Kick Detection at the Bit: Early Detection using Borehole Geophysics
Presented by NETL at Symposium on the Application of Geophysics to Engineering and Environmental Problems, 03/16
SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Presented by NETL at Symposium on the Application of Geophysics to Engineering and Environmental Problems, 03/16
Despite the wide range of geologic settings in which geothermal systems reside, many systems share characteristics that can be targeted by geophysical techniques. Over the last few years, the authors have used potential field and magnetotelluric (MT) methods to jointly image and characterize numerous geothermal systems in various geologic and tectonic settings. This includes systems in the Basin and Range, Washington Cascade Range, Northeastern Oregon, and the Snake River Plain. The principal geothermal targets are areas of elevated temperatures near fault interactions that promote permeability and fluid flow. Gravity, magnetic, and MT data are all sensitive to different physical parameters that together form a complementary set of tools for characterizing subsurface structure and identifying favorable geothermal targets. In general, potential fields delineate important geologic structures and MT locates fluids and zones of hydrothermal alteration. Joint interpretation of these complementary methods, along with other geologic information, provides a robust framework for geothermal prospecting and understanding a hydrothermal system. Specifically, common geothermal targets lie along potential field gradients and are associated with an electrically conductive anomaly that resides in the less-dense, less-magnetic material.
Although recent research suggests that contaminant plumes behave as geobatteries that produce an electrical current in the ground, no associated model exists that honors both geophysical and biogeochemical constraints. Here, we develop such a model to explain the two main electrochemical contributions to self-potential signals in contaminated areas. Both contributions are associated with the gradient of the activity of two types of charge carriers, ions and electrons. In the case of electrons, bacteria act as catalysts for reducing the activation energy needed to exchange the electrons between electron donor and electron acceptor. Possible mechanisms that facilitate electron migration include iron oxides, clays, and conductive biological materials, such as bacterial conductive pili or other conductive extracellular polymeric substances. Because we explicitly consider the role of biotic processes in the geobattery model, we coined the term 'biogeobattery'. After theoretical development of the biogeobattery model, we compare model predictions with self-potential responses associated with laboratory and field-scale conducted in contaminated environments. We demonstrate that the amplitude and polarity of large (>100 mV) self-potential signatures requires the presence of an electronic conductor to serve as a bridge between electron donors and acceptors. Small self-potential anomalies imply that electron donors and electron acceptors are not directly interconnected, but instead result simply from the gradient of the activity of the ionic species that are present in the system.
It has been documented for thousands of years that earthquakes and other tectonic processes have hydrological effects. Fault ruptures and the seismic waves they produce deform the crust, so it is expected that crustal fluids respond. The magnitude of the hydrological response can be very large because small stresses can be greatly amplified as changes in water pressure. This is also the premise behind searching for earthquake precursors; the very small hypothesized strains leading up to rupture might have hydrological, geochemical or geophysical manifestations. The magnitude and spatial and temporal patterns of reported changes are not always straightforward to explain and hence, remain the subject of active research.
Groundwater has been identified as one of the major freshwater sources that can potentially meet the growing demands of Egypt’s population. Gravity data (from 381 ground gravity stations) were collected, processed, and analyzed together with the available aeromagnetic (800 line-km) data to investigate the hydrogeologic and structural settings, areal distribution, geometry, and water storage of the aquifers in El Qaa coastal plain in the southwest Sinai Peninsula, and to assess their longevity given projected extraction rates. Findings include (1) complete Bouguer anomaly and total magnetic intensity maps show two connected sub-basins separated by a narrow saddle with an average basin length of 43 km and an average width of 12 km; (2) two-dimensional modeling of both gravity and magnetic data indicates basin fill with a maximum thickness of 3.5 km; (3) using anomalous residual gravity, the volume of water in storage was estimated at 40–56 km3; and (4) progressive increases in extraction rates over time will deplete up to 40 % of the aquifers’ volume in 200–230 years and will cause the water quality to deteriorate due to seawater intrusion in 45 years. Similar geophysical exploration campaigns, if conducted over the entire coastal plains of the Red Sea and the Gulfs of Suez and Aqaba, could assist in the development of sound and sustainable management schemes for the freshwater resources in these areas. The adopted techniques could pave the way toward the establishment of sustainable utilization schemes for a much larger suite of similar aquifers worldwide.
Underground Coal Gasification (UCG) produces less surface impact, atmospheric pollutants and greenhouse gas than traditional surface mining and combustion. Therefore, it may be useful in mitigating global change caused by anthropogenic activities. Careful monitoring of the UCG process is essential in minimizing environmental impact. In this paper, we first summarize monitoring methods that have been used in previous UCG field trials. We then discuss in more detail a number of promising advanced geophysical techniques. These methods – seismic, electromagnetic, and remote sensing techniques – may provide improved and cost-effective ways to image both the subsurface cavity growth and surface subsidence effects. Active and passive seismic data have the promise to monitor the burn front, cavity growth, and observe cavity collapse events. Electrical resistance tomography (ERT) produces near real time tomographic images autonomously, monitors the burn front and images the cavity using low-cost sensors, typically running within boreholes. Interferometric synthetic aperture radar (InSAR) is a remote sensing technique that has the capability to monitor surface subsidence over the wide area of a commercial-scale UCG operation at a low cost. It may be possible to infer cavity geometry from InSAR (or other surface topography) data using geomechanical modeling. The expected signals from these monitoring methods are described along with interpretive modeling for typical UCG cavities. They are illustrated using field results from UCG trials and other relevant subsurface operations.
A main issue in the storage of large volumes of fluids, mainly water and CO 2 , in the deep subsurface is to determine their field-scale-induced displacements and consequences on the mechanical behavior of the storage reservoir and surroundings. A quantifiable estimation of displacement can be made by combining the robust, cost-effective, and repeatable geophysical techniques of micro-gravimetry, differential global positioning system (DGPS), and differential synthetic aperture radar interferometry (DInSAR). These techniques were field tested and evaluated in an active large-volume aquifer storage and recovery (ASR) project in Pendleton, Oregon, USA, where three ASR wells are injecting up to 1.9 million m 3 /yr -1 into basalt aquifers to a depth of about 150 m. Injection and recovery of water at the wells was accompanied by significant gravity anomalies and vertical deformation of the ground surface localized to the immediate surroundings of the injection wells as evidenced by DGPS and gravity measurements collected in 2011. At a larger scale, and between 2011 and 2013, DInSAR monitoring of the Pendleton area suggests the occurrence of sub-centimetric deformation in the western part of the city and close to the injection locations associated with the ASR cycle. A numerical simulation of the effect of the water injection gives results in good agreement with the observations and confirms the validity of the approach, which could be deployed in similar geological contexts to look at the mechanical effects of water and gas injections. The gravity signal reflects deep phenomena and gives additional insight into the repartition of fluids in the subsurface.
Time-lapse electrical resistivity tomography (ERT) is a popular geophysical method to estimate three-dimensional (3D) permeability fields from electrical potential difference measurements. Traditional inversion and data assimilation methods are used to ingest this ERT data into hydrogeophysical models to estimate permeability. Due to ill-posedness and the curse of dimensionality, existing inversion strategies provide poor estimates and low resolution of the 3D permeability field. Recent advances in deep learning provide us with powerful algorithms to overcome this challenge. This paper presents a deep learning (DL) framework to estimate the 3D subsurface permeability from time-lapse ERT data. To test the feasibility of the proposed framework, we train DL-enabled inverse models on simulation data. Each measurement in both synthetic and field data is standardized by removing the mean and scaling the time-series to unit variance. This pre-processing step is necessary to bring simulation data closer to field observations. Subsurface process models based on hydrogeophysics are used to generate this synthetic data. Training performed on limited simulation data resulted in the DL model over-fitting. An advanced data augmentation based on mixup is implemented to generate additional training samples to overcome this issue. This mixup technique creates weakly labeled (low-fidelity) samples from strongly labeled (high-fidelity) data. The weakly labeled training data is then used to develop DL-enabled inverse models and reduce over-fitting. As both time-lapse ERT (1133048 features/realization) and 3D permeability (585453 features/realization) data samples are from a high-dimensional space, principal component analysis (PCA) is employed to reduce dimensionality. Encoded ERT and encoded permeability are generated using the trained PCA estimators. A deep neural network is then trained to map the encoded ERT to encoded permeability. This mixup training and unsupervised learning allowed us to build a fast and reasonably accurate DL-based inverse model under limited simulation data. Results show that proposed weak supervised learning can capture salient spatial features in the 3D permeability field. Quantitatively, the average mean squared error (in terms of the natural log) on the strongly labeled training, validation, and test datasets is less than 0.5. The R 2 -score (global metric) is greater than 0.75, and the percent error in each cell (local metric) is less than 10%. Finally, an added benefit in terms of computational cost is that the proposed DL-based inverse model is at least O(10 4 ) times faster than running a forward model once it is trained. Data generation, DL model training, and hyperparameter tuning to identify optimal neural network architectures utilized high-performance computing resources while the DL inference is performed on a standard laptop. Approximately, O(10 5 ) processor hours are used for generating data and DL tuning and training. We acknowledge that the data generation and DL model development are expensive. But once a DL model is trained, it can be re-used for inversion rapidly for the given system, with set physics and domain. Note that traditional inversion may require multiple forward model simulations (e.g., in the order of 10 to 1000), which are very expensive. This computational savings ≈ O(10 5 ) – O(10 7 )) makes the proposed DL-based inverse model attractive for subsurface imaging and real-time ERT monitoring applications due to fast and yet reasonably accurate estimations of permeability field.
Phase 1 of a Play Fairway Analysis (PFA) of geothermal resource potential across the State of Hawaii was recently completed. The final products of this work include a statewide geothermal resource probability map, a map of confidence in this probability, an assessment of the viability of development in areas of interest, and a prioritized list of recommended future exploration activities. The intersection of subsurface heat (H), permeability (P), and fluid (F) is necessary for an optimum geothermal play. This study: (1) identified and compiled all legacy and current data relevant to Hawaii’s geothermal resource; (2) ranked these datasets in terms of their relevance to subsurface H, P, F; (3) developed a Bayesian statistical method to incorporate the data and their rankings, and produce a statewide resource probability map; (4) developed a method to assess confidence in the probability values; and (5) assessed what we term ‘development viability’ in resulting areas of interest across the state. This paper details the basic project workflow, and activities (1) and (2). It describes the rationale for including this study’s datasets: surface geologic mapping data (calderas, rift zones, volcanic vents, dikes, faults), groundwater data (temperature, Chloride:Magnesium, SiO 2 ) and geophysical data (gravity, magnetotelluric, seismic, geodetic strain), and justifies their relative rankings in terms of H, P, and F. A second paper describes activities (3) and (4) related to the statistical methodology, and a third paper describes activity (5), the development viability criteria, as well as our suggested roadmap for future exploration activities. Overall, we find that the likelihood of abundant of geothermal resources is highest on the youngest island of Hawaii, but groundwater indicators suggest there may also be resources on the other, older islands. Furthermore, the higher demand for renewable energy on the more populated islands of Maui and Oahu, as well as a high viability of development on Lanai, also motivate further exploration on these islands.
The immense computational cost of traditional numerical weather and climate models has sparked the development of machine learning (ML) based emulators. Because ML methods benefit from long records of training data, it is common to use data sets that are temporally subsampled relative to the time steps required for the numerical integration of differential equations. Here, we investigate how this often overlooked processing step affects the quality of an emulator's predictions. We implement two ML architectures from a class of methods called reservoir computing: (a) a form of Nonlinear Vector Autoregression (NVAR), and (b) an Echo State Network (ESN). Despite their simplicity, it is well documented that these architectures excel at predicting low dimensional chaotic dynamics. We are therefore motivated to test these architectures in an idealized setting of predicting high dimensional geophysical turbulence as represented by Surface Quasi-Geostrophic dynamics. In all cases, subsampling the training data consistently leads to an increased bias at small spatial scales that resembles numerical diffusion. Interestingly, the NVAR architecture becomes unstable when the temporal resolution is increased, indicating that the polynomial based interactions are insufficient at capturing the detailed nonlinearities of the turbulent flow. The ESN architecture is found to be more robust, suggesting a benefit to the more expensive but more general structure. Spectral errors are reduced by including a penalty on the kinetic energy density spectrum during training, although the subsampling related errors persist. Future work is warranted to understand how the temporal resolution of training data affects other ML architectures.
The absorbed dose rates in air due to terrestrial radionuclides were estimated from aerial gamma spectrometric data for an area of 48,600 km{sup 2} in Southern Brazil. The source data was the Serra do Mar Sul Aero-Geophysical Project back-calibrated in a cooperative work among the Geological Survey of Brazil, the Geological Survey of Canada, and Paterson, Grant and Watson Ltd. The concentrations of eU (ppm), eTh (ppm) and K (%) were converted to dose rates in air (nGy{center_dot}h{sup -1}) by accounting for the contribution of each element's concentration. Regional variation was interpreted according to lithotypes and a synthesis was performed according to the basic geological units present in the area. Higher values of total dose were estimated for felsic igneous and metamorphic rocks, with average values varying up to 119{+-}24 nGy{center_dot}h{sup -1}, obtained by Anitapolis syenite body. Sedimentary, metasedimentary and metamafic rocks presented the lower dose levels, and some beach deposits reached the lowest average total dose, 18.5{+-}8.2 nGy{center_dot}h{sup -1}. Thorium gives the main average contribution in all geological units, the highest value being reached by the nebulitic gneisses of Atuba Complex, 71{+-}23 nGy{center_dot}h{sup -1}. Potassium presents the lowest average contribution to dose rate in 53 of the 72 units analyzed, the highest contribution being obtained by intrusive alkaline bodies (28{+-}12 nGy{center_dot}h{sup -1}). The general pattern of geographic dose distribution respects well the hypotheses on geo-physicochemical behavior of radioactive elements.
Reduced rank nonlinear filters are increasingly utilized in data assimilation of geophysical flows, but often require a set of ensemble forward simulations to estimate forecast covariance. On the other hand, predictor-corrector type nudging approaches are still attractive due to their simplicity of implementation when more complex methods need to be avoided. However, optimal estimate of nudging gain matrix might be cumbersome. In this paper, we put forth a fully nonintrusive recurrent neural network approach based on a long short-term memory (LSTM) embedding architecture to estimate the nudging term, which plays a role not only to force the state trajectories to the observations but also acts as a stabilizer. Furthermore, our approach relies on the power of archival data and the trained model can be retrained effectively due to power of transfer learning in any neural network applications. In order to verify the feasibility of the proposed approach, we perform twin experiments using Lorenz 96 system. Our results demonstrate that the proposed LSTM nudging approach yields more accurate estimates than both extended Kalman filter (EKF) and ensemble Kalman filter (EnKF) when only sparse observations are available. With the availability of emerging AI-friendly and modular hardware technologies and heterogeneous computing platforms, we articulate that our simplistic nudging framework turns out to be computationally more efficient than either the EKF or EnKF approaches.
The formation of Earth’s core left behind geophysical and geochemical signatures in both the core and mantle that remain to this day. Seismology requires that the core be lighter than pure iron and therefore must contain light elements, and the geochemistry of mantle-derived rocks reveals extensive siderophile element depletion and fractionation. Both features are inherited from metal–silicate differentiation in primitive Earth and depend upon the nature of physiochemical conditions that prevailed during core formation. To date, core formation models have only attempted to address the evolution of core and mantle compositional signatures separately, rather than seeking a joint solution. Here we combine experimental petrology, geochemistry, mineral physics and seismology to constrain a range of core formation conditions that satisfy both constraints. We find that core formation occurred in a hot (liquidus) yet moderately deep magma ocean not exceeding 1,800 km depth, under redox conditions more oxidized than present-day Earth. This new scenario, at odds with the current belief that core formation occurred under reducing conditions, proposes that Earth’s magma ocean started oxidized and has become reduced through time, by oxygen incorporation into the core. As a result, this core formation model produces a core that contains 2.7–5% oxygen along with 2–3.6% silicon, with densities and velocities in accord with radial seismic models, and leaves behind a silicate mantle that matches the observed mantle abundances of nickel, cobalt, chromium, and vanadium.
We recorded acoustic, seismic and radio-frequency signatures of 70 solid charge (~2–12 kg) surface explosions (shots) at local distances (0.1–1.5 km) to determine if such signals could be fused for blast monitoring. We observed that each geophysical signature was sufficiently repeatable between similar shots to be identifiable with multichannel correlation detectors. Using template signals from a large explosion, we then processed heavily contaminated data recording a smaller shot with these detectors, and missed or marginally detected the resultant target signals. By then fusing the p-values of these statistics through Fisher's combined probability test, we clearly identified the same explosion signals at thresholds consistent with the false alarm on noise rates of the correlation detectors. This resulting Fisher test thereby provided high-probability detections, zero false alarms and higher theoretical detection capability.
A growing body of applied mathematics literature in recent years has focused on the application of fractional calculus to problems of anomalous transport. In these analyses, the anomalous transport (of charge, tracers, fluid, etc.) is presumed attributable to long–range correlations of material properties within an inherently complex, and in some cases self-similar, conducting medium. Rather than considering an exquisitely discretized (and computationally intractable) representation of the medium, the complex and spatially correlated heterogeneity is represented through reformulation of the governing equation for the relevant transport physics such that its coefficients are, instead, smooth but paired with fractional–order space derivatives. Here we apply these concepts to the scalar Helmholtz equation and its use in electromagnetic interrogation of Earth’s interior through the magnetotelluric method. We outline a practical algorithm for solving the Helmholtz equation using spectral methods coupled with finite element discretizations. Execution of this algorithm for the magnetotelluric problem reveals several interesting features observable in field data: long–range correlation of the predicted electromagnetic fields; a power–law relationship between the squared impedance amplitude and squared wavenumber whose slope is a function of the fractional exponent within the governing Helmholtz equation; and, a non–constant apparent resistivity spectrum whose variability arises solely from the fractional exponent. In geologic settings characterized by self–similarity (e.g. fracture systems; thick and richly–textured sedimentary sequences, etc.) we posit that these diagnostics are useful for geologic characterization of features far below the typical resolution limit of electromagnetic methods in geophysics.
The ocean and the atmosphere, and hence the climate, are governed at large scale by interactions between pressure gradient and Coriolis and buoyancy forces. This leads to a quasigeostrophic balance in which, in a two-dimensional-like fashion, the energy injected by solar radiation, winds, or tides goes to large scales in what is known as an inverse cascade. Yet, except for Ekman friction, energy dissipation and turbulent mixing occur at a small scale implying the formation of such scales associated with breaking of geostrophic dynamics through wave-eddy interactions or frontogenesis, in opposition to the inverse cascade. Can it be both at the same time? We exemplify here this dual behavior of energy with the help of three-dimensional direct numerical simulations of rotating stratified Boussinesq turbulence. Here, we show that efficient small-scale mixing and large-scale coherence develop simultaneously in such geophysical and astrophysical flows, both with constant flux as required by theoretical arguments, thereby clearly resolving the aforementioned contradiction.
The use of deep-learning algorithms for estimating the value of geophysical variables from remotely sensed information is rapidly expanding. The typical objective function minimized in such algorithms is the mean square error (MSE), which is known to lead to smooth estimates with compressed dynamical range as compared to the true distribution of the variable of interest. Here, we introduce and evaluate alternative objective functions, focusing on the retrieval of precipitation rates from satellite passive microwave radiometric measurements using a deep convolutional neural network. For this testbed application, the results show that explicitly imposing the preservation of the statistical distribution and spatial wavelet power spectrum of the target variable allows to accurately reproduce extreme values and sharp gradients across multiple scales in this study.
Management of the nation’s power grid over the coming decades will need to factor multiple climate-driven threats to the power system that can be stressful or even detrimental to its operations. During disasters, grid operators suffer from cognitive overload where the electric grid is severely impacted by the weather, yet grid operators have limited awareness of these factors. Meanwhile, the electric grid is facing an explosion of data coming from a variety of sources which can enable the operator to evaluate the risks and develop mitigation strategies against hazardous events. In this work, we processed heterogeneous environmental and power grid data to learn and model grid behavior caused by extreme weather events. In this study, we focused on two weather-driven hazards, hurricanes and wildfires, which were analysed for the electric grid of Texas. We the used this data to train a Reinforcement Learning (RL) agent by analysing and predicting future behaviour of the grid during those hazard events. All these parts are incorporated in one framework to have a geophysically informed power simulators that can be used to train RL agents.