China
Africa
Explore chapters and articles related to this topic
Programming Techniques for Finite Element Analyses
Published in Jie Shen, Radhey Lal Kushwaha, Soil-Machine Interactions, 2017
One former approach to solve this problem is to use the overlay technique by which it is the programmer’s responsibility to divide programs into pieces and then to identify the pieces that are mutually exclusive. These mutually exclusive groups of pieces constitute overlays that are loaded and unloaded under the control of user program during execution. The programmer has to ensure that the program never tries to access an overlay not loaded, and the overlays loaded is not larger than the size of memory. The shortcoming associated to this approach is that controlling the overlay process adds a substantial burden for programmers.
O
Published in Phillip A. Laplante, Dictionary of Computer Science, Engineering, and Technology, 2017
overlay (1) a mechanism by which sections of code are explicitly loaded into memory when required, replacing code which was already present. Although largely supplanted by virtual memory mechanisms, overlay techniques are still used on some embedded machines, real-time systems, and other specialized applications. Many languages for these machines have explicit support for overlay management.
Variable frequency drive harmonics and interharmonics exciting axle torsional vibration resulting in railway wheel polygonisation
Published in Vehicle System Dynamics, 2020
The interharmonics supplied by the VFD present a second excitation mechanism that is capable of causing torsional vibration of the wheelset. The locations of the harmonics and interharmonics predicted with Equations (1) and (2) are shown in Figure 10(a). The measured energy supplied to the traction motor is shown in Figure 10(b). Figure 10(c) overlays the predicted and measured harmonics and interharmonics. It can be seen that many large amplitude, well-defined interharmonics were measured and some of them correspond to the computed interharmonics. Many of these interharmonics are seen to originate at 200 Hz. The interharmonics decrease their frequency with an increase in vehicle speed in the first 34 s of running before the first change in VFD switching pattern. The interharmonics are less visible between 34 and 50 s between the first and third change in switching pattern. The interharmonics in this region resemble a continuous spectrum similar to white noise. The time from 50 to 100 s between the third and fourth change in switching pattern is shown by a rectangle and many distinctive interharmonics are visible in this region. The interharmonics predicted with Equation (2) that corresponded to the measured interharmonics in this region were identified. The interharmonic that was measured near the frequency of the coupled mode was identified from Figure 10 as the interharmonic. The interference of the interharmonic was predicted in the Campbell diagram of Figure 5.
Re-configurable, expandable, and cost-effective heterogeneous FPGA cluster approach for resource-constrained data analysis
Published in International Journal of Parallel, Emergent and Distributed Systems, 2022
Dulana Rupanetti, Hassan Salamy, Cheol-Hong Min, Kundan Nepal
The heterogeneous experiments consist of three experiments – MNIST handwritten image recognition as used in the homogeneous experiment, Linear Regression of COVID 19 dataset, and Stroke detection data to train a model to predict strokes. As the datasets used in this section are more complex and large in nature, we opted to use Linux as the base system on all FPGA boards, giving us the opportunity to use Python language to process the data. While Linux is used as the operating system, ML-IP blocks were still implemented similar to the homogeneous experiment and the functionality of the cluster did not change except for the new high-level layers of software added to simplify the implementation. Using Xilinx's Vivado suite and petalinux packages, we generated Linux images using PYNQ files available for both ARTY A7 and ZedBoard. For PYNQ-Z1, being an opensource development board for python productivity, the only requirement was to install the default PYNQ image on the board. With Linux as the operating system, accessing the hardware structure to program the logic cells requires software overlays that map python code into FPGA blocks for acceleration through bitstreams. For basic packages used in python, the bistreams are available in the default PYNQ distribution. However, for the work in this paper, we utilised a revised version of Scikit-Learn [25] named Pynq-Sklearn [26], a library designed by leveraging the PYNQ hardware libraries, which provides a rich environment for data modelling and analysis. With few modifications to the standard bit-streams, Pynq-Sklearn was able to train and test datasets in all the heterogeneous experiments in this paper. Graphs for the results were generated through Python's Matplotlib [27] which was executed after the data was collected and did not have any effect on the run-time and the execution of the cluster.
Wear behaviour of hardfacing ultra carbide steel grades
Published in Surface Engineering, 2020
Three hardfacing grades of hyper-eutectic and hypo-eutectic Cr3C2 overlays on steel substrates were investigated. The samples showed high wear resistance and good mechanical performance owing to the presence of primary carbides (M7C3) and formation of complex carbides (MO2C) in the eutectic austenite matrix. This resulted in an increase in wear resistance by approximately 14 times in comparison to that of carbon steels, and by 7 times in comparison to that of quenched and tempered alloyed steels.