Regarding the issue of excessive stress at the pre-tightening stage in welding seals or weld lip seals.
Thread Content
Recently, it was necessary to verify a seal welding device, as the flange is directly connected to the head. Therefore, analytical design is required. Modeling, bolt pre-tensioning, lock pre-tensioning, pressurization. Looking at the results, it’s not bad; the stress on the flange neck is above the limit, but the tie rods show no issues. Then, I suddenly thought, what is the stress like when there is no pressure? Then it was found that, due to the extremely high bolt preload calculated based on 20582, with 15 MPa plus DL of 1400 mm, the calculated bolt preload reached 700 kN per bolt; as a result, during the preload stage, the stress at the edges of the surfacing layer exceeded the allowable limits by far. The membrane bending stress has reached over 400 MPa. Later, I searched the Internet for articles on stress analysis of seal welding, but found no articles that discussed the pre-tightening stage; they all focused directly on the results after applying pressure. I’d like to ask the experts: how can we draw conclusions at this preload stage? Not approved, needs modification? The size used is the standard size from 20582.
Since it is a lip seal, there are no gaskets between the sealing surfaces; therefore, there is no need to consider a pre-tightening stage
You mean that since the sealing is achieved through welding, there is no need to apply a preload; the bolt load calculated in 20582 is used to verify whether the bolts meet the requirements, rather than being used as a value for the bolt preload. In this way, in practical applications, the bolts only require a relatively low pre-tightening force to be tightened.
The standard has been found; Appendix A of HG20530 addresses the issue of flange installation preload. A value slightly greater than the axial stress caused by internal pressure is required; in other words, it needs to be equal to Wp as specified in 7.5.2.2 of GB150, or it must be larger than F – in short, it’s a relatively high value.