Statistical error estimation methods for engineering-relevant quantities from scale-resolving simulations

Scale-resolving simulations, such as large eddy simulations, have become affordable tools to investigate the flow in turbomachinery components. The resulting time-resolved flow field is typically analyzed using first- and second-order statistical moments. However, two sources of uncertainty are present when recording statistical moments from scale-resolving simulations: the influence of initial transients and statistical errors due to the finite number of samples. In this paper, both are systematically analyzed for several quantities of engineering interest using time series from a long-time large eddy simulation of the low-pressure turbine cascade T106C. A set of statistical tools to either remove or quantify these sources of uncertainty is assessed. First, the Marginal Standard Error Rule is used to detect the end of the initial transient. The method is validated for integral and local quantities and guidelines on how to handle spatially varying initial transients are formulated. With the initial transient reliably removed, the statistical error is estimated based on standard error relations considering correlations in the time series. The resulting confidence intervals are carefully verified for quantities of engineering interest utilizing cumulative and simple moving averages. Furthermore, the influence of periodic content from large scale vortex shedding on the error estimation is studied. Based on the confidence intervals, the required averaging interval to reduce the statistical uncertainty to a specific level is indicated for each considered quantity.

Low-Latency Spiking Neural Networks Using Pre-Charged Membrane Potential and Delayed Evaluation

Spiking neural networks (SNNs) have attracted many researchers’ interests due to its biological plausibility and event-driven characteristic. In particular, recently, many studies on high-performance SNNs comparable to the conventional analog-valued neural networks (ANNs) have been reported by converting weights trained from ANNs into SNNs. However, unlike ANNs, SNNs have an inherent latency that is required to reach the best performance because of differences in operations of neuron. In SNNs, not only spatial integration but also temporal integration exists, and the information is encoded by spike trains rather than values in ANNs. Therefore, it takes time to achieve a steady-state of the performance in SNNs. The latency is worse in deep networks and required to be reduced for the practical applications. In this work, we propose a pre-charged membrane potential (PCMP) for the latency reduction in SNN. A variety of neural network applications (e.g., classification, autoencoder using MNIST and CIFAR-10 datasets) are trained and converted to SNNs to demonstrate the effect of the proposed approach. The latency of SNNs is successfully reduced without accuracy loss. In addition, we propose a delayed evaluation method (DE), by which the errors during the initial transient are discarded. The error spikes occurring in the initial transient is removed by DE, resulting in the further latency reduction. DE can be used in combination with PCMP for further latency reduction. Finally, we also show the advantages of the proposed methods in improving the number of spikes required to reach a steady-state of the performance in SNNs for energy-efficient computing.

Nature’s Pick-Up Tool, the Stark Effect Induced Gailitis Resonances and Applications

A simple universal physical mechanism hidden for more than half a century is unexpectedly discovered from a calculation of low excitation antihydrogen. For ease of reference, this mechanism is named Gailitis resonance. We demonstrate, in great detail, that Gailitis resonances are capable of explaining p+7Li low energy nuclear fusion, d-d fusion on a Pd lattice and the initial transient fusion peak in muon catalyzed fusion. Hopefully, these examples will help to identify Gailitis resonances in other systems.

Comments on ‘Type II migration strikes back – an old paradigm for planet migration in discs’ by Scardoni et al.

ABSTRACT In the conventional view of type II migration, a giant planet migrates inward in the viscous velocity of the accretion disc in the so-called disc-dominate case. Recent hydrodynamic simulations, however, showed that planets migrate with velocities much faster than the viscous one in massive discs. Such fast migration cannot be explained by the conventional picture. Scardoni et al. has recently argued this new picture. By carrying out similar hydrodynamic simulations, they found that the migration velocity slows down with time and eventually reaches the prediction by the conventional theory. They interpreted the fast migration as an initial transient one and concluded that the conventional type II migration is realized after the transient phase. We show that the migration velocities obtained by Scardoni et al. are consistent with the previous simulations even in the transient phase that they proposed. We also find that the transient fast migration proposed by Scardoni et al. is well described by a new model of Kanagawa, Tanaka & Szuszkiewicz. The new model can appropriately describe significant inward migration during the initial transient phase that Scardoni et al. termed. Hence, we conclude that the time variation of the transient migration velocity is due to the changes of the orbital radius of the planet and its background surface density during the migration.

Novel Wood Resistance Measurement Method Reducing the Initial Transient Instabilities Arising in DC Methods Due to Polarization Effects

A novel method for measuring the electrical resistance in wood is presented. It is based on applying an Alternating Current (AC) to two electrodes rammed into the wood. The method reduces the transient time for value stabilization. In case of Direct Current (DC) resistance measurement methods, typically used in wood measurement, an initial transient exists, invalidating the measured value during an initial transient period. This measurement method uses an electronic circuit based on a relaxation oscillator where the wood automatically sets the oscillation frequency depending on its electrical resistance. Compared to other AC methods, this circuit greatly simplifies the measurement process, not requiring any previous analysis for wood AC frequency estimation. Experimental results for four different wood species showed that, in all cases, the transient response of the measured wood resistance is improved when compared to the DC resistance method, reducing the stabilization time from 10–12 min in DC measurements to almost zero for the novel measurement method. The deviation between the initial wood resistance measurement and the stable value resulted in a range between 0.2% and 5% with the proposed method, while a range between 6% and 10% was obtained when using the DC method. Moreover, the proposed circuit is able to detect changes in wood resistance produced by small variations in temperature and environmental relative humidity during continuous long term measurements. For repeatability analysis, it was verified that similar ambient conditions of temperature and relative humidity (variations lower than 1 °C and 1%, respectively) in different moments provided variations lower than 1.5%. The method can also be applied to other fiber materials affectd by polarization effects when an electrical current is applied into them.