Search

A new database system for the JT-60SA tokamak has been developed. This system meets the following requirements: to hold all the JT-60SA data and prevent loss of data, to allow researchers to use data without knowing internal database configuration and structure, to manage data revision so that stored data can be traced back to the older revision anytime. Depending on the timing of data transfer for storage or reference, the experiment data are categorized into three types: (i) discharge data, (ii) plant monitoring data, and (iii) unprocessed data. In the first step in constructing a new database system, we have designed the basic structure of the JT-60SA database system to manage all three data types. The dedicated tools for data transfer or retrieval prevent unregistered computers from accessing the experiment data. The RAID6 configuration of a hard disk drive (HDD) and the magnetic tapes improve data reliability. This paper reports details about the functions of the new JT-60SA database system. Furthermore, the elapsed time of data transfer and storage is measured to evaluate the performance of the JT-60SA database system. It takes 20 s to transfer and store 6 GB of data, which is enough to meet the required performance.

The assembly scenarios and assembly tools of the major tokamak components for JT-60SA are studied in the following. (1) The assembly frame (with a dedicated 30-tonne crane), which is located around the JT-60SA tokamak, is adopted for effective assembly works in the torus hall and the temporary support of the components during assembly. (2) Metrology for precise positioning of the components is also studied by defining the metrology points on the components. (3) The sector segmentation for weld joints and positioning of the vacuum vessel (VV), the assembly scenario and tools for VV thermal shield (TS), the connection of the outer intercoil structure (OIS) and the installation of the final toroidal field coil (TFC) are studied, as typical examples of the assembly scenarios and tools for JT-60SA.

On JT-60U, two mass data acquisition systems have been acquiring 91 channels of fast-sampling diagnostic data. They are, however, based on rather outdated bus standards and now difficult to update. Therefore we have developed a new mass data acquisition system that has more flexibility and ability to perform a faster data acquisition based on a modern bus standard. The system consists of PXI-based (PCI extensions for instrumentation) digitizers connected to controller PCs, and a server for collecting and processing data from these PCs and transferring them to the main diagnostic computer. Up to now, we have verified that the system acquires 27 channels of data that amounts to 2.2 GB in total, from 3 controller PCs, at a sampling rate of 1 MS/s in a 45 s discharge.

The design activity of JT-60SA (JT-60 Super Advanced) addresses the remodeling of JT-60 to a superconducting tokamak device under the JA-EU collaborative ITER–BA project. For the JT-60SA control system, the existing system should be reused as much as possible from the viewpoint of cost-effectiveness. Initial considerations of the configuration of the advanced supervisory control system (SVCS) including the following systems are: (1) an ultimate flexibility real-time control system, (2) a precision timing system able to diagnose a plasma event having a time range of ∼1 μs, and (3) a safety shutdown control system. In this report, we present the design study of the JT-60SA SVCS with the focus on the above systems.

Since remote participation in ITER experiments is planned, it is expected to demonstrate that the JT-60SA experiment is controlled from a Japanese remote experiment center located in Rokkasho-mura, Aomori-ken, Japan as a part of the ITER-BA project. Functions required for this experiment are monitoring of the discharge sequence status, handling of the discharge parameter, checking of experiment data, and monitoring of plant data, all of which are included in the existing JT-60 Man–Machine Interfacing System (MMIF). The MMIF is now only available to on-site users at the Naka site due to network safety. The motivation for remote MMIF is prompted by the issue of developing and achieving compatibility with network safety. The Java language has been chosen to implement this task. This paper deals with details of the JT-60 MMIF for the remote experiment that has evolved using the Java language.

In JT-60U, a mainframe computer called inter-shot processor (MSP-ISP) had been supervising the data acquisition of diagnostic subsystems, managing the storage of obtained data, and establishing communication with the JT-60U's discharge control system. But due to the increased demand for advanced diagnostics, such as improving spatial and temporal resolutions or extending the data acquisition period, the old ISP had been becoming insufficient in power and less cost effective. To resolve this problem, we have gradually migrated ISP to a new UNIX server (UNIX-ISP), and thereby carried out the decentralization of the old ISP's functionalities. Also we are developing a new mass data acquisition system (MDAS) that has simple architecture but has greater performance and flexibility. Simultaneously, to handle the increased volume of data, we are renewing the data storage system. The renewed data processing system (DPS) is now in operation. The performance of the new ISP and the newly developed MDAS is presented.

JT-60 data processing system (DPS) possesses three-level hierarchy. At the top level of hierarchy is JT-60 inter-shot processor (MSP-ISP), which is a mainframe computer, provides communication with the JT-60 supervisory control system and supervises the internal communication inside the DPS. The middle level of hierarchy has minicomputers and the bottom level of hierarchy has individual diagnostic subsystems, which consist of the CAMAC and VME modules. To meet the demand for advanced diagnostics, the DPS has been progressed in stages from a three-level hierarchy system, which was dependent on the processing power of the MSP-ISP, to a two-level hierarchy system, which is decentralized data processing system (New-DPS) by utilizing the UNIX-based workstations and network technology. At present, replacing the MSP-ISP with a UNIX-based workstation is under way as a final phase in building the New-DPS project. This replacement will be accomplished by October 2005. In addition, remote experiment in JT-60 is considered to be an important issue in extension of the collaborative research among the nuclear fusion community. Technical environments for this task have been developed. The participant from a remote site can set discharge parameters with assistance of the dedicated machine operating status-information tools.

The large tokamak fusion device JT-60 is expected to explore more advanced tokamak discharge scenarios toward the ITER and a future power reactor. Since various experimental issues are to be adequately discussed, and possibly to be solved in JT-60, the plasma real-time control system has been drastically improved with remodelling in hardware as well as in software. To satisfy the requirements, a “multiple networks” structure is employed as a basic principle of the remodelling. Distributed processors for diagnostics, actuators, and supervisory controllers are linked through a reflective memory (RM) network for fast, real-time communication. Similarly, advanced and complex calculations to reproduce plasma shape and profiles are performed by several processors connected to the same RM network. In this report, we discuss the developments to improve the JT-60 plasma control and data processing systems. In addition, a future plasma control system leading to a standard design for a power reactor is outlined, which is based on the 20-year plasma operation experience.

In the JT-60 tokamak at JAERI, an environment for remote participation is planning to be developed by concentrating ideas of experts of nuclear fusion research from research organizations and universities distributed throughout the country. We are constructing a hierarchical collaborative remote research system, which consists of remote experiment, remote analysis and remote diagnostic. In remote collaboration, it is important to share information, atmosphere and presence between the participants as well as to maintain the security of the system. To implement a collaborative environment, a video conferencing system, a video streaming system and remote analysis system called “VizAnalysis” have been developed for the communication. To assist the remote participation for the fusion research, the document management system based on the web browser called “VizSquare” has been developed. In JT-60, security and authentication methods utilizing a computer network and a new communication tool have been developed, and they are applicable to the remote participation of ITER.

Data processing system for JT-60 is constructed by mini-computers, microcomputers and CAMACs for the control and the data acquisition of plasma diagnostics. However, it is difficult to apply this system to the remote diagnostics using the internet, because the old microcomputers used for over 15 years has no internet connectivity. Therefore, the data processing subsystem using mini-computers and microcomputers should be replaced, keeping the consistency with the main computer. The workstation based on UNIX-OS is a useful candidate for replacing those computers, since it is excellent in the network communication, the cost and the stability. We therefore the replaced microcomputers controlled and monitored the diagnostic CAMAC system, and executed its data acquisition. The new workstation-based system was simple system, which consisted of the serial highway driver module for controlling CAMAC and CPU module with SPARC CPU in VMEbus. By these replacements, the difficulty in the remote control has been solved by using the network and screen control of X Window. The maintainability is also improved and the function enhancement becomes easier. In this paper, basic design for replacing mini-computer and microcomputer with the UNIX workstation will be described.

The JT-60 data processing system is a large computer complex, gradually being modernized by utilizing progressing computer and network technologies. There were three major changes in our system recently. A VME-based fast data acquisition system (FDS) has been developed to enlarge a transient mass-data storage system (TMDS). It can acquire the data of 6 MB per channel with 1 or 5 μs sampling. Now there are three FDSs with 29 channels. The TMDS, which was composed of a mini-computer and 61 channels of 6 MB memory modules with 5-μs sampling, has been replaced with a new system based on the technology of the VME-based fast data acquisition system. One of the outdated auxiliary crate controllers with micro-processor (ACM), which controls CAMAC modules in the crate on the serial highway, acquires data from them and transfers data to the main computer, has been replaced by a workstation with a VMEbus byte-serial highway driver. The combination of a workstation and a VMEbus byte-serial highway driver has been developed and used to replace many mini-computers in our system. A workstation with a digital linear tape (DLT) library has been introduced as final data storage. The conversion of a few terabytes of raw data tapes for a cartridge tape library (CTL), which is attached to a mainframe computer, Fujitsu GS8300, will be started by using data conversion software prepared to handle mainframe data on the workstation.

Real time processor, RTP, is a basic subsystem in the JT-60 data processing system and plays an important role in JT-60 feedback control utilizing various diagnostic signals. An original mini-computer system was replaced in 1994 with a workstation. It has 2 CPU's with a real-time parallel UNIX. During the experiment, one CPU communicates with the main computer synchronized with the experimental sequence, and the other executes a real-time process with 1 ms clock. RTP has been upgraded to a compatible RISC workstation with a faster A/D converter in 1997 in response to increased requests of handling the increased signals in more complicated manners. After this upgrade, an elapsed time is reduced to ∼75% and an I/O time is now dominant. To reduce a CAMAC access time, we try to use a reflective memory for the data transfer. Now RTP handles six kinds of signals to control an electron density, a neutron emission, an electron density and a radiation in the divertor region, and a neutral pressure. RTP will handle five more kinds of signals this year and be used for an advanced control with more complicated processing.

The JT-60 data processing system is a large computer complex and gradually modernized by utilizing progressing computer and network technology. There are two major changes in our system. A main computer of FACOM M-780 has been replaced with compatible GS8300 using state-of-art CMOS technology, which results in lower power and space usage with nearly the same performance. Now it can handle ∼500 MB of data per discharge. A gigabit ethernet switch with FDDI ports has been introduced to cope with the increase of handling data. The switch will connect a tera-byte (TB) data server at the bandwidth of a gigabit per second with the main computer and many data acquisition workstations. Other developments in our system are the realization of three workstation-based plans, the TB data server, the VME-based fast data acquisition system and a CICU. The TB data server is basically a UNIX workstation with ∼100 GB RAID disks and ∼900 GB MO auto-exchangers. The VME-based fast data acquisition system has been developed to enlarge the present TMDS. The CICU, which has a function of interfacing the main computer with the CAMAC system, has been replaced with the workstation-based system after the fine tuning.

The JT-60 data processing system was originally a large computer complex system including a lot of micro-computers, several mini-computers, and a mainframe computer. Recently, several improvements have been made to the original system to modernize the system. Many sub-systems composed of aged mini-computers have been replaced with workstations by utilizing recent progress in computer and network technologies. The system can handle ∼300 MB of raw data per discharge, which is three times larger than the amount in the original system. These improvements have been applied to develop element technologies necessary to the remote participation in JT-60 experiments. A remote diagnostics control and monitoring system and a computer system for access to JT-60 data from the Internet are used together with video conferencing systems for a real-time communication. In 1996, the remote participation based on them was successfully demonstrated in collaboration with Japan Atomic Energy Research Institute, Los Alamos National Laboratory, and Princeton Plasma Physics Laboratory.

JT-60U's diagnostic systems have been using CAMAC systems for their control and data acquisition. But the transient recorders used to store the diagnostic data are rather old and lack in memory expansion. Therefore, for a long pulse discharge of more than 15 s, data sampling rate must be reduced to cover the whole discharge period. The acquired data has, however, low time resolution and is sometimes insufficient for detailed analysis. To solve this problem, the transient recorders should be replaced with those having larger memory, but without modification to the existing data processing infrastructure. For this purpose, we have developed a new data acquisition system with SPARC CPU on VMEbus and the associated software, to retain the functionality of CAMAC control. We have applied this system to the existing bolometric diagnostics. In this paper, detail of design will be described.

The sustainment of radiative divertor plasma in JT-60U is performed by the feedback control of the gas puffing rate into the divertor region with an actuator of the divertor radiation power. The feedback control system switches the piezoelectric gas valve every 10 ms, so that the proper amount of gas flux is fed for divertor radiation power to follow the reference waveform. The divertor radiation of 6 MW, which is about 40% of the heating power, is sustained for about 3 s by the hydrogen gas puffing with feedback control. The present logic of the feedback system, however, is not sufficient to fully control the divertor radiation which is modified by the change in heatings and plasma parameters. In order to prevent the loss of control, multi actuators for the feedback control, for example, the involvement of the heating power, and the selection of observed bolometer channels may be required.

Based on the design requirements for the plasma facing components in ITER/FER, we have performed design studies on the conductively cooled first wall armor and the divertor plate with sliding supports. The full-scale armor tiles were fabricated for heat load tests, and good thermal performances were obtained in heat load tests of 0.2–0.4 MW/m 2 . It is shown by the thermomechanical analysis on the divertor plate that thermal stresses and bending deformation are reduced significantly by using the sliding supports. The divertor test module with the sliding supports has been fabricated to investigate its fabricability and to verify the functions of the sliding supports during a high heat load of about 10 MW/m 2 .