The quality control of complex steel castings is a critical challenge in modern manufacturing. These components, often featuring intricate geometries and significant variations in wall thickness, are susceptible to internal discontinuities such as shrinkage cavities, gas porosity, and inclusions, which can compromise structural integrity. Non-destructive testing (NDT) is therefore essential. While ultrasonic testing is applicable, its effectiveness is often hindered in steel castings by complex geometries, coarse grain structures, and the nature of certain defects, leading to challenges in signal interpretation and potential blind spots. Conventional film-based radiography has been the traditional method but suffers from limitations including low inspection efficiency, high consumable costs, significant labor intensity, difficulties in long-term archival, and a limited thickness latitude. Digital Radiography (DR) with flat-panel detectors presents a transformative solution, offering a large dynamic range, excellent thickness latitude, rapid imaging, and the benefits of digital image processing and storage. This article details a comprehensive study on the development and validation of a DR inspection process tailored for a complex-shape steel casting, establishing its equivalence to conventional film radiography.
The foundational principle of Digital Radiography using amorphous silicon flat-panel detectors involves the direct conversion of X-ray photons into electrical signals. When X-rays pass through a steel casting, they are attenuated according to the material’s thickness and density. The transmitted radiation pattern, carrying information about internal structures, strikes the scintillator layer of the detector. This layer converts X-rays into visible light, which is then detected by a photodiode array made of amorphous silicon. Each photodiode (pixel) generates an electrical charge proportional to the intensity of the incident light. This charge is read out, digitized, and processed to form a two-dimensional digital image. The signal for a given pixel can be conceptually described by:
$$I_{output} = G \cdot \Phi \cdot \mu \cdot t \cdot e^{-\mu x}$$
Where $I_{output}$ is the output signal (digital number), $G$ is the system gain, $\Phi$ is the incident X-ray photon flux, $\mu$ is the linear attenuation coefficient of the steel, $t$ is the exposure time, and $x$ is the local thickness of the steel casting. The key advantage over film is the linear response over a very wide range of exposure, providing the large dynamic range crucial for inspecting steel castings with thickness variations.
Image quality in DR is assessed by several key metrics. Spatial resolution is often characterized by the Modulation Transfer Function (MTF) or practically by the smallest visible wire diameter in an image quality indicator (IQI). The normalized signal-to-noise ratio ($SNR_n$) is critical for defining the visibility of low-contrast details and is calculated from the image as:
$$SNR_n = \frac{\mu_{signal} – \mu_{background}}{\sigma_{background}}$$
Here, $\mu_{signal}$ and $\mu_{background}$ are the mean pixel values in a feature and the adjacent background, respectively, and $\sigma_{background}$ is the standard deviation of the pixel values in the background, representing noise. The basic spatial resolution ($SR_b$), often determined using a duplex wire IQI, is fundamentally limited by the detector’s pixel size ($p$), approximately: $$SR_b \approx \frac{1}{2p}$$ For a detector with a 100 µm pixel pitch, the theoretical limiting resolution is about 5 lp/mm.
The inspection object is a representative complex-shape steel casting with the following challenging characteristics:
| Feature | Dimension / Characteristic |
|---|---|
| Maximum Height | 380 mm |
| Maximum Length | 340 mm |
| Maximum Width | 250 mm |
| Wall Thickness Range | 11 mm to 28 mm |
| Key Features | Cylindrical sections, flanges, intersecting walls, and significant non-uniform thickness transitions. |
The primary inspection system configuration for the DR method consisted of the following components:
| Component | Model/Specification | Key Parameters |
|---|---|---|
| X-ray Source | YXLON / 200D | Max Voltage: 200 kV, Focal Spot Size: 1.0 mm |
| Flat-Panel Detector | VAREX / XRpad2 4336 (Amorphous Silicon) | Active Area: 320 x 410 mm, Pixel Pitch: 100 µm, Limiting Resolution: ~5 lp/mm |
| Image Processing | Proprietary Acquisition & Analysis Software | Functions: Frame Averaging, Contrast Adjustment, Filtering |
The conventional film radiography setup used for comparative analysis was configured as follows:
| Component | Model/Specification | Key Parameters |
|---|---|---|
| X-ray Source | X-ray Generator | Max Voltage: 350 kV, Focal Spot Size: 2.0 x 2.0 mm |
| Film & Processing | AGFA C7 Film with standard chemical processing | Film Class: C7 (Fine Grain) |
The detection sensitivity requirement for both methods was established according to the A-technique grade of GB/T 5677-2018 (equivalent to high sensitivity levels in ASTM E1030). For DR, additional image quality parameters including unsharpness (related to resolution) and normalized signal-to-noise ratio ($SNR_n$) were controlled as per the A-grade requirements of GB/T 35388-2017.
The primary challenge was the significant thickness variation (17 mm range) within a single casting. For film radiography, the limited latitude necessitates multiple exposures with different parameters to cover different thickness zones adequately. A total of 13 distinct exposure setups were required to inspect the entire steel casting using the film method. The parameters for a subset of these are summarized below:
| Position ID | Technique | Wall Thickness (mm) | Tube Voltage (kV) | Exposure (mA·min) | Required IQI Sensitivity (Wire) |
|---|---|---|---|---|---|
| 1, 4 | Single Wall | 28 | 230 | 17.5 | W10 |
| 2, 3 | Single Wall (Angled) | 19-23 | 220 | 17.5 | W11 |
| 12, 13 | Single Wall | 11 | 180 | 15.0 | W12 |
In contrast, the high dynamic range of the DR system allowed for a radical consolidation of views. The entire steel casting was inspected in only 4 primary exposures. The exposure parameters were optimized to ensure adequate $SNR_n$ and resolution across the thickness range in each view. The DR inspection parameters are detailed in the following table:
| View ID | Approx. Thickness Range (mm) | Tube Voltage (kV) | Current (mA) | Frames Averaged | Required Resolution (Duplex Wire) | Required $SNR_n$ | Target IQI Sensitivity |
|---|---|---|---|---|---|---|---|
| 1 | 20-28 | 150 | 5.0 | 32 | D7 | ≥ 70 | W11 / W10 |
| 2 | 19-28 | 150 | 5.0 | 32 | D7 | ≥ 70 | W11 / W10 |
| 3 | ~11 | 110 | 5.0 | 32 | D8 | ≥ 70 | W12 |
| 4 | ~20 | 140 | 5.0 | 32 | D8 | ≥ 70 | W11 |
The procedure involved meticulous setup for each view. The steel casting was positioned between the X-ray source and the detector. A 2mm thick copper filter was placed at the tube port to harden the beam and reduce scatter. Crucially, all areas of the steel casting and detector not involved in the primary beam path were shielded with lead to minimize the detrimental “flare” or “off-focus radiation” effect, which degrades image contrast. Frame averaging (32 frames per image) was employed to improve the $SNR_n$ by a factor of $\sqrt{32} \approx 5.66$, as per the relation: $$SNR_{averaged} = SNR_{single} \cdot \sqrt{N}$$ where $N$ is the number of averaged frames. Image processing involved standard optimization techniques like contrast stretching and application of noise-reduction filters to enhance defect visibility without creating artifacts.
The detection sensitivity was evaluated using wire-type Image Quality Indicators (IQIs). The results for both methods across comparable sections of the steel casting are summarized below. The achieved sensitivity met the stringent A-grade requirements of the standard.
| DR View | Corresponding Film Positions | Required Sensitivity (Film Std.) | Achieved Sensitivity (DR) | Achieved Sensitivity (Film) |
|---|---|---|---|---|
| 1 | 5,6,7,8,9,11 | W10, W11 | W11 (thin), W10 (thick) | W11, W10 |
| 2 | 1,2,3,4 | W10, W11 | W11 (thin), W10 (thick) | W11, W10 |
| 3 | 12,13 | W12 | W12 | W12 |
| 4 | 10 | W11 | W11 | W11 |
The measured image quality parameters for the DR inspection confirmed compliance with the digital standard. The basic spatial resolution was at or better than the required level, and the $SNR_n$ significantly exceeded the minimum threshold of 70 in all cases, ensuring excellent contrast sensitivity for the steel casting.
| DR View | Required Resolution | Achieved Resolution | Required $SNR_n$ | Achieved $SNR_n$ Range |
|---|---|---|---|---|
| 1 | D7 | D7 | ≥ 70 | 90 – 125 |
| 2 | D7 | D8 | ≥ 70 | 110 – 160 |
| 3 | D8 | D8 | ≥ 70 | 125 – 170 |
| 4 | D8 | D9 | ≥ 70 | 75 – 90 |
The most dramatic advantage of DR for this complex steel casting was in inspection efficiency. The total exposure and cycle time was calculated and compared.
- Film Method: Total exposure time = 30.5 minutes. With mandatory cooling time equal to exposure time, plus film handling, processing, and evaluation, the total inspection time per steel casting exceeded 90 minutes.
- DR Method: Total exposure time (4 views * 32 frames * 1s/frame) ≈ 128 seconds (~2.1 minutes). With setup and positioning, the total inspection time was approximately 15 minutes per steel casting.
This represents an efficiency gain, $E$, of: $$E = \frac{T_{film}}{T_{DR}} \approx \frac{90}{15} = 6$$ Thus, the DR process was more than 6 times faster for inspecting this specific steel casting geometry. This gain stems from: (1) the consolidation of views due to superior thickness latitude, (2) shorter single-exposure times, and (3) the elimination of chemical processing.
The superior thickness latitude of DR can be understood by considering the film’s characteristic (H&D) curve versus the detector’s linear response. Film has a narrow linear region between density 2.0 and 4.0 suitable for interpretation. The allowable thickness range $\Delta x$ for a constant exposure can be approximated from the attenuation law and film gradient $G_f$: $$\Delta x \approx \frac{\Delta D}{G_f \cdot \mu}$$ where $\Delta D$ is the allowable density range (~2.0). For steel at 180 kV, this $\Delta x$ is roughly 6 mm. For the DR detector, the usable signal range is limited by noise floor and saturation, typically corresponding to a thickness range $\Delta x_{DR}$ exceeding 20 mm for the same beam energy, as the system can utilize a much wider range of signal intensities linearly.
A direct comparison of defect detection capability was performed. Both methods reliably identified typical casting discontinuities such as shrinkage cavities and gas porosity. The digital images provided comparable defect characterization to the film radiographs. In some instances, particularly for clusters of small gas pores, the DR images revealed a slightly higher number of indications, attributable to the enhanced contrast manipulation capabilities and the ability to optimize different density regions within a single image of the steel casting. The consistency in detection between the two methods confirms that the DR system’s inspection capability and reliability for this application are equivalent to that of the AGFА C7 film system.
The following table summarizes the key performance metrics of the developed DR process for the complex-shape steel casting against the conventional benchmark:
| Metric | Digital Radiography (DR) Result | Conventional Film Result | Standard Requirement (A-Grade) | Status |
|---|---|---|---|---|
| Detection Sensitivity | W10 to W12 (per thickness zone) | W10 to W12 | GB/T 5677-2018 | Met |
| Basic Spatial Resolution | D7 to D9 | N/A (Film granularity limited) | GB/T 35388-2017 | Met/Exceeded |
| Normalized SNR ($SNR_n$) | 75 to 170 | N/A (Implied by density contrast) | GB/T 35388-2017 ($\geq$70) | Met/Exceeded |
| Inspection Time / Steel Casting | ~15 minutes | >90 minutes | N/A | >6x faster |
| Thickness Latitude per Exposure | >20 mm | ~6 mm | N/A | Significantly superior |
| Defect Detection Consistency | All relevant defects detected | All relevant defects detected | Technical agreement | Equivalent |
In conclusion, a robust and highly efficient Digital Radiography inspection process has been successfully developed and validated for a complex-shape steel casting. The process utilizes an amorphous silicon flat-panel detector and optimized exposure parameters to achieve image quality that meets or exceeds the relevant standards for both sensitivity and digital image parameters. The direct comparison with conventional film radiography demonstrates equivalent defect detection capability and reliability. The most significant operational benefits are the dramatic reduction in inspection time, enabled by the large thickness latitude and rapid imaging, and the elimination of film consumables and chemical waste. This DR process is therefore fully suitable for the quality control of such complex steel castings, offering a modern, efficient, and reliable alternative to traditional film-based methods.