I am trying to understand different methods / techniques used to calibrate heat exchanger (HX) models.
Say, I am building a HX model from scratch to predict performance prediction at multiple operating points e.g. changes in mass flow rates or entering fluid temperatures. I am using the relevant and standard heat transfer correlations (HTC) from textbook for hot and cold side fluids and applying them on each discretized elements of the HX. Since the HTCs are derived from experiments, and have certain assumptions they have an error band or uncertainty associated with them. Therefore, the prediction from the model will not exactly match with the actual performance. To factor in these uncertainties, a ‘correction’ or ‘adjustment’ factor could be tacked on to the model that will bring the performance to the measured value/s. Now my question is:
1. On which equation this correction should be applied e.g. should be applied at heat transfer correlations or final heat transfer rate equation or anywhere else? (In my opinion it should be the heat transfer correlation as it is the main source of uncertainty and in process it also corrects the temperatures at discretized elements of HX, but I am open to suggestion for better place to apply these corrections)
2. What is an acceptable value of these corrections? I understand it will vary for type of HX, if there is phase change in HX etc. I am trying to get a ballpark value. For example, is 1.5 on heat transfer correlations is acceptable? If not, at which point it starts getting into acceptable range below, 1.2 or 1.05?
3. Is there a standard or document that describes the procedure to calibrate the HX models so that if I have a single (or limited) data point/s, I can calibrate the model at a point to make it sufficiently predictive over its operating map?
The type (i.e., physical configuration) matters greatly, as do the fluids. If you are searching the Web for answers, don't look up the word "calibrate" as this means something else. What you're looking for is more like a "performance factor" or what amounts to an economic "safety factor" but don't search for "safety factor" because this also means something else. I have a lot of experience with steam surface condensers (water and air cooled) and also heat recovery steam generators. You need to be more specific about what you intend to do with this heat exchanger. Before you can analyze it, you must pick a configuration, as this will also guide the calculations you will want to use. I suggest the NTU-effectiveness method for anything other than a steam surface condenser. Consider, for instance, conventional feed water heaters. The design geometry (U-tube bundle inside a shell with a cap on one end) is driven entirely by piping considerations, not heat transfer. The same is true for water-cooled and air-cooled steam surface condensers--completely different physical designs because of factors other than the steam (i.e., getting the water or air in and out). Compact plate heat exchangers that are held together by long threaded rods with nuts on both ends (often used for oil coolers on a gas turbine skid) are designed so that they can be easily disassembled and cleaned. Rather than ignoring these physical considerations when designing a heat exchanger, you need to start with them, then get to the heat transfer coefficients. You can download this Ebook for free on 12/17, 12/25, 1/2, 1/10, or 1/18 http://www.amazon.com/dp/B078HM16GV
I suggest you to simulate your HX with a good CFD and multiphysics software and to compare the heat flux with the the heat transfer correlation applied for this sort of HX.
But I am not sure to help you while writing this.
hi, first of all think it is impossible to get a only key for every thing but in compact heat exchanger normally you will find error band of about 10% for Colburn factor and 12% for friction factor. i thimk you could apply your correction on this two factors.
The type (i.e., physical configuration) matters greatly, as do the fluids. If you are searching the Web for answers, don't look up the word "calibrate" as this means something else. What you're looking for is more like a "performance factor" or what amounts to an economic "safety factor" but don't search for "safety factor" because this also means something else. I have a lot of experience with steam surface condensers (water and air cooled) and also heat recovery steam generators. You need to be more specific about what you intend to do with this heat exchanger. Before you can analyze it, you must pick a configuration, as this will also guide the calculations you will want to use. I suggest the NTU-effectiveness method for anything other than a steam surface condenser. Consider, for instance, conventional feed water heaters. The design geometry (U-tube bundle inside a shell with a cap on one end) is driven entirely by piping considerations, not heat transfer. The same is true for water-cooled and air-cooled steam surface condensers--completely different physical designs because of factors other than the steam (i.e., getting the water or air in and out). Compact plate heat exchangers that are held together by long threaded rods with nuts on both ends (often used for oil coolers on a gas turbine skid) are designed so that they can be easily disassembled and cleaned. Rather than ignoring these physical considerations when designing a heat exchanger, you need to start with them, then get to the heat transfer coefficients. You can download this Ebook for free on 12/17, 12/25, 1/2, 1/10, or 1/18 http://www.amazon.com/dp/B078HM16GV
- Badges
- Science topic
- Similar topics
- Correction