Laser in vessel viewing system for activated areas: mechanical design, manufacturing and tests

Laser in vessel viewing system for activated areas: mechanical design, manufacturing and tests

Fusion Engineering and Design 51 – 52 (2000) 1001 – 1006 www.elsevier.com/locate/fusengdes Laser in vessel viewing system for activated areas: mechan...

281KB Sizes 0 Downloads 25 Views

Fusion Engineering and Design 51 – 52 (2000) 1001 – 1006 www.elsevier.com/locate/fusengdes

Laser in vessel viewing system for activated areas: mechanical design, manufacturing and tests C. Talarico *, M. Baldarelli, A. Coletti, S. Lupini, C. Neri, M. Riva, L. Semeraro Associazione EURATOM-ENEA sulla Fusione, 45 Via Enrico Fermi, 00044 Frascati, Rome, Italy

Abstract A new laser based inspection system (LIVVS) has been developed by ENEA for periodic in-vessel inspection in large fusion machines. The design has been taking into account the radiation levels, operating temperatures and pressures forseen in large European fusion machines such as JET (Joint European Torus) and ITER (International Thermo-nuclear Experimental Reactor). Tests on the LIVVS mock-up are in progress in order to assess the mechanical, optical and electronic performances, and so to measure the system accuracy. Following a detailed mechanical description, preliminary results are reviewed and discussed. © 2000 Elsevier Science B.V. All rights reserved. Keywords: ; Charged coupled device camera; International thermonuclear experimental reactor

1. Introduction Tokamak machines need inspection systems to detect damages on the plasma facing components. The equipment currently in use for in vessel inspection is mainly based on charged coupled device (CCD) cameras, which can not withstand heavy radiation in vessel operating conditions. In order to match the in vessel inspection needs of international thermonuclear experimental reactor (ITER), a new laser based inspection device (laser in vessel viewing system, LIVVS) has been designed [3]. The system will be able to take in vessel * Corresponding author. Tel.: +39-6-94005338; fax: + 396-94005799. E-mail address: [email protected] (C. Talarico).

2-D/3-D pictures by means of a laser beam deflected on the target via a mechanical scanning probe. The probe, designed in compliance with ITER specification, will be tested at joint European torus (JET), replacing one of the existing in vessel scanning probes (IVIS) [1].

2. General description and working principle A 840 nm laser beam, generated in an optical module placed in control room, is routed into an optical fiber to a scanning mechanism inside the vacuum vessel. With respect to the launched beam, both intensity and phase shifting of the backscattered light are detected. On the basis of the signal intensity,

0920-3796/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 3 7 9 6 ( 0 0 ) 0 0 4 1 5 - 4

1002

C. Talarico et al. / Fusion Engineering and Design 51–52 (2000) 1001–1006

the target picture is reconstructed, whereas on the basis of the signal phase shifting, the target ranging can be performed. The system performances are strictly related to the scanning system accuracy. The LIVVS for JET includes a 65 kg, 4.5 m scanning probe. The probe supports the laserscanning prism, makes it rotate, and is equipped with two optical encoders to measure the laser beam angles. During plasma pulses, the probe is retracted from the vessel and separated from the primary vacuum by means of a VAT valve. Only

Fig. 2. LIVVS mock-up.

Fig. 1. LIVVS probe.

Fig. 3. LIVVS scanning head.

C. Talarico et al. / Fusion Engineering and Design 51–52 (2000) 1001–1006

when an inspection is required the valve is opened and the probe is inserted into the JET vacuum vessel through a vertical port. In order to match the vessel movements, an appropriate beam alignment system is placed above the LIVVS probe. It should be pointed out that owing to the laser technology, no additional light sources are required. All the relevant components are rated for in vessel conditions, that is ultra high vacuum (10 − 9 mbar), 350°C (scanning head), neutron (1020 n/m2) and g-radiation (2×104 Sv). The required accuracy is 1 mm @ 10 m, the target distance ranging from 0.5 to 10 m. As the magnetic field is maintained only during plasma pulses, the probe movements are not affected by electromagnetic forces. Otherwise, in order to comply with the ITER specifications regarding the magnetic field, research and development activities are being performed on SiC as probe structural material. The mechanical design of the probe (Fig. 1) is based on two AISI 304 coaxial rotating pipes controlling the silica fused scanning prism — at the bottom end of the probe-around two orthogonal axis, i.e. the tilt rotation axis and the pan revolution axis. The probe coaxial pipes are operated by means of two high vacuum radiation hardened stepping motors [2]. At the bottom end of the probe (Figs. 2 and 3), the tilt outer pipe, shaped as a bevel gear, engages the tilt train mechanism of the prism. The pan inner pipe revolves the prism around the probe longitudinal axis. The back-and-forth motion of the pan axis is limited by means of electrical limit switches and mechanical stops. The tilt and pan movements allow the prism to scan a quasi-spherical scene (360° azimuth, 150° elevation). The rotation speed ripple is controlled by means of micro-steps control technique driving the motors up to 50 800 steps per shaft revolution, and by using very accurate gearing and a gear play recovery system. The prism position is detected by means of two commercial optical encoders modified to meet the vacuum and radiation specifications. The rotation encoder disk is fixed to the prism support while the revolution encoder disk is fixed to the inner pipe. The encoder disks are read by means of optical

1003

fibers, carrying the light signal 120 m away the probe where the active laser electronics is placed. The viewing system pointing accuracy is the difference between the actual spot coordinates and the spot coordinates as measured by the encoding system. In theory, the pointing accuracy depends on the motor step size, the motor and gears backlash, the gear runout and profile errors, the pipes vibrations and the pipes straightness. Owing to the fact that the encoder disks are fixed to the prism and to the revolution pipe rather than the motor shafts, errors due to driving shafts and gearing (motor step size and gear run-out and profile errors) are strongly reduced. The probe is allowed to expand along its longitudinal axis and, since it rotates and has an axis of symmetry, no differential thermal expansion effects are expected.

3. Components and materials The selection of the LIVVS components and materials for use in the JET vacuum vessel has been a very important part of the design. Not only must the components have adequate performances and the materials are capable of being fabricated into the required shapes, but they must withstand the environmental conditions of temperature, pressure and g-ray radiation imposed on it by the JET machine, without limiting the attainable pressure. In fact, the LIVVS probe does not need to be removed from the machine during deuterium –tritium experiments. Stepping motors and drives have been supplied by Empire Magnetics and Parker Compumotor, respectively. Operating a motor in a vacuum is often perceived as a design challenge because of outgassing, leakage, cooling and corona effects, but, on the other hand, it gives advantages avoiding mechanical feedthroughs which compromise the accuracy and the resolution of the motion and positioning system inside the vacuum chamber. The LIVVS motors are placed inside the probe upper head and are rated to work up to 155°C, well above the maximum head temperature (50°C). The vacuum rating is 10 − 9 mbar. As far as the radiation is concerned, these motors are

1004

C. Talarico et al. / Fusion Engineering and Design 51–52 (2000) 1001–1006

rated 5× 108 rads (g-radiation total absorbed dose). The bearings dry film lubricant has a vapor pressure of 10 − 14 mbar at 20°C. In order to reduce friction — which for mechanism operating in vacuum and at high temperature can lead to seizure — and to comply with dimensional specifications, high precision ceramic or stainless steel miniature bearings have been successfully used. These bearings have been supplied by Miniature Precision Bearings. They have stainless steel races, ceramic (silicon nitride) or stainless steel balls, dry or molybdenum disulfide lubricant. Since they undergo low dynamic load ratings, the lack of lubricant has not produced any significant effect or reduced their performance. Optical encoders are devices that convert a mechanical position into a representative electrical signal by means of a patterned disk, a light source and photosensitive elements. Standard optical encoders contain opto-electronic components that cannot be made rad-hard. In order to overcome this problem, the encoders active electronics has been placed far away the scanning probe, relying on optical fibers for the light transmission. Therefore, each of the LIVVS encoders is made up of four parts, the borofloat disk; the borofloat collimator; the remote electronics; and the aluminum coated optical fibers. The rotation encoder disk is fixed to the prism (Fig. 3) while the revolution encoder disk is fixed to the inner pipe in the probe head. Each disk is patterned with a double track of lines (channels A and B) near the outside edge of the disk (Fig. 4). Each optical encoder provides two quadrature signals. The decoding that has been performed is a 4 × full quadrature, exclusive-or process, thus yielding 10 000 and 72 000

states per revolution for the tilt and pan rotations, respectively (resolution multiplication). A higher resolution (50 × ) has been recently obtained by means of a dedicated electronics and a special treatment of the analogue signals. The LIVVS encoders which will be used at JET measure the prism rotation and the inner vertical pipe revolution to a precision of one count every 0.036° (360° divided by 10 000) and every 0.005° (360° divided by 72 000), respectively. When the prism and the pipe move, the number of counts read by the encoders are recorded and converted to degrees. Continuing this process for the entire viewing range, an angle versus encoder count relationship may be determined and used to process the scanning laser data. As far as the 600 mm multimode step index optical fibers is concerned, they have a high purity silica core, a doped silica cladding, and an aluminum coating. The metal coating is vacuumtight and the relevant feedthroughs do not require epoxy, thus avoiding outgassing problems. The fibers operating temperature is − 196– + 400°C. ENEA has designed and tested a compact optical fiber vacuum feedthrough, which provides a vacuum tight seal without epoxy and fiber cutting. The device consists of a double tapered vespel ferrule and a flange with a stainless steel compression housing with a threaded cap. It enables the aluminum coated optical fiber to feedthrough the vacuum flange without cutting the fiber. The fused silica prism has an anti reflective coating and a gold reflective coating on cathetus and hypotenuse, respectively. With the anti-reflective coating, light transmission reaches 99.5%. The gold reflective coating allows to scan beyond the prism critical angle.

Fig. 4. LIVVS encoder disk.

C. Talarico et al. / Fusion Engineering and Design 51–52 (2000) 1001–1006

1005

4. Results Tests on the LIVVS mock-up are in progress in order to assess the mechanical, optical and electronic performances, and to measure the system accuracy. The test campaign includes under-vacuum tests at 350°C. Preliminary results show that the LIVVS can take laser scanned pictures both in terms of signal amplitude, for a 2-D representation and in the phase dominion for range finding. Fig. 5 shows a 1100× 1800 pixels detail of a 1100× 50 000 pixels quasi-spherical image (ambient pressure and temperature), the subject being 5 m from the mock-up scanning head, Fig. 2. The acquisition lasted about 30 min. Apart from a scaling of the gray levels, which was necessary to deal with the high dynamic range of the image, no software treatment has been carried out. It should be noted that the image does not have any shadow. As previously mentioned, this is due to the fact that the incident and the collected lights propagate along the same optical axis (monostatic image). The order of magnitude of the measured accuracy complies with the required one (1 mm @ 10 m).

5. Conclusions Fig. 5. LIVVS mock-up image.

Blackening of the inner surface of the inner pipe is required to reduce the backscattered light and increase the signal-to-noise ratio. Besides optical requirements, it is important that the coating remains firmly bonded and free from cracks or flakes. Coating degassing should also be taken into account. In fact, during operation in vacuum and temperature, any gas trapped in the coating during fabrication would be released over a long period. The black chrome provides black, anti-reflective deposits on the inner surface of the pan pipe. It is a composite consisting of chromium oxide and a proprietary additive. Several tests have been successfully carried out in order to assess the black chrome outgassing.

The LIVVS is be able to take in vessel 2-D/3-D pictures by means of a laser beam deflected on the target via a mechanical scanning probe. The probe, designed in compliance with ITER specification, will initially be tested at JET, if possible during the tritium experimental phase DTE2. Tests are in progress at ENEA laboratories on a mock-up in order to demonstrate the system performances and to measure the viewing accuracy. These tests are confirming the system suitability as tokamak reference design for in vessel inspection systems.

References [1] T. Businaro, G. Dalle Carbonare, J.F. Junger, T. Raimondi, Key Features of the New In Vessel Inspection System at JET, Karlsruhe, Germany, 22 – 26 August 1994. [2] J.V. Draper, B.S. Well, J.B. Chesser, R.I. Crutcher, K.U.

1006

C. Talarico et al. / Fusion Engineering and Design 51–52 (2000) 1001–1006

Vandergriff, Component testing in a high-level g-radiation environment, Oak Ridge National Laboratory Technical Report no. ORNL/TM-12213. [3] Coletti A., M. Baldarelli, L. Bartolini, A. Bordone, M. FerrideCollibus, G. Fornetti, S. Lupini, C. Neri, C. Poggi,

.

M. Riva, L. Semeraro, C. Talarico, L. Zannelli, Amplitude-modulated laser viewing system in apparatus for controlled thermonuclear studies, American Nuclear Society Eighth International Topical Meeting, 25 – 29 April, 1999, Pittsburgh, PA.