middle of Core 3) with slight deepening up to Core 2 and Core 1. The relative abundance of coral fragments below 1635.5 m is taken to indicate general reef proximity. Abundant coral fragments may represent an original patch reef environment and describes the environment as broken reef or reef flat. Wackestones and packstones/wackestones with local red algal/foram rhodoliths represent a transition to back reef lagoonal environments. Clews et al. [
25] reported that above 1635.5 m, the depositional environment becomes progressively deeper with increasing abundance of deep water forams and sporadic colonization by platy corals (top of Core 3 and Core 2). Based on [
26], all six cores from PETRONS-1 consist entirely of limestone, mainly comprising algal foram packstones, wackestones and grainstones with locally common coral fragments more than 5 cm (floatstones). The limestone contains common argillaceous laminae in Core 2 and the top of Core 3 [
26]. Additionally, the porosity types for PETRONS-1 cores are predominantly mouldic with lesser vuggy porosity and significant microporosity. Intergrannular porosity is significant in the sample from 1645.23 m, which consists of abundant fragments of neomorphosed coral. Porosity here is enhanced by dissolution and brecciation. Furthermore, it was reported that microporosity is significant in samples below 1635.5 m, giving a characteristics chalky texture to the limestone. Microporosity has probably been preserved at an early stage of diagenesis during early lithification or neomorphic aggradation of micrite. It is also important to note that microporosity provides interconnection of mouldic and vuggy pores on a microscopic scale. Although permeability is significantly reduced, microporosity may be effective for the transmission of gas.
Similarly, a report on PETRONS-2 cores [
27] noted that the cores consist of limestone and are made up of either coral, algal or a combination of both, acting as building blocks within the carbonate buildup. Even though the carbonate buildup has undergone extensive diagenesis, the depositional environment still primarily controls the distribution of porous and tight carbonate reservoirs in the PETRONS field buildup. Limestone deposited in back reef and protected environments, as well as reworked fore reef talus and detritus, were preferentially converted to chalkified and mouldic/vuggy limestone due to early leaching and possibly mouldic-sucrosic dolomite (fresh water influence), forming porous zones. Both dolomitization and fresh water stabilization processes occurred mainly in the central part of the buildup and decreased with intensity towards the flanks. It was reported that limestone deposited in relatively deeper open marine off reef settings and carbonate banks that are enriched with non-carbonate impurities (argillaceous limestone) were subjected to porosity-destroying compaction processes, thus becoming tight zones. Periods of rapid sea-level transgression also augment the argillaceous enriching processes. Rim reefs, patch reef and main reef cores are also mainly tight due to early cementation of its rich carbonate ooze and lime mud. The apparent contrast in reservoir properties, mainly tied to sea level changes, form a layer-cake phenomenon as observed in wireline logs and supported by seismic data.
This study is focused on producing a relationship between Klinkenberg-corrected permeability against water permeability and to discuss factors that cause the differences between the permeability values using carbonate core samples from PETRONS-1 and PETRONS-2 wells (PETRONAS, Sarawak, Malaysia). The relationship established can be a baseline towards a better and more comprehensive correlation that can aid in determining the absolute water permeability value based on the respective Klinkenberg-corrected permeability values of core samples from the same region.