The reliable operation of critical infrastructure, such as hydropower plants, hinges on the integrity of their core components. Among these, large-diameter butterfly valves, responsible for regulating massive water flows, play a pivotal role. The valve body and disc, typically manufactured as heavy steel castings, are subject to immense hydraulic pressures and cyclical stresses. Internal defects within these castings, if undetected, can lead to catastrophic failures. Therefore, rigorous non-destructive testing (NDT) is indispensable, with Ultrasonic Testing (UT) being the primary method for volumetric examination of internal flaws. This article details a comprehensive ultrasonic testing methodology developed for large hydropower butterfly valve steel castings, focusing on probe selection, procedure development, and defect characterization.
Introduction to Cast Steel and Associated Testing Challenges
Cast steel castings offer the design flexibility and mechanical properties required for large, complex components like valve bodies. However, the casting process inherently introduces microstructural heterogeneity. Unlike wrought steels with a uniform, fine-grained structure, steel castings often exhibit coarse, dendritic grain structures, shrinkage porosity, gas pores, inclusions, and potential hot tears. These features present unique challenges for ultrasonic inspection.
The primary challenge is ultrasonic attenuation and scattering. Attenuation refers to the loss of sound energy as it propagates through the material. In steel castings, attenuation is significantly higher than in wrought steel due to two main factors: scattering at the boundaries of coarse grains and absorption within the material. Scattering redirects ultrasonic energy away from the original beam path, reducing the energy available to detect defects. This phenomenon is frequency-dependent, as described by the scattering coefficient, which generally increases with frequency. Consequently, the standard high-frequency (e.g., 4-5 MHz) probes used for wrought steel inspection are often unsuitable for steel castings, leading to poor signal-to-noise ratios (SNR) and excessive background “grass.”
Furthermore, the geometry of steel castings for butterfly valves often includes thick sections and complex contours. The inspection is typically performed on machined surfaces (e.g., flanges, valve bore), requiring a procedure capable of detecting both deep-seated and near-surface flaws. A single-probe approach is insufficient; a composite strategy utilizing both single-crystal and twin-crystal longitudinal wave straight probes is necessary to achieve the required inspection coverage and sensitivity.
Ultrasonic Testing System and Probe Selection
The development of an effective UT procedure begins with the selection of appropriate equipment. A digital ultrasonic flaw detector with excellent linearity (vertical linearity error ≤ 3%, horizontal linearity error ≤ 0.1%), wide bandwidth (0.5-30 MHz), and sufficient gain (0-120 dB) is essential for analyzing the complex signals from steel castings.
Selection of Single-Crystal Longitudinal Wave Straight Probes
Given the high attenuation in steel castings, lower frequency probes are preferred. The optimal frequency and crystal size are determined empirically based on the casting thickness and the measured attenuation characteristics of the material. Through comparative trials on reference blocks and representative casting sections, the following guidelines for single-crystal probe selection were established:
| Casting Thickness Range (mm) | Single-Crystal Probe Nominal Frequency (MHz) | Round Crystal Diameter (mm) |
|---|---|---|
| 50 – 100 | 3.0 – 3.5 | φ20 – φ30 |
| 100 – 200 | 2.5 – 3.0 | φ20 – φ30 |
| 200 – 400 | 2.0 – 2.5 | φ20 – φ30 |
| 400 – 600 | 1.5 | φ20 – φ30 |
This selection balances penetration power (better with lower frequency) and resolution (better with higher frequency) for different thickness ranges of the steel castings.
Selection of Twin-Crystal Longitudinal Wave Straight Probes
Inspecting for near-surface defects and in thin sections is a critical requirement. Single-crystal probes have a large dead zone and initial pulse width, making them ineffective for this purpose. Twin-crystal (TR) probes, with separate transmitting and receiving crystals angled towards a focal point, offer a very short dead zone and excellent near-surface resolution.
A comparative study was conducted between two 2.5 MHz, φ20 mm, 20 mm focal depth TR probes with different piezoelectric materials: conventional Lead Zirconate Titanate (PZT) and a 1-3 Piezocomposite (polymer matrix with piezoelectric ceramic rods). They were used to scan a set of flat-bottom holes (FBH) of φ3 mm at depths (L) from 5 mm to 45 mm.
The 1-3 composite probe demonstrated superior performance. It clearly resolved the FBH at 5 mm and 10 mm depth with a high signal-to-noise ratio. The conventional PZT probe struggled to identify the 5 mm FBH and showed a consistently lower SNR for all holes. The 1-3 composite’s high damping factor and lower mechanical Q result in a shorter pulse duration, reducing the initial pulse width and improving near-surface flaw detection. Therefore, 1-3 composite twin-crystal probes are specified for inspecting near-surface regions of butterfly valve steel castings.
UT Procedure Development and Key Technical Points
1. On-Site Measurement of Attenuation Coefficient
A fundamental step in inspecting steel castings is determining the material’s attenuation coefficient (α). This coefficient, measured in dB/mm or dB/m, is crucial for setting and compensating sensitivity during testing. It is measured on a sound, parallel, machined area of the casting using the backwall echo method. The difference in decibels between two consecutive backwall echoes is measured, and the attenuation coefficient is calculated using the formula:
$$ \alpha = \frac{(B_n – B_{n+1}) – 6 \, \text{dB}}{2 \cdot T} $$
Where:
– $B_n$ and $B_{n+1}$ are the gains (in dB) required to bring the $n^{th}$ and $(n+1)^{th}$ backwall echoes to the same reference height.
– $T$ is the thickness of the test area.
– The 6 dB term accounts for the theoretical 6 dB drop per reflection from a perfect planar reflector.
This measurement should be taken at three different locations on the casting, and the average value is used for subsequent calculations.
2. Inspection with Single-Crystal Straight Probes
The sensitivity setting method depends on the thickness (T) relative to the probe’s near-field length (N). The near-field length for a circular crystal is given by:
$$ N = \frac{D^2 f}{4c} $$
Where $D$ is the crystal diameter, $f$ is the frequency, and $c$ is the longitudinal wave velocity in the steel casting.
a) Distance Amplitude Correction (DAC) Curve Method (for T < 3N):
A DAC curve is established using a reference block with flat-bottom holes (e.g., φ3 mm or as per specification). The block material should closely match the steel castings in composition, heat treatment, and attenuation. The probe is placed over FBHs at various depths, and the gain is adjusted to bring each echo to 80% of the screen height. These points are connected to form the DAC curve. The reference sensitivity is this curve. The scanning sensitivity is set by increasing the gain by 6 dB above the reference curve. During evaluation, the gain is reduced by 6 dB, and compensation for the measured attenuation coefficient (α) is applied based on the flaw depth.
b) Backwall Echo Method (for T ≥ 3N):
For thick sections, sensitivity is set based on the theoretical reflectivity of a disc-shaped flaw. The gain ($G$) required to detect a FBH of a given diameter ($d$) at a depth equal to the casting thickness is calculated:
$$ G = 20 \log \left( \frac{2 \lambda S}{\pi d^2} \right) + 20 \log \left( \frac{V_{BWE}}{V_{Ref}} \right) – 2\alpha S $$
Where:
– $\lambda$ is the wavelength.
– $S$ is the sound path (casting thickness).
– $V_{BWE}$ is the amplitude of the backwall echo from a sound area.
– $V_{Ref}$ is a reference amplitude.
– $\alpha$ is the attenuation coefficient.
– The term $20 \log \left( \frac{2 \lambda S}{\pi d^2} \right)$ represents the reflection difference between a FBH and a large planar reflector. In practice, simplified formulas or nomograms from standards like NB/T 47013.3 are often used. The scanning sensitivity is this calculated gain increased by 6 dB.
3. Inspection with Twin-Crystal Straight Probes
The DAC method is exclusively used for twin-crystal probes. Using a set of FBHs at different depths in a suitable reference block, a DAC curve is plotted. The scanning sensitivity is set 6 dB above this curve. During manual scanning, the probe must be moved perpendicular to the separator between the crystals, with a maximum scanning speed of 150 mm/s.
4. Scanning and Defect Evaluation
All accessible machined surfaces are scanned. A raster scanning pattern is employed, with scans performed in two perpendicular directions. Each pass should overlap the previous by at least 15%. Defect evaluation combines analysis of the static A-scan waveform and the dynamic behavior of the echo as the probe is moved.
Defect acceptance criteria are defined by the applicable specification (e.g., customer requirement, ASTM A609/A609M). Typically, for DAC-based inspection, a defect is considered rejectable if its echo exceeds the DAC curve and is non-point in nature, or if its indicated area exceeds a specified limit. For the backwall echo method, a flaw causing a backwall echo amplitude loss of 12 dB or more is generally recorded for evaluation. The characterization of defect type, based on waveform analysis, is a critical part of the evaluation process for steel castings.
Characterization of Typical Clustered Defects in Butterfly Valve Steel Castings
Defects in steel castings often occur in localized clusters due to solidification dynamics. Identifying these clustered defects by their ultrasonic signature is vital. Below are four common types, their waveform characteristics, and underlying physical mechanisms.
1. Stomatal (Gas Pore) Aggregation
Waveform Feature: Multiple, relatively low-amplitude echoes with sharp, distinct peaks that often appear linked or clustered. The backwall echo is generally still present, though it may be slightly reduced.
Mechanism: Gas pores are typically spherical or near-spherical cavities with smooth inner walls. A spherical pore acts as a point reflector. The reflected sound pressure ($P_{sph}$) from a spherical pore of diameter $d$ is proportional to its diameter, whereas the reflection from a flat disc ($P_{disc}$) of the same diameter is proportional to the square of the diameter. This relationship can be approximated as:
$$ \frac{P_{sph}}{P_{disc}} \propto \frac{d}{d^2} = \frac{1}{d} $$
This explains the relatively low echo amplitude. Furthermore, the spherical shape scatters sound energy in many directions, so only a small portion returns to the receiver crystal. The sharp peaks result from the clean, smooth gas/metal interface.
2. Island-Type Lamination/Cold Shut
Waveform Feature: A high-amplitude, steep, and sharp echo, often accompanied by multiple repeating echoes (due to ringing within the planar discontinuity). The backwall echo is severely reduced or completely absent. The echo is continuous over an area with irregular boundaries, like an “island.”
Mechanism: This defect is a planar separation within the casting, often parallel to the surface. It acts as an excellent, large planar reflector, resulting in a very strong, mirror-like reflection. The high acoustic impedance mismatch (steel/air or steel/oxide) causes most of the sound energy to be reflected, severely attenuating or blocking the sound path to the backwall. The repetitive echoes are caused by the sound wave bouncing between the lamination and the front or back surface of the casting.
3. Shrinkage Cavity Aggregation
Waveform Feature: One or two main “bundled” or broad, branching echo clusters with high amplitude. Numerous smaller satellite echoes often surround the main cluster. The backwall echo is significantly reduced or lost.
Mechanism: Shrinkage cavities form during solidification as the metal contracts, leaving an irregular, dendritic cavity often filled with air. The rough, complex internal geometry creates multiple reflection paths from different facets of the cavity within the same wavefront, resulting in a broad, bundled main echo. The surrounding micro-shrinkage or gas pores associated with the main cavity cause the smaller satellite echoes. The large cavity volume effectively blocks ultrasound transmission.
4. Slag Inclusion Aggregation
Waveform Feature: A “bundled,” lower and broader main echo with a rounded top and multiple small, low-amplitude echoes in the vicinity. The echo amplitude fluctuates considerably with probe movement. The backwall echo shows a significant drop.
Mechanism: Slag inclusions are non-metallic materials (oxides, silicates) entrapped in the steel. The acoustic impedance of slag is different from steel but not as drastically different as air. This results in partial transmission and partial reflection, explaining the typically lower and broader echo compared to a pure cavity. The aggregated, irregular shape and the presence of numerous small inclusions lead to the fluctuating, “jumpy” echo behavior and the surrounding grass-like signals. The distributed nature of the inclusions scatters and absorbs sound energy, causing the backwall echo loss.
| Defect Type | Typical Waveform Features | Backwall Echo Condition | Primary Physical Cause |
|---|---|---|---|
| Stomatal Aggregation | Low, sharp, linked peaks | Slightly reduced or present | Entrapped gas bubbles (spherical reflectors) |
| Island Lamination | High, sharp, with repetitions | Severely reduced or absent | Planar solidification discontinuity |
| Shrinkage Cavity | High, broad, bundled main echo with satellites | Severely reduced or absent | Irregular cavity from volumetric contraction |
| Slag Inclusion | Low/Med, broad, fluctuating echo with grass | Significantly reduced | Aggregated non-metallic impurities |
Conclusion
Ultrasonic inspection of large butterfly valve steel castings demands a specialized approach tailored to the material’s challenging acoustic properties. A composite procedure utilizing both single-crystal and 1-3 piezocomposite twin-crystal longitudinal wave probes is essential for effective detection across the entire thickness range, from near-surface to deep interior. Critical to the procedure’s success is the empirical selection of probe frequency based on thickness and attenuation, the accurate on-site measurement of the attenuation coefficient, and the appropriate application of sensitivity setting methods (DAC or backwall echo). Furthermore, the ability to recognize and characterize the ultrasonic signatures of common clustered defects—stomatal aggregates, laminations, shrinkage cavities, and slag inclusions—is a crucial skill for inspectors. By implementing this rigorous, physics-based UT methodology, manufacturers and inspectors can ensure the internal integrity of these critical hydropower components, guaranteeing that only sound steel castings proceed to the final machining and assembly stages, thereby upholding the safety and reliability of the entire hydraulic system.