Search

Lifetime estimation of power semiconductor devices have been widely investigated to improve the reliability and reduce the cost of maintenance of power converters. However in most reported work, the aging effect is not considered in the lifetime evaluation process due to the omission or limitation of thermal cycle counting method. Additionally, the electrical/thermal simulation and lifetime estimation are usually implemented in different simulators/platforms, for the same reason. Thus, to tackle these problems, a concise but comprehensive MOSFET model that enables electro-thermal modeling, aging and lifetime estimation on LTspice® circuit simulator is proposed in this paper. The idea comes from the fact that, MOSFET on-state resistance Rds,on is not only temperature dependent, but also widely accepted as the device failure precursor. In other words, as it carries critical information about instantaneous temperature and aging progress. Hence, co-simulation can be achieved by constructing electrical, thermal, and aging and lifetime sub-modules exclusively first, and using Rds,on, to build linkages among them. Averaged modeling technique is adopted due to the ease of establishing links among these three sub-modules, and fast simulation speed as compared to a switched converter model. Behavioral models are employed to realize the thermal cycles counting, stress accumulation and degradation evaluation. This paper demonstrates that it is possible to use a single simulation software to monitor performances of devices and circuits, and their lifetime estimation simultaneously. High-stress thermal cycling and long-term random mission profiles are applied to verify the correctness of the model and to mimic a 10-year load respectively. An accelerated aging trend can be observed in the long-term mission profile simulation, which is in agreement with the theory. Facilitated by the employment of averaged circuits, the proposed method is a good simulation/analytical tool to implement a long-term mission profile that requires reliability assessment.

Zero-sequence injection (ZSI) techniques are widely adopted in Wye- and Delta-connected cascaded H-bridge (CHB) three-phase systems to cope with the unbalanced power generation from each phase. Recent studies have shown that a ZSI can allow a $\triangle-$connected CHB converter to cope with a large range of power imbalances among phases when compared to a standard Y-connected CHB converter. The superiority in the electrical performance under unbalanced input power has been well investigated, however the comparative thermal performance for both configurations under same power imbalance condition has not been investigated. In this work, the thermal performance of a CHB converter connected in both $\triangle-$configuration and Yconfiguration operating under an unbalanced power condition is studied. For this, the enquired ZSI for both $\triangle-$ and Y-connected CHB converters is firstly analyzed. Then, the impact of ZSI on the thermal performance and lifetime expectancy of the power switches in each phase is evaluated and compared for both configurations. The thermal analysis carried out in this work shows that even though a $\triangle-$connected CHB converter offer a wider power imbalance operation than Y-connected counterpart from the electrical viewpoint, this is achieved by thermally stressing the power switches in each phase in an uneven manner, and in some case overstressing them. Consequently, there is a trade-off between electrical performance and thermal stress when dealing with power imbalances in $\triangle$-connected and Y-connected CHB converters.

Free spatial optical (FSO) communication has been widely studied in recent years for its fast data rate, ease of implementation and low cost. Due to the small divergence angle of lasers, the technology of acquisition, tracking and pointing is particularly important. This study focuses on the design of tracking systems for FSO communications and inter-satellite laser communications. We build two models in time domain and spatial domain to characterize the tracking problem. On this basis, we design a tracking algorithm based on Gaussian quadrics fitting to estimate the orientation of the target, where less feedback is needed. Compared with current tracking algorithms, this algorithm can greatly improve the tracking speed of the system. Simulation results show that the tracking accuracy highly depends on width, number, location and separation of laser beams. The results also show that the FSO tracking system can still be efficient and reliable in a noisy environment.

Web service recommendation based on the quality of service (QoS) is important for users to find the exact Web service among many functionally similar Web services. Although service recommendations have been recently studied, the performance of the existing ones is unsatisfactory because: 1) the current QoS predicting algorithms still experience data sparsity and cannot predict the QoS values accurately and 2) the previous approaches fail to consider the QoS variance according to the users and services' locations carefully. A Web service recommendation method based on the QoS prediction and hierarchical tensor decomposition is proposed in this paper. The method is called QoSHTD that is based on location clustering and hierarchical tensor decomposition. First, the users and services of the QoSHTD cluster into several local groups based on their location and models local and global triadic tensors for the user-service-time relationship. The hierarchical tensor decomposition is then performed on the local and global triadic tensors. Finally, the predicted QoS value through local and global tensor decomposition is combined as the missing QoS values. The comprehensive experiment shows that the proposed method achieves a high prediction accuracy and recommending quality of Web service, and can partially address data sparsity.

Due to the ongoing pursuit of high power density power supplies, thermal management becomes one of the most critical aspects to consider during the design phase for their stable and efficient operation. There are some reported electro-thermal models (ETMs) of power semiconductors and passive components, which help to estimate their electrical performance and temperature simultaneously. However, with the increase of the switching frequency for higher power density design, the simulation time increases accordingly. This can be effectively solved by adopting an averaged model of the converter. As the conventional average model is neither frequency nor temperature dependent with which are the two key parameters in ETM, modification is needed. Therefore, a modified electro-thermal average model (ETAM) of a boost converter considering self-heating phenomenon of all devices is proposed in this paper. This is achieved by a) adding additional behavior models to calculate the device losses; b) replacing the fixed resistance of each component with a temperature dependent one; c) using variable inductor and capacitor instead of a fixed value counterpart to obtain an accurate electrical model and precise losses estimation, and d) forming electrical and temperature feedback loops for each component. The advantages of the proposed research work are fast simulation speed, fairly accurate temperature prediction, and ease of implementation. Both the simulation and experimental results are given and compared to verify the proposed solution.

Thermal management is one of the most critical aspects of the design of high power density converters. Recently, significant efforts and progress have been made in developing electro-thermal models for power semiconductor devices as they are very sensitive to temperature changes. Passive components such as inductors and capacitors are also investigated, since they are temperature-dependent. However, most published work focuses on electro-thermal model of a single power device or a module only instead of a whole converter, which is more realistic in terms of converter design. Hence, in this work, a datasheet informed electro-thermal model is proposed for a boost converter. It is achieved by adding additional behavior models to the existing electrical model of each power device to reflect the temperature incurred electrical behavioral change. Loss model and RC network are used to estimate the temperature change. This forms a power loss and temperature feedback loop. The advantages of this work are ease of integration with existing electrical models in the SPICE library, and elimination of the complicated physical properties of the power devices, but fully utilizes the device datasheet information and mathematical method.

A dynamically tunable polarization converter has been proposed in this paper, which has a multi-layer structure consisting of an orthogonal cross metallic, graphene strips, spacer substrate and a gold back ground layer. In this paper, the proposed polarization converter can transform the linear polarization incident waves into its cross polarization in reflection mode in terahertz (THz) frequency within an ultra-wideband, and the polarization conversion ratio (PCR) of the proposed structure has over 97% from 4.1 THz to 7.1 THz. Furthermore, by adjusting the chemical potential (or Fermi energy) of the graphene strips, the operating frequency band can be changed dynamically.

Because of an ability of modelling context information, Recurrent Neural Networks (RNNs) or bi-directional RNNs (BRNNs) have been used for beat tracking with good performance. However, there are two problems associated with RNN-based beat tracking. The first problem is the imbalanced data: usually only around 2% frames are labelled as ‘beat’. The second one is the disagreement on the precise positions of beats in human annotations or the delay of annotations caused by human tapping. In order to tackle these problems, we propose to convolve the original ground truth with a Gaussian kernel as the target output of the network for a more robust training. We conduct a comparison experiment using five different Gaussian kernels on five individual datasets. The results on the validation sets show that we can train a better or at least competitive model in a shorter time by using the convolved ground truth with a proper Gaussian kernel.

We propose a music structure analysis method that converts a path-enhanced self-similarity matrix (SSM) into a block-enhanced SSM using non-negative matrix factor 2-D deconvolution (NMF2D). With a non-negative constraint, the deconvolution intuitively corresponds to the repeated stripes in the path-enhanced SSM. Then the block-enhanced SSM is constructed without any clustering technique. We fuse block-enhanced SSMs obtained using different parameters, resulting in better and more robust results. Discussion shows that the proposed method can be a potential tool for analysing music structure at different scales.

In order to search channel of Higgs bosons in high energy physics (HEP), in this paper we propose Deep Extreme Feature Extraction (DEFE), a new ensemble multi-variate Analysis method (MVA). DEFE is a deep ensemble learning scheme that trains a strongly diverse set of neural feature learners without explicitly encouraging diversity and penalizing correlations, which is achieved by adopting an implicit neural controller that directly controls and distributes gradient flows from higher level deep prediction network. Based on the construction and approximation of the extreme selection region presented in this paper, the DEFE model is able to be trained efficiently, and extract discriminative features from multiple angles and dimensions, hence the improvement of the selection region of searching new particles in HEP can be achieved. With the application in this model, the selection regions full of signal process can be obtained through the training of a miniature collision events set. In comparison of the classic deep neural network (DNN), DEFE shows a state-of-the-art performance: the error rate has decreased by about 37%, the accuracy has broken through 90% for the first time along with the discovery significance which has reached a standard deviation of 6.0. Experiment shows that DEFE is able to train an ensemble of discriminative feature learners that boosts the over-performance of final prediction. Furthermore, among high-level features, there are still some important patterns that are unidentified by DNN and are independent from low-level features, while DEFE can identify these significant patterns more effectively and efficiently.

Prediction of Quality of Service (QoS) values plays an important role in service selection, discovery, and recommendation. Previous works show that the QoS values would be influenced by the location information. However, these researches do not consider the fact that some users may provide untrustworthy QoS values even though they are in the same location region. QoS values from these unreliable users could significantly affect the QoS prediction accuracy. To address this issue, this paper proposes an alternative and efficient approach to predict the missing QoS values, referred as the Location and Reputation aware Matrix Factorization based Location Information (LRMF). LRMF combines both the user's reputation and location information to achieve more accurate prediction results. Experiments are conducted on a real-world Web service QoS dataset, and results show that the proposed method outperforms many other existing QoS prediction methods.

Cross regulation of single-inductor multiple-output (SIMO) boost dc/dc converter is usually considered as a key factor that degrades the voltage regulation of individual outputs. Analysis of the cross coupling of a single-switch single-input dual-output (SIDO) boost converter operating in continuous conduction mode (CCM) reveals that it is an automatic current feedback mechanism during converter operation, which, however, can improve steady-state voltage regulation of both outputs. It is possible to use this natural cause to simplify the PWM controller design. That is, only one output needs to be sensed and regulated. The other output will be regulated accordingly. The drawback is the transient response of one output will share by another output Low-frequency model of a SIDO boost converter is derived and analyzed to reveal such phenomenon. A single Type-in PWM controller is designed for one of the outputs but is able obtain both regulated outputs. The analysis and design are verified by both simulation and experimental results.

© 2015 IEEE. Combining a storage device and renewable generation is one of the viable solutions to deal with the intermittency of renewable energy sources. Conventional methods of combining these components involve multistage and single-stage converters. However, there is a trade-off between efficiency and component count. A novel non-isolated three-port DC-DC converter which combines a dual-input single output (DISO) boost and a single-input single output (SISO) buck converters for a photovoltaic (PV)-battery-load system is proposed in this paper. There are five different operation modes to achieve different scenarios for maximum power point tracking, battery charging and load regulation requirements. Experiment results are given to confirm this proposed converter.

In this paper we improve piano note tracking using a Hid den Markov Model (HMM). We first transcribe piano music based on a non-negative matrix factorisation (NMF) method. For each note four templates are trained to represent the dif ferent stages of piano sounds: silence, attack, decay and re lease. Then a four-state HMM is employed to track notes on the gains of each pitch. We increase the likelihood of staying in silence for low pitches and set a minimum duration to reduce short false-positive notes. For quickly repeated notes, we allow the note state to transition from decay directly back to attack. The experiments tested on 30 piano pieces from the MAPS dataset shows promising results for both frame-wise and note-wise transcription.

We investigate piano acoustics and compare the theoretical temporal decay of individual partials to recordings of real-world piano notes from the RWC Music Database. We first describe the theory behind double decay and beats, known phenomena caused by the interaction between strings and soundboard. Then we fit the decay of the first 30 partials to a standard linear model and two physically-motivated non-linear models that take into account the coupling of strings and soundboard. We show that the use of non-linear models provides a better fit to the data. We use these estimated decay rates to parameterise the characteristic decay response (decay rates along frequencies) of the piano under investigation. The results also show that dynamics have no significant effect on the decay rate.

This paper presents the development and performance of a new generation of Direct Attach middle power 6V High Voltage Flip Chip (HVFC) LED designed to be directly attached using AuSn or SnAgCu solder by surface mount technology (SMT). Comparison is made against a standard 3V Flip Chip (FC) LED of equal chip dimension (780um × 380um). At 0.5W operation, both HVFC and FC have comparable Wall Plug Efficiency of 37% and Luminous Efficacy of 120lm/W at 6500K CCT, 70 CRI.