Hydrostatic stretch testing of gas cylinders to IS 5844:2014 what is measured, the permanent-expansion limits, and why DA cylinders are examined instead

Every seamless or welded cylinder for compressed gas in India is proved, when it is made and at intervals through its life, by filling it with water and raising it to a test pressure well above its working pressure. The method is IS 5844:2014, Hydrostatic Stretch Testing of Compressed Gas Cylinders — Recommendations, and the pass limits come from the Gas Cylinders Rules 2016 and the cylinder's own design standard. This page explains what the test measures, how each method works, the limits, the calibration a testing station needs — and why a dissolved acetylene cylinder is the exception, tested hydraulically on the empty shell before its porous mass goes in and not at periodic inspection.

IS 5844 was first published in 1970 and revised in October 2014 by the Gas Cylinders Sectional Committee MED 16. Its scope is cylinders of more than 500 ml and not more than 1,000 litres water capacity. It recognises two methods and allows others that measure total and permanent expansion reliably and repeatably; the revision added the levelling-burette method, explicit use of a calibration cylinder, and a modified compressibility calculation.

Water-jacket method. The cylinder, filled with water, is sealed inside a water-filled jacket connected to a graduated burette. Pressure is raised to about two-thirds of test pressure and held while burette and gauge are checked steady for 10 to 30 seconds; it is then raised to full test pressure, isolated, and held for 30 seconds. The burette reading gives total expansion; after release, the residual reading gives permanent expansion. In the levelling-burette variant the burette is lowered or raised to a datum for each reading, so the jacket water head is the same every time; in the fixed-burette variant the levels are read directly against a zero mark. The burette must be fine enough to show a permanent change of about 1/20,000 of the cylinder water capacity, and long enough to hold the whole expansion.

Non-jacket method. The water pumped into the cylinder to reach test pressure is measured and corrected for the compressibility of water; no pressure fall is permitted during the 30-second hold. IS 5844 gives the correction as C = mP(K − 0.68P/10⁵), with C in cm³, m the total mass of water under compression in kg, P in bar and K a temperature factor from Table 1, and requires the water in the pump and high-pressure pipework to be added to the cylinder's water mass. CGA C-1's older direct-expansion method used a graphical Bureau of Explosives factor; a comparison of the two found the IS formula reproduces that chart to within about 0.06 percent at both of C-1's worked examples.

C-1 makes the point that the two methods do not measure the same thing: the jacket gives exterior volumetric expansion, the non-jacket method interior. A cylinder's history should not mix them without noting which was used.

IS 5844 itself does not state a percentage. Each of its result clauses says the ratio of permanent to total expansion shall not exceed the percentage given in the design specification. The numbers come from elsewhere:

Any reduction of pressure during the 30-second hold, any leak, visible bulge or deformation is a failure regardless of the ratio. A cylinder that fails, or loses more than 5 percent of its tare weight, is condemned and destroyed (Rule 36(1)). The full statutory frame is on the Gas Cylinders Rules 2016 page and the retest marking on cylinder testing and retest.

C-1's analysis of error direction is the most useful paragraph on the subject. Readings that are too high — from an incorrect gauge, for instance — reject sound cylinders. Readings that are too low — typically from air trapped in the jacket or a level taken at a different height — pass cylinders with dangerously thin walls. The small fraction of cylinders that should fail are exactly the ones a careless rig lets through. Its list of common faults includes gauges and burette zero not at eye level, the air vent not at the very top of the jacket, light burette hose that changes volume when bent, valves that put test pressure under the stuffing box instead of the seat, dirty or temperature-variable water, and jackets not cleaned — dirt compresses and traps air, and organic matter ferments and makes gas.

A DA cylinder holds a monolithic porous mass and acetone. Once the mass is in, the shell cannot be filled with water or inspected inside. IS 8433:2017, the periodic-inspection code for DA cylinders, says so in its foreword: neither a hydraulic nor a pneumatic pressure test nor an internal visual inspection of the shell can be carried out. The hydraulic test is therefore done on the empty shell, before the mass is introduced: Rule 37(vii) of the Gas Cylinders Rules 2016 requires not less than 60 kgf/cm², reducible to 53 kgf/cm² where a fusible plug is fitted, with permanent stretch not above 7.5 percent — a figure the cylinder standard IS 7312:1993 set independently in clause 15.1.2 (that edition is withdrawn; IS 7312:2018 is current and its text was not available for this page). ISO 10462:2013 cl 6.4 adds that a shell whose test interval has lapsed is hydraulically tested at its stamped test pressure before new porous material is introduced, so "never retested" is too absolute. The C.P.I. code Safety in the Production and Use of Acetylene (Paris, 1968) described the same practice: hydraulic test before filling with porous material.

In service, the DA cylinder is protected by examination instead: a visual examination to IS 8433 at every filling (Rule 39(1)), removal of the valve and inspection of the porous mass at the neck where no examination has been recorded in two years (Rule 39(2)), weighing at every fill to IS 16271, and the IS 8433 rejection criteria for dents, gouges, corrosion, fire damage and mass clearance. The procedure is on the IS 8433 periodic inspection page, and the construction that makes it necessary on inside a dissolved acetylene cylinder.