Problems of thermal stress

POWER PLANT

Problems of thermal stress in metal reinforcements of large-dimensional objects with elevated service temperatures

For the insulation of hot objects - especially large-dimensional  ducts as found in nuclear power plants, flue gas desulphurization, and de­nitronisation systems -, it must be considered that reinforcing stiffeners on the duct wall always constitute thermal bridges. Two problems result:

-      The increased thermal transmission through the thermal bridge leads to reduced tempera­tures at the inner surface of the duct wall. This may lead to going below the dew point tem­perature of the flue gas on that inner surface. This problem is not considered in this paper.

-      With unacceptably high-temperature dif­ferences between the inner and the outer edge of the stiffeners, thermal stress may re­sult leading to distortion of profiles resulting in cracking of welded seams. For this reason, it is common to calculate maximum admissible temperature differences that are to be main­tained through an appropriate construction and dimensioning of the insulation.

The limitation to the temperature differences demanded does not constitute a problem in steady-state service, i. e. with flue-gas tempera­tures not changing over time as long as the re­quired insulation material coverage on the out­side of the stiffeners - 1/3 s for stiffeners up to 100 mm, 2/3 s for stiffeners over 100 mm - is observed.

Difficulties, however, may occur in the non-steady-state service - where flue-gas tempera­tures change over time as the installation is started up or shut down.

On starting up the installation, the temperature on the inner surface of the wall and the inner edges of the reinforcing stiffeners follows the increasing flue-gas temperature, whilst the outer edges of the stiffeners remain cold, and the tem­perature there increases only after a con­siderable delay. This may lead to temperature differences substantially above those in steady-state service.

The magnitude of these „non-steady-state tem­perature differences“ is dependent upon a variety of factors:

-      The speed of temperature increase in the flue gas: the faster the installation the startup, the higher the temperature differences.

-      Size of the stiffeners: with big profiles and large masses, the temperature differences are higher than with smaller profiles.

-      Shape of the reinforcing stiffeners.

-      Thermal the conductivity of the materials used.

-      Thermal transmission conditions.

To lower the temperature differences, measures must be taken to allow for the movement of as much heat as possible through radiation and convection from the duct wall to the outer edge of the reinforcing stiffeners. This may be achieved - if technically feasible - by leaving an ample por­tion of the duct wall uninsulated.

These and other measures in the area of insula­tion are, however, of limited effect. With big re­inforcing stiffeners, the steady-state temperature differences cannot be reduced to acceptable values even through „the best possible insula­tion“. Therefore, other measures - outside of the control of the insulation trade - are required. Such measures could be e. g. to use several smaller stiffeners instead of one large one, or to reduce the rate of temperature increase when starting up the installation.

2.    Principal considerations concerning the non-steady-state temperature distribu­tion in reinforcing stiffeners

Depending upon the individual design, the tem­peratures in reinforcing stiffeners are influenced by the shapes, and the appropriate insulation material design values.

Some observations of principle can be made for the design examples

The simple reinforcing fin is shown in Figure 1 (steel sheet; generally smaller than 100 mm) would generally have roughly equal temperatures at the inner and outer edges, providing the insu­lation material coverage d was sufficiently exten­sive (see chapter 1). In this case, no elevated thermal stress occurs. The „dew point tempera­ture problem“ on the inner surface of the duct wall, however, must also be considered in this case. Contrary to this example, the temperature on the outer flange of the normally bigger I-profile (double T-profile - generally with webs exceeding 100 mm) definitely be lower than the inner flange, since bigger masses must be heated on the outer edge and the heat transport requires more time due to the length of the web.   Frequently, the insulation contractor is required to prove mathematically the temperature dif­ferentials to be expected - normally calculated against known warming-up conditions in the start phase of the installation. Such calculations can be computed with numerical procedures such as the finite difference or the finite element method. However, it must be remembered that with these methods the thermal transmission inside the stiffener can be calculated satisfactorily exactly, however, assumptions must be made regarding the movement of heat through radiation and con­vection, the precision of which is frequently very difficult to assess. This applies especially to ra­diation. Here, the surface conditions of the duct wall and the reinforcing stiffener are of decisive importance. They are not known to the insulation contractor with the precision required. Therefore, the declaration of warranties on the basis of such calculations should be cautioned against.    3.    Examples   For insulation following the surface of a profile IPE 360 as in Figure 2a, some results of finite element calculations are given below. Figure 3 shows the temperature increase over time at an uninsulated duct wall, the inner and outer edges of the reinforcing flange when the warming-up transients are 1,6 K/min and 0,4 K/min.   The maximum occurring temperature differences for a  profile IPE 400 insulated according to Figure 2a is  given for different warming-up tran­sients and for the  steady-state service in Table 1. Instationär Ausgangstemperatur = +40 °C non-steady-state Initial temperature = +40 °C Profil Profile Temperaturdifferenz der Flansche = DJ [K] Temperature differ- ence in stiffeners = DJ [K] Anfahrgeschwindigkeit K/min Initial temperature K/min (temperature transient) IPE 400 ca. (about) 50-60 1,6 - „ - ca. (about) 40 0,8 - „ - 90 2,0 IPE 460 75 1,6 - „ - 50 0,8 HEA 300 57 1,6 - „ - 45 0,8 IPE 370 53 1,6 - „ - 34 0,8 IPE 300 45 1,6 IPE 270 42 1,6 Table 2: Temperature differences in flanges with „air-gap“-insulation (according to Figure 2b)   The results show that especially whilst warming up the installation, critical stress maxima must be expected. The warming-up gradient has a decisive influence.   A comparison of the two designs considered here makes it obvious that the air-gap-insulation com­pared to the surface-following insulation results in smaller temperature differences for both the steady-state and the non-steady-state conditions. These observations, however, only hold true when uncontrolled convection influence can be prevented.     4.    Conclusions   For the insulation of large-dimensional, hot ob­jects, special thermal conduction considerations are required. Additionally, investigation of the possible deformation in the stiffeners as a result of temperature differentials is needed. This applies specifically to non-steady-state service conditions such as start-up and shut-down phases and accidents.   A mathematical proof of the maximum occurring thermal stresses in the steel construction of an object is not within the area of responsibility of the insulation contractor. The static system se­lected and the static and dynamic stresses to be born by the construction are in the area of responsibility of the installation contractor.   Nevertheless, this problem should be addressed when discussing contracts and the builder should be made aware of it. It could be possible that there is a duty to caution against possible damages, if the thermal stresses to be expected as result of the layout and size of the reinforce­ments and the temperature differences to be expected could lead to damages.   In critical cases, the necessity may even occur to ensure an even distribution of heat at the outer edges of the reinforcement in the warming-up phase of the installation by installing an extra heating system.