We have demonstrated a record 3 ms switching time wavelength selective switch using the proposed Fabry-Pérot-enhanced liquid crystal on silicon (LCoS) pixels and thin-cell liquid crystal design, achieving over two-order magnitude speed improvement versus conventional WSS, enabling telecom-grade sub-50-ms restoration in long-haul optical networks.

Numerous discontinuities, such as bedding planes and natural fractures, occur in reservoir rocks and significantly influence the propagation behavior of hydraulic fractures during reservoir stimulation. To elucidate the mechanisms underlying the influence of natural fracture networks in reservoir rocks on the propagation of hydraulic fractures, a finite element–discrete fracture model is employed to establish a fluid–solid coupled finite element–discrete fracture model. This model is used to investigate the propagation behavior of hydraulic fractures in fractured reservoirs and their underlying mechanisms. The results indicate that when distant from natural fracture networks, hydraulic fractures typically propagate along the direction of the maximum principal stress. Upon approaching natural fracture networks, the propagation path of hydraulic fractures is altered, leading to localized deflection. The mechanical properties of the rock matrix versus those of natural fracture networks, in situ stress, fracturing fluid viscosity, and injection rate significantly influence the propagation of hydraulic fractures in fractured reservoirs. Increased mechanical disparity between the rock matrix and natural fractures promotes deflection along natural fracture networks, resulting in the formation of complex fracture networks. However, increased in situ stress, fracturing fluid viscosity, and injection rate facilitate direct penetration of natural fractures by hydraulic fractures, yielding the formation of simple, long, straight primary fractures. Furthermore, the propagation distance of hydraulic fractures along the direction of the maximum principal stress is positively correlated with the in situ stress, fracturing fluid viscosity, and injection rate. The findings of this study provide theoretical guidance for optimizing fracturing design.

The interaction process among hydraulic fractures and natural fractures, bedding planes, and other discontinuities during shale fracturing determines the complexity of the fracture network that is formed. However, the current conclusions and understanding of the mechanisms underlying the interaction between hydraulic and natural fractures, as well as their primary controlling factors, fail to meet the requirements of hydraulic fracturing operations, thereby restricting the efficient development of shale gas resources. Therefore, in this study, a coupled thermal-hydraulic-mechanical finite element numerical model that is based on the maximum tensile stress and the Mohr-Coulomb criterion is established, thereby considering rock deformation, fluid flow, and heat transfer. The reliability of this model is validated on the basis of previous research. This model is subsequently employed to simulate the propagation behavior of hydraulic fractures in shale with well-developed bedding. The results indicate that when hydraulic fractures propagate to the bedding, five propagation modes may occur: arrest, diversion, diversion and crossing, crossing and diversion, and direct crossing. These modes are controlled by factors such as the mechanical properties of the shale matrix and bedding, geostress, bedding dip angle, temperature, and fracturing fluid injection rate. During fracture propagation, increases in the elastic modulus ratio between the rock matrix and the bedding, the bedding dip angle, and the temperature are favorable for hydraulic fractures turning along the bedding, whereas increases in the difference in vertical stress and the injection rate are favorable for hydraulic fractures directly crossing the bedding. Second, on the basis of four influencing factors, namely, the shale matrix and bedding elastic modulus ratio, bedding dip angle, difference in vertical stress, and temperature, propagation criteria for hydraulic fractures along the bedding under various combinations of influencing factors are established. The results provide theoretical reference data for the design and optimization of fracturing in shale with well-developed bedding. (c) 2025 The Authors. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

We propose a low-complexity digital carrier recovery algorithm that supports multi-modulation formats and is implemented on the field-programmable gate array (FPGA) platform. Based on an iterative fast Fourier transform (IT-FFT) architecture and a polar-coordinate blind phase search mechanism, the algorithm significantly reduces the number of multiplication operations by optimizing the phase estimation process, thus effectively lowering the hardware implementation complexity. Asimplified CORDIC algorithm is proposed, which combines CORDIC pre-convergence with small-angle linear approximation to effectively reduce computational complexity and hardware resource consumption while ensuring accuracy. Meanwhile, an additional feedback loop is introduced to enhance the phase tracking performance of the system under high frequency drift scenarios and improve the robustness of the algorithm. To verify the engineering feasibility of the algorithm, a hardware implementation of 32 GBaud signal processing is completed based on the FPGA platform, and the bit error rate performance between floating-point simulation and fixed-point hardware implementation is compared. Experimental results show that the performance loss of the fixed-point hardware implementation compared with floating-point simulation is negligible, which fully demonstrates its practicability and effectiveness in high-speed optical communication systems and can provide a reliable method for real-time processing of high-speed optical communication systems.

Unconventional oil and gas reservoirs naturally have low porosity and low permeability, which necessitate reservoir stimulation during production to achieve commercial exploitation. Therefore, to improve reservoir stimulation effectiveness, this study established a thermal-hydraulic-mechanical coupled numerical model suitable for hydraulic fracturing experiment scales based on rock mechanics, elasticity mechanics, damage mechanics, and flow mechanics theories, combined with maximum principal stress and Mohr-Coulomb damage criteria. The model was numerically solved within a finite element framework and used to simulate the reservoir hydraulic fracturing process. The results indicate that the propagation behavior of hydraulic fractures is controlled by reservoir rock mechanical properties, geostresses, reservoir temperatures, fracturing fluid viscosities, and injection rates. Among these, the increase in principal stress difference, reservoir temperature, fracturing fluid viscosity and injection rate promotes the propagation of hydraulic fractures along the direction of the maximum horizontal principal stress, whereas an increase in the rock's elastic modulus reduces the propagation length of the hydraulic fractures. During fracturing, the fracturing fluid fractures the reservoir rock, significantly improving its porosity and permeability. This not only enhances the mobilization of unconventional oil and gas resources but also provides effective flow pathways for their migration, thereby ensuring the commercial viability of unconventional oil and gas resource extraction. Additionally, selecting a fracturing process that matches the geological characteristics of the study area during fracturing design is a prerequisite for improving the reservoir stimulation effect. The results of this study provide a reference for fracturing design and optimization.

To reduce the power consumption and hardware complexity of coherent transceivers, enabling symbol-rate sampling is desirable but requires eliminate the local oscillator frequency offset (FO) prior to analog-to-digital converter (ADC) to avoid anti-aliasing-induced spectrum truncation. Conventional optical phase-locked loops (OPLL) require removing the modulation information of the received signal to achieve accurate frequency and phase locking. In this paper, we propose an opto electronic collaborative real-time frequency locking scheme for coherent receivers. The residual-carrier modulation technique is employed to facilitate accurate FO tracking. Meanwhile, by combining a low-point Fast Fourier Transform (FFT) with a chirp z-transform (CZT), the complexity can be reduced while maintaining accuracy. Frequency locking performance and transmission are experimentally validated in dual-polarization 32 GBaud QPSK/16QAM systems. The results show that the residual FO after locking fluctuates around 0 MHz, with a root mean square value of 4.2 MHz, which is within the tolerance of the blind phase search carrier recovery algorithm. In optical back-to-back and C-band 40 km transmission, the BER performance is essentially same with and without receiver-side DSP-based FO compensation, and the frequency locking module can operate properly even with very low received optical power. These results indicate that the proposed scheme offers an efficient and reliable solution toward low-cost and low-power coherent optical communication.

We have demonstrated a scalable optical network architecture integrating 800 Gbit/s or beyond coherent transmission, optical circuit switching, and telecom-grade ultra-fast rerouting technologies to enable highly reliable, low-latency and large-scale inter-and intra-data center networks, validated in both lab and field.

Natural fractures play a crucial role in controlling shale reservoir productivity, emphasizing the need for thorough characterization and predictive modeling to improve hydrocarbon extraction efficiency. In this study, a comprehensive workflow integrating borehole core observations, imaging logging interpretations, and data-driven predictive modeling was employed to investigate natural fracture systems in the Longmaxi Formation shale gas reservoir located in the Luzhou area of the southern Sichuan Basin. First, the mechanical types, filling degree, and occurrence characteristics of natural fractures were quantitatively characterized. Subsequently, leveraging these insights, a neural network model trained on conventional logging data was developed to predict fracture intensity in individual wells, demonstrating broad applicability where imaging log data are limited. Furthermore, correlation analysis identified tectonic activity, curvature attributes, and rock mineral composition as the primary factors influencing fracture development. Finally, by integrating multi-attribute fracture intensity probability volumes (with spatial constraints) and statistical analyses of dip-azimuth data, a 3D multiscale fracture model was constructed, which achieved a prediction error rate of 7.32–18.18

This paper studies the energy efficiency (EE) optimization of rate-splitting multiple access (RSMA) enabled integrated sensing and communications (ISAC) systems under imperfect channel state information at the transmitter (CSIT). Existing approaches typically optimize precoders directly from uncertain CSIT, relying on worst-case or statistical treatments that do not improve the channel estimates. To this end, we develop a two-stage framework that combines data-driven channel refinement with model-based optimization for a monostatic multiple-input single-output (MISO) RSMA-ISAC platform. A convolutional block attention module (CBAM) based network is first employed to refine the channel estimates by capturing inter-user correlations and spatial characteristics, yielding representations whose residual error is lower than the available imperfect estimates. Using these refined channels as input, the non-convex EE maximization problem, subject to communication rate, transmit power limit and a sensing performance requirement expressed through the Cramér–Rao bound (CRB) constraint, is then solved through a successive convex approximation (SCA) procedure. Extensive numerical results across different antenna sizes, channel-error SNRs, and circuit power levels show that the proposed refinement stage significantly improves EE compared with raw imperfect-CSIT design, worst-case robust optimization, and a weighted-sum-rate WMMSE (WSR-WMMSE) baseline. The comparison with an independently implemented transformer-based refinement network further shows that the proposed CBAM-based refiner achieves higher EE, confirming its effectiveness. The framework maintains high sensing performance and exhibits stronger resilience to channel uncertainty than conventional space-division multiple access (SDMA)-ISAC schemes under the same refinement pipeline.

This study analyzed the present-day in situ stress field characteristics in the Longmaxi Formation of Luzhou area using methods of comprehensive logging calculations, numerical simulations, etc. Further, this paper analyzed and quantitatively evaluated the mechanical effectiveness of natural fractures in the target layer based on the relationship between in situ stresses and natural fracture development. The results reveal that the dominant orientation of present-day in situ stress in the Layer S1l1-1 is WNW-ESE-trending in the Luzhou area, affecting by local structures. The horizontal maximum principal stress in the Layer S1l1-1 ranges from 70 MPa to 125 MPa, with the majority between 85 MPa and 110 MPa. The horizontal minimum principal stress varies from 50 MPa to 100 MPa, predominantly falling between 75 MPa and 90 MPa. The stress distribution exhibits strong heterogeneous characteristics, significantly influenced by tectonic activity. Under the present-day in situ stress state, mechanically effective natural fractures constitute a small portion of the total fractures. The effectiveness is mainly influenced by factors such as fracture inclination, fault zone development, and the orientation of stress relative to fracture strike. The results are expected to support shale gas exploration and development in the Longmaxi Formation of Luzhou area, Sichuan Basin.

Near-field channel estimation has become a key challenge in extremely large-scale multiple-input multiple-output (XL-MIMO) systems due to the spherical-wave propagation and the large number of antenna elements. In practical deployments, antenna failures and hardware impairments often lead to incomplete array observations, which significantly degrade channel estimation accuracy. Existing sparse recovery approaches, such as Orthogonal Matching Pursuit (OMP)-based methods, typically rely on high-dimensional dictionary searches and exhibit high computational complexity under near-field conditions. In this paper, we propose a robust near-field channel estimation framework for XL-MIMO systems that leverages the discrete Chirplet transform to handle missing antenna data. First, we show that near-field channel components exhibit strong peak sparsity in the Chirplet time-frequency domain, where each propagation path forms a dominant energy peak. We further analyze the impact of missing antenna samples and derive the Cramér–Rao bound (CRB) under incomplete observations. Based on this property, a two-stage channel estimation algorithm consisting of coarse peak detection and fine parameter refinement is proposed to achieve low-complexity and robust parameter estimation. Simulations confirm that the proposed method matches the accuracy of OMP-based approaches at a lower computational cost. Furthermore, it remains robust against severe antenna data loss, proving its practicality for real-world large-scale arrays with imperfect measurements.

Distributed edge-AI workloads now require both low end-to-end latency for multi-step interactive services and scalable capacity that can grow beyond a single edge site's computing or communication resources. Service providers' metro networks provide many of the building blocks for supporting such workloads: distributed central offices (COs), reconfigurable optical add/drop multiplexers (ROADMs), and optical transport network (OTN) equipment capable of transparent and reliable transport of a massive number of services. What is missing is a network architecture that natively supports these distributed edge-AI workloads to effectively serve as the computing power interconnect on a metro scale. In this paper, we propose a multi-ring AI-oriented OTN architecture that supports two complementary operation modes: Type I for point-to-point (P2P) connectivity and Type II for many-to-many (M2M) direct-path mesh connectivity. Type I uses parallel wavelengths to realize high-bandwidth connections between adjacent edge-AI sites, while Type II uses ROADM-enabled wavelength express-throughs to build low-latency lightpaths among all the edge-AI sites. In addition, optical circuit switching (OCS) enables scale-out that serves four purposes: (i) pooling computing and wavelength resources beyond a single-ring limit, (ii) restoring services onto edge-disjoint cycles under a link failure, (iii) federating rings without re-cabling, and (iv) reconfiguring the network topology to optimally fit the current workload phase. Moreover, trace-driven simulations of mixture-of-experts (MoE) inference (DeepSeek-V3-style) on $M{=}8$ COs at 100 Gb/s per wavelength show that Type II gives a ${\sim }2.2\times$ session-level speedup over Type I in decode-heavy phases by removing multi-hop O/E/O latency. Furthermore, the OCS scale-out provides ${\sim }8\times$ throughput for Type I and ${\sim }3\times$ for Type II at ${>}96\%$ scaling efficiency. The key contribution of this work is a novel network architecture that re-architects already-deployed metro optical rings to serve as a metro-scale computing power interconnect to natively support distributed edge-AI inference, positioning service providers' metro networks as a foundation to realize the ION-2030 vision of distributed intelligence.

As the demand for high-reliability and high-throughput optical transport networks intensifies with the rise of artificial intelligence (AI) and large-scale distributed applications, conventional wavelength-switched optical network (WSON) protection schemes, which rely on slower restoration times, become inadequate. In this work, we propose a novel, to our knowledge, sub-50-ms WSON protection scheme that integrates fast protocol processing, high-speed wavelength selective switching (WSS), and ultra-fast optical digital signal processing (DSP) reconstruction. By pre-configuring protection paths and parallelizing protocol processing, the proposed scheme eliminates key latency bottlenecks, ensuring rapid, real-time protection switching. We demonstrate its performance through both laboratory and field trials, achieving end-to-end protection switching within 50 ms, thus meeting the stringent latency requirements of AI applications. This work introduces a scalable and efficient solution for next-generation optical networks, addressing the challenges of real-time, fault-tolerant communication in a rapidly evolving digital landscape.

This study addresses the persistent challenge of balancing interpretability and robustness in black-box deep learning models for automatic modulation recognition (AMR), a critical task in wireless communication systems. To bridge this gap, we propose a novel explainable AI (XAI) framework that integrates symbolic feature interaction concepts into communication signal analysis for the first time. The framework combines a modulation primitive decomposition architecture, which unifies Shapley interaction entropy with signal physics principles, and a dual-branch XAI mechanism (feature extraction + interaction analysis) validated on ResNet-based models. This approach explicitly maps signal periodicity to modulation order in high-dimensional feature spaces while mitigating feature coupling artifacts. Quantitative responsibility attribution metrics are introduced to evaluate component contributions through modular adversarial verification, establishing a certified benchmark for AMR systems. The experimental validation of the RML 2016.10a dataset has demonstrated the effectiveness of the framework. Under the dynamic signal-to-noise ratio condition of the benchmark ResNet with an accuracy of 94.88%, its occlusion sensitivity increased by 30% and stability decreased by 22% compared to the SHAP baseline. The work advances AMR research by systematically resolving the transparency–reliability trade-off, offering both theoretical and practical tools for deploying trustworthy AI in real-world wireless scenarios.

We consider a rate-splitting multiple access (RSMA) assisted dual-functional integrated sensing and communications (ISAC) system, where the ISAC base station (BS) has the dual capability to simultaneously communicate with downlink users and to probe detection signals to a target. For this system, we focus on the problem of energy efficiency (EE) maximization and propose a new algorithmic framework that aims to optimize the beamforming matrices of RSMA such as to maximize the EE. Our framework is applicable to the optimization of both common and private streams' beamforming matrices, and it accounts for a variety of constraints which includes power consumption constraint, communication rate constraints and a sensing quality constraint which is expressed with the aid of Cramer-Rao bound (CRB). Finally, the performance of our framework is compared to that of SDMA and NOMA based ISAC, and the superiority of RSMA-ISAC to SDMA-ISAC and NOMA-ISAC is revealed.

The interaction mechanism between hydraulic fractures (HF) and natural fractures (NF) constitutes a critical research focus in hydraulic fracturing optimization. This study systematically investigates the influence of NF curvature on HF propagation behavior through large-scale true triaxial hydraulic fracturing physical simulations. The experiments were conducted on artificial rock specimens containing prefabricated fractures with varying curvature parameters. Results show that the curvature of natural fractures have important effects on the interaction between hydraulic fractures and natural fractures. When the injection rate is constant and the approximation angle is 90°, with the curvature of the natural crack gradually increasing (increasing curvature), the interaction between the hydraulic fracture and the natural fracture shows that the hydraulic fracture passes through the natural fracture and also partially extends along the natural fracture, and gradually changes to the hydraulic fracture extending only along the natural fracture, and then finally exists the natural fracture and extends along the direction of the maximum horizontal principal stress. In addition, the increase in curvature of the natural fractures leads to a decrease in fluid pressure as the hydraulic fractures interact with the natural fractures. The experimental methodology and results contribute to fundamental understanding of fracture propagation mechanics in heterogeneous media, with direct applications to stimulation design in naturally fractured reservoirs.

Hydraulic fracturing technology is important in developing low-porosity and low-permeability oil and gas reservoirs. However, the presence of discontinuities such as natural fractures (NF), lithological interfaces, and bedding (B) in rocks seriously affects the propagation paths of hydraulic fractures (HF) and their geometrical morphology. Therefore, this study proposes a numerical model based on the finite element method (FEM) for indoor hydraulic fracturing scale and simulates the interaction between hydraulic fractures and bedding during the fracturing process. The results show that when the hydraulic fracture extends to the bedding, five types of propagation behaviors may occur: arresting, turning, turning and then crossing, turning while crossing, and crossing directly, which are controlled by factors such as the geostress, the mechanical properties of the rock matrix and bedding, the bedding inclination, and its density, etc. During the fracturing process, the increase of elastic modulus ratio between the rock matrix and bedding and bedding inclination angle is conducive to the hydraulic fracture turning along the bedding; the increase of principal stress difference and bedding density is conducive to the hydraulic fracture directly passing through the bedding. When the inclination angle of the bedding distributes from 15° to 60°, the hydraulic fractures will first turn along the bedding when they extend to the bedding, and then extend for a certain distance before passing through the bedding and the distance of extension is positively correlated with the inclination angle of the bedding. The hydraulic fracture extension criterion in the bedding was established by using the ratio of elastic modulus of the rock matrix and bedding, geostress, and bedding inclination, and the results of the study will guide the design and optimization of fracturing in shale with bedding development.

Tensor convolution is a fundamental operation in convolutional neural networks, especially for processing tensors, which are prevalent in real-world applications. Current methods often convert tensor convolutions into matrix multiplications, leading to data replication, additional memory usage and increased hardware complexity. Here, a high-bit-efficiency optical tensor convolution accelerator with reduced data redundancy and lower memory consumption is presented. The bit-efficiency of the optical tensor convolution accelerator is first explored, significantly improving its effective computing power by utilizing the spatial dimension. Consequently, the optical tensor convolutional accelerator operates at speeds exceeding 3 Tera Operations Per Second (TOPS)—the fastest single-kernel optical convolutional accelerator to date, to the best of authors' knowledge. Its performance is validated on handwritten digit recognition and histopathologic cancer detection tasks, achieving 93.8% and 77% accuracy, respectively, closely matching in-silico results. This approach simultaneously multiplexes the physical dimensions—wavelength, time, and space—and leverages the parallelism and high throughput of light, enabling efficient optical processing of tensor data with significant computational power.

Hollow-core fiber (HCF) is developing rapidly, with its minimum attenuation already surpassing that of traditional solid-core fiber and continually setting new records. However, recent reports have pointed out that the CO2 absorption peaks present in the attenuation spectrum of HCF can severely affect the performance of high-speed signal transmission. Using a high-precision insertion loss test platform and a swept-frequency light source, we measured the attenuation spectrum of an HCF sample and found that under high-precision wavelength interval scanning, certain wavelength positions in this sample exhibited up to 0.5 dB/km extra loss due to CO2 absorption. We imported the experimental test attenuation results into VPITransmissionMaker for simulation analysis and found that in the wavelength regions where the CO2 absorption peaks are severe, the bit-error rate (BER) performance of both conventional single-carrier (SC) signals and digital sub-carrier multiplexing (DSCM) signals after 100 km transmission deteriorates significantly. To address this issue, we proposed a spectrum-optimized DSCM modulation scheme with sub-carrier center frequency optimization to circumvent the overlap between the CO2 absorption peaks and the effective components of the DSCM signal spectrum. Simulation results show that within a 150 GHz grid at a center frequency of 186.05 THz, for the 100 km HCF fiber measured in this study with pronounced CO2 absorption peaks, the proposed novel, to the best of our knowledge, DSCM scheme achieves a >10 × BER improvement compared to conventional SC and DSCM schemes, when the optical signal-to-noise ratio of the transmission system is above 26 dB.

Multi-band transmission (MBT) technologies, particularly the combination of S-, C-, and L-bands, offer significantly greater potential for improving the transmission capacity of optical communication systems compared with single-band or dual-band technologies. However, the adoption of S + C + L band typically faces more severe nonlinear effects (such as stimulated Raman scattering induced power transfer), immature device performance (S-band), and other issues, thus requiring more refined power management mechanisms (such as transmitter power pre-tilt, transmitter power optimization), device performance optimization (S-band optical gain and noise figure), and system end-to-end optimization (data rate, baud rate, modulation format, fiber type, etc.) We first conduct numerical simulations to evaluate the performance of the S + C + L triple-band scenario with an 18-THz spectral width, and a 144 Tbit/s transmission over 90-km optical fiber and a 120 Tbit/s transmission over 120-km fiber are performed. Then, we have further experimentally demonstrated a record real-time 128.7 Tbit/s transmission over 75 km ultra-low loss and large-effective area G.654.D fiber together with S + C + L-band 17-THz bandwidth using only lumped doped fiber amplifiers. The experimental results show that a record real-time single carrier bit rate of 1275.93 Gbit/s using digital coherent transceivers with a symbol rate up to 135 GBaud over the widened C-band with 6-THz bandwidth is successfully demonstrated.