Nuclear Decommissioning with Yb-Fibre Laser Tube Cutting

Paper presented at 2010 ICALEO Proceedings Anaheim, CA, USA, 26 - 30 Sept. 2010, Paper 707

Abstract

An essential part of nuclear decommissioning is to ensure that all the radioactive waste arising during the dismantling process is safely, effectively and economically managed. Although 99% of the radioactivity is associated with the fuel, which is removed following permanent shutdown, significantly large infrastructures remain, consisting of various materials, of which a considerable portion is metallic tubing of various diameters. These tubes need to be reduced in size to be packaged as nuclear waste. Amongst the available thermal cutting processes for such tubes is the multi-kilowatt fibre laser, offering high beam quality and power transmission over large distances through optical fibres. Because such a laser can be operated remotely from the cutting operation, there is no risk of contamination of the high value laser asset, which can therefore be reused on a number of decommissioning tasks. In this paper, the single-sided cutting capability of a 5kW multi-mode Yb-fibre laser is presented for various 304L stainless steel tubes of diameter up to 170mm and wall thickness up to 11mm. The results show that with the right optical configuration, it is possible to laser cut tubes from one side, at long, and thus safe standoff distances, if the main objective is simply to sever the tube.

Introduction

The first generation of nuclear reactors and their associated infrastructures were built in the 1950s and 1960s. Now, an increasing number of reactors have reached the end of their useful life and decommissioning is a topic of public concern. The dismantling and decommissioning of a nuclear site requires a variety of technologies, each with a specific application in mind. An average nuclear facility can contain significantly large metallic assemblies, including vessels, flasks, support structures and pipe-work, made from nickel, iron and aluminium alloys. All these structures have different shapes, sizes, and wall thicknesses. Some contaminated cells contain several kilometres of tube networks, of varying diameters and wall thicknesses, positioned in many orientations. Such structures need to be cut into pieces small enough to be packed into low or medium level containers for long term storage.

In view of the wide range of decommissioning tasks, many different cutting techniques have been developed so far.^[1] In some cases techniques already used in the sheet metal manufacturing industry have been adapted. Cutting techniques can be classified according to the physical principle in action for destroying inter-atomic bonds.^[2] These include thermal, strain and chemical energy techniques. In the thermal cutting technique, solid material is melted and then flushed away with a gas jet. One such thermal technique, which surpasses all others, is the laser cutting process, which generates minimum reactive force and fume.^[3] Unlike standard laser cutting used in many manufacturing industries, no requirements are imposed on the cut edge quality (i.e. surface roughness of the cut edge) when the process requirement is for decommissioning. The performance of the process is assessed by the capability to sever the part into separate pieces with minimum secondary emissions (aerosols, sedimented dross, attached slag etc). ^{[1,3 & 4]}

In the past, various high power lasers have been used to cut thick-section metallic materials for nuclear decommissioning applications where constant power density and nozzle standoff distance to the substrate were usually maintained. These included CO₂, CO, COIL and Nd:YAG lasers.^[5] All lasers offer unique capabilities, but the flexibility offered from solid-state lasers, employing optical fibre delivery of the laser power, reduces complexity and risks. Development of high power disc and fibre lasers, coupled with improvement in beam delivery and thermal management of the system, have further enhanced cutting capability by providing scalable power in the multi-kilowatt regime.^[6] However, the biggest challenge encountered for decommissioning arises due to the profile of the tubes and their juxtaposition, with respect to each other.^[7] They could be bundled, multi-layered, and concentric, in various orientations and sizes. From the deployment consideration, conventional laser cutting around the tube is almost impossible, and a method of single-sided tube cutting needs to be developed.

Unlike conventional laser cutting of flat plates or orbital laser cutting of tubes, where the beam focus and the nozzle standoff distance is maintained constant with respect to the tube surface,^[8] in the single-sided laser cutting described here, both the laser focus diameter and the standoff distance vary relative to the tube surface. Schematics of the process set up and the laser cutting head used for the trials are shown in Figure 1.

In March 2009 the UK's Nuclear Decommissioning Authority awarded TWI a contract to develop prototype equipment, in order to effectively demonstrate the twin processes of concrete scabbling^[9] and tube cutting and how these technologies might be implemented for remote use in nuclear decommissioning. In this paper, single-sided laser tube cutting using 5kW multi-mode and 1kW single mode fibre lasers is addressed. The process performances on a number of different 316L stainless steel tube diameters and wall thickness are evaluated.

Methodology

Single-sided laser cutting trials on 316L stainless steel tubes were performed using an IPG 5kW multi-mode (MM) laser, and for comparison, a 1kW single mode (SM) fibre laser. The beams from these lasers were focused to approximately 420µm and 70µm diameters respectively, by using collimating and focusing optics, coaxially aligned with a tailored lens design and a cutting nozzle assembly designed to operate at pressure of 8bar. Table 1 provides details of the equipment and parameters used to perform the laser cutting operations.

Table 1. Equipment and parameters used in single-sided laser tube cutting trials.

Lasers, powers and wavelength	IPG 5kW MM (BPP = 6) IPG 1kW SM (BPP = 0.3) 1070 - 1080nm
Fibre core diameter (µm)	150 & 17
Collimator focal length (mm)	120
Optical focal length (mm)	500
Tube diameters (mm)	60, 155, & 170
Tube wall thicknesses (mm)	1.5 to 11.1
Gas pressures (bar) (comp.air)	2 to 10
Nozzle diameters (mm)	3.25 & 3.65
Max. cutting speeds (mm/min)	10 to 2000

Laser cutting trials for a given laser power, gas pressure and cutting speed, were performed by traversing the laser beam across the tube in a straight line, while maintaining the focal position along the centre of the tube. Laser cutting on tubes with different diameters was achieved by extending or reducing the nozzle position, but always keeping the minimum standoff distance (Figure 1) of 10mm.

Single and double pass cutting techniques were examined. Maximum cutting speeds reported here were for a complete severing of the tube. Most of the time, if there was any lack of separation it was encountered at the sides of the tube. At these positions not only the standoff distance, but also the material thickness is at a maximum. Table 1 shows estimated maximum material thickness for particular tube and wall thickness combinations.

Table 2. Estimated maximum material thickness for various tube diameters and wall thickness.

Tube diameter DP	Tube wall thickness T (mm)	Maximum cut thickness H (mm)
60.0	1.5	18.73
60.0	4.0	30.0
60.0	5.44	34.5
60.0	8.71	42.3
60.0	11.1	46.6
155.0	1.5	30.4
170.0	7.0	67.6

Results

It is clearly desirable to cut large diameter steel tubes with just a single pass of the laser beam. Therefore, several simple linear cutting techniques were addressed and it was found that with a single pass technique, the maximum cutting speed was very slow and complete separation of the tube was limited to small tube diameters and small wall thicknesses. A double pass technique produced the best results enabling higher cutting speeds and cleaner cut surfaces. The double pass cutting was initiated and terminated at the centre of the tube, as shown in Figure 1. In one series of experiments, using the MM laser, samples were produced with a constant laser power, for each particular tube diameter and wall thickness, by varying cutting speed, standoff distance and gas pressure, to determine the maximum cutting speed. Figure 3 shows the maximum cutting speed for a 155mm diameter tube with 1.5mm wall thickness, at various gas pressures and laser power settings. The focal position of the laser beam was fixed, as shown in Figure 1, with the appropriate cutting nozzle and extension. The standoff distance between the nozzle exit and the closest approach to the tube was maintained at 10mm. Pressure dependent laser cutting trials were performed with a constant laser power of 4.6kW and laser power dependent cutting trials were performed at a constant gas pressure of 8bar.

It can be seen that the maximum cutting speed is proportional to both the laser power and the gas pressure. However, there appears to be a higher dependency on the laser power. This is to be expected due to a significant variation in the available laser power density at the top and the bottom edges of a large diameter tube and it is the lower edge of the tube which is more susceptible to adhering dross for variations in both the laser power density and the gas pressure, as seen in Figure 4. The best cut quality, in terms of speed, was attained with higher laser power, for the same gas pressure. Similarly, the cut quality, in terms of speed, was also better with higher gas pressure for the same laser power.

Trials on 60mm diameter tubes of various wall thicknesses, with a constant laser power of 4.8kW were also carried out, to determine the effect of assist gas pressure. As for the 155mm diameter tube, the focal position of the laser beam was fixed as shown in Figure 1, using a different nozzle extension tube and the standoff distance was again maintained at 10mm. The maximum cutting speed obtained for each wall thickness is shown in Figure 5.

As would be expected, the smaller the tube wall thickness, the faster the cutting speed and this reduces exponentially with an increase in the tube wall thickness. The laser cut edge quality for five different wall thicknesses with 8 bar gas pressure and a power of 4.8kW is shown in Figure 6. It can be seen that the edge quality in the lower half of all tubes is always worse than in the upper section and is worst at the highest thickness. Interestingly, all these tubes show poor quality cut edges, particularly at each end of the cut, and there appears to be a transition between a poor and a reasonable cut surface, as shown in the boxed regions in Figure 6. This transition region appears to be at an elevated angle to the horizontal axis, which increases as the wall thickness increases.

A comparison was also made between the SM and MM fibre lasers on the 60mm diameter tube with a 1.5mm wall thickness. A constant laser power of 1kW, for various gas pressures and three standoff distances, was used. The focal position relative to the nozzle exit was always maintained at distance of 55mm, while the three standoff distances of 10mm, 25mm and 40mm, were varied between the nozzle exit and the top surface of the tube. This meant that the laser focal position was varied ±15mm either side of the tube centre line. Figure 7 shows the maximum cutting speeds achieved with both systems.

It is interesting to note that below 6bar of nozzle gas pressure, higher cutting speed is obtained with the MM beam, even though the laser power density with the SM, under identical set up conditions, was significantly higher. Remarkably, there also appears to be little or no difference in the cutting speeds at standoff distances of 10 and 25mm for both beam qualities, considering there is an approximate three times difference in the maximum laser power density available for the two configurations. Generally the cutting speed at 40mm standoff distance was lower for both systems. The best performance was obtained at 8bar gas pressure for the SM laser at either 10 or 25mm stand off distance. However, this performance showed a sudden reduction in speed, for all standoff distances, at gas pressure of 10bar.

Single-sided laser tube cutting methods were also developed for selectively removing sections of much larger tubes, and a demonstration was also setup to simulate the effectiveness of remote deployment of this technology by sectioning various pipes of different sizes, wall thickness and orientations in one continuous robot program.

Figure 8 shows the largest diameter tube, at 170mm and with wall thickness of 7mm, used in this work. This was cut with a laser power of 4.8kW, at a linear speed of 100mm/min and using 8bar of compressed air assist gas. In this case a three pass technique was used; the first two cuts removed a segment from the front of the tube, thereby providing more energy at the rear of the tube during the third pass, which produced complete separation. A total cutting time of 7min was required for this tube. Sections of tube can be removed easily (should access into the interior be required) and tubes can be cut with the beam incident (at least) up to 45 degrees to the tube axis. Indeed, using the current equipment, it was also possible to cut arrangements of concentric tubes and tube bundles.

Tube networks or 'the nuclear jungle' as it is described in nuclear circles, have collections of tubes of various sizes, thicknesses and orientations. In order to simulate real conditions inside a contaminated cell, a collection of tubes of various diameters and orientations were mounted on a support framework. The support framework was designed to be re-used after each demonstration. The back wall of the support structure was constructed from graphite sheets, used to absorb any laser beam propagating through or part the tubes. Tubes of diameters from 25mm to 155mm, including arrangements of concentric tubes, were fixed to the support frame using, the type of fixtures commonly employed in practice. Using a laser power of 4.8kW and an assist gas pressure of 8 bar, all tubes were cut (using the same nozzle assembly) with a single robot program. Over 50 cuts were made on tube and fixtures to dismantle this tube network in an elapsed time of 15min. Figure 9 shows before and after images of the demonstration exercise.

Discussion

All current data on tube cutting with lasers relates to profile cutting around the circumference of the tube with the cutting nozzle always perpendicular to the tube surface. Fibre laser cutting of stainless steel tubes with similar wall thicknesses to those used here and using similar laser powers, using this technique, showed significantly higher available cutting speed in comparison to the results obtained for the single-sided tube cutting. However, the cut quality reported for profile tube cutting was reduced with increasing tube wall thickness, even though substantially higher gas pressure was used to remove the molten material from the kerf.^[8] For a linear traverse of the laser beam in a single-sided tube cutting process, it can be expected that variations in the tube surface angle, the inclination angle of the cut front, the tube wall thickness, the beam focal point position and the standoff distance, will have significant effects on the overall cutting speed achievable.

At the tube edge, where the laser absorption is zero, as soon as the laser beam interacts with the curvature of the tube, a fraction of the laser radiation is absorbed, raising the localised temperature to melting point. This, combined with the shear stress generated by the coaxially aligned gas jet, causes a cutting front to initiate. At this point a sudden increase in absorption at the cutting front becomes possible. As the laser beam progresses further across the tube, the amount of material available for cutting increases and this will depend on the tube diameter and the wall thickness. At the same time an increase in the distance between the laser focal position and the beam interaction point on the tube surface, increases the cutting front or the kerf angle, where absorption of laser radiation is the key for an efficient laser cutting process.^[10] The absorption of laser radiation within the material depends on the angle of incidence, the temperature dependent refractive index and the extinction coefficient, known as Fresnel absorption.^[11] The average theoretical absorption of circularly polarised or un-polarised laser wavelengths of 1.07µm and 10.6µm radiation for various angles of incidence are shown in Figure 10, which indicates that the absorptivity reaches a maximum at a distinct angle of incidence, referred to as the Brewster angle. For a Yb-fibre laser wavelength this is calculated to be 79.9°, compared to a value of 87.3° for the CO₂ laser wavelength.

In the case of a single-sided laser tube cutting process, the cutting front angle will vary as the beam moves in from the edge of the tube and this change will be greater the smaller the diameter of the tube. The largest depth to be cut occurs just before the beam penetrates into the core of the tube. This will depend on tube diameter and wall thickness. The cutting front angle will increase, and at some point will reach the Brewster angle and subsequently move above this value. Cutting performance will thus vary significantly because of the variations in cutting front angle. This will affect the cutting speed of a single-sided laser tube cutting process, and is evident in the photographs in Figure 6.

Once the laser beam penetrates into the core of the tube, where the melt expulsion from the kerf becomes easier, a sudden increase in absorption also takes place, due to a decrease in the cutting front angle, which is directly proportional to the depth of material being cut. This effect can be clearly noticed in the photographs in Figure 6. The cut quality attained in the upper section of the tube begins to improve and at a certain location the cutting front angle will match the Brewster angle, where the highest laser absorption (aided by the high assist gas pressure) results in an optimum cutting performance. The location of this optimum cutting point of course will depend on the tube diameter and the wall thickness, as can be seen by comparing the photographs in Figures 4 and 6.

Comparison between the 1kW SM and MM results indicated higher maximum cutting speeds for the MM laser beam below a gas pressure of 6bar. It is highly likely that the combination of the gas jet and the larger beam diameter of the MM system (for all impingement positions across the tube) plays a significant role. A larger laser beam on the surface will increase the kerf width, which will allow the lower half of the tube to be more influenced by higher gas mass flow-rate. A gas pressure higher than 6bar may result in increased gas jet momentum, which could be sufficient to penetrate through the narrower kerf width formed by the 1kW SM laser beam and help to flush away the molten material from the lower half of the tube. A higher gas pressure, especially with air (20% oxygen) may also lead to slight increase in the kerf width, which when coupled with higher laser intensity, might explain the increase in cutting speed seen with the SM system. The best performance with the SM laser was achieved with a nozzle pressure of 8bar, at the two smallest stand-off distances. Isentropic gas expansion at high pressure with correct nozzle design can produce jets with minimum shock behaviour. However, pressure below or above the design pressure can result in an under or over-expanded gas jet, with multiple shock behaviour and variable strength. This can produce a drastic effect on the gas dynamic conditions on the material surface, which can influence the melt ejection condition inside the kerf. ^[12]

 

Interestingly, a significant reduction in cutting speed for the SM laser at 10bar was noticed. This may be predominantly due to the under-expanded gas jet dynamics associated with kerf size. As indicated, a larger kerf size would allow higher mass flow-rate towards the bottom of the tube but a narrow kerf, especially at a higher and dynamically unstable gas pressure, may induce flow separation. A flow separation could result in a change in the flow direction, which is most likely to promote re-circulating eddies, especially towards the bottom half of the tube. Under these conditions the gas flow direction will be the same as the cutting direction, i.e. perpendicular to the kerf. Such a phenomena would deposit a larger volume of ejected oxidised material due to higher laser intensity from the top of the cut onto the bottom half of the tube, in to the path of un-processed material, which would then require additional laser power to re-melt.

The difference in the divergence of the two beams could also play a role on the cutting process, especially at the tube edges with respect to the formation of the cut front angle and laser absorption characteristics. A lower divergence beam would produce a larger cutting front angle, and coupling this with increased available material near the tube edges, will result in lower absorption of the SM laser beam. However, with the low divergence of the SM laser, the power density available during the cutting process is significantly higher than that available from the MM laser, which should compensate for lower laser absorption. When operating with the smaller kerf width obtained with the SM system, the effectiveness of the gas jet becomes very important. This perhaps explains the lower cutting speeds seen below 6bar and the sudden reduction at 10bar. However, use of a SM system for decommissioning applications may not be a real option, due to limits on the delivery fibre length (max 5m) for the SM lasers.

Conclusions

A very effective and efficient system for decommissioning of nuclear stainless steel tube structures has been developed. The cutting head, tailored for this application, is light and has a significant standoff tolerance and so is relatively simple to deploy and operate remotely. For all tubes of different diameters and wall thicknesses, the most critical regions are the tube sides, which require both high laser power and assist gas pressure, for clean separation. As a result, a double pass technique was preferred to a single pass method as the optimum cutting configuration. Even using the double pass technique, the cut quality at the tube edges and the lower surface, was always worse than at the top of the tube, and the cut quality also decreased with an increase in the tube wall thickness, for a constant diameter. Comparison between single and multi-mode lasers revealed the influence of the gas jet to be more significant than the divergence of the laser beam and that in single-sided laser tube cutting a larger kerf width is very important.

Acknowledgements

The authors gratefully express their thanks to the Nuclear Decommissioning Authority in (UK) for financial support of this project. They also would like to thank Sellafield Ltd for their support and guidance in this project.

References

1. Newton, G. J. et al, (1987): 'Collection and characterisation of aerosols from metal cutting techniques typically used in decommissioning nuclear facilities', American Industrial Hygiene Association, Vol. 48(11), pp922-932.
2. Al-Hassani, S. T. S., Marinetech North West Report ABN 6 R 17039.
3. Pilot, G et al, (2008): 'Measurement of secondary emission during laser cutting of steel equipments', Nuclear Engineering and Design, Vol. 238, pp2124-2134.
4. European Commission, nuclear science and technology, Handbook on decommissioning of nuclear installations, Report EUR 16211 EN, ISSN 1018-5593.
5. Nuclear Decom. (2001): Ensuring safe, secure and successful decommissioning, IMechE Conference Transaction.
6. Stano. S, (2007): 'New solid-state lasers and their applications in welding as generators of laser radiation', Welding International, Vol. 21 (3), pp174-179.
7. Beck, R, (1989): Durchführbarkeit ausgewählter Arbeitsverfahren für die Beseitung radioactiver Komponenten stillzulegender Kernkraftwerke, Batelle-Institut, Frankfurt D, Teil IV.
8. Purtonen, T, Salminen, A, (2009): 'Fibre laser orbital cutting of stainless steel tubes', paper 605, ICALEO.
9. Hilton, P., (2010): 'Fibre lasers for surface removal of concrete in the nuclear sector', ICALEO, paper 1305.
10. Mahrle, A., and Beyer, E., (2009): 'Theoretical aspects of fibre laser cutting', J. Phys D: Appl. Phys. Vol. 42,pp1-9.
11. Modest, M. F., (2001): 'Reflectivity and absorptivity of opaque surfaces', LIA Handbook of Laser Material Processing, Orlando FL, Magnolia Publishing, pp95-103.
12. Fieret, J. et al, (1987): 'Overview of flow dynamics in gas-assisted laser cutting,' SPIE High Power Lasers, vol. 801, pp233-250.