SEARCH · Engineering Papers

Results for “power boost”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 595 records · Page 33

All sources NASA NTRS DOE OSTI

DOE OSTI2023

Quantum Efficiency Gain in 2D Perovskite Photo and X‐Ray Detectors

Abstract The perovskite polycrystalline thin film detector fabricated by solution method is a promising low‐cost, scalable technology for radiation imaging, but its thin volume limits the sensing efficiency for high‐energy X‐ray photons. This work reports 2D perovskite thin film photo‐diodes with a detection gain when sensing visible and X‐ray photons. Detailed power and temperature‐dependent device characterizations reveal a charge multiplication effect to be responsible for the observed high gain. This is caused by a disparity in electron and hole transport, where electron transport is retarded by a trap/de‐trap process via shallow trap states, whereas the hole transport is fast enough to produce a photoconductive gain in satisfying the charge neutrality. The 2D perovskite made with butylamine spacers is also found to exhibit a larger efficiency gain than those made with phenylethylamines because of the higher likelihood of forming shallow traps in the former. The thin film diodes feature a high temporal response over 1 MHz due to the fast charge collection across a thin volume, and the discovery provides device physics mechanisms in connection to material structure‐function relationship for future optoelectronics design that can boost the efficiency of X‐ray sensing and dim light detection.

2D perovskites↗

DOE OSTI2026

Enhancing mobility and interface state engineering via UV-ozone treatment in BEOL-compatible ultrathin TiO 2 transistors

It has emerged as a potential candidate to improve the performance of monolithic-three-dimensional (M3D) integration of fused logic and memories through low-temperature in situ synthesis of high-performance metal–oxide–semiconductor (MOS) transistors. Here, we report the demonstration of the BEOL-compatible low thermal budget (350 °C) fabrication process of ultrathin-TiO2 transistors by the combination of RTA and UV-ozone treatment (RTA-UVOz). UV–ozone (UVOz) treatment of TiO2 anatase films significantly enhances stability, boosting ION current and field effect mobility (μ FE ) by two times of magnitude in TiO2 TFTs. UV ozone treatment helps to eliminate pre-existing oxygen vacancies and carbon contamination on TiO2 channels even at low temperature (100 °C), resulting in high-quality channel/dielectric interfaces with low interface states (D it ). During UVOz treatment, oxygen species (O x ) passivate the oxygen vacancies ($V$$^{2+}_{o}$), and hence low concentration of $V$$^{2+}_{o}$ would be left for ionization/de-ionization under PBS/NBS, leading to improved bias stress stability. Furthermore, the TiO 2 TFTs with thin ZrO 2 gate dielectric exhibited excellent performance including a lower subthreshold swing (SS) of 98 mV/dec with high drive I ON current ∼ 4.5 μA/μm, a high I ON /I OFF > 10 9 , and mobility μFE of 7 cm 2 /V-s under a battery powered voltage of 1 V. UV–ozone treatment enables high-performance, CMOS-compatible TiO 2 transistors with a low thermal budget, ideal for next-generation flexible, energy-efficient electronics.

BEOL↗

NASA NTRS1966

The Lunar Orbiter: A Spacecraft to Advance Lunar Exploration

The film describes the Lunar Orbiter's mission to photograph landing areas on the Moon. The Orbiter will be launched from Cape Kennedy using an Atlas Agena booster rocket. Once it is boosted in a trajectory toward the Moon, the Orbiter will deploy two-way earth communication antennas and solar panels for electricity. Attitude control jets will position the solar panels toward the sun and a tracker for a fix on its navigational star. The Orbiter will be put in an off-center orbit around the Moon where it will circle from four to six days. Scientists on Earth will study the effects of the Moon's gravitational field on the spacecraft, then the orbit will be lowered to 28 miles above the Moon's surface. Engineers will control the Orbiter manually or by computer to activate two camera lenses. The cameras will capture pictures of 12,000 square miles of lunar surface in 25 and 400 square mile increments. Pictures will be sent back to Earth using solar power to transmit electrical signals. The signals will be received by antennas at Goldstone, CA, and in Australia and Spain. Incoming photographic data will be electronically converted and processed to produce large-scale photographic images. The mission will be directed from the Space Flight Operations Facility in Pasadena, CA by NASA and Boeing engineers. After the photographic mission, the Orbiter will continue to circle the Moon providing information about micrometeoroids and radiation in the vicinity.

Source record↗

NASA NTRS1991

Evolving the SP-100 reactor in order to boost large payloads to GEO and to low lunar orbit via nuclear-electric propulsion

In striving to reduce exploration cost and exploration risks, a crucial aspect of the plans is program continuity, i.e., the continuing application of a given technology over a long period so that experience will accumulate from extended testing here on Earth and from a diversity of applications in space. An integrated view needs to be formed of the missions SEI will carry out, near term as well as far, and of the ways in which these missions can mutually support one another. Near term programs should be so constituted as to provide for the long term missions both the enabling technologies and the accumulation of experience they need. In achieving this, missions in Earth orbit should both evolve and show the technologies crucial to long term missions on the lunar surface, and the program for the lunar labs should evolve and show the enabling technologies for exploration of the surface of Mars and for flights of human beings to Mars and return. In the near term, the program for the Space Station should be directed and funded to develop and demonstrate the solar Brayton power plant that will be most useful as the power generator for the SP-100 nuclear reactor.

English, Robert E.↗

NASA NTRS1991

Evolving the SP-100 reactor in order to boost large payloads to GEO and to low lunar orbit via nuclear-electric propulsion

In striving to reduce exploration cost and exploration risks, a crucial aspect of the plans is program continuity, i.e., the continuing application of a given technology over a long period so that experience will accumulate from extended testing here on earth and from a diversity of applications in space. An integrated view needs to be formed of the missions SEI will carry out, near term as well as far, and of the ways in which these missions can mutually support one another. Near term programs should be so constituted as to provide for the long term missions both the enabling technologies and the accumulation of experience they need. In achieving this, missions in earth orbit should both evolve and show the technologies crucial to long term missions on the lunar surfaces, and the program for the lunar labs should evolve and show the enabling technologies for exploration of the surface of Mars and for flights of human beings to Mars and return. In the near term, the program for the Space Station should be directed and funded to develop and demonstrate the solar Brayton power plant that will be most useful as the power generator for the SP-100 nuclear reactor.

English, Robert E.↗

DOE OSTI2026

Integration of a thermoelectric subcooler with an air‑source heat pump to enhance heating capacity and dehumidification

This study presents an innovative configuration of a thermoelectric subcooler (TES) integrated with an air-source heat pump (ASHP). The TES is installed in the liquid line, where it absorbs heat from the refrigerant. On its hot side, it discharges heat to the return air in heating mode or to the supply air in cooling mode, enabling year-round operation. The ASHP–TES system was evaluated under various conditions to reduce supplemental heating needs and improve dehumidification. Experimental heating tests revealed that activating the TES increased the heating capacity by 12%–20%. However, this occurred with a 7.6%–10% reduction in heating coefficient of performance (COP). In cooling mode, TES operation significantly improved moisture removal. The TES reduced the sensible heat ratio (SHR) and enhanced dehumidification, removing 20%–50% more moisture than a conventional system when meeting the same sensible load. The TES can be turned off to use the refrigerant subcooler for reheat supply air or turned on to substantially boost dehumidification. When meeting the same latent load, the ASHP–TES system achieved up to 33% energy savings compared with a conventional ASHP using resistance heating. Relative to a conventional ASHP paired with a whole-house dehumidifier, the ASHP–TES system provided up to 20% energy savings. These findings underscore the potential of TES integration to improve latent load control and enhance overall energy efficiency in hybrid ASHP systems, especially in cold and humid climates.

Hu, Yifeng [ORNL] (ORCID:0000000242875185)↗

DOE OSTI2024

Reliability and resilience of environmental flows under uncertainty: reconsidering water year types and inconsistent flow requirements in California

Abstract Environmental water allocation in California is a complex legal process involving various government agencies and stakeholders. E-flow requirements can be based on annual runoff typologies called water year types (WYTs), which dictate water volume, timing, and duration. In this study, we examined hydropower licensing documents of the major water and power projects in the Central Sierra Nevada to catalog e-flow requirements by WYT. In this study case, we identify how WYT classification systems and categories vary across and within different basins. Additionally, we assessed the impacts of climate change on hydrology, the frequency of WYTs identified, and the reliability and resilience of e-flows using future projections (2031–2060) of 10 Global Circulation Models (GCMs). We then propose a potential adaptation strategy using a 30 year moving percentiles approach to recalculate WYTs. We identified eight WYT classifications systems were identified, and their WYT distributions statistically significantly changes across all GCMs, even though most GCMs indicate no statistically significant change in hydrology. Disparities in future impacts are observed among and within hydropower projects, with some river reaches showing negative impacts on reliability and resilience. The adaptation strategy can generally boost resilience and improve reliability, but simply updating existing WYT thresholds without flexible regulatory frameworks reconsidering WYTs and e-flows thresholds, may not yield substantial improvements. Challenges in managing e-flows in California within regulatory and hydroclimatic contexts are intricate due to the lack of standardized approaches, leading to inconsistencies and potential conflicts among stakeholders, that will likely be exacerbated by climate change. Thus, we emphasize that targeted, site-specific, and adaptive management strategies are crucial, besides the need for a harmonized and consistent approach to defining and applying WYT categories and methods and/or e-flow assessments.

Environmental Sciences & Ecology↗

DOE OSTI2020

3D‐Printed Structure Boosts the Kinetics and Intrinsic Capacitance of Pseudocapacitive Graphene Aerogels

Abstract The performance of pseudocapacitive electrodes at fast charging rates are typically limited by the slow kinetics of Faradaic reactions and sluggish ion diffusion in the bulk structure. This is particularly problematic for thick electrodes and electrodes highly loaded with active materials. Here, a surface‐functionalized 3D‐printed graphene aerogel (SF‐3D GA) is presented that achieves not only a benchmark areal capacitance of 2195 mF cm −2 at a high current density of 100 mA cm −2 but also an ultrahigh intrinsic capacitance of 309.1 µF cm −2 even at a high mass loading of 12.8 mg cm −2 . Importantly, the kinetic analysis reveals that the capacitance of SF‐3D GA electrode is primarily (93.3%) contributed from fast kinetic processes. This is because the 3D‐printed electrode has an open structure that ensures excellent coverage of functional groups on carbon surface and facilitates the ion accessibility of these surface functional groups even at high current densities and large mass loading/electrode thickness. An asymmetric device assembled with SF‐3D GA as anode and 3D‐printed GA decorated with MnO 2 as cathode achieves a remarkable energy density of 0.65 mWh cm −2 at an ultrahigh power density of 164.5 mW cm −2 , outperforming carbon‐based supercapacitors operated at the same power density.

asymmetric supercapacitors↗

DOE OSTI2026

Weak Lensing Mass Calibration of the ACT DR5 Galaxy Clusters with the DES Year 3 Weak Lensing Data

We use weak gravitational lensing measurements from Year 3 Dark Energy Survey data to calibrate the masses of 443 galaxy clusters selected via the Sunyaev-Zel'dovich effect from Atacama Cosmology Telescope Data Release 5 maps of the cosmic microwave background. We incorporate redshift and SZ measurements for individual clusters into a hierarchical model for the stacked lensing signals and perform Bayesian analyses to constrain the hydrostatic mass bias of the clusters. Our treatment of systematic uncertainties includes a prescription for measuring and accounting for the weak lensing boost factor, consideration of a miscentering effect, as well as marginalization over uncertainties in the source galaxy photometric redshift distributions and shear calibration. The resultant constraints on the normalization of the mass-observable relation have a precision of approximately 7%, with the mean WL halo mass of M $_{500c}$ = 5.4 × 10$^{14}$ M $_{⊙}$. We measure the bias between the true cluster mass and the mass estimated from the SZ signal based on an X-ray-calibrated scaling relation assuming hydrostatic equilibrium, to be 1 - b = 0.74$^{+0.06}$ $_{-0.05}$ over the full sample. When splitting the clusters into high (z = 0.43-0.70) and low (z = 0.15-0.43) redshift bins, we measure 1 - b = 0.58$^{+0.06}$ $_{-0.05}$ and 0.81$^{+0.08}$ $_{-0.06}$, respectively. When introducing additional freedom in redshift and mass to the hydrostatic bias model, we find that 1 - b decreases with redshift (with the power law of -1.8$^{+0.5}$ $_{-0.6}$, 99.95% confidence), consistent with findings from other recent studies, while we do not find any significant trend in mass. We also demonstrate that our result is robust against various systematics such as a scale cut, priors on baryonic and miscentering parameters, and degree of scatter in mass-observable relation. The weak-lensing mass calibration presented in this study will be a useful tool for using the ACT clusters as probes of astrophysics, and as a step towards using their abundance as a cosmological probe.

Shin, T. [Carnegie Mellon U.] (ORCID:0000000263895↗

DOE OSTI2023

Hybrid Attack Graph Generation with Graph Convolutional Deep-Q Learning

Critical infrastructures such as power grids have become increasingly complex, connected, and vulnerable to adverse scenarios, including cyber and physical attacks and faults. Effective risk mitigation for such cyber-physical energy systems (CPES), requires preemptive knowledge of likely adversarial attack scenarios. Hybrid Attack Graph (HAG) is a structured way to represent an adversarial scenario as an attack sequence using a threat model. However, the scarcity of documented attack sequences hinders analysts and CPES planners’ ability to identify credible attack scenarios for a given CPES. We propose a data-driven Graph Convolutional Deep-Q Network (GCDQ) to address this data challenge through generating HAGs. By leveraging limited real-world observations from the MITRE ATT&CK knowledge base, our GCDQ model synthesizes realistic graphs with the targeted attribute of minimum detectability via reinforcement learning. This generative model is the first step in creating a tool to substantially boost the attack sequence dataset and enhance the performance of CPS defense-related tasks by providing insights into likely attack sequences with given attributes.

deep learning, artificial intelligence↗

DOE OSTI2024

A Comprehensive Strategy for Grid Forming Control in DC Coupled Photovoltaic and Battery Energy Storage Inverters

This paper presents an integrated DC-DC and DCAC grid-forming control strategy for DC-coupled photovoltaic (PV) plus battery energy storage systems, considering the effect of DC link voltage variations caused by direct PV connections. A power reference algorithm determines power distribution between the PV and battery to the grid while observing device power ratings to prevent the over-rating of components and keep the battery's state of charge within an acceptable range. The simulated utility-scale model in MATLAB/Simulink illustrates its ability against extreme phase angle variation contingencies in the grid while controlled through grid-forming control with a fast dynamic on DC link voltage. The simulation results confirm the effectiveness of the proposed control in integrating PV plus battery configurations with grid forming control and maintaining reliable grid operation under severe grid disturbances.

battery, boost, control, energy storage, grid form↗

DOE OSTI2022

H α emission in local galaxies: star formation, time variability, and the diffuse ionized gas

ABSTRACT The nebular recombination line H α is widely used as a star formation rate (SFR) indicator in the local and high-redshift Universe. We present a detailed H α radiative transfer study of high-resolution isolated Milky-Way and Large Magellanic Cloud simulations that include radiative transfer, non-equilibrium thermochemistry, and dust evolution. We focus on the spatial morphology and temporal variability of the H α emission, and its connection to the underlying gas and star formation properties. The H α and H β radial and vertical surface brightness profiles are in excellent agreement with observations of nearby galaxies. We find that the fraction of H α emission from collisional excitation amounts to fcol ∼ 5–$10{{\ \rm per\ cent}}$, only weakly dependent on radius and vertical height, and that scattering boosts the H α luminosity by $\sim 40{{\ \rm per\ cent}}$. The dust correction via the Balmer decrement works well (intrinsic H α emission recoverable within 25 per cent), though the dust attenuation law depends on the amount of attenuation itself both on spatially resolved and integrated scales. Important for the understanding of the H α–SFR connection is the dust and helium absorption of ionizing radiation (Lyman continuum [LyC] photons), which are about $f_{\rm abs}\approx 28{{\ \rm per\ cent}}$ and $f_{\rm He}\approx 9{{\ \rm per\ cent}}$, respectively. Together with an escape fraction of $f_{\rm esc}\approx 6{{\ \rm per\ cent}}$, this reduces the available budget for hydrogen line emission by nearly half ($f_{\rm H}\approx 57{{\ \rm per\ cent}}$). We discuss the impact of the diffuse ionized gas, showing – among other things – that the extraplanar H α emission is powered by LyC photons escaping the disc. Future applications of this framework to cosmological (zoom-in) simulations will assist in the interpretation of spectroscopy of high-redshift galaxies with the upcoming James Webb Space Telescope.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2026

Distortion of a Relativistic Jet Echoing a Magnetic Flux Eruption

Magnetized accretion onto spinning black holes (BHs) can accumulate a large magnetic flux across the event horizon and launch a pair of relativistic jets via the Blandford–Znajek mechanism. In the magnetically saturated (arrested) state, excess magnetic flux is ejected from the BH in episodic magnetic flux eruptions, which result in a significant yet temporary reduction of jet power. We analyze results of a high-resolution 3D general-relativistic magnetohydrodynamic numerical simulation of geometrically thick magnetically saturated accretion onto a high-spin Kerr BH for a single cycle of magnetic flux eruption and accumulation. We show that following an eruption, a weakened jet develops a strong helical distortion with a distinct structure of magnetic fields—the poloidal field along the jet core is unaffected by the eruption; while toroidal field lines, ejected from the BH during the eruption and later readvected onto it, form poloidal “bypasses” along the inner jet sheath. Such a distortion may appear in sources fed by geometrically thick accretion flows as an asymmetric superluminal knot, strongly interacting with the jet sheath along an oblique working surface. The jet section repowered by magnetic flux reaccumulated on the BH is tilted by a few degrees, implying significant variations in radiation boost towards observers of BL Lac blazars. The intrinsic structure of the jet spine is consistent with axisymmetric semi-analytical models.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2022

Relativistic Alfvén Waves Entering Charge-starvation in the Magnetospheres of Neutron Stars

Instabilities in a neutron star can generate Alfvén waves in its magnetosphere. Propagation along the curved magnetic field lines strongly shears the wave, boosting its electric current j A . We derive an analytic expression for the evolution of the wavevector k and the growth of j A . In the strongly sheared regime, j A may exceed the maximum current j0 that can be supported by the background e ± plasma. We investigate these charge-starved waves, first using a simplified two-fluid analytic model, then with first-principles kinetic simulations. We find that the Alfvén wave is able to propagate successfully even when κ ≡ j A /j 0 $\ll$ 1. It sustains j A by compressing and advecting the plasma along the magnetic field lines with an increasing Lorentz factor, γ ≳ κ 1/2 . The simulations show how plasma instabilities lead to gradual dissipation of the wave energy. Our results suggest that an extremely high charge-starvation parameter κ ≳ 10 4 may be required in order for this mechanism to power the observed fast radio bursts (FRBs) from SGR 1935+2154. However, cosmological FRBs with much higher luminosities are unlikely to be a result of charge-starvation.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2022

Foundation Models of Scientific Knowledge for Chemistry: Opportunities, Challenges and Lessons Learned

Foundation models pre-trained on large corpora demonstrate significant gains across many natural language processing tasks and domains e.g., law, healthcare, education, etc. However, only limited efforts have investigated the opportunities and limitations of applying these powerful models to science and security applications. In this work we develop foundation models of scientific knowledge for chemistry to augment scientists with the advanced ability to perceive and reason at scale previously unimagined. Specifically, we build large-scale (1.47B parameter) general-purpose models for chemistry that can be effectively used to perform a wide range of in-domain and out-of-domain tasks. Evaluating these models in a zero-shot setting, we analyze the effect of model and data scaling, knowledge depth, and temporality on model performance in context of model training efficiency. Our novel findings demonstrate that (1) model size significantly contributes to the task performance when evaluated in a zero-shot setting; (2) data quality (aka diversity) affects model performance more than data quantity; (3) similarly, unlike previous work (Luu et al., 2021) temporal order of the documents in the corpus boosts model performance only for specific tasks, e.g., SciQ; and (4) models pre-trained from scratch perform better on in-domain tasks than those tuned from general-purpose models like Open AI’s GPT-2.

Foundation Models, Chemistry↗

DOE OSTI2022

Enhanced convection for higher module and system efficiency (Final Technical Report)

The goal of this work is to develop new solar photovoltaic (PV) modules and solar system-scale designs that promote an increase of the convective heat transfer coefficient of at least 40%, reducing the operating temperature of the solar PV panels, and leading to a boost on the annual energy yield by at least 5%. By achieving this goal, we further foresee decreasing solar panel degradation by +0.3%/year, reducing the LCOE by 2.9-4.5¢/kWh. Currently, this question and/or approach is not considered though it is a relevant factor in performance/efficiency. The technology proposed will be at a mature stage by the end of the project and could be aligned to be proven on a full-field test. PV users, installers and manufacturers will be interested in the technology due to the aforementioned benefits. The work was conducted through a suite of wind tunnel experiments, numerical simulations and field experiments. These were then evaluated through a correlation encompassing the entire data set and implementing findings into NREL SAM. The work resulted in eight peer reviewed publications with three others in preparation. The key results of the project are 1.) Showing the capability of increasing the solar farm power production by increasing the heat transfer coefficient; 2.) Providing strategies for achieving this through large scale arrangement manipulation (heights, spacings, angles), flow deflectors and vortex generators; 3.) Demonstrating scalability of results by comparing numerical simulations, wind tunnel experiments and field experiments; 4.) Providing a correlation to be able to capture the various arrangements while exploiting the heat transfer coefficient; and 5.) Proving the feasibility of such strategies and their capacity through a comprehensive techno-economic analysis. For the heat transfer coefficients, gains of 40% were attained and increases in power in the excess of 5% were observed. For heat mitigation strategies, consideration at the PV array scale were inter-panel spacing, panel height and panel angles while at the module level, wind breaks such as solid walls and selected gaps were evaluated as well as vortex generators on the panel. All proposed strategies saw an increase in heat transfer coefficient. For scalability, results were compared against field experiments and a correlation for the heat transfer quantity was attained leading to a parametric understanding of the variations and impact in the heat transfer coefficient independent of tools implemented (wind tunnel/numerical simulations). The strategies were also further assessed by a technoeconomic analysis finding improvements based on LCOE with respect to location (several states with varies atmospheric conditions/geographic locations) yielded different results. Improvements on this end were also observed especially towards the colder climates and when considering bi-facial modules. The quantitative demonstrations of such conclusions are provided in the main portion of this document with complementary results and their narration. Milestones and go/no-go points were successfully met in the project.

14 SOLAR ENERGY↗

DOE OSTI2022

When Do Extended Physics-Informed Neural Networks (XPINNs) Improve Generalization?

Physics-informed neural networks (PINNs) have become a popular choice for solving high-dimensional partial differential equations (PDEs) due to their excellent approximation power and generalization ability. Recently, extended PINNs (XPINNs) based on domain decomposition methods have attracted considerable attention due to their effectiveness in modeling multiscale and multiphysics problems and their parallelization. However, theoretical understanding of their convergence and generalization properties remains unexplored. In this study, we take an initial step towards understanding how and when XPINNs outperform PINNs. Specifically, for general multilayer PINNs and XPINNs, we first provide a prior generalization bound via the complexity of the target functions in the PDE problem and a posterior generalization bound via the posterior matrix norms of the networks after optimization. Moreover, based on our bounds, we analyze the conditions under which XPINNs improve generalization. Concretely, our theory shows that the key building block of XPINN, namely, the domain decomposition, introduces a tradeoff for generalization. On the one hand, XPINNs decompose the complex PDE solution into several simple parts, which decreases the complexity needed to learn each part and boosts generalization. On the other hand, decomposition leads to less training data being available in each subdomain, and hence such a model is typically prone to overfitting and may become less generalizable. Empirically, we choose five PDEs to show when XPINNs perform better than, similar to, or worse than PINNs, hence demonstrating and justifying our new theory.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Skip-Connected Self-Recurrent Spiking Neural Networks with Joint Intrinsic Parameter and Synaptic Weight Training

As an important class of spiking neural networks (SNNs), recurrent spiking neural networks (RSNNs) possess great computational power and have been widely used for processing sequential data like audio and text. However, most RSNNs suffer from two problems. 1. Due to the lack of architectural guidance, random recurrent connectivity is often adopted, which does not guarantee good performance. 2. Training of RSNNs is in general challenging, bottlenecking achievable model accuracy. To address these problems, we propose a new type of RSNNs called Skip-Connected Self-Recurrent SNNs (ScSr-SNNs). Recurrence in ScSr-SNNs is introduced in a stereotyped manner by adding self-recurrent connections to spiking neurons. The SNNs with self-recurrent connections can realize recurrent behaviors similar to those of more complex RSNNs while the error gradients can be more straightforwardly calculated due to the mostly feedforward nature of the network. The network dynamics is enriched by skip connections between nonadjacent layers. Moreover, we propose a new backpropagation (BP) method called backpropagated intrinsic plasticity (BIP) to further boost the performance of ScSr-SNNs by training intrinsic model parameters. Unlike standard intrinsic plasticity rules that adjust the neuron's intrinsic parameters according to neuronal activity, the proposed BIP method optimizes intrinsic parameters based on the backpropagated error gradient of a well-defined global loss function in addition to synaptic weight training. Here, based on challenging speech, neuromorphic speech, and neuromorphic image datasets, the proposed ScSr-SNNs can boost performance by up to 2.85% compared with other types of RSNNs trained by state-of-the-art BP methods.

97 MATHEMATICS AND COMPUTING↗