ASEG Extended Abstracts - 2nd Australasian Exploration Geoscience Conference: Data to Discovery, 2019

Logging-while-drilling (LWD) is the preferred logging method in large boreholes and shallow unconsolidated formations, as the borehole condition in such formations deteriorates soon after drilling, and wireline data is thus adversely affected. This paper demonstrates the high-quality results of the latest generation of LWD sonic tool, in shallow, unconsolidated formations and large borehole size. The results enabled accurate wellbore analysis computation and prevented drilling problems while also delivering high quality data at high rates of penetration (ROP). Comparison with wireline results further clarifies the situation.

The mineral system concept is a valuable framework to use for mineral exploration because it allows the user to interpret their data with more flexibility than for traditional ore deposit models. Better context is provided by a well constructed mineral deposit model because fundamental processes can be adapted to a broad range of systems.

Geochemistry adds much needed detail to any mineral deposit model in three main areas: (1) immobile element geochemistry to better constrain lithological units and define a more detailed stratigraphy, (2) alteration geochemistry to quantify mass balance leading to a better understanding of hydrothermal fluid flow and potential trap sites, and (3) metal enrichment signatures and how they vary across a mineral system so that exploration can be focussed on those parts of the mineral system that have the highest likelihood of exploration success.

The Brecknock, Calliance and Torosa (BCT) gas fields are located offshore approximately 425 km north of Broome, Western Australia. The fields have undergone an extensive appraisal drilling and seismic acquisition campaign between 2005-2012, with the main reservoirs intersected between 3600-4400 mTVDss in water depths of less than 700 mMSL. The integration of the associated subsurface studies has underpinned reservoir modelling for field development planning.

The gas is primarily hosted in a sequence of fluvial and coastal/deltaic siliciclastic sediments from the Early to Middle Jurassic Plover Formation. These sediments onlap and thin onto eroded Triassic structural highs along the outboard side of the Caswell Sub-basin, and are overlain by more strongly marine Late Jurassic and Early Cretaceous sequences. A thick interval of Late Cretaceous to Tertiary carbonates blankets the region.

Extensive seismic and well-based studies were undertaken to better understand the spatial distribution and properties of the reservoirs and seals, which vary considerably between the fields. Some key controls include pre-and syn-Jurassic tectonics and possibly by syn and post-depositional volcanism/igneous activity. More than 1000 metres of conventional coring in the appraisal drilling campaign have underpinned these interpretations along with many of the other studies described here. High resolution palynology was critical for defining the chronostratigraphic framework and relating the sequences to key regional sequence stratigraphic surfaces that can be recognised between the basins of the North West Shelf. Sedimentology and petrology studies, and the evaluation of potential modern and ancient analogues provided insights into the spatial arrangement of depositional systems and the present controls on reservoir quality. Findings from this regional to reservoir scale work have been integrated with the analysis of geochemical and pressure data to better understand reservoir connectivity, all of which are integral for supporting the path to development.

Measurement While Drilling (MWD) and Logging While Drilling (LWD) tools acquire geophysical and operational data from directly behind the bit during drilling operations. Traditionally, such tools have only been integrated into mud rotary drilling systems used in the oil and gas industry and are not available to the mining industry, where reverse circulation (RC) is the prevalent drilling method.

A joint venture between Wallis Drilling and Qteq was formed to develop tools that acquire geophysical data while using RC drilling systems. In this paper, we present drillMAX, the world’s first MWD tool for use in RC drilling.

This paper describes a novel approach to solving borehole rugosity for single detector gamma-gamma density logging. Modelling the space investigated by the gamma-gamma density tool provides the basis for deriving an applied rugosity correction.

This paper also describes how the same correction models when applied to dual density tool detectors can increase the vertical resolution of the tool revealing otherwise hidden formation heterogeneity.

Measurements made with single detector gamma-gamma density tools are accurate non-rugose (smooth) boreholes whereas measurements made in rugose (rough) boreholes are typically biased low, due to the undesirable effects of low-density borehole fluid, water or air sitting between the formation and the source or detector and in the path of gamma rays.

Accurate and precise identification of lithological facies is vital to understand geological variation in a proven reservoir. Four specific different machine learning (ML) classification algorithms are implemented to predict facies on an open dataset in the Panoma gas field in southwest Kansas, USA. The objective is improvement of facies classification accuracy with robust application of ML technique compared to previous published work on the same dataset. A total of 4,149 data samples are available for analysis with known facies from the core data where each sample point contains four or five measured properties (wire-line logs), and two derived geological properties (geological constraining variables). Facies classification is addressed with four well-known classification algorithm which are artificial neural network (ANN), support vector machine (SVM), decision trees and gaussian process classifier (GPC). High dimensionality, non-linear correlation and overlapping feature space of facies classes make the non-parametric method more suitable to handle complex datasets. Among the presented classifiers, ANN perform better relative to others on validation dataset. It is observed that our present approach of adding more input features, increasing number of training dataset and efficient implementation of algorithm have improved facies prediction accuracy significantly on a blind well.

The availability of relatively inexpensive UAVs in the form of drones provides many opportunities to seismic exploration.

While government regulations still provide many restrictions in this space, this trial has demonstrated that significant advantage can be achieved by using these tools. These have the potential to improve the safety of personnel and reduce the QC costs of seismic acquisition.

Recently, there was an increase of costs to explore new economical mineral occurrences, while the success of discovery of new deposits has diminished. Therefore, Metal Earth was conceived with the objective to improve understanding of the processes controlling the distribution of mineral endowment, with a focus on Archean greenstone belts of the Canadian Superior Province as a global representative of these belts. Part of Metal Earth is to study thirteen different transects with a total length of >1000 km associated with Archean greenstone belts with various degrees of mineralisation (from less endowed to well-endowed) using high-resolution multidisciplinary geological and geophysical data. The goal is to reliably identify deeper underlying crustal/mantle conditions controlling mineralisation and also to validate existing ore deposit models by better identifying components contributing to mineral endowment.

In this paper, we discuss acquiring, processing and preliminary modelling of multidisciplinary geophysical data (i.e. seismic, magnetotelluric, potential field data and petrophysical characterisations) to reliably model subsurface features along the transects. Across the Swayze transect, seismic sections clearly show sub-horizontal and dipping geological boundaries and magnetotelluric model reveals pathways as narrow upper crustal sub-vertical zones of low resistivity. In addition, joint inversion of gravity and magnetic data on initial models constrained by surface geology, seismic and magnetotelluric models and petrophysical characterisations have improved the models by better detecting at depth the limits of lithological units, low-density batholiths, dykes and fault zones.

New models reveal there are significant crustal difference between the well-endowed Abitibi and the less prospective Wabigoon subprovinces. The final models will map the lithospheric architecture and constrain large-scale intracratonic controls on magmatism, crustal and tectonic evolution, and subsequently mineralisation.

Gumai Formation is often regarded as merely regional seal in South Sumatra Basin as it is dominated by shale, claystone and siltstone. However, a variety of geological and geophysical approach has been conducted in Northeast Betara (NEB) field to boost the confidence of hydrocarbon bearing reservoir distribution in the untapped formation. A proprietary seismic attribute screening approach is developed for this study in order to optimise the match between modelled and actual porosity logs. Multi-attribute analysis, neural properties, density, gamma rays, and porosity are utilized to discriminate sand, thin sand and shale. Neural network analysis is done to improve the pseudo volume results including pseudo gamma-ray and pseudo porosity which are then used for geological interpretation. Validation is done by examining existing mud log data on several wells. Good result from multi-attribute and neural network analysis is exploited to map the distribution of lithology and porosity. The result is then overlaid on depth structure map and shows a good match between predicted porosity and actual porosity of the well. Training results and validation values also shows good correlation and validation. Correlation value of 0.6005 and 0.8060 are achieved by multi-attribute linear regression and Probabilistic Neural Network (PNN), respectively. Northwestern part of the field is considered having promising reservoir distribution and porosity. Average density of 2.0 to 2.25 g/cc and average porosity of 14 to 16% are inferred to be present in the reservoir based on this study.

The Patchawarra Formation is a Permian age fluvial sand and coal measure system deposited in the Cooper Basin of Central Australia. Fluvial sand channels up to 20 m thick form conventional gas reservoirs and are inter-bedded with seal and hydrocarbon source play components of overbank silts, clays and coal seams. This stacked play system presents a challenge to completion optimization and efficiency, which are critical components for achieving an economic well. A key driver to the success and economics of a well is the presence of liquid hydrocarbons (LPG; Propane and Butane and condensate; Pentane plus). As the liquid yields composition varies significantly throughout the Cooper Basin, estimating the liquid yield in new gas zones is vital to valuations of any potentially commercial development in this area.

This paper will present a new quantifiable method to forecast the liquid yields of specific gas zones utilizing mud gas logs. The method uses simple formulas applied to mud gas ratios to provide estimates of liquid yield per individual gas zone. Examples demonstrate the application of the method, calibration of fluid estimates to real PVT samples from corresponding gas zones and how the results have successfully optimized completion strategies in the Beach Energy operated Western Flank gas area. It is anticipated that with continued application, this simple method will become a useful tool in assessing the commerciality of gas wells and assist in identifying future exploration/appraisal and development targets.

We investigated a remanent anomaly in the main part of a Brazilian intracontinental Paleo/Mesoproterozoic rift using different geophysical methods. To enhance the anomaly signature at different depths we used RTP, DZ, THDR, 3D-AS, and UpCon techniques, generating a magnetic-structural interpretation of the main non-mapped features of the region. To identify the presence of non- induced magnetization in this particular anomaly and its total magnetization direction, we used the Maxi-Min, Magnetic Vector Inversion and Helbig’s Magnetic Moment Analysis techniques. We found total magnetization inclination and declination (I, D) of (49.10, -25.10), (46.30, -10.05) and (43.94, -19.53), respectively.

Based only on the TMI, we modelled this anomaly considering remanence using Magnetic Vector Inversion technique. Two bodies of great extension were identified at depths ranging between ~9 and 20 km. The anomaly was interpreted as a within-plate mafic pluton representing a magma chamber of continental tholeiitic association with associated feeder conduits. These results are crucial to understand the tectonic evolutional framework from this region affected by several superimposed events, including the Neoproterozoic Brasiliano Orogeny.

The prediction of facies and permeability in complicated reservoirs represents a major challenge that necessitates the usage of advanced statistical techniques. A new approach is presented in this paper to generate a log based continuous facies log and permeability index in the shaly sand of the Irwin River Coal Measures Formation.

An integration between the density log and the NMR free fluid index has been carried out. This was done to establish a comprehensive electrofacies model and accommodate the large permeability variations. A new Reservoir Index, equivalent to the Flow Zone Indicator (FZI), has been calculated to allow a quantitative reservoir characterisation that includes facies and permeability independent of any core measurements. The modelled electrofacies and permeability from this analysis showed very effective results of high resolution. In order to calibrate the calculated permeability, the repeat formation tester mobility was used as a reference for the generated model. To further validate the results, the created models have been correlated with the core thin sections and core permeability.

Four hydraulic flow units have been identified from the Reservoir Index, hence four different permeability models. From the log dependent results, an excellent match to the formation mobility has been achieved despite the reservoir heterogeneity. Furthermore, the outputs showed very encouraging results with the core data, upon which the methodology can be used in uncored wells.

The workflow described in this paper shows a new methodology for reservoir characterisation through electric logs. The methodology allows an independent technique for rock typing in the absence of any core data. Regardless of the degree of the reservoir heterogeneity, a high-resolution facies and permeability index can be generated for advanced formation evaluation.

The results presented in this paper draws on a regional pressure analysis of the offshore areas of the Browse Basin and the southern Vulcan Sub-Basin.

The presented study focuses on Late Permian to Recent stratigraphy and a new stratigraphic scheme consisting of 11 main sequences has been developed as part of the study.

Vp-Rho cross plot analysis conducted for wells in the study area, indicates that undercompaction (disequilibrium compaction) is the main overpressure generating mechanism present. Although no clear deviation from a normal compaction/disequilibrium compaction trend is evident in the analysed wells, densities can be very high at depth with densities up to 2.65 g/cm³ and above. This indicates that some cementation and possible clay mineral transformations have taken place in the deeper (and older) shales posing a challenge to conventional porosity/effective stress related pore pressure prediction.

For the purpose of this study, a model describing (shale) overpressures due to “primary” and “secondary” disequilibrium has been developed. The developed geological pressure model shows an overall good match with shale pressure predictions and/or forms the upper bound of the observed shale pressure/drilling data for the majority of the analysed wells across the study area. The model is particularly useful when planning to drill in areas with few offset wells for calibration and may also form a supplement to pore pressure predictions from seismic velocities away from well control and thereby significantly reduces the risk of encountering unexpected high pressures.

In a previous paper, we described how a small electromagnetic vibrator could be manufactured using commercially available components for less than US$2,000. The basis of the unit was a set of four low-frequency actuators designed for use in home-theatre systems. In this paper we describe a new unit with the number of actuators increased from four to ten and improved quality electronic components.

The maximum output of the vibrator, as measured using load-cells, was more than 10 kN; when operating the unit in the field we noticed that the weight of the vehicle was not sufficient to prevent it decoupling. Variation between the load-cell and accelerometer measurements, consistent with similar studies conducted using hydraulic vibrators, suggests that the new unit has considerable potential as a research tool looking at issues such as baseplate flexure. A VSP acquired using distributed acoustic sensing showed signal to a depth of 850 m for a single sweep.

Carbon Capture and Storage (CCS) remains one of the most viable short-term options for halting the increase in atmospheric concentration of greenhouse gases. Assurance of storage safety is an essential part of the process required to maintain a community operations licence. Ability to understand the migration behaviour of CO2 in the shallow subsurface and faults is essential for leakage detection and mitigation. Here we present the results of the successful detection and monitoring of a shallow injection of extremely a small quantity of carbon dioxide into a fault zone using a borehole seismic technique with fibre optic sensors. The experiment was conducted at the South West Hub In-Situ Laboratory in Western Australia.

We present results of 334 ultra-wide band MT stations across an area of 100 km × 100 km in the prospective eastern Gawler Craton. The survey area is situated ~100 km south of the supergiant Olympic Dam IOCG deposit, across an area of several IOCG deposits and prospects, including Carrapateena, Oak Dam, and Khamsin. Station spacing varies between 5 km and reduces to 1.5 km around areas of known IOCG prospects. The 3D resistivity models show a north-south oriented conductor in the upper crust. Known IOCG prospects are situated along its margins. These results expand the previously only 2D defined signatures of IOCG deposits, such as Olympic Dam to the full 3D domain. Together with the wider-spaced

AusLAMP deployment and a 1.5km to 3 km spaced AEM survey, the survey is unique for imaging the whole-oflithosphere footprint of IOCG deposits under cover.

Seismic interferometry, commonly known as empirical Green’s function retrieval in seismology, has been widely applied to extract the impulse response of Earth. The conventional approach based on cross-correlation of long-term ambient seismic wavefield relies on the simultaneous recording of noise signals at seismic receivers. Recent studies have demonstrated observationally that the correlation of coda of (ambient noise) cross-correlation function (C³) enables the reconstruction of inter-station Green’s function regardless of the operating time (i.e., synchronous or asynchronous) of stations. Here we extend the C³ scheme to a more general framework that involves the correlation of cross-correlation function (C²). This new approach exploits the deterministic energy of the wavefield and is more robust than C³ that may suffer from incoherent coda wave energy due to less ideal (e.g., sparse, noisy, short duration) network configurations. We apply this technique to the recently deployed ALFREX seismic network in southwestern Australia. We show that the Green’s function between asynchronous stations can be robustly recovered using the C² approach whereas this is not feasible from C³. The proposed technique can effectively bridge the temporal gaps between temporary networks and demonstrate great potential for improving the spatial coverage of data and resolution in seismic imaging of crustal structures.

Barnacle growth on seismic streamers creates noise on the recorded signals. Keeping a low noise level in barnacle areas is challenging to seismic crews: Frequent workboat trips are required for streamer cleaning, and workboat operations are weather-dependent and considered undesirable from an HSE perspective. In 2007 a purely mechanical and autonomous streamer cleaning unit (SCU) was introduced that is launched and recovered from the workboat. In 2016 we subsequently developed a remotely operated streamer tool (ROST) that is capable of launching and recovering SCUs to/from the streamer without the use of a workboat, and which is therefore less affected by weather. The ROST is operated from a support vessel that is independent of the seismic vessel. We present experiences from four surveys that were subject to different operating conditions. Operation in high sea states is demonstrated on a survey offshore Namibia, and operation in extreme currents is demonstrated on another survey east of South Africa. A noise removal method is also presented that allows operation of the ROST while online during seismic recording. The method was first applied on the offshore South African survey, and later on two surveys in offshore Angola. It is demonstrated that work boat exposure hours can be reduced by 70-80%, and a 14 streamer spread can be cleaned twice a week while acquiring seismic in all types of operating conditions.

In minerals exploration, routine drilling is performed and the data logged from these drillholes, including lithological composition, assays, and downhole geophysical measurements such as natural gamma logs, are used to create geological interpretations of the strata within each drillhole. A 3D geological model can be created by identifying corresponding stratigraphic boundaries within multiple drillholes. These models can be used for understanding the formation and the mineral endowment of a deposit.

We introduce a system for producing stratigraphic interpretations of iron ore exploration drillholes in the Pilbara region in Western Australia. The algorithm firstly classifies each data modality independently for each geological interval, for example 2m, with classification results for each stratigraphic unit as output. These classifiers, for geological logging, assays, gamma logs, were trained on historical datasets over a wide range of strata in the Pilbara. The influence of each classifier can be adjusted according to the user’s preference, and a novel optimisation algorithm incorporates known geological features such as dykes, faults and thicknesses of various stratigraphic units, to objectively create the best fit interpretation of the geology. A geologist can then adjust this interpretation to include local knowledge.

Manual interpretations of 396 drillholes from a high-grade iron ore deposit are compared to interpretations of the same hole prepared by the algorithm. An interval-byinterval comparison of these interpretations demonstrates that without any human input, similar interpretations are produced while reducing manual effort.