Why Dimples on Jacketed Fermenters?

[doug293cz]

I work in turbomachinery and develop heat transfer technology, so a sectional dimension is the only element missing to understand where this design sits in HT space. The principle is correct that turbulent trips enhance heat transfer, but one can also design in an unintelligent manner and achieve higher HTCs but higher pressure drop, thereby slowing flow and achieving lower overall heat flux.

In this application, the dimples look sized for establishing cylindrical concentricity rather than helping to maintain turbulent flow at lower flow rates.

The real metric is total energy cost required to cool the vessel. This has to include the cost of cooling the glycol, and the cost of running the pump. The overall rate of cooling is easily measured. You just need to know the coolant inlet, temperature, the coolant outlet temperature, the coolant heat capacity, and the coolant flow rate. The heat removal rate is then:

Cooling Rate = Heat Capacity * Flow Rate * (Inlet Temp - Outlet Temp)
​

Heat capacity is in units of energy / (mass * temp delta), Flow rate is in units of mass / time, thus cooling rate is in units of energy / time or power. If using Imperial units then heat capacity is BTU / lb / degree F, flow rate is lb / hour and cooling rate is BTU / hour. So, you can get the same cooling rate at higher flow and lower delta T, or lower flow rate and higher delta T. To get the total power required for cooling you convert the pump power required from horsepower to BTU/hr, or the cooling rate from BTU/hr to horsepower. Then you add the cooling power, the coolant chiller power, and the pump power to get the total power required.

Until you know the total power required for one type of jacket & vessel vs. another type, you can't say which one is better than the other. Comparisons based on heat transfer coefficient between the fluid and vessel, or flow rate thru the jacket are meaningless.

Brew on

 

[Beholder]

The real metric is total energy cost required to cool the vessel. This has to include the cost of cooling the glycol, and the cost of running the pump. The overall rate of cooling is easily measured. You just need to know the coolant inlet, temperature, the coolant outlet temperature, the coolant heat capacity, and the coolant flow rate. The heat removal rate is then:

Cooling Rate = Heat Capacity * Flow Rate * (Inlet Temp - Outlet Temp)
​

Heat capacity is in units of energy / (mass * temp delta), Flow rate is in units of mass / time, thus cooling rate is in units of energy / time or power. If using Imperial units then heat capacity is BTU / lb / degree F, flow rate is lb / hour and cooling rate is BTU / hour. So, you can get the same cooling rate at higher flow and lower delta T, or lower flow rate and higher delta T. To get the total power required for cooling you convert the pump power required from horsepower to BTU/hr, or the cooling rate from BTU/hr to horsepower. Then you add the cooling power, the coolant chiller power, and the pump power to get the total power required.

Until you know the total power required for one type of jacket & vessel vs. another type, you can't say which one is better than the other. Comparisons based on heat transfer coefficient between the fluid and vessel, or flow rate thru the jacket are meaningless.

Brew on

You have the principles right, though based on your assessment, one would do this assessment empirically, since I presume you do not have detailed knowledge of all the system component characteristics on which to develop a model for assessment. (Since naturally you would need the correlations within each system in order to accurately model it...)

Instead, comparisons based on heat transfer can be used to assess pump cycling time, which is a reasonable assumption given that your pump does not vary in power demand over flow resistance. Thus, lower power consumption is achieved via shorter duty cycle of the pump.

In addition turbulent transition will shift somewhat based on fluid temperature, but not significantly ~10 C, as the typical range of optimized glycol fluid temp.

In any case, no need to go to first engineering principles to solicit agreement - not sure your background and experience, so happy to simply agree to disagree if my arguments aren’t compelling.

 

[doug293cz]

You have the principles right, though based on your assessment, one would do this assessment empirically, since I presume you do not have detailed knowledge of all the system component characteristics on which to develop a model for assessment. (Since naturally you would need the correlations within each system in order to accurately model it...)

Instead, comparisons based on heat transfer can be used to assess pump cycling time, which is a reasonable assumption given that your pump does not vary in power demand over flow resistance. Thus, lower power consumption is achieved via shorter duty cycle of the pump.

In addition turbulent transition will shift somewhat based on fluid temperature, but not significantly ~10 C, as the typical range of optimized glycol fluid temp.

In any case, no need to go to first engineering principles to solicit agreement - not sure your background and experience, so happy to simply agree to disagree if my arguments aren’t compelling.

Not trying to describe how to model the system, just describe a pretty simple way to measure the actual performance at a "black box level." This avoids having to actually know thermal transfer coefficients, Reynolds number in the jacket, flow distribution in the jacket, and other complex things like that.

Yes, shorter duty cycle means less average power for a particular pump.

Brew on

 

[SanPancho]

The reason for the design is cost efficiency. These are single wall tanks with jackets, but no insulation.

A quick recap of a nano unit from china-
2bbl domed brite
Single wall- 1300
Add dimple- 1400
Jacketd, insulated, shelled- 2200

Thats a +50% premium for insulation and external shell. And quite a bit more shipping weight.