Jan Kasprzyk, Chairman of the Management Board of SPETECH^®
Radosław Sieczkowski, PhD Eng., Manager of SPETECH^® Technical Department
 

The technical development of pressure equipment is associated with increasing operating temperature, increasing diameter dimensions and increasingly higher pressures. At the same time, users expect increasingly higher reliability and safety in the operation of this equipment. An observed trend is also the extension of periods between overhauls and repairs. This corresponds with the obvious drive to reduce operating costs, as well as increase productivity by improving system availability.

At the same time, meeting the aforementioned expectations challenges equipment manufacturers and operators to improve design methods, use better materials, expand diagnostic methods or enhanced manufacturing and operating technologies. This is also accompanied by an economic calculation dictating that the hazards that most affect the continuity of production and safety should be eliminated first , while at the same time being able to be minimized at the lowest possible cost. It is clear that while the identification of these hazards is difficult, as it requires the involvement of advanced methods, the cost of minimizing them can most often be determined with a high degree of accuracy.

In the case of process systems, one of the highest costs and hazards are leaks resulting from both weld joint failures and the unsealing of detachable flange joints. Today, flanged joints of process systems can be arranged as extremely reliably thanks to modern calculation methods (EN 1591-1, FEA) and usage of high performance sealing materials. Flanged connections of large diameters, for high temperatures or high pressures, remain a challenge.

What is the challenge based on?

With large diameters and elevated temperatures, often temperature distribution both in time and in space of a flange-screw connection can be extremely non-uniform. It is influenced by the thermal insulation conditions of the unit, the space in which it is located (including additional heat sources in the vicinity), the vertical or horizontal position of the joint, and, if located outdoors, air temperature, wind, precipitation, as well as the proximity of other equipment.

Figure 1. Uneven temperature distribution around the perimeter of the flanged connection of a horizontal heat exchanger (example)

The temperature differences in a single flange joint connection of the unit also result from temperature distribution inside the unit itself. For example, the temperature will be different near the inlet ferrule compared to the outlet ferrule of the heating medium in the heat exchanger.

Analyzing the resulting thermal expansions of the individual components of the flange joint and, consequently, the resulting stresses and thermal deformations, we expose the entire complexity of the operation of the connection. It involves the conditions of load-deformation of bolts, deformation of flanges, loads and deformation of the sealing. In the case of a gasket, it is always necessary to consider both the elastic and plastic nature of its deformation.

The result is a stress pattern, of which those involved in sealing technology are most interested in contact pressures acting on the seal across all its surface, lateral displacements between flanges and permanent deformation of the gasket. For designers, of course, the stress in the bolts and flanges is also important.

The issue becomes more complicated in transient situations, when temperature gradients and differences resulting from the dynamics of starting or stopping the system must be taken into account. Both detailed analysis and operating experience show that this is when both stresses, strains and displacements between system components can reach critical levels.

For joints using flat seals in the traditional design process, the designer focuses on the necessary minimum and acceptable maximum contact pressures acting on the seal. In very few cases, it analyzes the magnitude of the angular bending of the flange orifices and the effect of this bending on the local stress concentration on the outer circumference of the seal.

The authors of this article, however, have not encountered an analysis conducted at the design stage of the mutual displacement of the flanges resulting from different thermal expansions of the flanges or from deformations due to pressure effects.

While flat gaskets are tested (since 2014 according to EN 13555) for resistance to loads parallel to the face (the gasket ‘friction factor’ is determined), there are no codified rules for describing the behavior of gaskets subjected to the forced displacement of flanges in contact with them. To describe how the process looks, the flanges moving along the seal, pressed down with high pressure, cause the effect of ‘rubbing’ against the gasket on the faces. This is accompanied by the formation of circumferential cracks perpendicular to the surface of the seal, changing its width and, consequently, its thickness. This kind of process intensifies after several heat up and cool down cycles, with the ultimate consequence being the unsealing of the flange joint connection.

Observations of operation show that among flat gaskets, grooved gaskets cope best with the situation described above. Spiral wounds gaskets as well as soft-material gaskets are much less resistant to the described loads. Unfortunately, grooved gaskets also do not provide the reliability expected by the user in every unit.  

Figure 2. Spiral wound gasket destroyed due to large differences in the thermal expansion of mating flanges

Solving the problem of unsealing the main flange connection of a horizontal steam superheater at a nitrogen plant

The problem involved a connection with an internal diameter of 2600 mm, operating at a pressure of 3.8 MPa and a temperature of 450°C.

As mentioned earlier, this is a horizontal heat exchanger, standing in the open air and historically creating problems precisely in the area of the main flange joint despite the fact that it had already been adapted and equipped with a simple weld ring gasket yet in the factory.

Stress analysis was done both analytically according to EN1591-1 and by FEA. Complete data on the materials and geometry and incomplete data from exchanger temperature measurements were available. 

The exchanger was equipped with flanges adapted to weld ring seals, and was originally equipped with an RM-2 type gasket, which, however, did not meet reliability expectations. Accordingly, optimizations were made.

Stress calculations were made in the RM-4 profile using FEA, and the thickness of the toroidal part was optimized, reducing the maximum stress level to approx. 35 MPa for a bolt-flange temperature difference of 50°C. 

Figure 3. Difference of the design of RM-2 (left) and RM-4 (right)

Of course, any model, including those used in FEA, is to some extent a simplification of the actual object, taking into account the factors considered most relevant.

The case described above does not take into account, for example, shape, manufacturing inaccuracies, strength changes within the weld seam, variable circumferential and radial temperature at the flange and, of course, transients. These factors can affect local stress build-up and consequent loss of cohesion. The observed cracks in the sealing weld seam may have originated from such accumulations. Another reason, which is difficult to verify at this time, may also have been a welding problem resulting from both the difficulty of making a relatively delicate weld with difficult access to the weld area through the 90 mm deep gap between the flanges.

Due to wear, the exchanger was replaced with a new unit. Taking advantage of the experience described above, the new exchanger has been equipped from the start with an RM-4 type weld seal with an optimized design. Additionally, 32pcs of M80 bolts at the joint were replaced with 104pcs of M42 bolts, which shortened the gap between the flanges by about 20 mm, making it easier to weld the sealing seam connecting the weld ring seal halves. The weld seal was made of the same material as the flanges so as not to introduce stress into the welds that attach the seal halves to the flanges.

At the moment, after six months of operation, the heater shows no problems on this connection.

Bibliography:
WUDT- UC - Technical specifications for pressure equipment
REPORT OF ANALYSIS OF THE MAIN CONNECTION OF TWO ELEMENTS OF THE E-117 STEAM SUPERHEATER (Customer's property).
Materials of the National Forum for Technical Equipment - online conference, SPETECH Bielsko-Biala 2020
Materials of the National Forum for Technical Equipment - online conference, SPETECH Bielsko-Biala 2021