Cold box separator cylinder leak detection is a critical non-destructive testing process used to verify the integrity of cryogenic gas-liquid separation vessels installed inside insulated cold boxes in air separation, LNG, and petrochemical facilities. By applying techniques such as helium mass spectrometry, pressure decay testing, and bubble emission testing in accordance with ISO, EN, ASTM, and GB standards, accredited third-party laboratories help operators prevent product loss, insulation failure, and serious safety incidents. This article explains the sample types, test items, detection methods, acceptance criteria, and typical applications of professional cold box separator cylinder leak testing.
Technical Overview
A cold box is a sealed, heavily insulated enclosure—typically filled with expanded perlite, mineral wool, or vacuum insulation—that houses cryogenic process equipment such as distillation columns, heat exchangers, and separator cylinders. The separator cylinder (also called a separator drum or knock-out vessel) performs gas-liquid phase separation at extremely low operating temperatures, often between -160 °C and -196 °C, in services such as liquid oxygen, liquid nitrogen, liquid argon, methane, ethylene, and hydrogen.
Because these vessels operate at cryogenic temperatures under internal pressure, even microscopic leaks can have severe consequences. Internal leakage between process passages degrades product purity and reduces plant efficiency, while external leaks into the insulation space can saturate the perlite, form ice plugs, corrode the outer carbon-steel casing, and in oxygen-enriched environments create a significant fire and asphyxiation hazard. Undetected leaks frequently force unplanned shutdowns, as repairing internal cold box components requires complete removal of insulation, warm-up, and re-purging—a process that can take weeks.
Leak detection is therefore performed at multiple life-cycle stages: during fabrication of new separator cylinders, after welding repairs or modifications, during commissioning of new plants, and as part of periodic turnaround inspections of in-service equipment. Results are quantified as leak rates, typically expressed in Pa·m³/s or mbar·L/s, and compared against design specifications and relevant standards such as ISO 20484, EN 1779, EN 13185, ASTM E432, ASTM E493, GB/T 15823, and the GB/T 18443 series for cryogenic vessels.
Testing Samples
Newly Fabricated Separator Cylinders
Complete vessels delivered from the manufacturer prior to insulation installation and cold box closure. Testing at this stage is the most effective, since all welds, nozzles, and flanged connections remain directly accessible for both global and local leak measurement.
In-Service and Repaired Vessels
Separator cylinders already installed inside operating cold boxes, examined during planned turnarounds, after internal inspections, or following weld repairs. Testing may be performed through access manways, insulation removal windows, or by tracer gas injection at accessible connection points.
Weld Joints and Subassemblies
Circumferential and longitudinal shell welds, nozzle-to-shell welds, head-to-shell joints, boss attachments, instrument tappings, and repair welds. Individual weld seams can be tested using sniffer probes or vacuum hoods to localize any leakage path.
Sealing Elements and Connections
Flanged joints, gasketed manway covers, threaded instrument ports, drain and vent connections, and pipe penetrations passing through the cold box casing. These mechanical joints are statistically the most common leak sources in cryogenic service.
Vacuum-Insulated Configurations
For vacuum-jacketed separator cylinders and annular spaces, the insulation vacuum itself serves as a leak indicator, and vacuum decay or leak-up rate measurements are applied in addition to tracer gas methods.
Testing Items
Professional leak detection programs for cold box separator cylinders typically cover the following measurable items:
- Overall (global) helium leak rate of the complete vessel or closed system, expressed in Pa·m³/s
- Local leak rates of individual weld seams, joints, and seals using probe or hood techniques
- Pressure decay rate during pneumatic strength and tightness hold tests
- Vacuum leak-up rate for vacuum-insulated annuli and jacketed sections
- Bubble-emission tightness verification of accessible external surfaces and connections
- Background tracer gas concentration and instrument calibration verification against certified reference leaks
- Temperature and pressure stabilization data required to validate test sensitivity
- Comparison of measured values against design leak-rate criteria and applicable acceptance standards
Acceptance criteria depend on the service class of the vessel. Typical specifications for cryogenic separator cylinders require local leak rates no greater than 1 × 10⁻⁶ Pa·m³/s for general service, with tighter classes down to 1 × 10⁻⁹ Pa·m³/s for high-purity, vacuum-insulated, or oxygen service applications.
Testing Methods
Helium Mass Spectrometry Leak Detection
Helium mass spectrometry is the most sensitive and widely used method for cold box components, offering detection thresholds down to 1 × 10⁻¹² Pa·m³/s. Two principal modes are applied. In the vacuum (inside-out) method, the vessel is evacuated and connected to the mass spectrometer while helium is sprayed onto external surfaces; any helium entering through a leak path is detected with maximum sensitivity. In the sniffer (outside-in) method, the vessel is pressurized with helium or a helium-nitrogen mixture and a handheld probe is scanned along welds and joints to both detect and precisely locate leaks. A hood or bag technique, in which an entire weld cluster or component is enclosed in a helium-filled envelope, provides a rapid global leak rate. Relevant standards include EN 13185, ASTM E493, ASTM E498, ASTM E499, and GB/T 15823.
Pressure Change Testing
Pressure decay testing involves pressurizing the separator cylinder with dry nitrogen or air to a specified test pressure, isolating it, and recording the pressure drop over a measured time interval after temperature stabilization. Corrections for ambient temperature variation are applied to distinguish true leakage from thermal contraction effects. The corresponding vacuum decay (leak-up) method is used for evacuated insulation spaces. These methods are simple, economical, and suitable as screening tests, with sensitivity typically around 1 × 10⁻³ to 1 × 10⁻⁵ Pa·m³/s depending on volume and hold time.
Bubble Emission Testing
For accessible external surfaces, the vessel is pressurized and a surfactant solution is applied, or small components are immersed, so that escaping gas forms visible bubbles at leak locations. Vacuum box testing reverses the principle: a transparent-sided box evacuated to a partial vacuum is sealed over a weld zone that has been wetted with the soap solution, and bubbles forming inside the box reveal leaks. These techniques comply with EN 1593 and ASTM E515 and are valuable for rapid localization before or after helium testing.
Alternative Tracer Gas Methods
Where helium background levels are high, hydrogen-nitrogen tracer mixtures (typically 5 % hydrogen in nitrogen) detected by thermal conductivity or semiconductor sensors can be used, as can halogen detectors for certain services. Method selection follows the decision framework of EN 1779 and ASTM E432, considering required sensitivity, accessibility, vessel geometry, and process fluid compatibility.
Test Sequence and Quality Assurance
A rigorous program normally combines methods: pressure decay as a gross screen, followed by helium mass spectrometry for quantitative measurement and localization. Instruments must be calibrated before and after each test using certified reference leaks traceable to national standards, per ISO 20486. All tests on vessels in oxygen or oxygen-enriched service require strict cleanliness and oxygen-safety controls consistent with pressure vessel codes such as GB/T 150 and applicable national special equipment safety regulations.
Applications
Cold box separator cylinder leak detection is applied across every industry that operates cryogenic separation and liquefaction equipment:
- Air separation units producing gaseous and liquid oxygen, nitrogen, and argon, where separator integrity directly affects product purity and safety
- Ethylene plants and other petrochemical complexes with cryogenic demethanizer, deethanizer, and cold-box separation trains
- Natural gas processing and LNG facilities, including NGL recovery columns, LNG flash drums, and boil-off gas separators
- Hydrogen and helium liquefaction plants, which demand the highest leak-tightness classes due to small molecule size
- Cryogenic storage and loading systems, including transfer lines, pump vessels, and vaporizer separators
- Rare gas recovery units (neon, krypton, xenon) and cryogenic carbon capture pilot installations
- Fabrication shops performing pre-delivery acceptance testing and plant operators conducting turnaround and post-repair verification
FAQ
What leak rate is acceptable for a cold box separator cylinder?
Acceptance limits are set by the design specification and process service. Common criteria are 1 × 10⁻⁶ Pa·m³/s per local weld for general cryogenic service and 1 × 10⁻⁹ Pa·m³/s or lower for vacuum-insulated, high-purity, or hydrogen service. The certified test report must state the measured value, the required limit, and the demonstrated sensitivity of the method.
Can in-service testing be done without removing the perlite insulation?
Partial testing is possible by pressurizing the vessel with tracer gas and probing at manways, casing penetrations, and casing weld seams, or by monitoring the cold box annular space with sniffers. However, if leakage is confirmed, insulation removal is generally required to locate and repair the defect precisely, which is why thorough pre-shipment testing is strongly recommended.
What is the difference between sniffer mode and vacuum mode helium testing?
Sniffer mode pressurizes the component with helium and searches for escaping gas externally; it locates leaks but is roughly two to three orders of magnitude less sensitive. Vacuum mode evacuates the component and detects helium sprayed from outside, achieving the highest sensitivity and quantitative global rates but requiring more setup. The two are frequently combined in one test campaign.
How often should separator cylinders be leak tested?
New vessels are tested at fabrication and again during commissioning. In-service units should be tested after any weld repair, at major turnarounds (typically every three to five years), and immediately if operating data such as purity drift, cold spots on the casing, or abnormally high insulation-space oxygen content indicates a possible leak.
How long does a typical test take?
Duration depends on vessel size, method, and accessibility. A sniffer survey of accessible welds may take a few hours, while a full vacuum-method global test on a large separator cylinder—including evacuation, stabilization, calibration, and measurement—can require one to three days. Test planning should account for drying and purging after any hydrostatic testing.
Are special safety precautions required?
Yes. All testing on vessels in oxygen or oxygen-enriched service must follow strict oxygen cleanliness and ignition-control procedures. Pressure testing must use regulated gas sources with relief protection, confined-space entry rules apply when working inside cold boxes, and asphyxiation hazards from nitrogen and helium require continuous atmosphere monitoring.