Inter-shot Analysis Research Articles

Simulation of neutron emission in neutral beam injection heated plasmas with the real-time code RABBIT

In plasmas heated with deuterium beams a deficit of the expected fusion neutron rate is an indicator of the deterioration of the fast-ion confinement, caused, for instance, by magnetohydrodynamic instabilities. The capability of predicting this deficit during the discharge relies on the availability of real-time estimates of the neutron rate from NBI codes which must be fast and accurate at the same time. Therefore, the recently developed real-time RABBIT code for neutral beam injection (NBI) simulations has been extended to output the distribution function and calculate the neutron emission. After the description of this newly installed diagnostics in RABBIT, benchmarks with NUBEAM, a massively used and validated Monte Carlo NBI solver, are discussed on ASDEX-Upgrade and JET cases. A first application for control-room intershot analysis on DIII-D is presented, and the results are compared on a large database with a slower NUBEAM analysis. Further application possibilities, e.g. for real-time control of Alfvén eigenmodes, are outlined.

Improved framework for the maintenance of the JET intershot analysis chain

At the JET experiment data from routine diagnostics is analysed automatically by a suite of codes within minutes after operation. The maintenance of these interdependent codes and the provision of a consistent state of the physics database over many experimental campaigns against a backdrop of continuous hardware and software updates, requires well defined maintenance and validation procedures. In this paper, the development of a new generation of maintenance tools using distributed version control and a work-flow following the principle of continuous integration [1] is described.

The JET Intershot Analysis: Current infrastructure and future plans

The JET Intershot Analysis (Chain1) generates processed data following a pulse. Maintaining the pulse repetition rate is one of JET's key success factors, so performance of Chain1 is crucial. This paper will describe JET's experience of managing Chain1, including a description of the control system used to ensure the analysis chain runs as quickly as possible, and a discussion of JET's experience of integrating externally developed codes into a standard analysis framework. The current Chain1 infrastructure was developed in 1999 and although reliable and efficient is starting to prove costly in terms of flexibility and extensibility to meet JET's current and future needs. For this reason JET is planning to re-implement the Chain1 system. The paper will outline the work done towards this aim, and present a model of the proposed new system. Finally, possible future steps towards an integrated data production chain for JET will be discussed, and the potential applicability to next generation fusion devices will be outlined.

Database techniques at TEXTOR as a store for evaluated data

During the last years the need for a flexible handling of evaluated and validated data at TEXTOR was recognized. Therefore, a commercial relational database has been installed, which stores the analyzed data and makes it accessible to the users. Each signal is stored in one to three tables, depending on the dimension of the signal. The TEXTOR physics database (TPD) allows for a version management. The data retrieval is based on the HTTP protocol and is embedded within the retrieval architecture of TEXTOR. The database allows queries with constraints on the signals, which can be combined with the pulse-oriented logbooks of the physicist- and engineer-in-charge. This helps in identifying pulses with special events, e.g. disruptions. An automatic inter-shot analysis of the performed plasma pulse was brought into operation. It allows online as well as off-line storage of analyzed data in the database. It is very modular and based on a script that calls user written subroutines for the data analysis. In addition to the inter-shot analysis, diagnosticians can store evaluated data via HTTP based procedures.

A generic remote method invocation for intensive data processing

Based on the Extensible Markup Language (XML) and the Remote Method Invocation (RMI) standards, a client/server remote data analysis application has been developed for intensive data processing. This GRID oriented philosophy allows a powerful tool to maintain updated code and centralized computational resources. Another major feature is the ability to share proprietary algorithms in remote computers without the need of local code and libraries installation and maintenance. The 16 CPU Orionte cluster in operation at Centro de Fusão Nuclear (CFN) is currently used to provide remote data analysis. The codes running in languages such as Octave, C, Fortran or IDL are called through a script remote invocation and data is released to the client as soon as available. The remote calculations parameters are described in an XML file containing the configuration for the server runtime environment. Since the execution is made by calling a script any program can be launched to perform the analysis, the only requirement is the implementation of the protocol described in XML. Some plasma properties of the CFN tokamak (ISTTOK) that require heavy computational resources are already obtained using this approach, allowing ready inter-shot analysis and parameterization decisions.

Scene extraction in motion pictures

This paper addresses the challenge of bridging the semantic gap between the rich meaning users desire when they query to locate and browse media and the shallowness of media descriptions that can be computed in today's content management systems. To facilitate high-level semantics-based content annotation and interpretation, we tackle the problem of automatic decomposition of motion pictures into meaningful story units, namely scenes. Since a scene is a complicated and subjective concept, we first propose guidelines from film production to determine when a scene change occurs. We then investigate different rules and conventions followed as part of film grammar that would guide and shape an algorithmic solution for determining a scene. Two different techniques using intershot analysis are proposed as solutions in this paper. In addition, we present different refinement mechanisms, such as film-punctuation detection founded on film grammar, to further improve the results. These refinement techniques demonstrate significant improvements in overall performance. Furthermore, we analyze errors in the context of film-production techniques, which offer useful insights into the limitations of our method.

Automatic analysis of JET charge exchange spectra using neural networks

The analysis of charge exchange recombination spectra represents a very challenging problem due to the presence of many overlapping spectral lines. Conventional approaches are based on iterative least-squares optimization and suffer from the two difficulties of low speed and the need for a good initial approximation to the solution. This latter problem necessitates considerable human supervision of the analysis procedure. The authors show how neural network techniques allow charge exchange data to be analysed very rapidly, to give an approximate solution without the need for supervision. The network approach is well suited to the fast intershot analysis of large volumes of data, and can readily be implemented in dedicated hardware for real-time applications. The neural network can also be used to provide the initial guess for the standard least-squares algorithm when high accuracy is required.

Application of function parametrization to the analysis of polarimetry and interferometry data in TEXTOR

Function parametrization (FP) provides a way to perform complex data analysis in a fast and reliable manner that allows inter-shot analysis. The method has been used to analyse polarimetry and interferometry data in TEXTOR with the purpose of obtaining spatial distributions of the electron density and the toroidal plasma current. A standard TEXTOR discharge is investigated and the results from FP are compared with results obtained by conventional methods of data analysis. Agreement between the two methods is generally good. For the central safety factor the value of q0 = 0.8 ± 0.1 is in accordance with previous calculations. The new method can easily be modified to incorporate additional data.