Flow Bursts Research Articles

Substorm Current Wedge: Energy Conversion and Current Diversion

AbstractUsing a magnetohydrodynamic simulation of magnetotail reconnection, flow bursts, and dipolarization, we further investigate the current diversion and energy flow and conversion associated with the substorm current wedge (SCW) or smaller‐scale wedgelets. Current diversion into both Region 1 (R1) and Region 2 (R2) sense systems is found to happen inside (that is, closer to the center of the flow burst) and equatorward of the R1 and R2 type field‐aligned currents. In contrast to earlier investigations the current diversion takes place in dipolarized fields extending all the way toward the equatorial plane. An additional FAC system with the signature of Region 0 (R0) (same sense as R2) is found at higher latitudes in taillike fields. The diversion into this system takes place in layers equatorward of the R0 currents but outside the equatorial plane. Whereas the diversion into R1 and R2 systems is pressure gradient dominated, the diversion into the R0 system is inertia dominated and may persist only during flow burst activity. While azimuthally diverging flows near the dipole contribute to the buildup of R1 and R2 systems, converging flows at larger distance contribute to the buildup of R0 and R1 systems. In contrast to the current diversion regions inside the current wedge, generator regions are found on the outside of the wedge, similar to earlier results. Within the tail domain covered, these regions are overpowered by load regions, such that additional generator regions must be expected closer to Earth, not covered by the present simulation.

SmartFCT: Improving power-efficiency for data center networks with deep reinforcement learning

Reducing the power consumption of Data Center Networks (DCNs) and guaranteeing the Flow Completion Time (FCT) of applications in DCNs are two major concerns for data center operators. However, existing works cannot realize the two goals together because of two issues: (1) dynamic traffic pattern in DCNs is hard to accurately model; (2) an optimal flow scheduling scheme is computationally expensive.In this paper, we propose SmartFCT, which employs the Deep Reinforcement Learning (DRL) coupled with Software-Defined Networking (SDN) to improve the power efficiency of DCNs and guarantee FCT. SmartFCT dynamically collects traffic distribution from switches to train its DRL model. The well-trained DRL agent of SmartFCT can quickly analyze the complicated traffic characteristics using neural networks and adaptively generate a action for scheduling flows and deliberately configuring margins for different links. Following the generated action, flows are consolidated into a few of active links and switches for saving power, and fine-grained margin configuration for active links avoids FCT violation of unexpected flow bursts. Simulation results show that SmartFCT can guarantee FCT and save up to 12.2% power consumption, compared with the state-of-the-art solutions.

The increase in the curvature radius of geomagnetic field lines preceding a classical dipolarization

Abstract. Based on assumptions that substorm field line dipolarization at geosynchronous altitudes is associated with the arrival of high-velocity magnetotail flow bursts referred to as bursty bulk flows, the following sequence of field line dipolarization is proposed: (1) slow magnetoacoustic wave excited through ballooning instability by enhanced inflows in pre-onset intervals towards the equatorial plane; (2) in the equatorial plane, slow magnetoacoustic wave stretching of the flux tube in dawn–dusk directions resulting in spreading plasmas in dawn–dusk directions and reduction in the radial pressure gradient in the flux tube. As a consequence of these processes, the flux tube assumes a new equilibrium geometry in which the curvature radius of new field lines increased in the meridian plane, suggesting an onset of field line dipolarization. The dipolarization processes associated with changing the curvature radius preceded classical dipolarization caused by a reduction of cross-tail currents and pileup of the magnetic fields. Increasing the curvature radius induced a convection surge in the equatorial plane as well as inductive westward electric fields of the order of millivolts per meter (mV m−1). Electric fields transmitted to the ionosphere produce an electromotive force in the E layer for generating a field-aligned current system of Bostrom type. This is also equivalent to the creation of an incomplete Cowling channel in the ionospheric E layer by the convection surge.

Impact of Flow Bursts in the Auroral Zone on the Ionosphere and Thermosphere

AbstractHigh‐latitude ionospheric plasma flows often contain embedded localized velocity enhancements with associated plasma density structures. Localized strong flows are significant energy sources to the thermosphere that are not contained in present empirical models of plasma convection. In this study, Global Ionosphere‐Thermosphere model, a nonhydrostatic model with flexible resolution, has been utilized to evaluate the influence of such localized strong enhancements of forcing on the global dynamics of the upper atmosphere. Specifically, three idealized configurations of flow burst have been examined, one flow burst with the typical spatial scale seen in the auroral zone, two neighboring flow bursts with the same spatial scale, and one flow burst with doubled spatial scale. The comparison between the cases with and without the flow burst illustrates that one regular‐sized flow burst at midnight can cause the neutral density to increase by 5% and the horizontal wind to increase by ∼30 m/s at 300 km altitude. The associated enhancement in the electron precipitation can increase total electron content by 15%( 1 TECU) during the first 15 min. Two simultaneous flow bursts can cause a rich wave structure with multiple wavelengths in traveling atmospheric disturbances due to wave‐wave interactions. A flow burst with doubled spatial size can increase the impact on the ionosphere/thermosphere by 2 times or even more. The outcome of this study illustrates substantial local and nonlocal responses of the ionosphere‐thermosphere system to the mesoscale energy and momentum inputs from the magnetosphere.

Storm-time convection dynamics viewed from optical auroras

A series of statistical and event studies have demonstrated that the motion of patches in regions of Patchy Pulsating Aurora (PPA) is very close to, if not exactly, E×B convection. Therefore, 2D maps of PPA motion provide us the opportunity to remotely sense magnetospheric convection with relatively high space and time resolution, subject to uncertainties associated with the mapping between the ionosphere and magnetosphere. In this study, we use THEMIS ASI (All Sky Imager) aurora observations combined with RBSP electric field and magnetic field measurements to explore convection dynamics during storm time. From 0500 UT to 0600 UT on March 19 2015, auroral observations across ~4 h of magnetic local time (MLT) show that increases in the westward velocities of patches are closely related to earthward flow bursts in the inner plasma sheet. Together with the meridian scanning photometer (MSP) data, this suggests that the increase in the westward velocities of PPA patches is caused by earthward-moving ion injection structures carried by the fast earthward flows.

Observational Evidence for Throat Aurora Being Associated With Magnetopause Reconnection

AbstractThroat auroras have been suggested to be related to indentations on the subsolar magnetopause. However, the indentation generation process and the resulting ionospheric responses have remained unknown. An EISCAT Svalbard Radar experiment was designed to run with all‐sky cameras, which enabled us for the first time to observe the temporal and spatial evolution of flow reversals, Joule heating, and ion upflows associated with throat aurora. The high‐resolution data enabled us to discriminate that the flow bursts and Joule heating were concurrent and co‐located, but were always observed on the west side of the associated throat auroras, reflecting that the upward/downward field‐aligned currents associated with throat aurora are always to the east/west, respectively. These results are consistent with the geometry of Southwood (1987) flux transfer event model and provide strong evidence for throat aurora being associated with magnetopause reconnection events. The results also support a conceptual model of the throat aurora.

On the Contribution of Dipolarizing Flux Bundles to the Substorm Current Wedge and to Flux and Energy Transport

AbstractBursty bulk flows and dipolarizing flux bundles within them play an important role in the transport of mass, energy, and magnetic flux in the magnetotail. On the basis of an magnetohydrodynamic simulation of magnetotail reconnection and dipolarization, we investigate the contribution of individual bursts and flux transport events to the buildup of the substorm current wedge, as well as to the earthward transport of magnetic flux and energy. Individual events, defined by increased flow speed (flow bursts), increased cross‐tail electric field, or increased (or increasing) magnetic field Bz, are found to be closely related but not identical. Multiple individual magnetic flux transport events collectively contribute to tailward and azimuthal expansion of dipolarization in the inner tail and to an increase of total field‐aligned currents toward or away from the ionosphere. In contrast, the current closure across midnight, estimated from the surface currents at the inner (earthward) boundary of the simulation box, was found to remain only a fraction (∼10% or 0.2 MA) of the total Region 1 current into to ionosphere. The simulation showed dipolarization everywhere earthward of the near‐Earth x‐line, amounting to ∼2.3 ×108 Wb, commensurate with substorm estimates. This can appear at a satellite in various ways, through either classical earthward transport and pileup (outward moving accumulation) or lateral (azimuthal) or tailward (vortical or recoiled) convective motion of dipolarized flux tubes, or a combination of these.

Electron‐Driven Dissipation in a Tailward Flow Burst

AbstractTraditionally, the magnetotail flow burst outside the diffusion region is known to carry ions and electrons together (Vi = Ve), with the frozen‐in condition well satisfied (E + Ve × B = 0). Such picture, however, may not be true, based on our analyses of the high‐resolution MMS (Magnetospheric Multiscale mission) data. We find that inside the flow burst the electrons and ions can be decoupled (Ve ≠ Vi), with the electron speed 5 times larger than the ion speed. Such super‐Alfvenic electron jet, having scale of 10 di (ion inertial length) in XGSM direction, is associated with electron demagnetization (E + Ve × B ≠ 0), electron agyrotropy (crescent distribution), and O‐line magnetic topology but not associated with the flow reversal and X‐line topology; it can cause strong energy dissipation and electron heating. We quantitatively analyze the dissipation and find that it is primarily attributed to lower hybrid drift waves. These results emphasize the non‐MHD (magnetohydrodynamics) behaviors of magnetotail flow bursts and the role of lower hybrid drift waves in dissipating energies.

Flow Context and Host Behavior Based Shadowsocks’s Traffic Identification

Cloud Virtual Private Server (VPS) services provide the chance of rapid deployment of anonymous proxy services, becoming an important part of many anonymous proxy solutions. The anonymous system represented by ShadowSocks (SS), through proxy services deployed on VPSs provided by different cloud service providers, has become an important mean for illegal network activists to engage in illegal network activities such as cyber-attacks and darknet transactions. It is difficult for local network administrators to supervise SS traffic from the cloud. While from the local network, the task faces the challenges of Invisible Negotiation Process and Data Transparent Transmission. In this paper, we present a novel SS detection method based on flow context and host behavior. The method can not only accurately identify SS flows, but also be applicable to a large-scale network environment. In this method, we extract 12-dimensional features from three aspects: the relationship between flows, hosts' flow behavior, and hosts' DNS behavior to build the detection model. Among them, the four features about flow burst and the feature of unassociated domain names' number are innovatively proposed in this paper. Moreover, the big data statistical and association techniques are used in the method. To verify the effectiveness of the method, we first built a real SS running environment based on the campus network and two VPSs on two different public cloud platforms. Moreover, we conduct a series of experiments on the NTCI-BDP data platform which is a big data platform built by our team. The experimental results show that our method achieves 93.43% accuracy on experimental data sets and can effectively identify SS traffic.

Fast plasma sheet flows and X line motion in the Earth's magnetotail: results from a global hybrid-Vlasov simulation

Abstract. Fast plasma flows produced as outflow jets from reconnection sites or X lines are a key feature of the dynamics in the Earth's magnetosphere. We have used a polar plane simulation of the hybrid-Vlasov model Vlasiator, driven by steady southward interplanetary magnetic field and fast solar wind, to study fast plasma sheet ion flows and related magnetic field structures in the Earth's magnetotail. In the simulation, lobe reconnection starts to produce fast flows after the increasing pressure in the lobes has caused the plasma sheet to thin sufficiently. The characteristics of the earthward and tailward fast flows and embedded magnetic field structures produced by multi-point tail reconnection are in general agreement with spacecraft measurements reported in the literature. The structuring of the flows is caused by internal processes: interactions between major X points determine the earthward or tailward direction of the flow, while interactions between minor X points, associated with leading edges of magnetic islands carried by the flow, induce local minima and maxima in the flow speed. Earthward moving flows are stopped and diverted duskward in an oscillatory (bouncing) manner at the transition region between tail-like and dipolar magnetic fields. Increasing and decreasing dynamic pressure of the flows causes the transition region to shift earthward and tailward, respectively. The leading edge of the train of earthward flow bursts is associated with an earthward propagating dipolarization front, while the leading edge of the train of tailward flow bursts is associated with a tailward propagating plasmoid. The impact of the dipolarization front with the dipole field causes magnetic field variations in the Pi2 range. Major X points can move either earthward or tailward, although tailward motion is more common. They are generally not advected by the ambient flow. Instead, their velocity is better described by local parameters, such that an X point moves in the direction of increasing reconnection electric field strength. Our results indicate that ion kinetics might be sufficient to describe the behavior of plasma sheet bulk ion flows produced by tail reconnection in global near-Earth simulations. Keywords. Magnetospheric physics (magnetospheric configuration and dynamics; plasma sheet) – space plasma physics (numerical simulation studies)

Cytocompatibility and mechanical properties of surgical sealants for cardiovascular applications

ObjectivesThe present study compared physical, mechanical, and biologic characteristics of 4 clinically available surgical sealants for cardiovascular repair. MethodsBioGlue (Cryolife Inc, Kennesaw, Ga), PreveLeak (Mallinckrodt Pharmaceuticals, St Louis, Mo), Tridyne VS (BD, Franklin Lakes, NJ), and Coseal (Baxter Healthcare Corporation, Westlake Village, Calif) were compared for the following properties: hydrated swelling, cytocompatibility, burst strength, biaxial stretching (elasticity), and in vitro degradation. ResultsSealants showed a wide range of swelling upon hydration. By gravimetric and volumetric measurement, swelling was greatest for Coseal followed by Tridyne VS, BioGlue, and PreveLeak. Tridyne VS was the most cytocompatible based on Alamar Blue assay results, supporting 85% cell survival compared with 36% to 39% survival with the other sealants. All sealants withstood pressure above mean arterial pressure (70-110 mm Hg) and physiologic systolic blood pressure (90-140 mm Hg) in an ex vivo arterial flow burst model; lowest peak pressure at failure was PreveLeak at 235 ± 48 mm Hg, and highest peak pressure at failure was BioGlue at 596 ± 72 mm Hg. Biaxial tensile testing showed no differences in elasticity between ex vivo porcine aorta and carotid arteries and Tridyne VS or Coseal, and BioGlue and PreveLeak were significantly stiffer. In vitro degradation time for Coseal was 6 days and 21 days for Tridyne VS. No degradation was observed in BioGlue or PreveLeak for 30 days. ConclusionsAlthough all sealants withstood supraphysiologic arterial pressure, there were differences in characteristics that may be important in clinical outcome. Coseal degradation time was short compared with other sealants, whereas BioGlue and PreveLeak showed a significant compliance mismatch with native porcine carotid artery. Tridyne VS was significantly more cytocompatible than the other 3 sealants.

Clogging and transport of driven particles in asymmetric funnel arrays

We numerically examine the flow and clogging of particles driven through asymmetric funnel arrays when the commensurability ratio of the number of particles per plaquette is varied. The particle–particle interactions are modeled with a soft repulsive potential that could represent vortex flow in type-II superconductors or driven charged colloids. The velocity-force curves for driving in the easy flow direction of the funnels exhibit a single depinning threshold; however, for driving in the hard flow direction, we find that there can be both negative mobility where the velocity decreases with increasing driving force as well as a reentrant pinning effect in which the particles flow at low drives but become pinned at intermediate drives. This reentrant pinning is associated with a transition from smooth 1D flow at low drives to a clogged state at higher drives that occurs when the particles cluster in a small number of plaquettes and block the flow. When the drive is further increased, particle rearrangements occur that cause the clog to break apart. We map out the regimes in which the pinned, flowing, and clogged states appear as a function of plaquette filling and drive. The clogged states remain robust at finite temperatures but develop intermittent bursts of flow in which a clog temporarily breaks apart but quickly reforms.

Stormtime substorm onsets: occurrence and flow channel triggering

Bright auroral emissions during geomagnetic storms provide a good opportunity for testing the proposal that substorm onset is frequently triggered by plasma sheet flow bursts that are manifested in the ionosphere as auroral streamers. We have used the broad coverage of the ionospheric mapping of the plasma sheet offered by the high-resolution THEMIS all-sky-imagers (ASIs) and chose the main phases of 9 coronal mass ejection (CME) related and 9 high-speed stream (HSS)-related geomagnetic storms, and identified substorm auroral onsets defined as brightening followed by poleward expansion. We found a detectable streamer heading to near the substorm onset location for all 60 onsets that we identified and were observed well by the ASIs. This indicates that substorm onsets are very often triggered by the intrusion of plasma with lower entropy than the surrounding plasma to the onset region, with the caveat that the ASIs do not give a direct measure of the intruding plasma. The majority of the triggering streamers are “tilted streamers,” which extend eastward as their eastern tip tilts equatorward to near the substorm onset location. Fourteen of the 60 cases were identified as “Harang streamers,” where the streamer discernibly turns toward the west poleward of reaching to near the onset latitude, indicating flow around the Harang reversal. Using the ASI observations, we observed substantially less substorm onsets for CME storms than for HSS storms, a result in disagreement with a recent finding of approximately equal substorm occurrences. We suggest that this difference is a result of strong non-substorm streamers that give substorm-like signatures in ground magnetic field observations but are not substorms based on their auroral signature. Our results from CME storms with steady, strong southward IMF are not consistent with the ~ 2–4 h repetition of substorms that has been suggested for moderate to strong southward IMF conditions. Instead, our results indicate substantially lower substorm occurrence during such steady driving conditions. Our results also show the much more frequent occurrence of substorms during HSS period, which is likely due to the highly fluctuating IMF.

Ultrasonic vibration-strengthened adhesive bonding of CFRP-to-aluminum joints

The strengthening mechanism of an ultrasonic vibration-strengthened adhesive bonding method was studied via experimental and numerical analyses. This processing method was found to enhance the bonding strength and stability of CFRP-to-aluminum joints by 40% and 60%, respectively. Ultrasonic vibration could promote the flow of adhesive to fill the bonding gap through the vibration conversion and transmission in the CFRP laminate. The high-frequency vibration of the adhesive caused large bubbles in the adhesive layer to stretch, deform towards the nearby flow front and burst on account of the asymmetric resistance offered by the bubbles to the vibration-induced flow. Ultrasonic vibration made it easier for the adhesive to permeate into the fine structure of the surface, because the permeation was driven by the hydraulic pressure difference produced by the prompted flow of the adhesive. The processing method improves the adhesion strength and bonding stability by utilizing external energy to effectively facilitate the adhesive bonding process; hence, it is an active enhancement method.

A Case Study of Connection Between Ground Magnetic Field Perturbations and Tail Current Sheet Bursty Flows at X = -60 RE.

We report the first observations that connect a burst of fast flows and Pi2 fluctuations in the midtail current sheet at X ~-60 RE with ground magnetic negative bays and Pi2 pulsations at high latitudes near 70°. Both the flow burst and negative bays occurred around midnight and were isolated and lasted for ~10 min. The flow burst had the characteristics of a low-density plasma bubble. It was moving earthward at ~340 km/s and was accompanied by field-aligned currents. The timing and equatorward motion of magnetic bays can be accounted for by this earthward moving bubble. Enhanced Pi2 fluctuations were observed within both the bubble and the magnetic bays. Pressure anisotropy with higher parallel pressure was observed associated with the flow burst and was firehose unstable. We discuss firehose instability as one of the possible mechanisms for the Pi2 fluctuations.

Influence of Auroral Streamers on Rapid Evolution of Ionospheric SAPS Flows

AbstractSubauroral polarization streams (SAPS) often show large, rapid enhancements above their slowly varying component. We present simultaneous observations from ground‐based all‐sky imagers and flows from the Super Dual Auroral Radar Network radars to investigate the relationship between auroral phenomena and flow enhancement. We first identified auroral streamers approaching the equatorward boundary of the auroral oval to examine how often the subauroral flow increased. We also performed the reverse query starting with subauroral flow enhancements and then evaluated the auroral conditions. In the forward study, 98% of the streamers approaching the equatorward boundary were associated with SAPS flow enhancements reaching ~700 m/s and typically hundreds of m/s above background speeds. The reverse study reveals that flow enhancements associated with streamers (60%) and enhanced larger‐scale convection (37%) contribute to SAPS flow enhancements. The strong correlation of auroral streamers with rapid evolution (approximately minutes) of SAPS flows suggests that transient fast earthward plasma sheet flows can often lead to westward SAPS flow enhancements in the subauroral region and that such enhancements are far more common than only during substorms because of the much more frequent occurrences of streamers under various geomagnetic conditions. We also found a strong correlation between flow duration and streamer duration and a weak correlation between SAPS flow velocity and streamer intensity. This result suggests that intense flow bursts in the plasma sheet (which correlate with intense streamers) are associated with intense SAPS ionospheric flows perhaps by enhancing the ring current pressure and localized pressure gradients when they are able to penetrate close enough to Earth.

Limitations of demand- and pressure-driven modeling for large deficient networks

Abstract. The calculation of hydraulic state variables for a network is an important task in managing the distribution of potable water. Over the years the mathematical modeling process has been improved by numerous researchers for utilization in new computer applications and the more realistic modeling of water distribution networks. But, in spite of these continuous advances, there are still a number of physical phenomena that may not be tackled correctly by current models. This paper will take a closer look at the two modeling paradigms given by demand- and pressure-driven modeling. The basic equations are introduced and parallels are drawn with the optimization formulations from electrical engineering. These formulations guarantee the existence and uniqueness of the solution. One of the central questions of the French and German research project ResiWater is the investigation of the network resilience in the case of extreme events or disasters. Under such extraordinary conditions where models are pushed beyond their limits, we talk about deficient network models. Examples of deficient networks are given by highly regulated flow, leakage or pipe bursts and cases where pressure falls below the vapor pressure of water. These examples will be presented and analyzed on the solvability and physical correctness of the solution with respect to demand- and pressure-driven models.

Driving of strong nightside reconnection and geomagnetic activity by polar cap flows: Application to CME shocks and possibly other situations

Previous studies have shown that dynamic pressure impacts (e.g., shocks initiating CME storms) with southward IMF promptly lead to strong auroral nightside activity and concurrent poleward expansion (indicating strong nightside reconnection), and strong enhancements in convection and currents. Here we use a combination of ground-based ASI and radar observations to further describe this response, to address what is driving the strong activity, and to suggest similar driving in other situations. Consistent with some previous studies, we find that shock driven auroral activity and poleward expansion resembles a substorm, but starts from an already broad MLT sector without much subsequent azimuthal expansion and without classical brightening of the equatorward-most arc. We furthermore find a large enhancement of meso-scale ionospheric polar cap flows heading towards the nightside separatrix immediately after shock impact. Recent studies have shown that such enhanced flows often cross the separatrix leading to plasma sheet flow bursts, poleward boundary intensifications (PBIs), streamers, and poleward motion of the polar cap boundary from reconnection. Thus these flow enhancements, which must extend outward along field lines from the ionosphere, are an attractive candidate as the driver for the almost immediate strong auroral, current, and reconnection activity resulting from shock impact. We also discuss and present some evidence that this phenomenon may be more general, leading to similar oval responses without a shock impact, including during and following the expansion phase of some substorms. These suggestions could lead to some possibly fundamental questions, such as when do polar cap convection enhancements lead to a substorm growth phase versus leading directly to strong polar expansion of, and strong activity along, the auroral oval field line region?

Episodic surface intrusions in the Yellow Sea during relaxation of northerly winds

AbstractThe surface currents over the Yellow and East China Seas are mapped from the Geostationary Ocean Color Imager (GOCI). Based on a composite of six intrusion events in January–April, the strong northward surface current in the Yellow Sea is shown to be concentrated along the deep trough, accompanied by a broad northward surface current over the East China Sea. From the corresponding surface winds, the episodic northward surface flow bursts appear to be associated with abrupt changes from the strong northerly winds to weak southerly winds during cold front passages. A three‐dimensional model driven with observed surface winds is used to simulate the observed shelf‐wide response to northerly winds. There is an outstanding agreement between the simulated and observed surface currents. The surface intrusion in the Yellow Sea is shown to be driven primarily by a barotropic longitudinal surface slope, while the strong northward current in the East China Sea is associated with a coastal trapped wave. Moreover, the surface intrusion is associated with a large volume transport, suggesting that the transient intrusions could be important in the northward heat transport. The unprecedented capability of GOCI satellite in providing a regional circulation pattern, in conjunction with complementary model simulations, could contribute greatly to understanding of the dynamics of the Yellow and East China Seas.

Forecasting short-term subway passenger flow under special events scenarios using multiscale radial basis function networks

Reliable and accurate short-term subway passenger flow prediction is important for passengers, transit operators, and public agencies. Traditional studies focus on regular demand forecasting and have inherent disadvantages in predicting passenger flows under special events scenarios. These special events may have a disruptive impact on public transportation systems, and should thus be given more attention for proactive management and timely information dissemination. This study proposes a novel multiscale radial basis function (MSRBF) network for forecasting the irregular fluctuation of subway passenger flows. This model is simplified using a matching pursuit orthogonal least squares algorithm through the selection of significant model terms to produce a parsimonious MSRBF model. Combined with transit smart card data, this approach not only exhibits superior predictive performance over prevailing computational intelligence methods for non-regular demand forecasting at least 30min prior, but also leverages network knowledge to enhance prediction capability and pinpoint vulnerable subway stations for crowd control measures. Three empirical studies with special events in Beijing demonstrate that the proposed algorithm can effectively predict the emergence of passenger flow bursts.