Detection of High-Temperature Gas Leaks in Pipelines Using Schlieren Visualization

Abstract

This paper investigates the application of Schlieren flow visualization for detecting leaks in pipelines carrying high-temperature fluids. Two experimental setups were constructed: one with a 25 mm PTFE tube featuring a 2 mm diameter perforation, and another with a 100 mm diameter pipe insulated with an aluminum jacket and featuring a 12 mm leak gap. A single-mirror-off-axis Schlieren system, employing a 150 mm diameter parabolic mirror, was used to visualize the leaks. The temperature of the leaking air varied between 20 and 100 °C, while the ambient temperature was maintained at 14 °C. To quantify the leaks, the coefficient of variation for pixel intensity within the leak region was calculated. Results showed that for the PTFE tube, leaks became detectable when the temperature difference exceeded 34 °C, with the coefficient of variation surpassing 0.1. However, in the insulated pipe, detecting clear leak patterns was challenging. This research demonstrates the potential of Schlieren visualization as a valuable tool in enhancing pipeline leak detection.

1. Introduction

Accidents at nuclear power plants (NPPs) can have significant environmental and human health impacts, cause power outages, and increase plant maintenance costs. To prevent such occurrences, it is crucial to implement proactive measures. We can significantly reduce the risk of major accidents by detecting and addressing elemental component failures through diverse monitoring and diagnostic techniques. Monitoring water and steam pipelines, which constitute the largest portion of the plant, plays a pivotal role in this prevention. Effective monitoring not only helps prevent significant accidents, such as the release of radioactive materials or high vibrations in rotating machinery (e.g., turbine generators and main feedwater pumps), but also enhances the economic and operational efficiency of the plant [1,2,3].

Technologies such as acoustic detection, ultrasonic inspection, and infrared thermography are employed to monitor pipe leaks in NPPs [4,5,6,7,8]. However, these methods can be diminished in high-noise environments, making precise leak localization difficult. The installation of ultrasonic sensors directly on pipes represents a significant challenge, particularly in the case of insulated pipes, and is also a costly undertaking. Infrared thermography cannot detect leaks in their early stages when only small amounts of fluid escape. To address the limitations of existing technologies, techniques that utilize image processing of pipeline photographs for the visual and optical detection of leaks can be employed.

The Schlieren flow visualization technique has been the subject of extensive study due to its capacity to discern flows with varying temperatures through visual means. Schlieren visualization represents the refraction of light due to density differences, thereby enabling high-resolution observation of fluid movements with temperature variations [9,10,11,12,13]. Precise optical system design and alignment enable the detection of gas leaks with temperature differences of just a few degrees. Schlieren technology has been employed in many fields for the visualization of flows, particularly in detecting leaks of flammable or toxic gases such as hydrogen, methane, and ammonia, which are less dense than air [14,15,16]. Recent studies have reported the visualization of heat leaks through cracks in building walls or facades using this technique [17].

The application of Schlieren technology to the monitoring of pipes in NPPs allows for non-contact observation of extensive areas in real time. When integrated with artificial intelligence [18], Schlieren technology has the potential to significantly enhance the precision of leak detection. This advancement in technology instills confidence in the safety of NPPs. This study establishes an experimental configuration for the simulation of pipe leak scenarios in NPPs, with the objective of conducting Schlieren flow visualization experiments. The experiments are designed to encompass two scenarios: leakage of hot air through perforations and gaps in insulated pipes. The temperature of the leaking air is controlled, and Schlieren images are analyzed in accordance with the surrounding temperature disparities. To quantify the leaks, the coefficient of variation in pixel brightness values within the visible flow regions is calculated. An analysis of leak detection based on the coefficient of variation is conducted for both experimental setups. This study is expected to make a significant contribution to the safety management of NPPs by presenting a precise, efficient, and novel approach for pipe leak detection.

2. Methods

The Schlieren flow visualization technique is a method for observing changes in the refractive index resulting from variations in density within a medium through which light is passing. The refractive index is defined as the ratio of the speed of light in a vacuum to the speed of light in the medium. For an ideal gas with density ρ, the refractive index n is expressed as follows:

$$n-1=K\rho ={\displaystyle \frac{KP}{RT}}$$

(1)

where K is the Gladstone–Dale constant, which is a property of the medium and the wavelength of light [17]. P represents pressure, R is the specific gas constant, and T is the absolute temperature of the medium. A precisely aligned Schlieren optical system is capable of observing natural convection due to air temperature differences of less than 10 °C on a palm surface [10].

In this study, the Schlieren flow visualization technique was employed for the simulation and detection of pipe leaks. As illustrated in Figure 1, the pipe with leakage was situated in front of a mirror, and Schlieren images were captured using a knife edge and a camera. The light source and knife edge were positioned at twice the focal length of the mirror. When hot air leaks from the pipe, the path of light from the source is refracted and reflected by the mirror. Unrefracted light is blocked by the knife edge and thus not recorded by the camera, while refracted light is captured and appears as shadows in the flow visualization images.

In order to quantify the flow of gas in Schlieren images based on temperature differences, it is possible to utilize the brightness values of pixels representing the shadows [19]. Specifically, as the temperature difference between the leaking gas and the surrounding air increases, the shadow of the leaking flow becomes more distinct, leading to a more significant variation in the brightness of these pixels. Consequently, the coefficient of variation (I_E) in the flow region can be calculated using the following equation [20]:

$$I_{\mathrm{E}}={\displaystyle \frac{1}{\overline{I}}}\sqrt{{\displaystyle \frac{1}{N}}\sum (I-\overline{I}{)}^2}$$

(2)

where I and $\overline{I}$ are the brightness and average brightness of the observation area, respectively, and N is the number of pixels. I_E is the ratio of the standard deviation of brightness values to the average brightness value. This value increases as the clarity of the shadows, caused by the temperature difference between the leaking gas and the surrounding air, increases. Therefore, calculating I_E can serve as an indicator to determine the presence of gas leakage on the pipe surface.

The experimental setup for simulating leaks in high-temperature pipes comprises optical components, an image-recording device and a pipe that simulates the leak. To obtain Schlieren flow visualization images, a parabolic mirror with a focal length of 1.22 m and a diameter of 150 mm, an 11 W LED fiber optic light source, a knife edge, and a digital camera (GH6, Panasonic (Osaka, Japan)) equipped with a 200 mm focal length lens were utilized. The light source and knife edge were positioned at a distance of twice the mirror’s focal length. An iris was positioned before the LED light source to direct the light through a 1 mm diameter aperture onto the mirror.

To ensure precise alignment and calibration of the Schlieren system, we first verified that the iris of the light source and the knife edge were equidistant from the mirror. This was achieved by positioning the knife edge where the reflected light from the mirror appeared the smallest. Additionally, the camera position was finely adjusted to ensure that the beam path reflected from the mirror was evenly blocked by the knife edge, with the remaining half of the beam reaching the camera lens and sensor. The light source and knife edge were aligned within 100 mm for obtaining clearer Schlieren images.

Two distinct types of pipes were employed during the leak simulation experiments. The initial configuration, illustrated in Figure 2, comprises a pipe arrangement intended to emulate the phenomenon of hot gas leakage through perforations. A 2 mm hole was drilled in a 25 mm internal diameter polytetrafluoroethylene (PTFE) tube. Air with a controlled temperature was introduced using a hot air blower. The unperforated sections of the PTFE tube were covered with 10 mm thick cylindrical urethane insulation, which was employed to exclude the natural convection occurring on the surface of the PTFE tube from the Schlieren images.

The second leak simulation experiment employs a pipe assembly analogous to those utilized in operational NPPs. This assembly comprises a 100 mm outer diameter aluminum pipe encased in an aluminum jacket with insulation. The experimental setup is illustrated in Figure 3. The interior of the 250 mm outer diameter aluminum jacket is filled with fiberglass insulation. It is divided into two half-concentric cylinders that encase the internal pipe, secured together with clips. To simulate a gas leakage from the pipe and jacket, a 12 mm diameter hole was drilled into the pipe, and the aluminum jacket was configured with a gap approximately 12 mm wide, as illustrated in the right side of Figure 3. In contrast to the preceding PTFE tube leak simulation configuration, the thickness of the pipe and insulation necessitated an adjustment in the height of the Schlieren optical system by approximately 250 mm while maintaining the exact positioning. The pipe was positioned 200 mm in front of the mirror for both experimental setups.

The hot air supplied to the experimental apparatus was delivered using a hot air blower (Hotwind Premium, Leister AG (Kägiswil, Switzerland)). The airflow rate was fixed at the minimum setting of 200 LPM. The temperature of the leaking air was adjusted from 20 to 100 °C in 10 °C increments, with the temperature monitored by a K-type thermocouple installed at the perforated hole. The experiment was conducted during an unheated winter period, at an ambient temperature of 14 °C. Schlieren images were recorded at one-second intervals using a camera, resulting in 20 images for each temperature of the leaking air. The brightness values from the regions exhibiting leakage in the acquired images were extracted to calculate the coefficient of variation. The brightness analysis was conducted within a rectangular region measuring 25 × 12.5 mm, situated at the center of the flow emanating from the bottom center of the mirror, approximately 25 mm in height. Approximately 5000 pixels were analyzed to compute the mean and standard deviation, thereby facilitating a comparison of I_E. The subsequent section presents the Schlieren flow visualization images and image processing results based on the leakage temperature in the simulation setup.

3. Experimental Results and Discussion

Two experimental setups were devised using a Schlieren optical system to observe and detect leaks in high-temperature gas flow pipelines, such as those found in power plants. The initial configuration entailed a PTFE tube with a direct perforation. The outcomes, documented at varying temperatures of the leaking air, are illustrated in Figure 4. Since the laboratory air temperature was maintained at 14 °C, the actual temperature difference between the ambient environment and the leaking air ranged from a minimum of 6 °C to a maximum of 86 °C. At a leaking air temperature of 20 °C, the leak was only discernible to a slight degree. However, at temperatures exceeding 30 °C, the leak could be unambiguously identified.

As the temperature of the leaking air increased, the width of the leak expanded, and the shadows of the leaking flow became more distinct. While the flow patterns remained similar between 30 °C and 50 °C, more apparent shadow differentiation was noted at temperatures above 60 °C. The flow area of the leaking gas approximately doubled when comparing the flow at 60 °C to that at 100 °C. Given that the volumetric flow rate of the ambient air entering through the hot air blower was kept constant, the 40 °C temperature increase corresponds to an approximate 10% increase in volumetric flow rate and leakage velocity. However, as the temperature and velocity of the leaking air increase, the surrounding air is heated and entrained, resulting in a broader flow area and a more extensive leakage pattern.

Furthermore, as the temperature of the injected air increases, flows presumably due to natural convection on the insulated surface were observed. This phenomenon can be found in images where the injected air temperature is above 60 °C, as rising flows to the left and right of the main flow. Despite the PTFE tube being covered with urethane insulation, visualization of natural convection occurs as the surface temperature of the insulation rises above that of the surrounding environment due to the increased temperature of the injected air.

To confirm the presence of leaks from the images at varying temperatures, the brightness values of individual pixels were extracted to calculate the I_E, as shown in Figure 5. The region where the brightness value was calculated is visible in the experimental image with an injected air temperature of 100 °C, which is inserted at the top left of Figure 5 (red dotted region). As expected, the I_E value shows a steady increase with the rise in the leaking air temperature. A sudden expansion of the flow width is observed when the leaking air temperature exceeds 60 °C. These results lead to the conclusion that a leak is present in the pipe if the temperature difference between the hot leaking air and the surrounding air is 16 °C or more, or if the I_E is 0.1 or higher.

The experiment in Figure 6 involved imaging leaks from a perforated pipe surrounded by insulation and a cover jacket with gaps. Similar to the previous PTFE leak experiment, this test was carried out in an environment with ambient air at 14 °C. In this experiment, the temperature of the air injected into the aluminum pipe was incrementally increased from 20 °C to 100 °C in 10 °C intervals, and Schlieren flow visualization images were captured. The leak gap in this experiment had a larger area compared to the PTFE tube experiment. As a result, the flow velocity was relatively lower, making it more difficult to observe distinct shadows due to the insulation and aluminum jacket. When the temperature of the leaking air surpassed 40 °C, faint rising flows became noticeable. Due to the low flow velocity, the flow tended to deviate to the left or right, depending on the surrounding air currents. Transparent shadows that were visually distinguishable were observed only when the leaking air temperature reached 70 °C or higher.

While the leakage flow in the PTFE pipe exhibited turbulent characteristics at all temperatures, the flow depicted in Figure 6 can be classified as laminar, regardless of the leakage temperature. Despite the leakage hole being six times larger, the flow rate was maintained constant, reducing the leakage velocity and Reynolds number to 1/36 and 1/6, respectively. Consequently, the low flow velocity made the leakage highly susceptible to surrounding air currents. This led to a flow pattern distinctly different from that observed in the PTFE experiment.

As in the preceding experiment, an area for observing pixel brightness values was established at the bottom-center section. Although the area was set to the exact dimensions as in the PTFE experiment, most of the leakage flow was observed to be biased, approximately 15 mm to the left. Consequently, the observation area was shifted by this distance from the center to the left. The observation area (red dotted region) and the calculated I_E for this area are illustrated in Figure 7. In all leakage temperature conditions, the calculated values were below 0.1, which is lower than those observed in the PTFE tube experiment. This can be attributed to the significantly reduced flow velocity resulting from the larger size of the leakage gap in comparison to the 2 mm hole observed in the PTFE tube experiment. Additionally, direct leaks result in high turbulence intensity, which causes more complex shadow patterns due to the presence of local lateral and vertical flow components. In contrast, the low-velocity laminar flow observed in this experiment resulted in less distinct shadow differentiation between pixels.

Experimental errors are presumed to stem primarily from the blower’s flow rate and ambient air current. The blower’s flow rate was adjusted by rotating the knob without further verification, potentially introducing an error of about 10%. Conversely, the leakage temperature was measured using a thermocouple with an error margin of 0.2 °C, resulting in relatively precise control. These errors are reflected in the experimental results as deviations in I_E. The standard deviation of I_E used to quantify leakage in both pipes was less than 10% of the mean value. Analyzing 20 images for each leakage temperature case, we found that the deviation in the calculated I_E across all images matched this standard deviation level. The I_E deviation was nearly 10% in the high-turbulence PTFE pipe experiment. In contrast, it was under 4% in the aluminum pipe experiment.

The leak experiment involving a pipe covered with insulation and an aluminum jacket may present challenges in determining the presence of a leak through Schlieren flow visualization, in contrast to the previous PTFE tube experiment. This is because the shadows generated by the flow emerging from the gaps in the aluminum jacket are similar to those caused by natural convection on the insulation surface in the PTFE tube experiment when the air temperature exceeds 90 °C. Suppose the aluminum jacket surface is subjected to direct sunlight. In that case, it may prove challenging to distinguish between the rising flow of hot, leaking air and natural convection.

Most leaks occur through narrow cracks rather than circular drilled holes. Therefore, the leakage flow patterns observed in this study are expected to differ from those of actual pipe leaks. However, the most significant factor affecting the leakage pattern in Schlieren visualization is the leakage temperature and velocity. Consequently, in distinguishing the differences between actual case to current experiments, the outflow velocity determined by the cross-sectional area of the leakage hole is considered the most critical variable rather than the shape of the leakage hole.

To obtain more precise observations of the leaks from the pipe, further considerations can be made regarding the Schlieren optical system and image processing. The optical system utilized in this experiment, the single-mirror off-axis Schlieren system, plays a crucial role. Its simplicity, employing a single reflective mirror, makes it well-suited for monitoring leaks in pipes situated in less accessible locations. The system’s capability to detect leaks if the leaking flow has a temperature difference of approximately 10 °C or more from the surrounding air, provided there are no external heat sources, underscores the significance of the Schlieren optical system in the experiment.

Suppose pipes are situated in more expansive areas or less accessible locations. In that case, it may be advantageous to consider a background-oriented Schlieren (BOS) system, which does not require mirrors. In contrast to traditional Schlieren optical systems, which position mirrors and imaging devices around the subject, BOS implements Schlieren flow visualization using only image recording from a camera positioned on one side of the subject. The subject is positioned against a background with a random pattern, and flow fields are calculated by extracting x- and y-directional flow motion through brightness changes in background pixels, thereby providing more detailed flow information [11,12].

Pipes that transport high-temperature fluids, which are prone to leakage, are typically insulated and encased in jackets. In the event of a leak caused by a crack or perforation, the temperature differential between the leaking gas and the surrounding environment is likely to diminish considerably as the gas traverses the insulation and jacket. These factors have the potential to interfere with natural convection on the pipe covering surface due to surrounding heat sources, thereby making it more challenging to ascertain the presence of a leak. To enhance leak sensitivity through Schlieren flow visualization, it may be beneficial to employ flow field calculations using pixel-by-pixel analysis, as applied in BOS image processing. By calculating velocity fields, it may be possible to distinguish between intermittent or random rising flows due to natural convection and consistent flows due to leaks [17]. Additionally, the potential of combining Schlieren flow visualization image processing with artificial intelligence and big data learning is expected to significantly improve accuracy and sensitivity in leak detection and assessment, providing reassurance and confidence in the system’s capabilities.

4. Conclusions

This study employed Schlieren optics to visualize leaks in pipes carrying gases at temperatures exceeding those of the surrounding environment. Two experimental setups were constructed to simulate gas leaks. The first setup involved a PTFE tube with a 25 mm outer diameter and a 2 mm diameter hole, while the second setup used a 100 mm outer diameter pipe covered with insulation and an aluminum jacket. The Schlieren system employed a single-mirror off-axis configuration with a 150 mm diameter parabolic mirror. The temperature of the leaking air was varied between 20 °C and 100 °C, while the ambient temperature was maintained at 14 °C. To quantify the leaks, the leakage areas observed in the flow visualization images were analyzed, and the coefficient of variation was calculated, representing the standard deviation of pixel brightness values in the defined regions.

In the PTFE tube, leaks were discernible when the temperature difference exceeded 34 °C, with a coefficient of variation reaching 0.1 or higher. On the other hand, in the second experimental setup, where the pipe was covered with insulation and a jacket, distinct leak flows were more challenging to detect due to the increased insulation thickness and gap size, with all experiments exhibiting a coefficient of variation below 0.1. This study underscores the complexity of the research and the efficacy of the Schlieren optical system and flow visualization technique in detecting leaks in high-temperature pipes. These findings are expected to significantly improve the safety and efficiency of high-temperature gas pipelines, with potential applications across a wide range of industrial sectors.