原文：Figure 4 shows the instrumentation of the test specimen (test
object). The wall of the specimen was 12.5 mm thick.The temperature was measured by 1.5 mm thermo-elements of Ktype. The location is shown in the figure. In some experiments water was filled into the specimen and boiled dry.
Several tests were performed with different heat loads and
different filling of water inside the specimen. These are reported in [5] and [6]. In this presentation three cases will be presented.
Simulations have been performed using VessFire. This is a
system for simulation of fire response of process equipment. It
simulates heat conduction and performs stress calculations of a
3-dimesional shell. Simultaneously the system simulates the
inventory by treating the gas phase and liquid phase separately.
The two phases are linked through evaporation, condensing,
heat transfer and evacuation (for blowdown simulations). The
whole system is linked together to a multi-physics simulation.
See [1] and [2].
VessFire is assuming exposure from a flame in kW/m2. In
general the flux can vary in time and space, but in this case the
heat load is constant in space. The heat flux includes both
radiated heat and convective heat and is defined to be the net
flux transferred to the exposed object while the object is at its
initial conditions. Figure 14, Figure 17 and Figure 20 show the
heat load applied in VessFire for the different cases. The load is
found by taking the average measured temperatures of the heating foil and apply the Stefan-Boltzmann law. The simulations assume the inventory gas zone initially to be filled with air, 78% N2 and 22% O2. The emissivity of the specimen is set to 0.7.
Figure 15, Figure 18 and Figure 21 show the results of
measured and calculated inventory temperatures. When
thermocouples are used to measure gas temperatures there is a
risk of having influence from the surrounding temperatures. In
this situation the surrounded steel was glowing and obviously
influenced the thermocouple by radiation. This is in general a
problem and should be noticed when results are published. The
influence can be quite strong and is here estimated by using the
calculated gas temperature to calculate a corresponding
thermocouple temperature. In the figures this is called “Calc.
temp. thermocouple”. The calculation is done stepwise by use
of: （ 此处一个公式）where ΔT is the temperature increase during the time Δt, mt is the mass per m and cp is the specific heat for the thermocouple (Inconel). The convective heat transfer is calculated as （此处一个公式）where At is the surface area of the thermocouple per m and Tl and Tt is the temperature for gas and thermocouple respectively. λ is the thermal conductivity for the actual gas and dt is the outer diameter of the thermocouple. Nu and Re is respective the Nusselt and Reynolds numbers.
The net radiation heat transfer is calculated as （此处一个公式） where As is the area of the enclosing specimen per m, εt = 0.3 is the emissivity thermocouple (Inconel),εs = 0.7 is the emissivity for specimen inside and σ is the Stefan-Boltzmann constant for black radiation. Ts is the specimen temperature on the inside.
The correction is an estimate, but it gives an idea of the magnitude of the influence on the thermocouple from the specimen surface.
Figure 16, Figure 19 and Figure 22 show the comparison between the measured and calculated steel temperatures. They also show the time where all the water is evaporated. The steel temperature at the bottom of the specimen remains constant as long as there is water present. When the steel is dry the temperature increases rapidly. As can be seen from the figures the boiling time is reasonably well predicted and so are the steel temperatures.