Hello all,
I have a problem on establishing the simple 1D model for multi-tubular reactor with exothermic reactions. The packed-bed is cooled by coolant to dissipate the reaction heat.
Since this 1D model will be employed to a hybrid-approach that covering another data-driven deep learning ANNs, the 1D model must be ultimately simplified.
In general, the constraints for that simple 1D model are:
1. Minimize the number of equations (only mass and energy balance. pressure drop, diffusion, dispersion are neglected)
2. Minimize the number of empirical correlations: I need to find the good overall HTC for the packed-bed reactor. Also, I need the good method for estimating the mass-transfer limitation.
3. Maximize the validity range of model: We need a wide range of application.
Can you help me to find:
1. The simplest governing equation
2. The simplest correlation
for such kind of problem?
Thank you in advance,
Hello
Please go through : https://core.ac.uk/download/pdf/213394653.pdf
You will complete insight and answers of your questions related simple 1D model for multi-tubular reactor.
Thanks for your suggestion,
However, this 1D model is not the simplest one, I think.
It is not possible to bring this model to my approach. Training the ANNs will be very complex.
Regards,
A Hydrogen plant furnace is a multi-tubular reactor that often has hundreds of tubes. It is often modeled be assuming a single average tube and multiplying the product rate by the number of tubes.
Thank you all for your help.
I already found my solution through this book:
G.G.F. Froment, K.B.K. Bischoff, J. De Wilde, J. De Wilde, Chemical reactor analysis and design, third, John Wiley & Sons, Inc., New York, 2011.
I think the most simple 1D model for such problem can be:
1. Mass + species balance
2. Energy balance + effective Radial and near wall HTCs. Dixon model for those HTCs are correct for wide range of Re.
The fluid properties such as viscosity, heat capacity, thermal conductivity can be calculated via polynomial fitted equations.
I solved the above 1D model correctly using only few lines of code. The difference between rigorous 1D model ant this one is less than 10%.
- Badges
- Science topic