Hydraulic Fracture Monitoring Research Articles

Continuous seismic monitoring of hydraulic fracturing reveals complex subsurface dynamics: observations using distributed acoustic sensing and surface orbital vibrators

Understanding hydraulic fracturing is crucial to improving the stimulation of unconventional reservoirs and increasing fluid production. This study proposes a novel seismic monitoring technology, using distributed acoustic sensing (DAS) and surface orbital vibrators (SOV), to capture fracture seismic response and mechanical properties at high temporal intervals. We analyze continuous time-lapse Vertical Seismic Profiling (VSP) data acquired every hour during the first nine days of treatment of an unconventional reservoir in the Austin Chalk/Eagle Ford Shale Laboratory. The VSP data contains clear seismic signals scattered from the activated fractures. The spatiotemporal changes of the fracture reflectivity revealed by the SOV/DAS data correlate well with the observations of fracture locations inferred from low-frequency DAS data. These results capture the fracture opening and closure processes, as well as highlighting potential pre-stage activations of the fractures due to hydraulic connectivity with pre-existing fracture systems. Therefore, analysis of the presented data set provides a unique opportunity to understand fracture initiation and subsequent evolution, not only in the context of unconventional resources, but also in enhanced geothermal systems.

Grounded source transient electromagnetic 3D forward modeling with the spectral-element method and its application in hydraulic fracturing monitoring

A long wire with large current source transient electromagnetic (TEM) monitoring, with a large detection depth, low cost, safety, and environmental protection, has unique advantages in the testing and identification of unconventional reservoir fluid and the evaluation of stimulated reservoir volume. So, the TEM 3D forward modeling method has become a research hotspot. Although the finite-element method (FEM) is a type of numerical algorithm that has been widely applied in three-dimensional (3D) electromagnetic field forward modeling, the efficiency and accuracy of FEM require further improvement in order to meet the demand of fast 3D inversion. By increasing the order of the basis function and adjusting the principle of mesh discretization, the precision of the mixed-order spectral-element (SEM) result will be increased. The backward Euler scheme is an unconditionally stable technique which can ignore the impact of the scale of the time step. To achieve a better description of the nonlinear electromagnetic (EM) response of the grounded source TEM method and to optimize the efficiency and accuracy/precision of the 3D TEM forward modeling method significantly, we proposed the use of 3D TEM forward modeling based on the mixed-order SEM and the backward Euler scheme, which can obtain more accurate EM results with fewer degrees of freedom. To check its accuracy and efficiency, the 1D and 3D layered models are applied to compare the SEM results with the semi-analytical and FEM solutions. In addition, we analyzed the accuracy and efficiency of the SEM method for different types of order basis functions. Finally, we calculated the long-wire source TEM response for a practical 3D earth model of a shale gas reservoir for fracturing monitoring and tested the feasibility of the TEM method in a hydraulic fracturing monitoring area to further demonstrate the flexibility of the SEM method.

A feasibility study on time-lapse controlled-source electromagnetic method for hydraulic fracturing monitoring of Well Eyangye-2HF in Yichang, Hubei Province, China

Abstract Hydraulic fracturing plays a crucial role in enhancing reserves and production of unconventional oil and gas resources. Injecting fracturing fluids into the ground to improve reservoirs also introduces the risk of inducing earthquakes; thus, monitoring the migration of these fluids is crucial. The microseismic positioning method determines the fracturing fluid by locating microseismic events generated by the fractured rock strata; however, this method is susceptible to errors. Low-resistivity subsurface fluids can directly change electromagnetic field signals, making the electromagnetic method a technically advantageous approach for monitoring the migration of hydraulic fracturing fluids. The monitoring test data of Well Eyangye-2HF show that the time-lapse controlled-source electromagnetic (CSEM) method is suitable for hydraulic fracturing and has good monitoring effects. The results of CSEM method can also compensate for deficiencies in microseismic monitoring. The electric field (Ex) observed using the CSEM method can directly predict the distribution edge of the fracturing fluid, and the anomalous zone of the Exrate of change is consistent with the fracturing-fluid injection parameters and microseismic monitoring results. The analysis of field data and forward simulation, based on electrical logging results, led to the conclusion that hydraulic fracturing operations can cause changes in the resistivity of the target layer and surrounding strata. These changes are attributed to the synergistic effects of formation stress, temperature, and the fracturing fluid. The electric-field changes observed using the CSEM method may be caused by multiple factors; however, use of time-lapse CSEM for monitoring hydraulic fracturing is still feasible.

Noise Characteristics and Denoising Methods of Long-Offset Transient Electromagnetic Method

The advantages of the long-offset transient electromagnetic method include deep detection and sensitive response to resistivity anomalies. It is widely used in underground mineral resources exploration, fluid identification in petroleum reservoirs, hydraulic fracturing, and dynamic residual oil and gas monitoring. After the primary field signal is turned off, grounded electrodes or coils are used to observe the secondary eddy field. The secondary field signal decays quickly and has a large dynamic range and a wide frequency band but is easily affected by various natural and human electromagnetic interferences. Therefore, noise reduction and distortion correction are important issues in the processing of transient electromagnetic data. This paper proposes a systematic noise interference suppression process. Multi-period and positive–negative bipolar signal stackings were used to remove random noise and suppress DC offset signals. Then, a time-domain inverse digital recursive method was applied to remove characteristic frequency signals, e.g., power frequency signals and their harmonic interference. A standard noise-free signal was constructed through forward modeling simulation and verified by adding different types of noise. Finally, high-quality transient electromagnetic secondary field attenuation signals were obtained through overlapping windowing technology. We applied this algorithm to obtain electromagnetic data from dynamic monitoring of hydraulic fracturing in Fuling shale gas and from a copper–iron metal mine in Daye City, demonstrating its effectiveness.

Physics-informed deep learning to forecast {widehat{{varvec{M}}}}_{{varvec{m}}{varvec{a}}{varvec{x}}} during hydraulic fracturing

Short-term forecasting of estimated maximum magnitude ({widehat{M}}_{max}) is crucial to mitigate risks of induced seismicity during fluid stimulation. Most previous methods require real-time injection data, which are not always available. This study proposes two deep learning (DL) approaches, along with two data-partitioning methods, that rely solely on preceding patterns of seismicity. The first approach forecasts {widehat{M}}_{max} directly using DL; the second incorporates physical constraints by using DL to forecast seismicity rate, which is then used to estimate {widehat{M}}_{max}. These approaches are tested using a hydraulic-fracture monitoring dataset from western Canada. We find that direct DL learns from previous seismicity patterns to provide an accurate forecast, albeit with a time lag that limits its practical utility. The physics-informed approach accurately forecasts changes in seismicity rate, but sometimes under- (or over-) estimates {widehat{M}}_{max}. We propose that significant exceedance of {widehat{M}}_{max} may herald the onset of runaway fault rupture.

Multi-stage hydraulic fracture monitoring at the lab scale

We introduce a novel experimental and modeling approach to perform lab-scale multistage hydraulic fracturing in Poly-Methyl-Meth-Acrylate (PMMA) and study pressure-based methods to monitor hydraulic fracturing. The approach combines an internal circular V-notch method for fracture initiation with a method for creating hydraulic isolation between subsequent stages. Using high-resolution models of fracturing experiments with injection and monitoring wells, we present novel results of displacement, pressure, and stress evolution around the wells. We identify injection-induced rotational deformation in the domain and asymmetric growth of the hydraulic fracture. We quantify the role of the monitoring well on the rotational deformation and the fracture asymmetry. We analyze the role of undrained deformation in generating the monitor well’s pressure response to individual fracture stages at the injection well. Our results show that lab-scale multistage hydraulic fracturing can be a useful tool to estimate the geometry and property of a hydraulic fracture using pressure data from offset wells in the field.

Concurrent Active Acoustic and Deformation Monitoring of Hydraulic Fracture in Laboratory Experiments

The results of laboratory experiments on concurrent active acoustic and deformation monitoring of a hydraulic fracture are presented. The experiments were carried out in a model material based on gypsum. For comparison, reference experiments were carried out to study ultrasonic waves propagation through a liquid-filled gap of controlled width between two precision glass plates. The purpose of the experiments was to study the dependence of the amplitude of the ultrasonic wave passing through the crack on the magnitude of its opening. In these experiments, a circular hydraulic fracture was created, the plane of which was perpendicular to the axis of the cylindrical sample. A cased injection well was located along the same axis, ending at the middle of its height. The sample was placed between two aluminum alloy disks equipped with piezoceramic transducers built into them, operating both in the emitter and receiver modes. Through the channel in the lower disk, the working fluid was supplied to the fracture. The sample was saturated with pore fluid through the upper disk. The entire assembly was placed in a hydraulic press providing a constant compressive force. The crack opening value varied depending on the flow rate of the fluid supplied to the center of the crack and was measured by the relative change in the distance between the disks of the assembly using inductive displacement transducers. Experiments were also carried out simulating a fracture filled with proppant. In this case, the crack aperture varied depending on the applied vertical force on the sample. Based on the results of experiments carried out under various conditions, the dependences of the amplitude of the ultrasonic wave that passed through the crack were plotted. It has been experimentally established that sound attenuation in the hydraulic fracture, which has a natural roughness, is two times lower than in the gap between precision glass plates. The obtained results will make it possible to estimate the value of hydraulic fracture opening in laboratory experiments conducted on larger samples using active acoustic monitoring.

Electromagnetic monitoring with local mesh for time-lapse monitoring of hydraulic fracturing

The controlled-source electromagnetic method (CSEM) is a useful tool for monitoring underground fluid movement due to hydraulic fracturing. To understand the characteristics of time-lapse monitoring, an efficient three dimensional(3D) forward modeling is a premise. However, 3D finite element method (FEM) employs the normal mesh that applies fine grids to model the whole sensitive area, resulting in a large degree of freedom (DOF) of the equation that is associated with expensive computation cost. Moreover, time-lapse monitoring requires continual expensive calculations, which lead even an unacceptable computational burden. Inspired by the fact that during the hydraulic fracturing, the resistivity changes occur in a limited domain, we propose a method called as the FEM with local mesh, which is based on secondary field equation to model the EM field perturbation due to hydraulic fracture, and uses fine grids on the fracturing area with resistivity change and as large grids as possible to represent other areas without resistivity change. Therefore, the DOF can be reduced significantly. The numerical tests demonstrate that our method is accurate and has only 10% of the DOF of conventional method.

A De-Noising Method for Microseismic Signal Based on the Synchrosqueezing Transform in the Time-Frequency Domain

The denoising for the microseismic signals in the strong background interferences during the microseismic monitoring of hydraulic fracturing is very important for events identification and first-arrival picking. As characteristics of microseismic signals are normally random, non-stationary, and time-varying, a powerful tool for weak signal extraction of microseismic data is still a great challenge. In this study, a denoising method for microseismic signals is proposed based on the synchrosqueezing transform in the time-frequency domain. Firstly, the microseismic signals are denoised by the adaptive threshold method in the time-frequency domain by the synchrosqueezing transform. Then, the frequency center of the effective signal is integrated and extracted in the time-frequency domain. At last, the weak signal is extracted by the reconstruction of synchrosqueezing transform. The results from the synthesized non-stationary signal with different noise intensities and the real microseismic single of a track record show that this method has better anti-noise ability and higher signal extraction accuracy. Moreover, according to the processing and analysis of microseismic data in field wells, it is proved that this denoising method has a good practical value and can extract the effective and high-precision signals in the strong noise background.

ФОРМУВАННЯ ШВИДКІСНОЇ МОДЕЛІ І ТЕОРЕТИЧНІ ОСНОВИ МЕТОДІВ ВИЗНАЧЕННЯ ГІПОЦЕНТРУ МІКРОСЕЙСМІЧНОЇ ПОДІЇ ПРИ ПРОВЕДЕННІ ГІДРОРОЗРИВУ ПЛАСТА

The article is devoted to the analysis of the characteristics of the determining stages of data processing of microseismic monitoring of hydraulic fracturing. The importance of having an accurate velocity model and the stages of its creation are substantiated. A classification of velocity models and the input data required for their creation are provided. Groups of methods for determining the hypocenter of a microseismic event, their advantages, limitations and factors affecting the accuracy of calculations are considered. A detailed description of specific methods is given. Among the absolute methods are grid search methods, linearized inversion (Geiger) and methods based on the extension of wave fields into the medium. Among the relative methods are the master-event method and the double-difference method. Algorithms for the application of specific methods are given. An example of the comparative characteristics of various methods when choosing a method for determining the hypocenter of a microseismic event is provided.

Optimum sensor configuration of surface arrays for retrieving accurate source mechanisms of microseismic events

Hydraulic fracturing and induced networks are significant for effective production of oil and gas from unconventional resources. The knowledge of fracturing source mechanisms is helpful for optimizing hydraulic fracturing treatments to maximize production. For source monitoring, the moment tensor inversion is commonly used and the source mechanisms are interpreted by the radiation patterns of microseismic waves. The accuracy of source interpretation is significantly influenced by sensor configurations, which still need further researches. In this study, the mechanism of sensor arrangements to suppress noise effect is analyzed and clarified mathematically, then an optimization method of searching for proper sensor configurations is proposed. For superior sensor configurations, errors caused by noise can be allocated evenly to the 6 moment tensor components and the source interpretation based on the moment tensor decomposition is till accurate, but errors can not be completely eliminated by optimizing sensor configurations. Generally, high-precision inversion results can be calculated by the sensor configuration that one sensor is at the center and the others are around at the same angular intervals. Compared with tradition sensor configurations, this new one can achieve similar inversion accuracy by less than a third of the sensors. Sensor numbers are not the more the better and dependent on the sizes of the regions of sensor arrangement. The conclusions arrived in this study are helpful for evaluating and designing sensor configurations for hydraulic fracturing monitoring.

Integration analysis of 3D fractures network reconstruction and frac hits response in shale wells

The frac hits affected a drilling platform in the Changning region, Sichuan Basin, China. The microseismic events of the three wells coincided in the adjacent fracturing stages. When the intermediate well fracturing, the production pressure of two adjacent wells decreases, and gas lift occurs. This study first reconstructed the 3D discrete complex fracture network through microseismic events using the local RANSAC algorithm and quantitatively evaluated the interaction and connectivity of hydraulic fracture networks between wells using kernel density estimation. An integrated multi-physical fields model was established under the dual-porosity/dual permeability concept and embedded discrete fracture models, including poroelastic deformation, desorption/adsorption, Knudsen diffusion, and viscous flow. The finding showed that frac hits during the fracturing operations could change the occurrence frequency, moment magnitude, and spatial distribution of microseismic events. The maximum dip direction of hydraulic fracture changed by 53° because of the pore pressure connection. The stress deflexion induced by the old well exploitation inward deflected between 13° and 45°, which means that the hydraulic fractures preferentially would grow asymmetrically and propagate to the pressure depletion zone. This research integrates geology and engineering and can give direct and practical guidance to oilfield fracturing construction, hydraulic fracture monitoring, and gas production.

Downhole microseismic data interpretation for media anisotropy evaluation with limited acquisition geometry in Western Siberia

We have evaluated the results of processing and interpretation for one of the first projects in Russia for downhole microseismic monitoring of hydraulic fracturing in Western Siberia. The data are characterized by large distance to the acquisition array and noisy recordings. We illustrate that preliminary quality control of microseismic array installation is a useful tool that may predict some of the issues and help to promote better acquisition quality. Therefore, it is important to reiterate the significance of the supervision for microseismic field measurements, especially when the technology is tested in new regions. Despite limited acquisition aperture and a comparatively small number of observed microseismic events, it is possible to derive valuable interpretation results from the data set. The geometry of the event locations suggests significant unplanned development of the hydraulic fractures, including a significant shift of the fracture from the initiation point. Records with clear S-wave splitting allowed us to determine arrival times of P waves and two S waves for several microseismic events. Arrival times were then used for simultaneous kinematic inversion for microseismic-event location and anisotropic velocity-model update (layered transversely isotropic with the vertical axis of symmetry model). We estimate significant anisotropy from microseismic data. Thomsen parameters [Formula: see text] and [Formula: see text] are well constrained, whereas parameter [Formula: see text] is constrained only for one of the layers.

Development and Consideration of 3D Transient Electromagnetic Forward Modelling for the Hydraulic Fracturing Monitoring

Hydraulic fracturing monitoring is an essential technical means for unconventional oil and gas development. Hydraulic fracturing monitoring can be applied to evaluate the degree of development and scale of fractures in target reservoirs and to diagnose the extent of fracturing fluid. It can then be used to estimate the stimulated reservoir volume (SRV). Therefore, it is critical to forecast reservoir yields and unconventional oil and gas resource utilization. Conventional hydraulic fracturing monitoring is predominantly based on microseismic monitoring, which can effectively delineate the range and main direction of the fracturing fractures. However, microseismic fractures is difficult to guarantee that the volume of fracturing in the seismically active fractures is equal to that in the fracturing reservoirs. Recently, the transient electromagnetic (TEM) based on the substantial conductivity difference between the fracturing fluid and the background earth has been used as an emerging technology in the field of fracturing monitoring. It can then be used to evaluate the degree of fracturing development quantitatively. Based on the background of using TEM in Hydraulic fracturing monitoring, this paper gives a short review of 3D forward TEM modelling developments and its applications for hydraulic fracturing monitoring. And the challenges of 3D forward modelling for hydraulic fracturing monitoring application have been discussed. We found that current TEM forward modeling algorithm should meet the challenges of considering fracturing medium, anisotropy and using a multi-scale forward method. Solving these technical difficulties will promote hydraulic fracturing monitoring technology for unconventional oil and gas exploration.

Roof weighting and support of a largely mined shallow coal seam

To study roof weighting and the support of a largely mined shallow coal seam, the hydraulic support resistances in both small and large periodic weightings were derived by different theoretical formulae. Support working resistances of the 12,401 mining face in small and large periodic weighting were 10,442 and 17,064 kN, which represented a loading-increase coefficient of up to 1.64. Mining cracks were formed up to land surface and were visible to the naked eye, and numerical simulation of roof stratum structure by 3DEC has been proven to be realistic. A 1.2-MPa supporting intensity represented a critical threshold for effectively reducing horizontal displacement and inhibiting rib spalling using FLAC3D. Hydraulic fracturing and micro-seismic monitoring were used to ensure the safety of the 12,401 mining face. These findings can explain the mechanism of mining a shallow coal seam for effective prevention and controlled design. The results show high accuracy and are consistent with reality.

3D Microseismic Monitoring Using Machine Learning

AbstractMicroseismic source localization is important for inferring the dynamic status of the subsurface stress field during hydraulic fracturing. Traditional deterministic methods for 3D microseismic source localization require either ray tracing or full waveform modeling, thus are computationally expensive. We propose a very efficient (e.g., within 1 s) microseismic source localization method based on machine learning. First, three‐dimensional (3D) ray tracing is performed with hypothetical event locations and realistic acquisition geometry to calculate the theoretical travel‐times. The theoretical travel‐time differences and the spatial locations of the stations are treated as the input features of the training data set, and the corresponding source locations are used as the labels. The manually or automatically picked arrival time differences between different stations and a reference station after a microseismic event and the actual station locations are fed into the well‐trained model for a fast and accurate location prediction. The proposed method is efficient enough to be widely applied for the real‐time monitoring of hydraulic fracturing. The machine learning model is analogous to 3D grid search but performs 3D ray tracings before the actual localization, and needs to be retrained when applied to a new study area. We use several synthetic tests and a real data example from the Weiyuan hydraulic fracturing experiment in Sichuan, China, to demonstrate the effectiveness of the proposed method. The application of the proposed method to earthquake localization is also demonstrated to be straightforward.

GroningenNet: Deep Learning for Low-Magnitude Earthquake Detection on a Multi-Level Sensor Network.

Automatic detection of low-magnitude earthquakes has become an increasingly important research topic in recent years due to a sharp increase in induced seismicity around the globe. The detection of low-magnitude seismic events is essential for microseismic monitoring of hydraulic fracturing, carbon capture and storage, and geothermal operations for hazard detection and mitigation. Moreover, the detection of micro-earthquakes is crucial to understanding the underlying mechanisms of larger earthquakes. Various algorithms, including deep learning methods, have been proposed over the years to detect such low-magnitude events. However, there is still a need for improving the robustness of these methods in discriminating between local sources of noise and weak seismic events. In this study, we propose a convolutional neural network (CNN) to detect seismic events from shallow borehole stations in Groningen, the Netherlands. We train a CNN model to detect low-magnitude earthquakes, harnessing the multi-level sensor configuration of the G-network in Groningen. Each G-network station consists of four geophones at depths of 50, 100, 150, and 200 m. Unlike prior deep learning approaches that use 3-component seismic records only at a single sensor level, we use records from the entire borehole as one training example. This allows us to train the CNN model using moveout patterns of the energy traveling across the borehole sensors to discriminate between events originating in the subsurface and local noise arriving from the surface. We compare the prediction accuracy of our trained CNN model to that of the STA/LTA and template matching algorithms on a two-month continuous record. We demonstrate that the CNN model shows significantly better performance than STA/LTA and template matching in detecting new events missing from the catalog and minimizing false detections. Moreover, we find that using the moveout feature allows us to effectively train our CNN model using only a fraction of the data that would be needed otherwise, saving plenty of manual labor in preparing training labels. The proposed approach can be easily applied to other microseismic monitoring networks with multi-level sensors.

Quantitative prediction of the drilling azimuth of horizontal wells in fractured tight sandstone based on reservoir geomechanics in the Ordos Basin, central China

Hydraulic fracturing and horizontal well technology can effectively improve reservoir production but can affect reservoir reconstruction through many factors, including the reservoir lithology, reservoir stress state and natural fractures. Considering tight sandstone as an example, the finite element method (FEM) and the optimal trajectory design scheme for a horizontal well are used in this study to determine the influence of the stress state, mechanical parameters and natural fractures of rock on the effects of hydraulic fracturing. Core reorientation via a paleomagnetic method is used to determine the natural fracture orientations in different well blocks. The in-situ stress magnitudes and directions are determined using microseismic monitoring of hydraulic fracturing, borehole breakout analysis, difference strain analysis and acoustic emission experiments. A heterogeneous geomechanical model is formulated based on log interpretations of the rock mechanical parameters of 111 wells and the results of triaxial mechanical experiments. The FEM is used to determine the in-situ stress field in the Dongrengou block. A predictive model for the hydraulic fracture propagation direction is formulated considering the in-situ stress, geomechanical model and natural fracture orientation. The results show that the mechanical parameters of the Dongrengou block are characterized by planar anisotropy and three-dimensional heterogeneity, where the rock mechanical parameters vary considerably across the plane and are vertically stratified. Using these results in conjunction with the drilling directions and oil production of 14 horizontal wells shows that oil and gas recovery is maximized by orienting horizontal wells perpendicular to the propagation direction of hydraulic fractures.

Quantitative Hydraulic-Fracture-Geometry Characterization with Low-Frequency Distributed-Acoustic-Sensing Strain Data: Fracture-Height Sensitivity and Field Applications

Summary Low-frequency distributed acoustic sensing (LF-DAS) has been used for hydraulic fracture monitoring and characterization. Large amounts of DAS data have been acquired across different formations. The low-frequency components of DAS data are highly sensitive to mechanical strain changes. Forward geomechanical modeling has been the focus of current research efforts to better understand LF-DAS signals. Moreover, LF-DAS provides the opportunity to quantify fracture geometry. Recently, Liu et al. (2021a, 2021b) proposed an inversion algorithm to estimate hydraulic-fracture width using LF-DAS data measured during multifracture propagation. The LF-DAS strain data are linked to the fracture widths through a forward model developed based on the displacement discontinuity method. In this study, we first investigated the impacts of fracture height on the inversion results through a numerical case with a four-cluster completion design. Then we discussed how to estimate the fracture height based on the inversion results. Finally, we applied the inversion algorithm to two field examples. The inverted widths are not sensitive to the fracture height. In the synthetic case, the maximum relative error is less than 10% even when the fracture height is two times the true value. After obtaining the fracture width, the fracture height can be estimated by matching the true strain data under various heights with a strong smooth weight. The error between the calculated strain and true strain decreases as the height approaches the true value. In the two field examples, the temporal evolutions of the width summation of all fractures and the width of each fracture show consistent behaviors with the field LF-DAS measurements. The calculated strain data from the forward model match well with the field LF-DAS strain data. The results demonstrate the robustness and accuracy of the proposed inversion algorithm.

Three-Dimensional Ultrasonic Imaging and Acoustic Emission Monitoring of Hydraulic Fractures in Tight Sandstone

Hydraulic fracturing is an important means for the development of tight oil and gas reservoirs. Laboratory rock mechanics experiments can be used to better understand the mechanism of hydraulic fracture. Therefore, in this study we carried out hydraulic fracturing experiments on Triassic Yanchang Formation tight sandstone from the Ordos Basin, China. Sparse tomography was used to obtain ultrasonic velocity images of the sample during hydraulic fracturing. Then, combining the changes in rock mechanics parameters, acoustic emission activities, and their spatial position, we analyzed the hydraulic fracturing process of tight sandstone under high differential stress in detail. The experimental results illuminate the fracture evolution processes of hydraulic fracturing. The competition between stress-induced dilatancy and fluid flow was observed during water injection. Moreover, the results prove that the “seismic pump” mode occurs in the dry region, while the “dilation hardening” and “seismic pump” modes occur simultaneously in the partially saturated region; that is to say, the hydraulic conditions dominate the failure mode of the rock.