Post by fiedia on Jun 24, 2018 5:36:59 GMT -8
The purpose of this thermal model is to scale bells for mass rocket stoves. It will compute different bell temperatures as well as heating power.
For experimental set up, please read this : donkey32.proboards.com/thread/3346/brick-prototype-improves-standard-efficiency
Disclaimer :
This thermal model is made to help scaling a multibell RMS. It will take into account bell geometry and materials, and run on an Excel sheet. It will not give such an accuracy as finite element models do, but it should help understanding the impact of the bell design on the system efficiency.
Very few meaurements were available.and they came from a unique bell prototype and heat source. This model is based on general thermal engineering and fluid mechanics. Some assumptions had to be made due to lack of experimental data. These assumptions work well with the present set up but the model may not stick to reality when varying bell geometry or heat source.
Approach :
- Phase 1 : study of the cooling phase. The bell has been heated to a given temperature, then its entry and exhaust closed and its skin temperature recorded. The bell behavior is in this case easier to describe as no heat source is involved. Exclusively cooling through radiation and convection takes place. This step gave radiative and convective coefficients for the bell prototype.
- Phase 2 : study of the heating phase. The bell has been heated on a high heat. It is much more complex to describe as it involves a mix of forced and natural convection inside the bell + conduction through the bell walls + radiation and convection outside the bell. In addition, the heat source is not perfect as its temperature and smoke flow may vary. This step gave a full model from fire start to end of bell cooling.
Results :
- The cooling model overestimates slightly the bell cooling rate. Nevertheless it is quite accurate compared to the measurements uncertainties. It will be used without calibration coefficient.
- In contrary, the heating model required calibration and many assumptions :
o Heat source flows and temperatures had to be estimated since no measurement was available and the stove supplier offers average values which do not fit to our high eat experiment.
o No description of the heat build up inside the bell was found. The forced convection in short pipe model really underestimates heat transfer. In fact, for this experiment, the temperature rise depends apparently on the bell thermal capacity exclusively. As if there were unconstrained heat transfer between smokes and walls. It is probably true for the present bell cross section / heat flow ratio. But it should be checked with higher heat flows or narrower bells.
Running the final model for the bell prototype (graph bellow) gives details about temperatures inside the bell as well as heat power absorption and emission. The prototype bell provides 2 kW max heating power. It is perceived as much more efficient as a standard 2kW heater as 55% of this heat is radiated.
To improve the model :
- It would be interesting to set a new experiment with several thermocouples at the bell entry, exit, skin at different heights, inside flues and measure the mass flow at he entry of the fireplace or at the chimenea exhaust (anemometer + thermocouple ?).
- Different bell sizes and heat sources should be tested.
- Study how heat builds up inside the bell.
- Study the heat transfer through a multiple skin wall. The air gap will probably introduce a radiative coupling.
This model is not perfect but it is interesting now to use it to design a multibell mass heater. This will be presented in a later post.
I would be interested to test this model with other set ups. Has anyone already done the same type of measurements (skin and smoke temp vs time) ?
For experimental set up, please read this : donkey32.proboards.com/thread/3346/brick-prototype-improves-standard-efficiency
Disclaimer :
This thermal model is made to help scaling a multibell RMS. It will take into account bell geometry and materials, and run on an Excel sheet. It will not give such an accuracy as finite element models do, but it should help understanding the impact of the bell design on the system efficiency.
Very few meaurements were available.and they came from a unique bell prototype and heat source. This model is based on general thermal engineering and fluid mechanics. Some assumptions had to be made due to lack of experimental data. These assumptions work well with the present set up but the model may not stick to reality when varying bell geometry or heat source.
Approach :
- Phase 1 : study of the cooling phase. The bell has been heated to a given temperature, then its entry and exhaust closed and its skin temperature recorded. The bell behavior is in this case easier to describe as no heat source is involved. Exclusively cooling through radiation and convection takes place. This step gave radiative and convective coefficients for the bell prototype.
- Phase 2 : study of the heating phase. The bell has been heated on a high heat. It is much more complex to describe as it involves a mix of forced and natural convection inside the bell + conduction through the bell walls + radiation and convection outside the bell. In addition, the heat source is not perfect as its temperature and smoke flow may vary. This step gave a full model from fire start to end of bell cooling.
Results :
- The cooling model overestimates slightly the bell cooling rate. Nevertheless it is quite accurate compared to the measurements uncertainties. It will be used without calibration coefficient.
- In contrary, the heating model required calibration and many assumptions :
o Heat source flows and temperatures had to be estimated since no measurement was available and the stove supplier offers average values which do not fit to our high eat experiment.
o No description of the heat build up inside the bell was found. The forced convection in short pipe model really underestimates heat transfer. In fact, for this experiment, the temperature rise depends apparently on the bell thermal capacity exclusively. As if there were unconstrained heat transfer between smokes and walls. It is probably true for the present bell cross section / heat flow ratio. But it should be checked with higher heat flows or narrower bells.
Running the final model for the bell prototype (graph bellow) gives details about temperatures inside the bell as well as heat power absorption and emission. The prototype bell provides 2 kW max heating power. It is perceived as much more efficient as a standard 2kW heater as 55% of this heat is radiated.
To improve the model :
- It would be interesting to set a new experiment with several thermocouples at the bell entry, exit, skin at different heights, inside flues and measure the mass flow at he entry of the fireplace or at the chimenea exhaust (anemometer + thermocouple ?).
- Different bell sizes and heat sources should be tested.
- Study how heat builds up inside the bell.
- Study the heat transfer through a multiple skin wall. The air gap will probably introduce a radiative coupling.
This model is not perfect but it is interesting now to use it to design a multibell mass heater. This will be presented in a later post.
I would be interested to test this model with other set ups. Has anyone already done the same type of measurements (skin and smoke temp vs time) ?
