7 Replies Latest reply on Jun 28, 2018 5:52 PM by Chris Phelan

    Heating water - fluid sub-domain temperature dominates results?

    Chris Phelan
      1. Greetings. I am trying to simulate the process of heating water on a stove, with the end goal being the heat flux through the sheet metal pan below the burner/flame. I already have empirical test results which capture pan temperature vs. time which I have been using to ground my simulations in reality. So far, they are not grounded.

       

      To "model" the flame, I have a solid that roughly emulates the flame shape. I have no material defined, so it defaults to air. I've defined the flame as a heat source (given that I know the BTU/hr output of the burner - at least per the spec sheet) and also as a fixed temperature source (1300degC). More details below.

       

      To model the pan, I've copied the pan geometry and distance from the flame, and set the materials/radiation surfaces (stainless steel, glazed porcelain).

       

      Otherwise, in general, all simulations are external with a defined domain enclosing the model. Heat conduction and radiation are on. The environment is set to 25degC, gravity is on, humidity is on (80%), and the default fluid type is air. I do not have any boundary conditions. Initial ambient is also set to 25degC and 101.3kPa. Goals are set to pan surface temperature (solid) and pan surface flux.

       

      All simulations are static except the last one

       

      First simulation - no pot or water. Flame modeled as a heat generator with a 17kbtu (4982W) output. Flame surfaces defined as blackbody walls. Pan temperatures exceed empirical results by about 2x. This raises the question - how is heat lost from a heat generating source? I'm assuming it can be shed to the environment - not just the solids defined.

       

      Second simulation - same as first, but with the radiation properties dropped from the flame surfaces. Same results as the first simulation - radiation property definition does not seem to impact the results – maybe not too surprising since it is defined as (I’m assuming) an ideal heat source

       

      Third simulation – pot added in, stainless steel material w/ radiation defined, inner surface split halfway, “real wall” defined on the lower half/bottom of the pot with a 100degC temperature (boiling water). Pan temperatures are elevated further past empirical measurements – I’m assuming this is because the pot is holding heat from the flame closer to the pan than without the pot, but I was expecting to see somewhat of a heat reduction due to the pot acting as a sink. Is using a “real wall” adequate considering I don’t really care about what is going on inside the pot?

       

      Fourth simulation – fluid (water) subdomain added in to the pot as the volume in the lower half. The subdomain’s temperature (thermodynamic parameter) is defined as 100degC. The same heat source is used. No boundary conditions are applied. The result – all surfaces are now 100degC. It seems like this new fluid subdomain is acting like an infinite heat sink at 100degC. How should I properly set this up?

       

      Fifth simulation – Same as the fourth, but with the fluid subdomain temperature set to 25degC. All surfaces are now 25degC, confirming that the fluid subdomain functions as an infinite sink.

       

      Sixth simulation – Same as fourth, except that the heat source is now defined as constant 1300degC temperature. Water and pan temperatures now rise. Water/pot elevates to 600 degC (not realistic of course), pan temperature is ~25% less than empirical. Interesting that we are getting closer on the pan temperature front, but given the water temperature, these results can’t be trusted.

       

      Seventh simulation – Same as sixth, but transient analysis. Based on a quick boil water test, I know that the water should hit ~124degC in 5 minutes (300s). I stopped at 30s to check in, I am seeing water regions hitting 500degC, so this isn’t working out well either. Pan temperatures are, not surprisingly, low.

       

      Questions (collected from above):

       

      1) How is heat lost from a heat generating source? I'm assuming it can be shed to the environment - not just the solids defined.

      2) Do I need to define the flame as a heat source with a fixed temperature? Or is the heat source itself sufficient? Should I include radiation properties on the surface of the flame “solid”?

      3) In many (all?) of my simulations, I do get an error that T>Tmax for the real gas parameters (air) – should I be concerned?

      4) How should I properly set up the fluid subdomain such that a heat generation source actually heats it? Based on my results, it seems like only a temperature source will heat the fluid subdomain.

      5) Can I set the fluid subdomain to never exceed 100degC (to more accurately simulate boiling water)?

       

      Any help appreciated - thanks!

       

      -Chris

        • Re: Heating water - fluid sub-domain temperature dominates results?
          Chris Phelan

          While looking at my results, I noticed that my fluid temperature result cut plot did not illustrate the fluid subdomain temperature - I could only see the subdomain temperature if I inserted a solid temperature cut plot. My fluid subdomain is defined by the surfaces on a solid representative of the water volume. My reaction to this was to omit this solid from the calculation using component control - but doing this creates an error since the solver can no longer find the faces to define the subdomain. So, some additional questions:

           

          1) Is there an issue with the fluid subdomain being defined by surfaces of a solid - does the solver treat the solid as the fluid subdomain, or does it double-count (solid overlapping fluid subdomain)? In this case, since the solid has no material properties, it will be defined by the air default.

           

          2) Should I insert surfaces for the fluid subdomain and omit the solid?

           

          -Chris

            • Re: Heating water - fluid sub-domain temperature dominates results?
              Chris Phelan

              Well I may be working towards answering my own questions. Selecting all of the faces on my water "solid" seems to convert the whole computational domain to water - its like fluid subdomain selection tool is selecting the inverse of what I want. How do I get around this? I can use the inner surface of the pot to define most of the subdomain, but I can't use a standalone surface to define the top of the subdomain. The top seems to require a solid lid. If I use this solid lid, the lid region is treated like a solid, and there is no air-water interface. If I omit the lid from the simulation, the computational domain again fills completely with water.

               

              So, how do I properly define the fluid subdomain??

            • Re: Heating water - fluid sub-domain temperature dominates results?
              Bill McEachern

              I have not read all your posts in detail but in Flow sim if you want to simulate the heating of water in a pot by conduction, convection and radiation it is going to be a bit tricky due to limitations in the flow package. Generally in these conditions with a well composed flame (i.e its blue) the radiation of the flame can be ignored. So the approach I would recommend is you put in your body that is the flame (I don't think it matter what the shape is but some volume that encloses the combustion zone). You need to disable it so it becomes a volume but is not anything but likely air in you case (it will be whatever the fluid is in the region). You can then apply a heat source to (BTU/hr) and it will heat the air passing through it. You may also want to include an surface that will inject a volume of gas to represent the input fuel.

                • Re: Heating water - fluid sub-domain temperature dominates results?
                  Chris Phelan

                  Bill-

                   

                      Thanks - I had the flame defined as a solid, and I mistakenly thought that it would default to the default gas - I forgot that a default solid was also defined in the input parameters. I've since updated the sim to omit my flame solid, now it is just a heat generation source.

                   

                  I still have an issue with the definition of the fluid subdomain. The only way I can get the subdomain to exist as a small volume within the computational domain is to define a "void" (no body within) using the pot floor and wall along with a cap - if I use a solid body to define the subdomain, the simulation assumes that everything but the solid body is the subdomain (the inverse from what I want). And the solid body is considered a solid in the analysis, further skewing my results.

                   

                  Of course, if I use a cap, the cap is defined as a solid body. To work around this, I've made the cap very thin (.010) and I've defined it as a gold sheet (high thermal conductivity). The issue with this configuration is that I've now built a pressure vessel. If I apply an environmental pressure boundary condition to the inner surface of my cap, I get a vortex crossing the boundary error. I'm not sure which is the lesser of two evils (pressure vessel vs. vortex error).

                   

                  Is it not possible to define a fluid subdomain within a computational domain without a solid between the two?

                   

                  -Chris

                  • Re: Heating water - fluid sub-domain temperature dominates results?
                    Chris Phelan

                    Bill-

                     

                       Following up on your comment about being able to ignore the radiation from the flame. Is this true considering that I am most interested in the pan temperatures below the flame? If I just have a solid defined as a heat generating source, but then omit the solid in the calculation (turning the heat generating source in to air), there will be no flame-based radiation heating of the pan, correct? Given that our burner pan is black, I think that this is a problem. Any thoughts about the best way to resolve?

                     

                       At this point, I'm thinking that heating the water may not even be all that important in this model.

                     

                    -Chris