As a key component of an aero engine, hollow turbine blades are known as the "jewel in the crown" due to their characteristics such as frequent temperature changes, complex stress conditions, harsh operating environments, high working reliability requirements, and long stable working life. Generally, hollow turbine blades are formed into a small and complex cavity with the help of ceramic cores, but the core removal process is difficult to ensure complete removal of the ceramic cores. The residual cores in the cavity will not only affect the normal flow of cooling airflow in the blades during operation, but also Reduce the cooling effect, and may block hundreds of air film holes (about 0.3~0.5mm in diameter) used for cooling on the blade, causing local high temperature and premature failure of the blade, bringing huge safety hazards to the engine.

Based on this research background, the researchers from the Beijing Aviation Materials Research Institute of China Aviation Development and the Institute of Nuclear Physics and Chemistry of the Chinese Academy of Engineering Physics are the basis for verifying the optimal spatial resolution of the combination of neutron source and single-sided emulsion film. Above, the DD5 nickel-based superalloy material was used to design and manufacture stepped simulation specimens with different thicknesses of residual cores and blade comparison samples. The detection sensitivity of the hollow turbine blade residual cores and the relationship with various influencing factors were studied, with a view to the residual cores The neutron detection provides a reference basis.

1

The interaction between neutrons and matter is very complicated. The interaction between neutrons and matter has nothing to do with electrons, but is related to the nucleus of matter and its microstructure. After neutrons pass through a certain substance and interact with the nucleus of the substance, the energy decays. From a macro perspective, the following mathematical relationship exists between the intensity of transmitted neutrons and the intensity of incident neutrons:

I =I0 e-∑t (1)

In the formula: I is the intensity of transmitted neutrons; I0 is the intensity of incident neutrons; t is the transilluminated thickness of the sample; Σ is the absorption coefficient of the sample for neutrons (macro cross-section).

Among them, the absorption coefficient of Gd is 1190/cm, and the absorption coefficient of Ni is 0.42/cm. The difference in the absorption coefficients of these two materials will form a significant contrast difference on the neutron image, which is conducive to high-sensitivity detection of residual cores.

2

试验设备为中国工程物理研究院核物理与化学研究所的反应堆冷中子照相装置，冷中子束流由反应堆内冷包将热中子冷却后获得，通过中子导管传输至成像装置，在满功率情况下，成像位置的中子注量率可达10⁶n·cm^-2·s^-1，准直比为300~12000。试验成像装置包括单面乳化胶片、Gd转换屏和暗盒等。所用胶片型号为Agfa D3-SC，与转换屏紧贴一起置于不漏光的暗盒内，被中子射线直接曝光。

空间分辨力是用来表征图像上分辨两个相邻细节特征的指标。采用瑞士PSI公司的标准空间分辨力试样，该试样材料为表面镀有Gd层的铜片，铜片呈放射状排列，可定量评价25，50，100μm等空间分辨力指标。

对比试样对模拟评价叶片残芯的检测灵敏度至关重要。科研人员设计制作了两种类型的对比试样：阶梯状模拟试样用于确定模拟残芯的检测灵敏度；叶片对比试样除了获得残芯的真实检测灵敏度外，还用于分析叶片结构对检测灵敏度的影响。

如图1所示，基体材料为DD5单晶高温合金，台阶厚度分别为2，3，4，5，6，7，8mm，每个台阶上采用电火花加工了一系列不同深度的Φ2mm圆孔，如图1(a)所示，孔深依次为0.10，0.15，0.20，0.25，0.30，0.35，0.40，0.50，0.60，0.70，0.80，0.90mm。所有圆孔经三坐标测量后，实际孔深约为0.14~0.95mm。采用掺杂法将纯度为99.9%的Gd2O3粉末与陶瓷型芯浆料混合均匀，Gd2O3的质量分数为3%，将浆料填充在不同深度的圆孔内，并在1150℃烧结0.5h，固化后通过机加工去除试样表面溢出的型芯材料，同时去除每个圆孔内一半的型芯，形成空腔，去除一半型芯后的试样如图1(d)所示，圆孔中白色物质为型芯材料。

如图2所示，叶片材料为DD5单晶高温合金。采用与制作阶梯状模拟试样相同的陶瓷型芯浆料，通过3D打印技术制成长5mm、宽5mm的薄片，厚度分别为0.2，0.3，0.4mm。将薄片破碎后，挑选尺寸较小的残芯，通过铝胶带黏贴在叶片叶身处，每3个为一组；选择高度分别为38mm和85mm的两种规格叶片(记为小叶片和大叶片)，两种叶片叶身的透照厚度都约为3.4mm，在两种叶片叶身的不同位置黏贴残芯，其中小叶片共3组(都位于叶身表面)，大叶片共4组(3组位于叶身表面，1组位于叶片内腔)。

Place the samples close to the cassette and place them in a cold neutron source and single-sided emulsified film imaging system for 30 min exposure. After the film is developed, fixed and dried in a dark room, it is digitized using an Array 2905HD laser scanner, and the image type is converted to 8bit using Image J software, and then the gray scale of different parts is extracted.

3

As shown in Figure 3, the image is processed by noise reduction and boundary sharpening. In Figure 3(a), draw two straight lines near the spatial resolution index of 50μm to obtain the gray distribution curve, as shown in Figure 3(b) and (c). Show. It can be seen from the test results that the peaks and troughs of the gray distribution curve can be clearly distinguished on the outer side of the index 50μm (line 1), while the crests and troughs cannot be distinguished on the inner side of the index 50μm (line 2), indicating that the cold neutron source and the single The combination of surface emulsified film has excellent spatial resolution, and the resolution can reach 50μm.

As shown in Figure 4, each step in the figure has a gray part, a lighter part, a darker part, etc. The gray part is the base material, the lighter part is the residual core, and the darker part is the cavity. It can be seen from Figure 4 that when the thickness of the residual core is the same, the contrast between the residual core and the cavity decreases as the thickness of the transillumination increases, and it is gradually difficult to distinguish on the image; when the thickness of the transillumination is the same, the residual core and the cavity The contrast of the cavity gradually decreases as the thickness of the residual core decreases, and it is also difficult to distinguish on the image. The above results indicate that when the residual core material is the same, the detection sensitivity of the residual core depends on the transilluminated thickness and the residual core thickness. Use the image average gray level function of Image J software to extract the gray levels of the core and cavity, and the contrast C between the two is calculated by the following formula:

C=(D1-D)/D=ΔD/D       (2)

Where: D1 is the gray level of the core; D is the gray level of the cavity; ΔD is the gray level difference between the core and the cavity.

In Figure 5, the black ball indicates that the remaining core is not visible (C<0.05); the red ball indicates that the remaining core is visible (0.05≤C<0.1); the green ball indicates that the remaining core is clearly visible (C≥0.1). Figure 5 further shows that the detection sensitivity of the residual core is closely related to the thickness of the transillumination and the thickness of the residual core. The change in the thickness of the transillumination has a greater influence. When the thickness of the transillumination is 5.8mm, the detection sensitivity of the residual core is 0.2mm. As shown by the arrow A in the figure, as the thickness of the transillumination gradually decreases, the detection sensitivity of the residual core increases (less than 0.2mm; as shown by the arrow B in the figure, when the thickness of the transillumination is 6.5mm, the detection sensitivity of the residual core decreases. As shown by the arrow C in the figure, when the thickness of the transillumination increases to 7mm, the detection sensitivity of the residual core drops to 0.95mm).

In summary, the quantitative indicators obtained by stepped samples can be used to guide the cold neutron photographic detection of the residual core of the blade. However, there are still differences between the above quantitative indicators and the true detection sensitivity of the residual core of the blade as follows:

① For the 0.95mm thick residual core on the thickest step in Figure 4, it can be distinguished by visual observation, but the quantitative results show that it is invisible. This contrast grading may lead to low detection sensitivity;

② This quantitative index does not consider the influence of blade structure;

③ The quality of the digitally processed image is lower than that of the film, and the evaluation result is lower than the detection sensitivity of the residual core in the actual film imaging;

④ The above quantitative indicators are obtained based on Gd2O3 with a mass fraction of 3%. If the mass fraction of the attenuating material decreases, the detection sensitivity of the residual core will be lower than the above quantitative indicators and needs to be re-evaluated.

图6中黑色虚线表示叶身表面的残芯，黑色实线表示叶片内腔中的残芯。从图6可以看出，在两种规格叶片叶身上预置的不同厚度残芯（0.2~0.4mm）均能清晰分辨，表明残芯的检测灵敏度达到了0.2mm。两种叶片的透照厚度（约3.4mm）及其变化范围几乎相同，如图7所示，因此也可以通过上文中的定量指标推断出两种规格叶片残芯的检测灵敏度要优于0.2mm。另外，由于透照厚度的变化，图像上较薄部位的残芯对比度略高于较厚部位。

叶片叶身的横截面为变厚度曲面，且不同叶片之间的结构差异较大，文中大叶片结构的复杂性及曲率均大于小叶片的，这会对残芯的检测灵敏度产生一定的影响。

As shown in Figure 8, from a quantitative analysis, when the thickness of the residual core is 0.2~0.4mm, the image contrast is greater than 0.1, the residual core in the two specifications of the blade can be clearly distinguished, and the detection sensitivity of the residual core is better than 0.2mm. In addition, for blades of the same specification, different transillumination thicknesses result in a more consistent range of contrast variation of the residual core with the same thickness. For example, the contrast change of the residual core in a small blade is about 0.8, and the contrast change of the residual core in a large blade is about 0.3. Among them, the 0.3mm thick residual core in the large blade has low contrast, and the contrast is lower than 0.3, which is caused by the position of the residual core in the blade stiffener, as shown by the arrow in Figure 6(b). It can be seen from Figure 8 that the contrast of large blades is lower than that of small blades as a whole. Considering the uniformity of the thickness variation of the transillumination, the main reason for the decrease in contrast is the poor image quality caused by the complicated structure and large curvature of the blades.

in conclusion