High-Temperature Joint Behaviour
The behaviour of bolted and mechanical joints at elevated temperature — preload relaxation, differential expansion, gasket creep, flange rotation, creep load redistribution in the joint, and the assessment of joint integrity.
Joints at elevated temperature
Bolted and mechanical joints at elevated temperature face multiple time-dependent effects: the bolt preload relaxes by creep, the clamped components may creep, the gasket (if present) may creep and lose sealing force, and the differential thermal expansion between components changes the load distribution. The joint that is adequately designed at room temperature may degrade over time at elevated temperature. The analysis must capture all these effects to predict the joint integrity over the service life. This is a coupled problem — the bolt relaxation, the flange deformation and the gasket creep interact through the joint stiffness.
Preload relaxation
The bolt preload relaxes at high temperature as the bolt material creeps. The preload reduction depends on the bolt creep properties, the temperature, the initial preload and the joint stiffness. The joint stiffness determines how much the bolt stress changes for a given creep strain — a stiff joint (thick flanges, short grip length) has a high stiffness and relaxes faster; a compliant joint (flexible flanges, long grip length) has a lower stiffness and relaxes more slowly. The preload relaxation analysis must model the bolt and the clamped structure as a system, not just the bolt in isolation.
Gasket creep and sealing force
In a gasketed joint, the gasket provides the seal. The gasket is compressed by the bolt preload. At elevated temperature, the gasket material may creep (in compression), reducing the gasket thickness and the compressive stress. If the gasket stress falls below the minimum sealing stress, the joint leaks. The gasket creep is often faster than the bolt creep (gasket materials are typically less creep-resistant than bolt materials). The analysis must include the gasket creep in the joint model. The gasket material should be selected for creep resistance at the service temperature, and the gasket stress at end-of-life must be above the minimum sealing stress.
Flange rotation and creep
In a bolted flange joint, the flange rotates under the bolt preload — the flange face tilts, changing the load distribution on the gasket. At elevated temperature, the flange material may creep under the bending stresses. The flange creep reduces the flange rotation over time (the flange "settles"), which may increase or decrease the gasket stress depending on the geometry. The flange creep also redistributes the stress in the flange. The analysis must model the flange as a creeping structure, not as a rigid body. A 3D FEA of the flange joint with creep may be needed for critical joints.
Creep load redistribution in the joint
In a multi-bolt joint, the load is distributed among the bolts. At elevated temperature, the bolts may creep at different rates (due to different temperatures, different stresses or different material conditions). The differential creep redistributes the load among the bolts — the bolts that creep faster lose load, and the load is transferred to the bolts that creep more slowly. This redistribution can overload the slower-creeping bolts. The analysis must model all the bolts in the joint and capture the differential creep. A single-bolt model cannot capture the load redistribution in a multi-bolt joint.
Assessment of joint integrity
The joint integrity must be assessed over the full service life, considering all the time-dependent effects. The assessment criteria depend on the joint function: for a sealing joint, the gasket stress must remain above the minimum sealing stress; for a structural joint, the clamping force must remain above the minimum for friction or shear transfer; for a pressure-containing joint, the preload must resist the pressure end-load with adequate margin. The analysis should predict the preload (or gasket stress, or clamping force) at end-of-life and compare it to the minimum required value. The assessment should consider the uncertainty in the creep properties, the initial preload scatter and the temperature distribution.
High-temperature joint behaviour is a coupled problem involving bolt creep, flange creep, gasket creep and differential thermal expansion. A single-component analysis (bolt only or flange only) cannot capture the interaction. Model the joint as a system with all the relevant components and their creep properties.
System-level modelling and load-path verification
A detailed joint model should demonstrate that the assembled load path is physically credible before creep is activated. Bolt force, flange contact pressure, gasket compression and support reactions should balance the applied mechanical and pressure loads in the initial state. The thermal solution should then be checked because even modest temperature differences between bolts, flanges and gasket can dominate the early preload change. During the creep dwell, tracking these same quantities as histories provides a much clearer picture of redistribution than reviewing stress contours alone. Simplified sector or axisymmetric models can be valuable for sensitivity studies, but their symmetry and load-sharing assumptions should be justified against the real joint geometry.
Repeated cycles and stabilisation
For joints that see repeated start-up, dwell and shutdown cycles, the first cycle may not be representative of later service. Residual deformation and altered contact conditions can change the preload at the start of each subsequent cycle. The analysis should continue for enough representative cycles to determine whether the joint response stabilises, progressively loses clamp load, or accumulates separation and slip. If cycle-by-cycle simulation of the full life is impractical, a justified acceleration or cycle-jump strategy may be used, but it should preserve the governing creep and contact state variables and be checked against explicitly simulated cycles.