Abstract

Wind turbines are becoming widely used as they are an environmentally friendly way for energy production without emissions; however, they are exposed to a corrosive environment. In addition, as wind turbines typically are the tallest structures in the surrounding area of a wind farm, it is expected that they will attract direct lightning strikes several times during their operating life. The purpose of this paper is to show that the radiography with a transportable unit is a solution to find defects in the wind turbine blade and reduce the cost of inspection. A transportable neutron radiography system, incorporating an Sb–Be source, has been simulated using the MCNPX code. The simulated system has a wide range of radiography parameters.

1. Introduction

At the end of 2008, worldwide nameplate capacity of wind-powered generators was approximately 121 GW [1]. Today the largest wind turbines generate 6 MW and involve glass fibre reinforced plastics (GRP) blades up to 126 m in length. These blades will be subject to enormous stresses, especially in storm conditions. Damages to the wind turbine blades are relatively serious since the cost for replacements is really high and time-consuming repair service is needed. It is considered as an obstruction factor to the expansion of the wind turbine installation [2]. Wind turbine blades must be inspected during and after manufacture, and often during their service life, to ensure that their condition is suitable for their purpose.

Radiography is a powerful nondestructive method. Gamma ray radiography technique is frequently used, either on its own or as complementary to neutron radiography (NR), for the analysis of objects. Both methods can be used either in 2D mode to take radiographs or with a tomography setup in 3D mode. Radiography allows the detection of internal flaws and defects such as cracks, corrosion, inclusions, and thickness variations; consequently, it is a technique of growing importance to science and industry in order to determine structural defects [3, 4].

The most common gamma rays sources, which are commonly used for radiography purposes, such as 137Cs and 60Co, are made artificially. The gamma rays are emitted as line spectra of discrete energies and different relative intensities, which are the characteristic of the particular source. In this work, a transportable unit for radiography, based on the use of Sb–Be isotopic source incorporated within it, has been simulated using the MCNPX Monte Carlo code [5]. The aim is to optimize the design of the unit with reference to its collimator and shielding, rendering it suitable for quality nondestructive testing, while ensuring appropriate radiation protection and safety standards for the personnel in the neighborhood of the system.

2. Materials and Methods

2.1. The Sb–Be Source

Fantidis et al. have showed that it is possible to design a transportable unit which incorporates an Sb-Be source, which is capable to provide a wide range of values for the parameters characterising the neutron and gamma radiographies. This means that the unit is capable to produce radiographies of variable quality [6]. For gamma radiography, the unit uses only the 124Sb source. The 124Sb source ( days) emits a large number of photons with energy higher than common isotopic source like 137Cs and 60Co. Just like other gamma sources it is portable; it can be operated without the electricity or cooling water. The gamma ray spectrum from the decay of 124Sb has been experimentally derived by Patil et al. and has been used for the intentions of this paper [7]. According to this spectrum, photons with energy 603 keV and 1691 keV have the highest intensity. The 1.691 MeV gamma-line of 124Sb can be suitable for gamma radiography for bulky dense objects, for example, the wind turbine blades. Gamma radiation is emitted continuously and cannot be switched off, which means that for reasons of safety the 124Sb source must be kept in a special container shielded usually with lead (Pb). The design offers the possibility to lift the 124Sb source and the lid () which covers it, by a simple remote control mechanism (Figure 1).

The Sb-Be neutron source effectively comprises two parts: the 124Sb isotope and a Be target. Neutrons are formed from the interaction of the photons emitted from 124Sb with the Be by means of the photoneutron production reaction 9Be () 8Be. The threshold energy of the reaction is approximately 1.667 MeV. The source emits neutrons with a soft spectrum having an average energy at 24 keV and a small fraction of fast neutrons. Bücherl et al. have simulated, using the MCNPX code, the neutron spectrum with an excellent agreement to experimental results [8, 9]. The thickness of the Be target for maximum neutron yield was calculated by Monte Carlo simulations using [10] where is the neutron yield per gamma quanta emitted in the 124Sb source, is the photoneutron production cross section, is the total photon scattering and absorption cross section, and are the density and thickness of the Be target [10]. The number of produced neutrons equals

In accordance with (1), the neutron yield is proportional to the thickness of the Be. Nevertheless, an augment of the thickness would be accompanied by an augment of the total dimensions and weight of the system. Hence, the choice of the Be is a compromise between beam intensity and transportability for insitu testing. The Sb–Be source has a low cost in setup and maintenance in comparison with low power (research) reactors or accelerator-driven neutron sources. In addition, the Sb–Be source has the advantage to offer an on/off switching capability (compared with radioisotopic sources, e.g., 252Cf) since the subtraction of the Be target ceases the production of the neutrons. This makes the transportation of the source easier since only shielding for the gamma emitted by 124Sb would be required. The source has one main disadvantage when used for NR: the high specific gamma activity of 124Sb. This can be overcame using gamma insensitive neutron detection plates [11, 12] and with image filtering in digital NR [9].

2.2. Simulation of the Radiography Unit

The radiography layout consists of two main units: the attachable transport and shielding container for the 124Sb source and the radiography source container. Both parts are designed in the form of a sphere with a radius of 30 and 13 cm, respectively. Side view through the centre of the unit is shown in Figure 1. The 124Sb part of the source has been simulated as a sphere with 1 cm radius and is symmetrically placed at the centre of the bigger sphere (POS1) which is the source container when not used for radiography purposes. This sphere is made of Pb providing the necessary shielding against the photons from the 124Sb source which is considered yielding a  Bq. When the unit is in use, the 124Sb source is placed in position 2 (POS2). During use of the unit, the conic shape shielding sectors S1 and S2 can be removed to allow placing collimators in order to provide a path for the photons and neutrons from the source towards the objects being imaged (Figure 2). The second sphere was a combination of two materials: an 11 cm thick layer of Pb in order to stop the gamma rays and a 1 cm layer of borated polyethylene (PE-B) with the purpose to eliminate any thermal neutrons diffusing out of the unit.

Figure 3 shows the unit with a neutron and a gamma collimator in positions S2 and S1, respectively. The use of the 124Sb source for gamma radiography requires a simple collimator which is shown in position S1. A divergent collimator made of 7 cm thick Pb walls was used. The first 11 cm of the collimator wall were provided by the Pb spherical unit itself while the total length could be extended in proportion to the application. A good quality image is required to be sharp, and the geometric unsharpness, , for significant object thicknesses can be expressed by the formula: where is the source focal-spot size, is the distance from the source to front surface of the object, and is the thickness of the object. Codes and standards used in industrial radiography require that the is as small as possible. In general, the allowable amount is 1/100 of the material thickness up to a maximum of 0.1 cm [13].

The collimator ratio () determines the quality of the NR imaging for a given design of the collimator, which is governed by: where is the image surface to object distance, is the source to object distance, is the inlet aperture diameter, is the neutron flux at the image plane, is the neutron flux at the aperture, and is the geometric unsharpness. The beam divergence, a significant measure of the usefulness of the beam near its periphery, is described by its half-angle () given by [14] where and are the maximum dimension of the image plane and the length of the collimator. The imaging quality of a system would be further characterized by the Thermal Neutron Content (TNC), which describes the number of thermal neutrons within the neutron beam and the relative intensities of the neutron () and the photon () components of the beam (). Hawkesworth [15] has shown that the least recommended value is about 500 n cm−2 mSv−1. However, while digital radiography would allow fast image-processing (spatial, average, median, and sharpening filters) [9] and with the development of new gamma insensitive neutron detection plates, a ratio less than this can also be justified. The use of the system for NR demands a dedicated collimator which is shown in position S2 (Figure 3). Next to the 124Sb source, there is the Be target (1) with dimensions 2.3 cm × 5 cm × 5 cm (in , , axis, resp.). According to (1), a  Bq 124Sb source is considered yielding  n/s. The photoneutrons formed, with an average energy of about 24 keV, are thermalised using 1 cm thick high density polyethylene (HDPE), providing a maximum thermal neutron flux at the collimator inlet aperture.

The collimator being used for NR comprises two parts. The first is a rectangular HDPE (3) with dimensions 5 cm × 11 cm × 5 cm (in , , axis, resp.). Inside of them, there is a void conic collimator with length 11 cm, radii 3 cm, and 0.5 cm, with the larger radius near to the source. Calculations with MCNPX were carried out in order to determine the void collimator length, which would give the maximum thermal neutron flux at the object. The length was determined to be 11 cm.

Next to the rectangular HDPE and the void collimator there is a divergent collimator, determining to a great extent the quality of the image for a given source type. The material used in the design of the collimator should prevent stray and scattered neutrons from reaching the object through absorbing them, hence improving the unsharpness of the obtained image. In this respect, the inner lining of the collimator is particularly important and it should be made of a neutron-absorbing material [14]. The lining is composed of a 1 cm layer of boron. The borated polyethylene (PE-B) with 3 cm depth was chosen among the available materials as a filling material in the collimator, since it provides a good shielding against stray neutrons. 1 cm thick Bismuth (Bi) layer was chosen as the collimator casing. The most commonly used types of composite material in the wind turbine industry are GRP. GRP dominates the market because it provides the necessary properties at a low cost. Glass fibre is manufactured from glass manufacturing materials such as silicate, colemanit, aluminum oxide, and soda. Glass fibre is the most commonly known and used one among fibre contributed composites. On wind turbine blades E-Glass type is used. The chemical composition of E-glass is SiO2 54.3%, Al2O3 and Fe2O3 15.2%, CaO 17.3%, Na2O/K2O 0.6%., and B2O3 8% to 10% [1618]. Properties that have made E-glass so popular in fibreglass include low cost, high production rates, high strength, high stiffness, relatively low density, non-flammable, resistant to heat, good chemical resistance, relatively insensitive to moisture, able to maintain strength properties over a wide range of conditions, and good electrical insulation. The advantageous properties of E-glass generally outweigh the disadvantages which include low modulus, self abrasiveness (if not treated appropriately), relatively low fatigue resistance, and higher density compared to carbon fibres and organic fibres [19].

3. Results and Discussion

Wind Turbines will be installed on a tower which is between 50–135 m tall, have diameters (blade span) from 40 to 126 m and are rated between 500 kW–6 MW [20, 21]. As a consequence, the minimum distance from blade to ground is about 30 m. Turbine towers are constructed from rolled steel plate and are normally about 4 to 5 m diameter at the base and about 2 to 3 m diameter at the top. A portable NDT scanning system is required to detected internal defects in the wind turbine.

3.1. Dose Results/Requirements

The system was designed under the constraint that the Dose Equivalent Rate (DER) would not overcome the annual occupational dose limit of 0.5 Sv (or 25 μSv·hr−1). When the 124Sb source is situated in the POS1, the total dose rate, due to the photons, was calculated with the MCNPX Monte Carlo code, using the F2, Fm2 tallies and the DE, DF cards. The F tallies describe the flux within a cell, while the D cards convert the absorbed dose to equivalent dose. Calculations were performed with histories yielding an accuracy <1%. The resulting total DER is 21.2 μSv·hr−1. When the 124Sb source is in POS2, the total DER is about 18.7 μSv·hr−1 and comprises the photon and neutron components with values of 18.7 μSv·hr and 0.02 μSv·hr−1, respectively. The induced photons from the interaction of the neutrons and the PE moderator material result in a dose which is about three orders of magnitude less than the dose from the 124Sb photons.

Figure 4 shows the resulting gamma dose rates at the surface of the system, and at a distance of 30 m, respectively, for different thicknesses of the Pb shielding surrounding the 124Sb source. Hence, the dimensions of the proposed unit, satisfying the occupational dose limit, constraint are approximately 60 cm × 60 cm × 105 cm (), and the overall weight is about 1600 kg. Is much easier, safer and inexpensive this unit with a combination of a simple detectors array to lift in the same height with turbine blade comparatively to move the blades to an inspection laboratory in the ground.

3.2. Gamma Radiography Results

The proposed system was simulated, with the MCNPX code, for gamma radiography, making use of the fact that the 124Sb is a strong photon emitter. Photon fluxes were calculated using the F2 tally which gives the required photon flux averaged over a surface in photons cm−2 per starting photon. Calculations were carried out with histories yielding an accuracy <1% (Table 1). The system has been considered with a variable collimator length ( cm), source to object distance ( cm), and divergence angle (). With the purpose to determine the ,the thickness of the object was assumed to be 20 cm. A source focal-spot size () of 1 cm was considered in order to keep below the recommended value of 0.1 cm. The collimation system has been designed to obtain a spot of 30–60 cm in diameter of the photon beam at the object position.

The variation of the photon flux at the field of view at the object position was less than 1%. The photon neutron flux varies from up to  p cm−2 s−1 for a maximum activity of the 124Sb ( Bq). In keeping with Table 1, the uncollided photon flux which characterizes the beam quality has excellent values from 99.22 up to 99.27%.

3.3. Neutron Radiography Results

In terms of the thermal NR () the proposed unit has collimator with a variable collimator length ( cm), variable diameter of its aperture next to the image plane ( cm), changeable divergence angle of the beam (), and an inlet aperture () of the collimator of 1 cm. The distance between the object and the imaging detector () was considered at 0.5 cm [15]. The variation of the thermal neutron flux at the field of view at the object position was less than 3%. The TNC and () parameters calculated are shown in Table 2 for different collimator parameters. The neutron flux was calculated with the aid of the MCNPX code using the F2 tally. Calculations were executed with histories for neutrons yielding an accuracy <1.5%. An energy boundary of 0.01–0.3 eV was used to score the thermal neutron flux. The gamma dose in the ratio was calculated with F2, Fm2 tallies and the DE, DF cards. Calculations were performed for yielding an accuracy in the calculations of <2%.

According to Table 1, the parameter remains in all circumstances above the recommended limit. The varies from up to  n cm−2 s−1 for a maximum neutron yield. The TNC varies from 11.7–14.9 % and the range from up to  n cm−2 mSv−1. Good-quality thermal neutron images require exposures of the order of 106 n cm−2 [22], with the exposure time being proportional to the thermal neutron flux. The exposure time varies between 3.2 and 68.7 min for 25 and 70, respectively.

Figure 5 shows two characteristics radiographic images of a typical turbine blade using the gamma rays (a) and neutrons (b) from the proposed unit. These figures were plotted from the MCNPX Visual Editor. In the blade, there are two defects, a small cylindrical void (1) with dimensions of 0.05 cm radius and 1.5 cm height and a small sphere from water (2) with radius 0.2 cm. The void is distinguishable in both figures but the water sphere is distinct only in the second figure; this is due to the much higher linear attenuation coefficient of hydrogen for thermal neutrons compared to gamma rays. From a number of radiography images obtained at different rotation angles of the sample, a 3D representation of the object can be computed. The method is known as neutron computed tomography [23, 24]. Computed tomography with its extremely high contrast resolution (ability to measure small density differences) and high spatial resolution (ability to resolve fine structural details) makes attractive for critical inspection requirements of turbine blades, where failure would result in extremely high costs and possibly lives.

4. Conclusions

A transportable system using an Sb–Be source has been simulated to provide access to wind turbine blades, with gamma and neutron radiography purposes, using the MCNPX Monte Carlo code. The Sb–Be source has some considerable advantages like low cost and easy transportation and on/off switching of the neutron source. Appropriate collimators have been simulated for the two radiographies options. Radiography with photons and neutrons made possible in extracting internal features of blade including inaccessible areas nondestructively, which is not possible through conventional method.