| Applications |
Back to Application Examples! |
|
|
 |
HEAT RECOVERY STEAM GENERATOR (HRSG)
I’ve been planning to come up with this application for some time now. First because it’s a very nice example of a heat exchanger and the method is generally applicable to almost any machine with heat exchange like the evaporator of coolers. Second, because it demonstrates how a machine with complicated physics can be handled easily and fast with a little bit of thinking plus CFD simulation. I said fast which by us means 16 hours from raw 2D drawings till converged solution.
HRSG and arrows showing inlet and outlet
If you plan to ask tons of questions about how I model and analyse all tube bundles of a heat recovery steam generator containing 4 blocks – 1 superheater, 2 evaporators and an economizer – well I would say, forget about it, I won’t. I don’t have to. At least for the pressure loss calculation on the flue gas side I don’t. However what we definitely needed is the thick thermal management calculation, which was done by my colleague Mr. Andras Szabo who is specialized to the design of such boilers.
The calculation contains all necessary data we need: flue gas volume flow, thermal data of flue gas as a function of temperature, inlet and outlet temperatures of flue gas for each and every tube bundle.
Applying the magical equation of specific heat times mass flow times temperature difference, the heat transmitted by each heat exchanger block can be calculated. This value in Watts can be set as boundary condition in the CFD software on the volume representing the specific block of tubes. For example the heat generated in the superheater is 4.3 megawatts.
But tube bundles mean certain resistance when flue gas flows in the block so this resistance must be applied to the volume too. I made some 2 dimensional simulations on partial geometry of tube bundles to calculate the pressure loss. In case of our steam generator these 2D simulations were needed anyway because the heat exchanger tubes were hanged on a steel structure and this way every second row of tubes had to be shifted a bit. Because of that we had a smaller and a bigger gap between tubes.
2D tube bundle simulation, velocity among tubes
From these simulations the pressure loss of the flue gas in the given tube bundle can be read in a jiffy. The pressure loss of the whole bundle of a block now is just a question of multiplication, for example it is 96 Pascals for the superheater.
Building the coupled fluid flow and heat transfer simulation this way and a proper mesh ensure the accurate velocity and temperature results. These are the keys to determine accurate pressure loss data of the steam generator. The flow field itself is not as complicated as a labyrinth, streamlines travel smoothly side-by-side, I could only find two separation bubbles in the upper turning chamber. I saw machinery with much more complex velocity patterns but in this case it is not the point. The thermal management calculation showed 263.9 °C flue gas temperature after the economizer, our simulation presented 266. The error is around 1%.
We can check all pressure losses on the heat exchanger blocks, the 96 Pascals of the superheater is also there, and the resistances of the other tube bundles match the results from the 2D calculations. Beside those the pressure loss of the whole steam generator can be obtained from a graph generated by the CFD code, it is 444 Pascals.
Video of the application example (you can put that to full screen by clicking the icon on the right of the player):
For systems with such dimensions (just the boiler itself is 12 meters high, not to mention the connecting machinery like centrifugal separators, flue gas channels and so on) determining pressure loss is extremely important. Just imagine how big the ventilator is that circulates the flue gas. Either it is too big or too small, that’s too bad.
If the HRSG is placed after a gas turbine, than in that case there is no ventilator in the system. But gas turbines are really sensitive in terms of maximum back pressure, so knowing the pressure loss of devices after it is essential.
But as you see the job can be done easily and fast with CFD simulation.
2010.03.22 - CFD Engineering Hungary Kft. |
|
|
|
|
|
|
