High Fault Research Articles

Estimation of polarity of slope of instantaneous current increment for power swing immune protection of series compensated line

Existing protection schemes for series-compensated transmission line (SCTL) lack either in reliability or speed during power swing (PS). A new ultra-fast reliable method is proposed in the present article using the slope of instantaneous current increment (SICI). The peaks of SICI (denoted by PKS) of a modal component of currents above a dynamic threshold are obtained to detect the sudden jump in SICI, which occurs after a transient disturbance. PKS of two terminal-currents are of same polarity for a fault within SCTL but they are of opposite polarity for any external event. This unique characteristic is mathematically verified and exploited to discern internal faults from external events during PS. Proposed method is found unaffected for high level of series compensation, position of compensation, close-in fault, fault inception point, high fault resistance, synchronization error and CT saturation. The results are finally compared to the incremental quantity-based methods to justify its merit.

A hybrid intelligent system for protection of transmission lines connected to PV farms based on linear trends

Conventional relays face challenges for transmission lines connected to inverter-based resources (IBRs). In this article, a single-ended intelligent protection of the transmission line in the zone between the grid and the PV farm is suggested. The method employs a fuzzy logic and random forest (RF)-based hybrid system to detect faults based on combined linear trend attributes of the 3-phase currents. The fault location is determined and the faulty phase is detected. RF feature selection is used to obtain the optimal linear trend feature. The performance of the methodology is examined for abnormal events such as faults, capacitor and load-switching operations simulated in PSCAD/EMTDC on IEEE 9-bus system obtained by varying various fault and switching parameters. Additionally, when validating the suggested strategy, consideration is given to the effects of conditions such as the presence of double circuit lines, PV capacity, sampling rate, data window length, noise, high impedance faults, CT saturation, compensation devices, evolving and cross-country faults, and far-end and near-end faults. The findings indicate that the suggested strategy can be used to deal with a variety of system configurations and situations while still safeguarding such complex power transmission networks.

Comparative analysis of tubular retractors and hook retractors in oblique lumbar interbody fusion at the initial stage of the learning curve

PurposeOblique lumbar interbody fusion (OLIF) still has a steep learning curve that many spinal surgeons who want to develop are hesitant. The purpose of this study is to provide reference for beginners through the comparative analysis of the application of two kinds of retraction devices in the early stage of learning curve.MethodWe prospectively included the first 60 patients with lumbar degenerative diseases treated with OLIF by a surgeon in our department. According to the application of different retraction devices during the operation, the patients were divided into hook retractor group and tubular retractor group. The clinical effects and complications of the two groups were compared.ResultThe average age of hook retractor group was 62 years old, the average age of tubular retractor group was 65 years old. There was no significant difference in age, sex, operative segment, follow-up time and blood loss between the two groups. The operation time in hook retractor group was less than that in tubular retractor group. The incidence of complications in hook retractor group (11.8%) was significantly lower than that in tubular retractor group (38.5%).ConclusionThe tubular retractor group has a higher risk of neurovascular injury in the initial stage of learning, as well as the risk of vertebral fracture. In contrast, the hook retractor group has the advantages of simple method, high fault tolerance and relatively low incidence of complications. Therefore, we believe that the application of hook retractor in the early stage of OLIF learning curve is easier to increase the operator’s confidence and make OLIF more acceptable.

Geochemical Characteristics and Origin of Natural Gas in the Middle of Shuntuoguole Low Uplift, Tarim Basin: Evidence from Natural Gas Composition and Isotopes

Multiple types of reservoirs, including volatile oil reservoirs, condensate gas reservoirs, and dry gas reservoirs, have been discovered in ultra-deep layers buried at depths greater than 7500 m. Understanding the genetic types of natural gas is of utmost importance in evaluating oil and gas exploration potential. The cumulative proved reserves of the super deep layer in the Shuntuoguole low uplift area of the Tarim Basin exceed 1 × 108 t (oil equivalent). The origin, source, and accumulation characteristics of natural gas still remain a subject of controversy. By analyzing the composition and carbon isotope of natural gas, a detailed investigation was conducted to examine the unique geochemical and reservoir formation characteristics of the Ordovician ultra-deep natural gas within different fault zones in the middle region of the Shuntuoguole low uplift. It was determined that most of the natural gas in this area is displaying a characteristic of wet gas with a drying coefficient ranging from 0.41 to 0.99. The carbon isotope composition of methane in the gas reservoir shows relatively light values, ranging from −49.4‰ to −42‰. The carbon and hydrogen isotopes of the components are distributed in a positive order. The natural gas is oil type gas, which is derived from marine sapropelic organic matter and has a good correspondence with the lower Yuertusi formation. The maturity of natural gas in Shunbei No. 1 and No. 5 fault zones is about 1.0%, which is the associated gas of normal crude oil, while the maturity of No. 4 and No. 8 fault zones is higher than 1.0%, which is the mixture of kerogen pyrolysis gas and crude oil pyrolysis gas. The variations in the drying coefficient and carbon isotope composition of the natural gas provide evidence for the migration patterns within the Shuntuoguole low uplift central region. It indicates that the Shunbei No. 5 and No. 8 fault zones have likely migrated from south to north, while the No. 4 fault zone has migrated from the middle to both the north and south sides. These migration patterns are primarily controlled by high and steep strike-slip faults, which facilitate the vertical migration of natural gas along fault planes. Consequently, the gas accumulates in fractured and vuggy reservoirs within the Ordovician formation.

Dynamic Byzantine Fault-Tolerant Consensus Algorithm with Supervised Feedback Mechanisms

Among the existing consensus algorithms, there are very few low-latency consensus algorithms that can simultaneously take into account node dynamics, fault tolerance, and scalability, and which are applicable to large-scale open scenarios where low latency is required. Therefore, this paper proposes a low-latency scalable dynamic consensus algorithm with high fault tolerance utilizing a combination of layered and threshold signature technologies, known briefly as LTSBFT. Firstly, LTSBFT achieves linear communication complexity through the utilization of threshold signature technology and a two-layer structure. Secondly, the mutual supervision feedback mechanism among nodes enables the attainment of linear complexity for reaching consensus on the faulty upper-layer nodes during the view-change. Lastly, a node dynamic protocol was proposed for the first time to support dynamic changes in the number of nodes during the consensus process. That is to say, consensus can still be reached when the number of nodes dynamically changes without interrupting the client’s request for consensus in the network. The experimental results indicate that LTSBFT exhibits lower communication latency and higher throughput compared to the classic HotStuff and PBFT algorithms. Furthermore, in comparison to double-layer PBFT, the LTSBFT has been demonstrated to have improved fault tolerance.

High impedance fault detection method based on improved Mayr Arc model and fitting curve coefficients

To solve the problem of weak transient fault currents and significant nonlinear characteristics caused by arcs in resonant grounding systems, proposing a high-impedance fault (HIF) detection method based on an improved Mayr arc model and fitting curve coefficients. Firstly, to address the issue of low accuracy in simulating fault waveforms in existing HIF models, construct an improved arc model based on Mayr and control the parameter values under different feature dimensions, achieving accurate simulation of arc waveforms. Secondly, based on the impedance characteristics of the feeder, analyze the phase frequency characteristics, determine the first capacitive frequency band, and then determine the cutoff frequency of the Chebyshev filter to improve the ability of waveform feature extraction and fault recognition. Finally, by fitting the voltage and current geometric characteristic curve within the characteristic frequency band, using the least squares method and the quadratic coefficient values of the fitting curve to construct detection criteria, accurate detection of HIF is possible. Research shows that the improved arc model can fully characterize the current arc waveform under various working conditions. Simulations and real-world measurements verify that HIF can be accurately detected using the quadratic coefficient value of the fitting curve.

Fault diagnosis of satellite power system based on unsupervised knowledge acquisition and decision-making

Fault diagnosis (FD) is an important foundation for the maintenance of complex aerospace systems, such as satellite power systems, in which the attribute reduction has essential effect to eliminating data redundancy and improving diagnostic results. However, due to the difficulty and high cost of obtaining labels in some situations, especially early failures, FD based on unsupervised methods is of great significance but less commonly-studied. Moreover, with respect to FD preprocessing, unsupervised attribute reduction (UAR) usually applying clustering methods suffers from the need for cluster number, randomness, inability to handle non-spherical clusters, etc. Therefore, this paper proposes an unsupervised FD strategy including a knowledge acquisition method to mine the rules from the unlabeled data, a decision-making method to process the acquired knowledge, and a diagnosis decision for the fault identification. As for the preprocessing part, this paper proposes a wrapper UAR method (named DPC-UAR) based on the density peak clustering (DPC) and heuristic method, which can automatically identify the cluster centers and deal with the nonspherical data. Finally, experiments of attribute reduction performance on UCI data show that compared with other UAR methods, DPC-UAR has the greatest effect to improve performance of unsupervised learning algorithms, and plays a relatively good role in the supervised algorithm. Experiments on satellite power system fault diagnosis illustrated that the proposed FD strategy based on DPC-UAR has high accuracy, a high fault detection rate, and a low false alarm rate.

Cross-domain manifold structure preservation for transferable and cross-machine fault diagnosis

To address the decline or failure in the autonomous learning capability of traditional transfer learning methods when training and test samples come from different machines, resulting in low cross-machine fault diagnosis rates, we propose a cross-domain manifold structure preservation (CDMSP) method for diagnosing rolling bearing faults across machines. The CDMSP method can induce the manifold space projection matrices of the source and target domains more effectively. This method maps high-dimensional features into a low-dimensional manifold, preserving non-linear relationships and aligning distribution differences while maintaining cross-domain manifold structure consistency. Additionally, highly confidently labeled target domain samples are selected from each mapping result and added to the training dataset to enhance subspace learning in subsequent iterations. The CDMSP method is both simple and effective at capturing the underlying structures and patterns in the data. The CWRU dataset and our self-built test platform dataset were used to validate this method. Experimental results show that CDMSP, as a non-deep domain adaptation method of transfer learning, outperforms similar methods in cross-machine fault identification, achieving a maximum fault identification accuracy of 100 % with excellent convergence performance. Furthermore, simulated diagnostic experiments under noise interference indicate that CDMSP maintains high fault identification accuracy, even in noisy environments. Overall, CDMSP is an efficient and reliable new method for diagnosing cross-machine bearing faults.

Research on the prevention and manage of rockburst in deep‐seam mining roadways

AbstractThis study presents a strategy for preventing and controlling rockbursts in deep and vast seams during the mining operation. We analyzed the main factors influencing rockburst by combining the geological structure of the rock mass with field‐collected microseismic monitoring data. Determined the disturbance strain sensitive threshold through mechanical experiments on coal's antidisturbance characteristics, the indoor pull‐out test of the supporting bolt was carried out. Studied the rockburst support technology and the antirockburst capability of the support system based on the energy method. The key governing variables of regional impact risk in the big buried deep coal seam operating face include high tectonic stress, goaf effect, fault, and fault protection coal pillar. The uniaxial compressive strength of the coal sample in the 3308 working face is 15.11 MPa, the antidisturbance strength is 10 MPa. When the extreme stress is 86.03%, the specimen fails; the ultimate failure strength under disturbance is 13 MPa. The support system's antishock ability is analyzed based on its energy absorption, under the premise of ensuring the support quality and support strength, the antiscour ability of 3308 Tailentry support meets the anti‐impact requirements. It is believed that the system offers adequate lateral confining pressure for the internal creep disturbance‐sensitive area, enhancing the area's resistance to disturbance. This is crucial in preventing and managing rockburst incidents.

Supervised classification and fault detection in grid-connected PV systems using 1D-CNN: Simulation and real-time validation

Photovoltaic (PV) systems are prone to various faults, including short-circuit, open-circuit, partial shading, and inverter bypass diode issues, which reduce power output and can damage components. This study presents an innovative fault detection and online monitoring technique for grid-connected PV (GCPV) systems, combining Internet of Things (IoT) technology with a one-dimensional convolutional neural network (1D-CNN) deep learning approach. The method involves developing a temperature-dependent PV system model using series resistance and ideality factor, capturing real-time data from a 15kWp GCPV system with optimally placed sensors to minimize sensor count while maintaining data accuracy, and validating the model through MATLAB/Simulink simulations and real-time experiments under various fault scenarios. The collected data is used to train the 1D-CNN model to classify different fault types. The model is then implemented on an IoT platform for real-time monitoring and fault detection, displaying system status and alerts via a dashboard. The proposed system achieves a high fault detection accuracy of 98.15 % and 93.12 % during cyberattacks, with an uncertainty of ±4 %, significantly enhancing fault detection reliability and efficiency compared to existing methods. The IoT dashboard provides an effective tool for monitoring system performance and issuing alerts under abnormal conditions.

A multi-period restoration approach for resilience increase of active distribution networks by considering fault rapid recovery and component repair

As the frequency of extreme weather events continues to rise, there is an urgent need to strengthen the safe and stable operation of active distribution networks (ADNs), and it is of great value to establish highly resilient ADNs to withstand multi-faults caused by extreme weather events. This paper proposes a multi-period restoration approach for the resilience increase of ADNs by considering fault rapid recovery and component repair under typhoon disasters. Firstly, based on the structural reliability theory, the failure rate model of the main components is established, and in light of the system information entropy, the typical fault scenario selection strategy is designed to determine the branches with high fault probability. Then, according to the fault islanding division and network reconfiguration, a fault rapid recovery method is suggested for the ADNs, where the impact of typhoon disasters on the output features of distributed generators (DGs) are taken into account, and meanwhile, the network structure and the output power of the DGs are jointly optimized to minimize the operating cost of the ADNs. Further, a fault component repair model is formulated by adopting the adaptive ant colony algorithm, and a multi-period restoration approach is proposed for the ADNs to fulfill a rolling optimization of the network reconfiguration and fault component repair. The improved IEEE 33-node and IEEE 118-node systems are used for the approach verification, and the results show that the proposed approach can effectively improve the overall load restoration level and increase the component repair efficiency. Following a multi-criteria resilience evaluation system, the proposed approach enables the ADNs to more effectively withstand typhoon disasters, offering a resilience increase of 6.93 % and 32.24 % regarding the 33-node and 118-node systems.

Faulty feeder selection method for high impedance grounding fault in resonant grounding system based on discrete Fréchet distance

Abstract When a high impedance fault (HIF) occurs in a resonant grounding system, the electrical signal is weak and the current distinction between the faulty feeder and non-faulty feeder is inconspicuous. To address this issue, this paper proposes a faulty feeder selection method based on discrete Fréchet distance (DFD). Firstly, the instantaneous value of the zero-sequence current flowing through each sampling point, along with the instantaneous values at its previous sampling moment, is accumulated to enhance the fault signal, and furthermore to magnify the distinction in the amplitude of the accumulated waveforms of zero-sequence current between the faulty feeder and non-faulty feeder. Then, the DFD of the accumulated waveforms of the zero-sequence current for each feeder is calculated, and the Fréchet distance matrix is established to acquire the average DFD (ADFD) value of the accumulated waveform for each feeder. Finally, the relative DFD (RDFD) value is defined as a criterion to distinguish the bus fault from the feeder fault according to the values of the two feeders with the maximum ADFD value and the minimum ADFD value. If the RDFD value is negative, it is identified as a bus fault, otherwise, the feeder corresponding to the maximum ADFD value is identified as a faulty feeder. MATLAB/Simulink simulation results and field measurement data show that the proposed method is not affected by various fault conditions, and the accuracy and reliability in faulty feeder selection are high when detecting an HIF, with good applicability.

An Internal-State-Variable-Based Continuous Dynamic Recrystallization Model for Thermally Deformed TC18 Alloy.

Continuous dynamic recrystallization (CDRX) is widely acknowledged to occur during hot forming and plays a significant role in microstructure development in alloys with moderate to high stacking fault energy. In this work, the flow stress and CDRX behaviors of the TC18 alloy subjected to hot deformation across a wide range of processing conditions are studied. It is observed that deformation leads to the formation of new low-angle grain boundaries (LAGBs). Subgrains rotate by absorbing dislocations, resulting in an increase in LAGB misorientation and the transition of some LAGBs into high-angle grain boundaries (HAGBs). The HAGBs migrate within the material, assimilating the (sub)grain boundaries. Subsequently, an internal state variable (ISV)-based CDRX model is developed, incorporating parameters such as the dislocation density, adiabatic temperature rise, subgrain rotation, LAGB area, HAGB area, and LAGB misorientation angle distribution. The values of the correlation coefficient (R), relative average absolute error (RAAE), and root-mean-square error (RMSE) between the anticipated true stress and measured stress are 0.989, 6.69%, and 4.78 MPa, respectively. The predicted outcomes demonstrate good agreement with experimental findings. The evolving trends of the subgrain boundary area under various conditions are quantitatively analyzed by assessing the changes in dynamic recovery (DRV)-eliminated dislocations and misorientation angles. Moreover, the ISV-based model accurately predicts the decreases in grain and crystallite sizes with higher strain rates and lower temperatures. The projected outcomes also indicate a transition from a stable and coarse-grained microstructure to a continuously recrystallized substructure.

A hybrid multi-port HVDC circuit breaker that improves interruption and reset time through shared energy dissipation and fault current regeneration

This paper proposes a Hybrid Multi-Port DC Circuit Breaker (HMPDCCB) design that addresses the limitations of existing hybrid DC circuit breakers (HDCCBs), such as high cost, complexity, and slow response to high fault current and voltage demands. The HMPDCCB is based on a common breaker, energy dissipation, and regeneration branch for all connected HVDC lines, reducing interruption time to 1.8 milliseconds. It effectively manages fault energy through Insulated Gate Bipolar Transistors (IGBTs) of the main breaker branch with parallel surge arrester, reducing energy losses and enhancing power system efficiency. The HMPDCCB provides selective protection for all connected DC lines at a lower investment cost and incorporates current-limiting inductors (CLIs) to slow the rise of fault currents. The proposed HMPDCCB has been rigorously modelled and simulated using MATLAB Simulink®/Simpowersystems, and the simulations show that it achieves a fault current interruption time of 1.8 milliseconds and a reset time of fewer than 3 milliseconds, outperforming existing models.

Domain adversarial transfer fault diagnosis method of an axial piston pump based on a multi-scale attention mechanism

Effective failure diagnosis of axial piston pumps is crucial to guaranteeing the security and steady run of hydraulic equipment systems. However, the current domain-adaptive methods have insufficient feature extraction and domain-adaptive capability for axial piston pumps across conditions and some unlabeled sample conditions. Therefore, this paper proposes a method of domain adversarial transfer fault diagnosis based on multi-scale attention mechanisms. The method enhances the ability of the model to adapt to the differences in various fault information by constructing a multi-scale residual feature extraction network. The domain adaptive capability of the network is improved by constructing a joint loss function to decrease the distributional discrepancies between the source and the target domain data. The experimental consequences demonstrate that the method can differentiate various fault types of axial piston pumps across conditions and can achieve a high fault identification rate without labeled samples in the target domain.

Dynamic recrystallisation: A quantitative study on grain boundary characteristics and dependence on temperature and strain rate in an aluminium alloy

Dynamic recrystallisation (DRX) usually occurs during hot forming of metallic materials, significantly impacting the mechanical properties of the final parts. Although DRX has been studied for decades, the types of DRX that occurs in high stacking fault energy (SFE) materials like aluminium alloys are still controversial, and their dependence on both temperature and strain rate is surprisingly missing. To fill these gaps, in the present study, uniaxial compression tests on an aluminium alloy AA6061 were carried out at different temperatures from 250 to 475 °C and strain rates from 0.01 to 1 s-1. The grain orientation and misorientation before and after the hot deformation were quantitively characterised using high-resolution Electron Backscatter Diffraction (EBSD) with a misorientation resolution of 0.05°, enabling high-angle grain boundaries (HAGBs), low-angle grain boundaries (LAGBs) and geometrically necessary dislocations (GNDs) to be accurately quantified and correlated against temperature and strain rate. Results reveal that both continuous dynamic recrystallisation (CDRX) and geometric dynamic recrystallisation (GDRX) occurred concurrently, resulting in the formation of discontinuous HAGBs and fine equiaxed grains, respectively. The misorientation angle distributions of the discontinuous HAGBs exhibit an opposite trend compared to those of the fine grains’ HAGBs. Both temperature and strain rate significantly affect the density of HAGBs formed by DRX, but have little effect on their misorientation angle distributions. This study sheds light on the fundamental understanding of DRX and provides comprehensive experimental data for future modelling of DRX processes in high SFE materials.

Continuous Dynamic Recrystallization and Deformation Behavior of an AA1050 Aluminum Alloy during High-Temperature Compression

Continuous dynamic recrystallization (CDRX) forms a new recrystallized microstructure through the progressive increase in low-angle boundary misorientations (LAGBs) during the hot forming of metallic materials with high stacking fault energy (SFE), such as aluminum alloys. The present work investigates the effect of deformation parameters on the evolution of the dynamic recrystallization microstructures of an AA1050 aluminum alloy during compression at elevated temperatures. The alloy microstructure is investigated at deformation temperatures and strain rates in the range of 300 °C to 500 °C and 0.001 to 0.8 s−1. A well-defined substructure and subsequent DRX grains provide indication that recrystallization can proceed with continued strain under high-temperature compression. At a strain rate of 0.1 s−1, the DRX fraction is observed to be 0.25 at a temperature of 300 °C. This fraction increases to 0.32 as the temperature rises to 400 °C. The recrystallization mechanism is identified by analyzing the flow stress, the evolution of the subgrain misorientation angle, and the distribution of recrystallized grains. The observations of discontinuous dynamic recrystallization (DDRX) and CDRX under various deformation parameters are discussed. Moreover, the main substructure evolution laws observed from the high-temperature compression of an AA1050 Al alloy are summarized.

Fixed-time fault-tolerant attitude tracking control for UAV based on fixed-time extended state observer

PurposeTo improve the robustness of carrier-based unmanned aerial vehicle (UAV) with actuator faults attitude tracking control system, this paper aims to propose a fixed-time backstepping (FXTBSC) fault-tolerant control based on a fixed-time extended state observer.Design/methodology/approachA fixed-time extended state observer (FXTESO) is designed to estimate the total disturbance including nonlinear, coupling, actuator faults and external disturbances. The integration of backstepping control and fixed-time technology ensures fixed-time convergence.FindingsThe simulation results of tracking the desired attitude angle show that the anti-interference, fault tolerance and tracking accuracy of FXTBSC-FXTESO are better than the BSC-ESO control method.Originality/valueDifferent from the traditional BSC-ESO, the convergence speed and convergence accuracy of FXTBSC-FXTESO proposed in this paper are better than conventional extended state observer. And the fixed time controller has the advantages of high tracking accuracy, fault tolerance and anti-interference ability.

Bayesian-optimized LSTM-DWT approach for reliable fault detection in MMC-based HVDC systems

As Europe integrates more renewable energy resources, notably offshore wind power, into its super meshed grid, the demand for reliable long-distance High Voltage Direct Current (HVDC) transmission systems has surged. This paper addresses the intricacies of HVDC systems built upon Modular Multi-Level Converters (MMCs), especially concerning the rapid rise of DC fault currents. We propose a novel fault identification and classification for DC transmission lines only by employing Long Short-Term Memory (LSTM) networks integrated with Discrete Wavelet Transform (DWT) for feature extraction. Our LSTM-based algorithm operates effectively under challenging environmental conditions, ensuring high fault resistance detection. A unique three-level relay system with multiple time windows (1 ms, 1.5 ms, and 2 ms) ensures accurate fault detection over large distances. Bayesian Optimization is employed for hyperparameter tuning, streamlining the model’s training process. The study shows that our proposed framework exhibits 100% resilience against external faults and disturbances, achieving an average recognition accuracy rate of 99.04% in diverse testing scenarios. Unlike traditional schemes that rely on multiple manual thresholds, our approach utilizes a single intelligently tuned model to detect faults up to 480 ohms, enhancing the efficiency and robustness of DC grid protection.

Subduction Zone Geometry Modulates the Megathrust Earthquake Cycle: Magnitude, Recurrence, and Variability

AbstractMegathrust geometric properties exhibit some of the strongest correlations with maximum earthquake magnitude in global surveys of large subduction zone earthquakes, but the mechanisms through which fault geometry influences subduction earthquake cycle dynamics remain unresolved. Here, we develop 39 models of sequences of earthquakes and aseismic slip (SEAS) on variably‐dipping planar and variably‐curved nonplanar megathrusts using the volumetric, high‐order accurate code tandem to account for fault curvature. We vary the dip, downdip curvature and width of the seismogenic zone to examine how slab geometry mechanically influences megathrust seismic cycles, including the size, variability, and interevent timing of earthquakes. Dip and curvature control characteristic slip styles primarily through their influence on seismogenic zone width: wider seismogenic zones allow shallowly‐dipping megathrusts to host larger earthquakes than steeply‐dipping ones. Under elevated pore pressure and less strongly velocity‐weakening friction, all modeled fault geometries host uniform periodic ruptures. In contrast, shallowly‐dipping and sharply‐curved megathrusts host multi‐period supercycles of slow‐to‐fast, small‐to‐large slip events under higher effective stresses and more strongly velocity‐weakening friction. We discuss how subduction zones' maximum earthquake magnitudes may be primarily controlled by the dip and dimensions of the seismogenic zone, while second‐order effects from structurally‐derived mechanical heterogeneity modulate the recurrence frequency and timing of these events. Our results suggest that enhanced co‐ and interseismic strength and stress variability along the megathrust, such as induced near areas of high or heterogeneous fault curvature, limits how frequently large ruptures occur and may explain curved faults' tendency to host more frequent, smaller earthquakes than flat faults.