I am looking into running a CFD simulation of a secondary chamber of an incinerator that is used for disposing of the bodies of infected cows.

The incinerator comprises a primary chamber, in which the cows are burnt, initially with some help from gas burners. They then burn under their own steam (if you pardon the pun). Volatiles and water vapour from the cows pass into a secondary chamber that is maintained at about 1000 deg C before exhausting to the atmosphere. What I want to do is predict the residence time of the gas in this secondary chamber.

I appreciate the complexity of this problem, but it must be possible to estimate the amount of carbon, hydrogen, etc in a cow and hence approximate the volatiles released and the energy released as they combust. We do have some test data from a similar incinerator to help us here.

Does anyone have any suggestions as to how this can be done? Are there any references out there on similar topics? It occurs to me that someone must have looked at combustion of similarly complex biological materials (Fluent includes wood combustion information in its database). Also, I'd be interested in any comparisons between CFD derived residence times and test data for incinerators and other vessels.

Thanks,

Neil

I guess you can use a feed back loop to control the residence time of gas as well as flow flux.

I don't know much of combustion, but there is some thing similar in waste water treatment. The waster water is introduced in to a chamber(a tank) for the chemical reactions, deposition processes and there is a outflow pipe to take the clear water out and still another pipe loop back to the inflow pipe to control the residence time of waster water in the chamber.

Your problem sounds interesting,

wen

Have you tried modelling this with CFD? Do you get a good comparison with test data?

Cheers,

Neil

Hallo

you need the volume of your secondary chamber. This is easy to find out. Under stationary conditions everything that is coming in, comes out, meaning that you don`t have any accumulation of gases. What you need is the amount of gas leaving your apparatus through the exhaust. You only need to find out the flow rate (m**3/sec) of the exhaust and measure the temperature at the same point. Then you find the volume of your gases in the temperature of your chamber, and by deviding this number with the volume of your chamber you get the residense time. Surely you have a residense time distribution, and the temperatures are not known exactly and the flow rate is difficult to estimate. But i bet; this method is more accurate than any CFD calculation. The input data you need for CFD are almost the same and through the calculation you don't gain in accuracy.

best regards Jannis

I agree that you can make a simple estimate of residence time based on vessel volume and volumetric inflow. However, this estimate will miss two important issues.

1. Flow around corners, baffles etc generates regions where the flow is stagnant. These dead areas reduce the effective volume of the vessel.

2. The gases are reacting. The mass flow of gas entering the chamber is the same as the gas mass flow leaving the chamber. However the volume of gas entering the chamber is not equal to the volume leaving the chamber. The gas entering the chamber will include fuel, oxygen and other gases. The gas leaving will be hotter and will include combustion products, less oxygen and possibly unburnt fuel. The CFD simulation will give us information on the chemical composition of the exhaust gas (assuming we adequately approximate the fuel combustion process).

Cheers,

Neil

OK, I am not sure if your problem might be solved via a CFD calculation. You need many input data, that you actually don't have, and where I don't see any possibility to get them. Don't forget: Bullshit in = Bullshit out.

So what do you need? Information about your chamber; you can get it, you can analyse and optimize your chamber with CFD. So you can get no stagnation points for example. Makes no sense, since you have a chamber.

You need detail information about the gases coming to your chamber. You can analyse your exhaust gases coming in to your chamber and find out which are the main products in it. In combustion exhaust gases you usually have a lot of nitrogen, CO, CO2, and you may use some components representing a larger group of comustion products. So the first task is to find these products! if might get a lot of time; and you need experience since your focus points define the product groups you are interested on. I mean, are you interested on minimizing the amount of NOx or the amount of hydrocarbons? This is the most difficult part.

Then you need detailed information about the chemical reactions taking place. This is a difficult part, since you usually have no idea, which components are in, and you approximate your system. If you use many equations you get a stiff system of differential equations. Don't try a coupling of CFD and Chemistry. This is always difficult, and no matters what CFD sellers say: it doesn't work always! You need a turbulence model, a solution for your chemistry and many many diskretisation points in order to get good results. And much computation time!

As you have seen you must make many assumptions. Every assumption is leading to inaccurities. So try to define what exacktly you want. Other questions like the residence time might be approximated. Try to find the flowrates coming in your chamber and try to define a middle temperature. This is a rough aproximation. The chemistry questions are more interestig, and more difficult to solve! If you have any questions; don't hesitate and ask.

best regards

Jannis

I agree with a lot of what you're saying.

The value I am primarily interested in determining is the residence time. Realistically, the most complex model I am likely to run will include one or two combustion reactions. This is obviously a gross assumption. There are probably thousands of complex reactions going on as proteins and fats break down to CO2, H2O etc. This is probably true of nearly all CFD combustion problems e.g. coal, oil, natural gas combustion, fire modelling etc. However, people still make assumptions and use CFD to produce useful Engineering data in combustion problems. A gross approximation is OK as long as you understand how large an error it generates in the predicted value.

For example, I could do this -

1. I know the mass of air and cow going into my process. I could work out how much carbon there is in a cow and approximate the volatiles coming off it to CH4 and assume all the carbon emitted is in the form of CH4.

2. I could then define a single reaction process as being -

xCH4 + yO2+ zN2 -> a CO2 + b H2O + z N2

3. Then all I need to do is define the mass flow of the other gases entering the vessel and run the CFD simulation.

Lets assume the actual residence time is 2 seconds. If the above method predicts 1.8 seconds then it's good enough for me. If it predicts 102 seconds then it's not much use. I'm sure that this will be case sensitive - if the reaction doesn't affect the flow much then the above approximation will probably be OK.

People have definitely used CFD to model complex combustion problems, such as waste incineration using single reaction approximation. They must make an intelligent (?) guess at a representative fuel oxidisation reaction. The question is how good are these guesses and what methods do people use? Does anyone know the chemical equation for vapourised cow? Are there any good papers out there?

Thanks,

Neil

If you are mainly interested in the residence time a simple experiments might do it. Tracer experiments would give you an answer.

Concerning the reactions; a cow is somehow a new material in the combustion technology. You know the amount of fuel and air coming in; you know the amount of exhaust gases leaving your system and the temperature. Don't overestimate the reaction part, since you have few information about it. CH4 represents carbonhydrates that are difficult to oxidise, but this is the best aproximation you have at the moment. I would try to find an other compound that releases a comparable amount of CO2 and heat. I would say cow burning is an exthermic process; try to estimate the amount of gas released from burning (big calorimeter and steak experiments!)

Articles about coal combustion would be very helpful.

Don't forget to use the DISI principal (Do It Simple Idiot), where possible; good luck. If you need more help, ask!

best regards

Jannis

Why bother modelling an incenerator - in the UK we just burn cows in a field.

You ever tried modelling highly buoyant sources issued at ground level into an oncoming atmospehric boundary layer to get good ground level concentration predictions? Not easy. Stick with the chamber Neil.

The DISI method may also require you to forget about reactions. As far a residency time is concerned (or effective ventilation rate) gross flow features such as recircs, stratification, plumes etc. are all that would be required to be predicted imho.

A good HVAC cfd code would be good enough. Flovent, Airpak etc.