PBE is far from standard practice, but FLUENT is not at all secretive about having it or about how it works - there's detailed documentation on the add-on modules. The basics of the model are also well explained and straightforward, but yes, mathematically the equations can be extensive and take some time to fully grasp. Engineers are being paid well for a reason ;)

Why tutorials don't include PBE's is simple, as I said, it's not standard practice to use them, and it's not necessary to use them in order to understand how to set up 2-phase flow problems. A single-bubble size approach works well for that (and is much faster). The reason the use of PBEs is not standard has several reasons. For one, they are computationally extremely expensive to run. So until recently, their use was simply not feasible for most problems.

Second is that they are practically not well developed. Yes, the theory behind population balance modeling is old and well known, but the kernels capturing breakup and coalescence are highly empirical, differ widely in their assumptions and simplifications, and are often very case dependent. A kernel that works well for solid particles may not do well for gases, one that does well for heterogeneous bubble columns may not do well for stirred tanks. Even two sets of kernels that both give good predictions for the same flow problem, may do so based on vastly different grounds. For example, Wang et al. in their 2005 and 2006 work showed the breakup and coalescence rates may differ orders of magnitude between different kernels! (but as long as they balance, they can still give similar mean diameters). So when using population balance models, you really have to know what you are doing - which kernels to pick for which problem, if there are any applicable to your problem at all. In my experience, the Luo kernels (in FLUENT) are poorly applicable for industrial-scale gas-liquid flows, for example, leading to strong over-predictions of bubble size (See Haringa et al., CES, 2017, 2018 for more details).

Third, there is often little data on bubble size distributions (surely for industrial-scale gas-liquid flows), so you have no idea of knowing whether your predictions are correct. Why bother taking a computationally expensive, highly specialized approach if you can't even validate? Again there, population balance can be a very useful tool for the experienced user. But at this point they can't be offered as a default function, because if you use them without the proper background knowledge and really considering their necessity and applicability, you have high odds at getting garbage out of your simulation.

There's not a lot of tutorials around for the population balance model., presumably for two reasons. One is that it's specifically stated as an add-on module, so it receives somewhat less attention than the basic capabilities of the program. The more important reason is that it's quite specific and advanced; so when you come to a point where you need the population balance solver, you are probably already an experienced FLUENT user, you know how the interface works, and how to set up a multiphase flow workflow. So a tutorial on "how to work with module X" is not that relevant anymore; you may need to know physical interpretations of certain parameters or differences between kernels, but tutorials often don't give you that information (the theory guide and its references are for that). They just show you how the interface works and where to find things.

So what do you need to know about the population balance model?

Respected

You need to know the physical nature of the defined parameters. The How to Work with Module Tutorial is not relevant.

All the best

Cees Haringa : Thank you for your response. I wonder, when we can get the step to step tutorials for complex 3D simulation, why don't we get at least an introduction about the PBE. I feel, most of the researchers keeping the solving techniques of PBE as a secret, with noted technical terms instead of how to do that.

I could find plenty of materials for multiphase modeling (for instance bubble column), but why most of them neglected to consider aggregation and coalescence when they have a tool to solve?

PBE is far from standard practice, but FLUENT is not at all secretive about having it or about how it works - there's detailed documentation on the add-on modules. The basics of the model are also well explained and straightforward, but yes, mathematically the equations can be extensive and take some time to fully grasp. Engineers are being paid well for a reason ;)

Why tutorials don't include PBE's is simple, as I said, it's not standard practice to use them, and it's not necessary to use them in order to understand how to set up 2-phase flow problems. A single-bubble size approach works well for that (and is much faster). The reason the use of PBEs is not standard has several reasons. For one, they are computationally extremely expensive to run. So until recently, their use was simply not feasible for most problems.

Second is that they are practically not well developed. Yes, the theory behind population balance modeling is old and well known, but the kernels capturing breakup and coalescence are highly empirical, differ widely in their assumptions and simplifications, and are often very case dependent. A kernel that works well for solid particles may not do well for gases, one that does well for heterogeneous bubble columns may not do well for stirred tanks. Even two sets of kernels that both give good predictions for the same flow problem, may do so based on vastly different grounds. For example, Wang et al. in their 2005 and 2006 work showed the breakup and coalescence rates may differ orders of magnitude between different kernels! (but as long as they balance, they can still give similar mean diameters). So when using population balance models, you really have to know what you are doing - which kernels to pick for which problem, if there are any applicable to your problem at all. In my experience, the Luo kernels (in FLUENT) are poorly applicable for industrial-scale gas-liquid flows, for example, leading to strong over-predictions of bubble size (See Haringa et al., CES, 2017, 2018 for more details).

Third, there is often little data on bubble size distributions (surely for industrial-scale gas-liquid flows), so you have no idea of knowing whether your predictions are correct. Why bother taking a computationally expensive, highly specialized approach if you can't even validate? Again there, population balance can be a very useful tool for the experienced user. But at this point they can't be offered as a default function, because if you use them without the proper background knowledge and really considering their necessity and applicability, you have high odds at getting garbage out of your simulation.

You can find information about the PBM in the Fluent Advanced Add-On Modules.pdf.

The population balance module is loaded into ANSYS Fluent through the text user interface (TUI): define/models/addon-module

There are various macros DEFINE_PB_* to be used in UDF to consider break-up, coalescence, nucleation and growth.

Cees Haringa Thank you for the detailed explanation.

Thank you Philippe.