Papers by Keyword: Gaussian Distribution

Abstract: Lithium-ion batteries play a critical role in the reliability and safety of a system. Battery health monitoring and remaining useful life (RUL) prediction are needed to prevent catastrophic failure of the battery. The aim of this research is to develop a data-driven method to monitor the batteries state of health and predict their RUL by using the battery capacity degradation data. This paper also investigated the effect of prediction starting point to the RUL prediction error. One of the data-driven method drawbacks is the need of a large amount of data to obtain accurate prediction. This paper proposed a method to generate a series of degradation data that follow the Gaussian distribution based on limited battery capacity degradation data. The prognostic model was constructed from the new data using least square support vector machine (LSSVM) regression. The remaining useful life prediction was carried out by extrapolating the model until reach the end of life threshold. The method was applied to three differences lithium-ion batteries capacity data. The results showed that the proposed method has good performance. The method can predict the lithium-ion batteries RUL with a small error, and the optimal RUL starting point was found at the point where the battery has experienced the highest capacity recovery due to the self-recharge phenomenon.

Abstract: Two characterization methods are compared in terms of their suitability for predicting the electrical behavior of non-uniform Ni/4H-SiC Schottky contacts up to 450°C, using data measured at lower temperatures. These techniques are based on the established Gaussian distribution of barrier heights model and a recently proposed discrete barrier distribution model, respectively. Two samples with different degrees of contact inhomogeneity are measured and their forward characteristics are fitted using both techniques. The Gaussian distribution approach is shown to accurately fit experimental data only for the nearly-uniform sample, while requiring the extraction of two separate barrier height values from measurements in the room-250°C range, only. When attempting to use this method to characterize the sample with strong non-uniformity, fitting accuracy (given by R²) drops under 90%. In contrast, the discrete barrier distribution technique is proven able to forecast the electrical behavior of both samples (with R² > 99% in most cases), over the entire room-450°C range, using a single Schottky barrier for each device (1.61V, corresponding to a Ni₂Si Schottky contact and 0.9V, afferent to a Ni metallization).

Abstract: Based on the coupling phenomena between the batch drilling process quality fluctuation and monitoring signal features changes, the acoustic emission signals of batch drilling are taken as research objects to solve the quality consistency control and detection problem of the high-precision batch drilling step. The average bi-spectrum amplitude of each signal is taken as an eigenvalue for the quantitative analysis of the deviation degree from Gaussian distribution under different conditions. The calculation and analysis results show that there are organic connections between the bi-spectrum feature of monitoring signals and drilling step quality, and the consistency quality testing of batch drilling step is realized by bi-spectrum features.

Abstract: The description of hysteresis is one of the classical problems in magnetic materials. The progress in its solution determines the reliability of modeling and the quality of design of a wide range of devices, the proposed approach has been applied to model the behavior of many samples and the results show the robustness and efficiency of Neural Network to model the phenomenon of hysteresis loop. The goal of this study is to optimize the parameters of hysteresis Loop by Preisach model with the Neural Network, the method developed is based on an analysis of two distribution functions. The modified Lorentzian function and Gaussian function have been analyzed. The implemented software and performances of the distributions are presented.

Abstract: Due to the distortion is always accorded when the laser facular reaches the four-quadrant detection photosensitive surface of the pulsed laser tracking system, resulting in the decline of the target measuring accuracy. Therefore the effect of facular size and energy distribution for the target azimuth measurement accuracy is researched. The calculation method for laser pot offset is discussed, and the formulas and principle circuit are given. The effect analysis of the laser pot size for measuring linear region and sensitivity are carried out, when the laser spot shows a uniform distribution as well as the Gaussian distribution. The performance of the laser pots under the two energy distribution situations are compared, the corresponding simulation curves are obtained. The simulation results show that the linear region of the Gaussian spot is larger than the uniform spot, but its sensitivity is relatively lower than the uniform spot within the effective dynamic range.

Abstract: According to Hill’s thermodynamics theory for small system, the effect of small system on the type I antifreeze protein ‘HPLC-6’ is discussed in this article. We conclude that when the solution is very dilute, the effect is not visible, and as the concentration increases, the effect becomes more visible than before, and the result also shows that the number of molecules on the ice surface becomes larger when the effect of small system is considered.

Abstract: In this paper a cellular automata model is proposed to describe driver behavior at a single-lane roundabout. Two truncated Gaussian distributions are used to model heterogeneous and inconsistent driver behavior. Numerical results show that our method is feasible and valid.

Abstract: Considering spatial heterogeneity of LAI and nonlinearity of its inversion model, a new spatial scaling method based on Gaussian distribution theory was proposed, aiming to quantitatively analyze scale effects and reveal scaling rules. In this method, higher spatial resolution data, obeying Gaussian distribution when the volume is large enough, were taken as baseline. Statistical parameters and Gaussian distribution forms were integrated into the process of spatial scaling. Barley was selected as experimental object. Firstly, multi-resolution data at 10m, 15m, 20m and 30m were constructed based on 5m resolution data through up-scaling algorithms. Secondly, extracting statistics of these data and constructing scaling data based on Gaussian distribution theory. Finally, quantitatively analyzing scale effects of LAI by introducing sensitive bands ordering and multivariate linear regression. Number of effective bands, R2 of observations and estimations, and MRA could fully confirm the feasibility and validity of this proposed method.

Abstract: Sintering temperature influences on sintering process essentially. Laser power and scanning speed determined sintering temperature. Preheat to powder is beneficial to improve the surface temperature uniformity. Stress concentration of part is various in different layer depth. Therefore, the influence of laser power, scanning speed, thickness of spreading layer and preheat temperature on part quality in selective laser sintering(SLS) are main factors. Based on laser energy in Gaussian distribution and mechanism of SLS, with manufactured specimen pieces by molding machine AFS-450, orthogonal experimental design and analysis of variance are adopted to post-treatment. The prototyping sintering parameters are optimized. The result and solution of the experiment are the preheat temperature of 100°C, the scanning speed of 2000 mm/s, the laser power of 24W, the thickness of single layer of 0.2mm for ABS resin. This work can provide optimized parameters in SLS for ABS resin. It will be of benefit to improve the part dimensional precision and strength.

Abstract: We implement the concept of the distributed electrode, which allows to improve the modal behavior of lasers and to reduce spatial-hole burning effects by preferentially localizing current injection in the center of the structure, hence discriminating the optical mode. We report the first realization of distributed electrode lasers emitting at 808 nm with the measured full width at half maximum (FWHM) angle of the minimal horizontal angle as 3.8° while the maximum continuous-wave output power is up to 4 W and high slope efficiencies as high as 0.95 W/A.