Industry & Background

The client in this case study is a chemical manufacturer that produces various chemical products. The company uses nitrogen gas in numerous stages of its manufacturing processes. Nitrogen is an essential utility gas for preventing oxidation, maintaining inert atmospheres, and preserving product quality.

In chemical plants, nitrogen gas is consumed in large quantities as a sealing gas to prevent oxidative degradation of raw materials or as a purge gas in reaction vessels. Therefore, the integrity of nitrogen supply systems is a critical management item directly linked to manufacturing quality.

At this plant, years of operation had caused aging of piping, valves, fittings, and other equipment, with sporadic reports of trace nitrogen gas leaks. However, because nitrogen gas is colorless and odorless, detecting leaks was extremely difficult, and responses tended to be reactive rather than proactive.

Challenge

On-site maintenance personnel had detected nitrogen gas leaking near flow meters by the faint hissing sound. While they could hear the leak sound when listening carefully, pinpointing the exact location was extremely difficult.

The biggest obstacle was that the piping in the suspected area was densely packed. In an environment where multiple pipes run in parallel with fittings and valves in close proximity, sound reflection and attenuation made it virtually impossible for human hearing alone to accurately determine the direction of the leak sound.

Furthermore, because the leak volume itself was very small, flow changes were extremely minor, making it impossible to narrow down the leak location from flow meter data. Conducting a pressure drop test on the entire piping system would have required shutting down the production line, which was not easily feasible.

Human hearing has physical limitations. In particular, sounds in the ultrasonic range generated by trace gas leaks often exceed the frequency range detectable by the ear, and there were likely many leaks that could not be found by relying on audible sound alone. With conventional inspection methods, the facility faced the dilemma of being able to recognize the existence of problems but unable to obtain the precise information needed to solve them.

Product Adopted

To solve this challenge, the acoustic camera "AlgoLeak AL64" was adopted. The AlgoLeak is an acoustic visualization device equipped with a 64-microphone array that displays the position of sound sources as a heat map overlaid on the camera image in real time.

Operation is very simple. Simply scanning the target area with the camera causes the source of leak sounds to be color-mapped on the display. Even in environments with densely packed piping, it is possible to intuitively confirm which fitting or valve is producing the sound directly on the visual display.

A key feature of the AlgoLeak is its ability to detect ultrasonic-range sounds that are imperceptible to the human ear. It captures ultrasonic components emitted by trace gas leaks with high sensitivity and pinpoints their location with centimeter-level accuracy. Even in noisy environments, the frequency filtering function enables visualization of only the leak sound.

Measurement data can be recorded as still images and videos, making it useful as evidence for report preparation and repair requests. It is a tool that simultaneously enables rapid on-site decision-making and accurate information sharing with management departments.

Results

The AlgoLeak survey revealed that nitrogen gas at 0.4 MPa was leaking from the seal tape joint of an internal tube. Being able to pinpoint the leak location even within densely packed piping was an achievement that would have been impossible with conventional inspection methods.

With the leak location precisely identified, repair work could be carried out accurately and quickly. The leak was completely stopped through the relatively simple repair of re-wrapping the seal tape, without requiring major piping work. Because the leak was discovered at an early stage, it was addressed before it could expand, minimizing losses.

Through this case, it was demonstrated that even extremely fine leaks that are typically overlooked during routine inspections can be reliably detected using the AlgoLeak. The difference from conventional visual and auditory patrol inspections was stark, with equipment maintenance accuracy dramatically improved.

Following this success, the company is planning regular AlgoLeak surveys covering the entire plant. By expanding its use to include not only nitrogen gas but also compressed air and other utility gas leak surveys, they aim to drive energy efficiency improvements across the entire factory.

  • Pinpointed trace nitrogen gas leaks within densely packed piping
  • Discovered leak from seal tape joint at an early stage
  • Enabled repair at a stage where simple fixes were still effective
  • Detected ultrasonic-range leaks impossible with conventional auditory inspection
  • Planning expansion to regular plant-wide surveys