The challenge of flange connections for steel-enamel appliances

Due to the manufacturing process, steel-enamel apparatus flange connections have considerable unevenness on the sealing surfaces - this can amount to several millimetres. Reliably sealing these unevennesses is a major challenge, which is also reflected in the limited selection of suitable gaskets. 

To date, there are no standardized installation instructions for the various seal types. Instead, there are many different specifications, although the gaskets often differ only slightly in technical terms. In addition, the new TA Luft [1] stipulates requirements for proof of tightness that could not be provided until now. One reason for this is the special design of the joints: Unlike steel-enamel pipelines, no loose flanges and screws are used here, but so-called clamp screws. 

Operators, equipment manufacturers and seal manufacturers have joined forces to develop a standardized installation specification. 

Subject of the investigations 

The following equipment, devices and individual parts were used: 

1. Enamelled steel apparatus flanges with weld neck collars to DIN 28139-2 [2], shape M, enamelled bases, covers and intermediate rings in nominal width DN600 
2. Clamp bolts to DIN 28152-2 [3], M24, strength class 5.6 
3. Envelope gaskets with inserts to DIN 28148 [4] 
4. Flat gaskets in punched form or designed as sealing tape 
5. Vacuum chamber, approx. Ø 1,000 x h 500 with helium leak detector 
6. Circulating air oven 1m³, max. 300 °C with charging trolley 

The bases of the test set-ups are available in two versions: On the one hand, components with standard geometric tolerances are used, and on the other, specially manufactured bases that fully utilize the maximum permissible tolerances in accordance with the DIN standard. The corresponding metal welding collars were processed accordingly - by milling to size and subsequent enamelling. 

The staple bolts are installed in accordance with the VCI installation instructions [5] using cordless torque screwdrivers. Four of the clamp screws used are designed as force-calibrated measuring screws. 

Carrying out the experiment 

The test procedure is closely based on the test scheme for type-based component testing in VDI2290:2024-05 (draft) [6]. 

Analogous to the VCI installation instructions for steel enamel pipes [7], after initial installation with the surface pressure QI, a warm storage at Ttest1=100 °C for 48 h takes place. The test assembly can then cool down to room temperature and is retightened to QI. After determining the leakage L1 at pB=10 bar, a warm storage at Ttest2=200 °C for 48 h is carried out. After cooling down again to room temperature, leakage measurement L2 is carried out at pB=10 bar. 

Compliance with the TA-Luft leakage limit value of L0.01=1.0∙10 2 mg/m∙s would now provide proof of tightness. In this project, the gasket is also to be pushed to its performance limit and the retightening to QI is to be repeated with subsequent warm storage at 200 °C. 

After final disassembly, the seal is inspected for possible damage. A recommendation for use under the assembly and test conditions applied can only be made if no damage is found. 

Damage includes material cracks, penetrations (e.g. holes) and sheared or detached inserts. 

Based on the experience gained from the first series of tests, the following procedure has proven to be expedient: 

Assembly begins with the lowest intended surface pressure and is then carried out with the highest. If both installations lead to positive results, the area in between is considered the permissible installation window. In the event of negative results, the respective surface pressure is increased or reduced in steps of 5 MPa. 

In contrast to the usual procedure, the leakage tests are terminated just a few hours after a stable leakage level is reached or when the limit value previously related to the circumference of the seal is exceeded, provided that no further significant change in the leakage is to be expected. 

Results 

The test results to date illustrate the sometimes-considerable differences in the performance of the gaskets tested: While some envelope gaskets have already successfully proven their suitability, flat gaskets have not yet been able to achieve sufficient tightness on either standardized or maximum-tolerance test setups. In addition, some of the flat gaskets suffered considerable mechanical damage. 

However, there are also clear differences within the group of envelope gaskets: in particular, gasket types with an integrated diffusion barrier in combination with a strongly pronounced metallic corrugated ring caused above-average leakages. 

Whether and to what extent the corrugated ring itself has a negative influence has not yet been conclusively clarified. However, previous findings indicate that only designs with a slightly pronounced corrugated ring and graphite smooth sheet support deliver consistently positive results. 

A similar picture emerges for the residual surface pressures achieved: Some envelope gaskets achieve stable values of at least 5 MPa, usually even significantly higher. In contrast, the residual surface pressures of the flat gaskets tested were significantly lower; in some cases, clamp screws were even loose after the test. 

The maximum tested surface pressure of 25 MPa led to the destruction of the PTFE envelope in all envelope seals, so it can be assumed that the upper limit of the permissible surface pressure is around 20 MPa. Considering both the lower limit (avoidance of impermissible leakage) and the upper limit (avoidance of envelope damage or insert failure), a mounting surface pressure of 15 MPa appears to be a sensible target value. 

Interim conclusion 

Based on the test results to date, initial recommendations for the future design of gaskets can already be derived: The use of diffusion barriers and - presumably - strongly pronounced metallic corrugated rings in combination with graphite coatings without smooth sheet metal inserts tends to have a negative effect on the sealing effect of envelope gaskets and should therefore be critically scrutinized. 

Flat gaskets made of pure or filled PTFE have not been convincing to date. This is mainly due to their structural design and their pronounced creep behaviour under load. Some manufacturers are currently working on corresponding optimizations. In conjunction with high-quality and precise installation, reliable sealing nevertheless appears realistic in the future. 

In conclusion, it can be said - as with many other flange/gasket combinations - that a permanently tight and reliable flange connection can only be achieved by selecting suitable gaskets, careful design and professional installation by qualified and experienced personnel.