Analysis the influence of pulse-to-pulse stability of modulator on high-power microwave output of pulsed klystron

X-ray free electron laser (XFEL) facility based on electron linear accelerator (LINAC) is regarded as one kind of the fourth-generation light source with the characteristics of high intensity, exceptional brightness, ultrashort pulse duration, and spatial coherence. In electron linear accelerator, energy of beam bunches is provided by high-power electromagnetic microwaves which are generated by a microwave tube called klystron. The stability of beam voltage of klystron occupies a key position in both the stability of output RF (Radio Frequency) power and the jitter of output RF phase, furthermore, it plays an extremely important role in beam energy stability of electron linear accelerator. In this paper, high power RF fluctuation and phase jitter of klystron output caused by beam voltage instability of klystron are analyzed and calculated. Influence of klystron beam voltage instability on beam energy gain in linear accelerator have also been further analyzed and calculated. The calculating procedure is particularly valuable for us to understand the relationship between pulse modulator stability and beam energy gain fluctuations. Finally, relevant experimental results measured by Shanghai Soft X-ray Free Electron Laser Test Facility (SXFEL-TF) is presented.

FTT-NAS: Discovering Fault-tolerant Convolutional Neural Architecture

With the fast evolvement of embedded deep-learning computing systems, applications powered by deep learning are moving from the cloud to the edge. When deploying neural networks (NNs) onto the devices under complex environments, there are various types of possible faults: soft errors caused by cosmic radiation and radioactive impurities, voltage instability, aging, temperature variations, malicious attackers, and so on. Thus, the safety risk of deploying NNs is now drawing much attention. In this article, after the analysis of the possible faults in various types of NN accelerators, we formalize and implement various fault models from the algorithmic perspective. We propose Fault-Tolerant Neural Architecture Search (FT-NAS) to automatically discover convolutional neural network (CNN) architectures that are reliable to various faults in nowadays devices. Then, we incorporate fault-tolerant training (FTT) in the search process to achieve better results, which is referred to as FTT-NAS. Experiments on CIFAR-10 show that the discovered architectures outperform other manually designed baseline architectures significantly, with comparable or fewer floating-point operations (FLOPs) and parameters. Specifically, with the same fault settings, F-FTT-Net discovered under the feature fault model achieves an accuracy of 86.2% (VS. 68.1% achieved by MobileNet-V2), and W-FTT-Net discovered under the weight fault model achieves an accuracy of 69.6% (VS. 60.8% achieved by ResNet-18). By inspecting the discovered architectures, we find that the operation primitives, the weight quantization range, the capacity of the model, and the connection pattern have influences on the fault resilience capability of NN models.

Validation of Voltage Instability Detection and Control Using a Real Power System Incident

This paper applies a voltage instability monitoring method based on voltage and current measurements from a transmission bus PMU on the Hellenic Interconnected System using both unstable and marginally stable scenarios, derived from the historical 12 July 2004 blackout of the Athens area. The effectiveness, selectivity and reliability of the proposed monitoring method is clearly demonstrated, allowing its integration into a System Protection Scheme with direct load shedding. It is shown that the proposed instability detection and control scheme could have prevented the voltage collapse if applied at the time of the event.

Alternative Criteria of Voltage Stability Margin for the Purpose of Load Shedding

Voltage instability takes on the form of a dramatic drop of transmission system voltages, which may lead to system disruption. During the past two decades it has become a major threat for the operation of many systems and, in the prevailing open access environment, it is a factor leading to limit power transfers. The objective of this paper is to present new method of under voltage protection with maximum utilization of system capabilities.

Optimal Shedding Against Voltage Collapse Based on Genetic Algorithm

The prevalent tendency in power transmission systems is to operate closer and closer to the energy limit, rendering system voltage instability a commonly widespread phenomenon. It is, therefore, necessary that certain remedial corrective controls need be undertaken whenever these systems tend towards failure. In this respect, load shedding stands as a major correction mechanism and such a failure can be prevented and nominal system voltage can be resumed. It is worth noting however that load shedding must be implemented very carefully to ensure the satisfaction of both the customer and the electricity-production company. In this context, our focus of interest is laid on load and machine shedding against voltage collapse as an effective corrective method. It is important to note that such a problem turns out to be commonly defined as an optimization problem under constraints. Using genetic algorithms as resolution methods, the application of the proposed methods was implemented on the 14-node IEEE test network, while considering a number of different case studies.