Close This website uses its own and third party 'cookies' to offer you a better experience and service. By browsing or using our services You accept the use we make of them.

News > In-vessel viewing system prototype performance measurements and simulation of measurement quality across the ITER in-vessel components


The In-Vessel Viewing and Metrology System (IVVS) is a fundamental tool for the ITER machine operations, tasked with inspection of plasma facing surfaces (PFS) of in-vessel components for both damage and erosion.The IVVS design has evolved to include a hybrid viewing and metrology system, using Frequency Modulated (FM) lidar for absolute distance measurements, and amplitude modulated (AM) lidar for surface viewing A full-scale prototype has been built, and has produced line-of-sight range precision of <2 μm in the optimum scenario. Error analysis indicates that 3D precision from 2 to 400 μm should be available over surfaces from 0.5 to 10 m with 0–70° incidence angles, exceeding the ITER requirements.Performance was found to be highly sensitive to received power and surface reflectivity. To parameterize the reflectivity, the Bidirectional Reflectance Distribution Function (BRDF) of a divertor monoblock prototype was measured and an analytic BRDF model was developed. An IVVS simulator was developed and used both to analyse the in-vessel geometry and to predict the coverage of the PFS. We have shown that 85% of the PFS area can be covered if the IVVS meets performance requirements up to incidence angles of 65° and distances of 8 m.1. Introduction The In-Vessel Viewing and Metrology System (IVVS) is a key ITER diagnostic for inspection of the plasma facing surfaces, as it provides the highest resolution coverage over the widest in-vessel field of view of the ITER diagnostic inspection systems. The development of the concept design of the entire IVVS system has been extensively documented in previous publications [[1], [2], [3], [4]]; only a brief overview is given here as this report focusses on one sub-system, covering the optical layout and measurement, referred to as the “IVVS probe”.

Figure 1 shows tge system as installed in ITER. It is permanently resident in a Vacuum Vessel (VV) port extension, and during plasma pulses it is retracted and shielded. For operation a push-chain deployment systems is used both to insert the so-called “mobile assembly” into the vessel and to elevate it for improved viewing angles on the target surfaces. Six IVVS units will be installed in port extensions spaced at close to 60º intervals around the Tokamak.

Figure 1. Left: the IVVS shown in the deployed and elevated position. Middle: monostratic and bistatic probe optical layouts; right, hybrid probe optical layout. Tx = transmission fibre, Rx= Receiver fibre, AM7FM= amplitude / frequency modulated lidar


The IVVS mobile assembly carries the IVVS probe containing all the optical and beam steering components. Dose rates in the deployed position reach hundreds of Gy/hr, so no electronics are present in the probe itself. All optical signals are transmitted through fibre optic transmission (Tx) and receiver (Rx) channels to electronics located in the port cell or diagnostic building.

Section 2 describes the concept design of the IVVS optical system, recent design modifications, and the justification for the changes. Section 3 reports on the IVVS prototype and measurements of its performance on laboratory targets. Performance is shown to depend critically on the power of the backscattered light from the surface, and Section 4 reports the determination of both measured and analytical bidirectional reflectance distribution functions (BRDF) for the ITER target surfaces.

Section 5 shows how these BRDF functions and measured performance can be extrapolated to the full ITER vessel geometry by using a custom IVVS simulator code to predict the coverage and accuracy of the IVVS across the full range of ITER PFS.


2. Evolution of the IVVS probe design

2.1 Monostatic optical layout for the ranging measurement

A critical design driver for the IVVS is the need to operate at both 20 °C and 70 °C, with a stability during scanning of +/−1 °C. The concept design bistatic layout (using separate channels for the Tx and Rx pathways), was identified as sensitive to misalignment, and requiring active elements within the probe to maintain alignment through temperature changes over the IVVS lifetime.To avoid this, a monostatic layout was proposed using a single fibre for transmission and reception and a single lens for focusing. If mounted in a cylindrical barrel, this layout would be self-aligning and near insensitive to thermal variations, with only the focus requiring adjustment. The Tx and Rx optical channels can be separated in the single fibre using a fibre optic circulator with low cross-talk between the channels.

2.2 Selection of FM lidar as the ranging technology

The IVVS concept design used Amplitude Modulated (AM) lidar in a bistatic layout, however this technology is not suitable for a monostatic optical layout.In a monostatic configuration, the −70 dB back-reflection from the fibre optic end facet unavoidably travels to the detection system. This reflection cannot be filtered out of the AM lidar signal and will distort the ranging measurements. A monostatic prototype AM lidar system has been built and tested, showing that the range measured was invalid when the back-reflected signal had a power above 1% of the real scattered signal. Due to the high dynamic range required from the IVVS, even a −70 dB end facet reflection leads to an unacceptable loss of coverage, making monostatic AM lidar not viable as an IVVS measurement technology.Our proposed solution is the use of Frequency Modulated (FM) lidar as the IVVS ranging technology. Frequency Modulated systems use a calibrated linear sweep of the laser frequency to measure range to target with micron precisions even at km distance, typically far better than corresponding AM systems [5]. FM lidar can resolve multiple reflections along the measured path, which is critical for prism-based scanning where internal reflections can disturb AM measurements. FM uses the end facet reflection as the zero point for range measurements, removing the need for a calibration target in the probe itself. FM systems also return an absolute distance to the target, while AM systems measure via deconvolution of phase, which suffers from an inherent 2π uncertainty in the measured distanceThe major disadvantage of FM systems is that they use coherent detection. This means no viewing signal can be extracted from the FM lidar signal, requiring a separate viewing channel to be included in the system design.

2.3 Hybrid AM/FM configuration for viewing and metrology

To retain the benefits of FM lidar technology while still providing a viewing channel for inspection, a hybrid architecture has been proposed and is shown in Fig. 2. This architecture retains the AM lidar bistatic optical system to create a viewing channel only, combined with a FM lidar monostatic optical system for temperature insensitive precision ranging measurements. The AM lidar system for viewing has high dynamic range and viewing is by nature less sensitive to temperature-induced misalignments. The AM Tx signals co-propagate with the FM Tx signals to the target, with the scattered light separated from the FM Rx beam by a dichroic mirror, and sent to the detector in a separate Rx fibre. A reflective parabolic mirror is used instead of the previously used refractive Rx optical element to allow for a single actuator to work for both the 1550 nm FM and 800 nm AM beam. Linear piezomotors on all focusing elements allow the system to optimize focus on the target during measurements. Note that dynamic range compensation is required, as the refracted optical path length in the rotating prism varies as a function of the rotation angle. Compensation is also required for the position of the parabolic focusing mirror. Both rotary and linear piezo actuators will be equipped with encoders to allow and correction of the optical path length via post-processing.


Fig.2. IVVS hybrid probe design. Schematic layout with beam path indicated in red for FM lidar and blue for AM lider (for interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

The IVVS concept design used a bundle of single-mode fibre-optics for the Rx channel, which suffered from high connector losses. The hybrid system uses SM fibres for the AM Tx pathway and the FM Tx and Rx pathways, and an MM fibre for the AM Rx pathway, allowing extension of the length of the Tx and Rx fibres to 200 m and movement of the IVVS cubicles from the port cell to the diagnostic building, greatly reducing their radiation dose.


3. IVVS prototype testing and performance

3.1 Prototype system description

Two full scale IVVS prototypes have been produced to allow verification of the design decisions made, one mono-static and one bistatic. These include linear piezoelectric actuators (PM Bearings) allowing both Tx and Rx pathways to be focussed at distances from 0.5 m to 10 m.Piezoelectric rotary actuators were proposed in the IVVS concept design, and tested for magnetic, temperature, vacuum and neutron and gamma compatibility. Our IVVS prototype implemented them for the first time for steering of the beam (PM Bearings), allowing characterisation of their performance to show that they are functionally suitable.

The current rotation optical encoders used have a resolution of 0.05 mrad, however these are not radiation compatible. Work is ongoing to develop a radiation hard optical encoder with a precision of 0.05 mrad.The AM lidar measurement system uses a diode laser system (Picoquant LDH-M-C-805, 808 nm, 15 mW) modulated at 550 MHz an APD detector (Hamamatsu C5658) and a commercial lock-in-amplifier (Zurich Instrument UHFLI) for measurement of the phase and amplitude of the Rx signal.

The FM lidar measurement system uses a commercial system (Bridger Photonics SLM-IM, customised for measurement over 0.5 m–10 m range) comprising a 1550 nm frequency-swept laser measuring at 2 kHz and internal detection electronics. The SML-IM was also modified to include 200 m Tx/Rx fibre length between the laser/detection unit and the probe optical head.All lenses used are custom-designed components whose surface profiles have been optimised using a multi-configuration optimisation routine to obtain the best focus at all working distances. The lenses have aspherical plano-convex surfaces with a conic constant added to the spherical surface. Custom design of this sort is required to tightly focus the scattered light to allow efficient coupling to the Rx fibres. The dichroic mirror and AM Rx fibre are in the shadow of the AM Tx mirror so that obscuration of the FM Tx/Rx path is minimized.

3.2 Testing and measurements results

The prototype line-of-sight (LoS) measurement precision and accuracy was measured for a single measurement location (no scanning) using a sanded aluminium test target. Measurements were taken at angles of incidence of 0°, 10°, 40° and 70° and for distances from 1 to 10 m. LoS accuracy and precision were determined by using a calibrated translation stage (Thorlabs PT1/M-Z8), to move the target along the direction of the beam at steps of 0.5 ± 1e-4 mm over a range of 2 mm, taking 4000 individual measurements at each point (see Fig. 3). Accuracy is calculated by comparing the mean step range measured to the known step moved, while precision is the standard deviation of the measurement set.


Fig.3Example measurement data for LoS accuracy and precision testing. Known distance changes are represented by a black line.


In this configuration the error of the LoS distance can be clearly separated from encoder or beam pointing errors, because no angular scanning takes place. In the case of scanning the beam, the small angular error of Δφ can generate a range error of ΔR by displacement of the beam across a suface with incidence angle of magnitude  θ, assuming the most conservative case when the angle of incidence is aligned with the angular axis φ:

ΔR = R Δφ tan(θ)

Results from the measurements are reported in Table 1, grouped by angle and distance according to the ITER requirements.

Table 1. LoS accuracy (LoS Δ) and precision (LoS σ) testing results for AM and FM as a function of angle of incidence (θ) and target distance (L), displaying the worst measured value for the θ,L ranges. Precision and accuracy of the AM system in this table are for a fibre length of 2 m. Propagation through 200 m of fibre puts both the precision and accuracy worse than 1 mm. Also shown is the range error taking into the FM LoS precision, adding 0.05 mrad standard deviation encoder error to create the system expected precision (σR). The final column shows the ITER required precision (σreq).


It can be clearly seen that in the laboratory environment the IVVS prototype system exceeds the ITER requirements for the relatively diffuse reflections from the test target. These results do not address the linearity of the measurement over large intervals, due to the lack of a suitable target with sub-10 μm accuracy over intervals from 0 to 10 m. Previous publications have measured the linearity of a SLM-IM system and found precisions of ˜1 μm and linearity errors < 10 ppm for distance measurements over 24 cm, using an interferometer for comparison [5].
The manufacturer has confirmed that they also expect linearity of <10 ppm over the full range of our customized SLM-IM.During testing the dependence of precision and accuracy of the IVVS measurements it was shown that the return power is the dominant parameter affecting the precision, with error increasing exponentially as return power decreases. Additionally, for the FM system, low return powers cause the system to produce null measurements, which can rise to over 50% of measured points for return powers of <7.5 nW, even if the LoS precision of the remaining measured points is <50 μm. Therefore, the reflectivity of in-vessel surfaces becomes a critical parameter for prediction of the IVVS performance.

4. Target reflectivity measurements

4.1. Target surface BRDF

To characterize the expected in-vessel surface reflectivity, an extensive set of BRDF measurements were taken on a prototype of the divertor baffle region tiles made from machined tungsten. The BRDF measurement data is available on demand.

The BRDF is plotted in Fig. 4 and shows variation in reflectivity of ˜3 orders of magnitude for incidence angles from 0 to 70 degrees, with a high degree of anisotropy depending on surface orientation. The BRDF has a narrower specular peak anticipated in the ITER PA requirements, with a 2° divergence angle. The diffuse component also varies from the current specification, with a more intense background easing detection at high angles.
Fig.4. BRDF results: measured data from the divertor target is shown in blue, the Ashikhmin-Shirley model is in orange. The left and right panes show photos of scatter from the surface, with the sample orientated along the principal anisotropic axes (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).
BRDF results are similar to those reported in [7], and by parallel with that paper, the anisotropy can probably be attributed to the microscopic grooved structure present on machined metal surfaces.


4.2. Analytic BRDF representation

The measured BRDF from the divertor monoblock does not necessarily represent the BRDF of the final in-vessel components. To allow parametric studies, we implemented an analytic BRDF model for anisotropic surfaces. The Ashikhmin-Shirley (A-S) model defined in [8] is an energy-conserving model capable of representing machined metallic surfaces with anisotropic reflectivity. This model was implemented in a Python script to allow calculation of the IVVS return power within the IVVS simulation software. Fig. 4 shows a comparison between an A-S model and the measured BRDF, indicating that with the A-S model we can get a good general agreement with the anisotropy and overall shape of the measured BRDF.

5. IVVS measurement simulation & performance analysis

An IVVS simulator was developed to allow generation of simulated point measurement sets over the full ITER vessel. The simulator defines a set of rays based on the deployed IVVS position and the probe rotation axes. Each ray is cast from the probe position towards the target surfaces; intersection coordinates and angles of incidence are found; reflected power is calculated; and from this the range measurement noise is calculated. The simulation takes into account the three dimensional accuracy by adding random white noise Δθ, Δφ with a specified standard deviation σθ, σφ to the direction of each IVVS ray along each of the two encoder axes. The measured coordinate in the simulator then corresponds to the coordinate (R,θ+Δθ,φ+Δφ), i.e. the ray-surface intersection along the (θ+Δθ),(φ+Δφ) but it is reported in the point cloud at the position (R,θ,φ).The simulator also calculates the effects of the effects of non-zero beam size, and slow drifts in the range calibration.The simulator is based on the Blender open source software package with a custom Python code used to create the set of IVVS rays to be simulated, apply the BRDF functions defined in section 4, apply an empirical noise function based on the prototype measurements, and extract results and statistical information.The ITER requirements specify that the precisions σreq defined in Table 1 should be maintained over 85% of the surface area. To analyse this, detailed models of the ITER divertor and First Wall panels were edited to include only the facets facing the plasma. Each individual blanket tile or divertor monoblock was represented by two triangular surface facets. Measurements were simulated over a 120 ° toroidal segment, including three IVVS probes to take into account the slight asymmetry in probe positioning around the Tokamak.

Four deployed probe positions were used to provide a compromise between operational complexity and minimisation of the angle of incidence on each facet. Facets were initially allocated to the closest IVVS probe, resulting in a max probe-facet distance of 8 m. Then, using all four positions for each IVVS probe, the minimum incidence angle on each facet was selected. A histogram of surface area vs incidence angle was generated, showing that 85% of the PFS area lay at incidence angles <65 deg and distances <8 m from the probe (Fig. 5). The key regions where measurement is difficult lie on a poloidal plane half-way between two IVVS probes. In particular, blanket modules on the inner wall lie at challenging incidence angles.

Fig.5 Right: percent of total surface area of the PFS falling below a defined incidence angle. Left: map of incidence angle in the tokamak, using the color code from the bar graph.


While current testing results have shown that the σreq precision requirements of Table 1 can be achieved at up to 70 °s for sandblasted targets (i.e. over >90% of the surface area), simulation of coverage performance is extremely sensitive to the BRDF parameters. If the final in-vessel surface reflectivity is highly specular, the required 85% coverage will become extremely challenging to achieve. Work is in progress to parameterize the surface reflectivity of actual in-vessel components and reliably estimate the IVVS coverage.

6. Summary

The IVVS design has evolved from the concept design, now using FM lidar technology to provide metrology ranging measurements in a hybrid architecture enables that enables both viewing and redundant ranging.A custom built simulator has been developed that allows IVVS measurement simulation using realistic geometry, beam profiles, beam scanning and surface reflectivity. Valuable information about measurement geometries can be extracted, but improved knowledge of the BRDF of the in-vessel components is essential for accurate performance and surface area coverage estimation.



The views expressed in this publication are the sole responsibility of the authors and do not necessarily reflect those of Fusion for Energy (F4E) or of the ITER Organization (IO). Neither F4E nor any person acting on behalf of F4E is responsible for the use which might be made of the information in this publication.

Direct source: Science Direct – Jorunals&Books.



ThomasSiegela , Efstathios Kolokotronisb, Andrés Cifuentesa , Paloma Matia Hernandoa , Adrià Sansaa, Parthena Symeonidoua , Philip Batesb , Carlo Damianib , Gregory Dubusb , Adrian Puiub , Roger Reichlec

a: ASE Optics Europe, Parc UPC Ed. K2M, Carrer de Jordi Girona, 1, 08034, Barcelona, Spain

b: Fusion for Energy, c/ Josep Pla, n°2 – Torres Diagonal Litoral – Edificio B3, 08019, Barcelona, Spain

c: ITER Organization, Route de Vinon-sur-Verdon, CS 90 046, 13067, St. Paul Lez Durance Cedex, France