Complete Guide To Designing A Heat Exchanger
Designing a heat exchanger is one of the most common tasks within chemical engineering. It encompasses a wide range of different principles, ranging from heat transfer, to energy balances, and unit operations. Not only is this an integral part of studying chemical engineering, but it is also a highly sought-after skill within many industries.
So how exactly do you design a heat exchanger? Well in this post you will get a step-by-step guide on the steps you need to take to design a heat exchanger from start to finish. You can also find some useful charts and correlations to help with your design.
A General Overview
When designing a shell and tube heat exchanger, it’s important to be familiar with the general steps, assumptions, and key variables to consider. If you are not interested in a detailed design methodology, but would rather simply select the most appropriate configuration, you can refer to the TEMA codes; as seen in the figure below. Here you can select the most appropriate heads and shells to suit your operation.
The Design Process
Now it’s time to fully design a shell and tube heat exchanger. Each step should be carried out in chronological order as many are pre-requisites of each other.
1. Gather initial data
To begin your design you must first have the physical properties of your fluids (both stream and utility). You must also know the inlet and outlet temperatures of your streams; or at least one of them, as this will tell you the amount of sensible heat that should be added or removed.
This is then followed by gathering some initial details about the dimensions of your tubes, i.e. the length, inner diameter, material, fouling resistances, etc. This is needed later in the design process.
2. Determine the suggested duty of the fluid
For the duty of your fluid, we first need to determine the Log Mean Temperature Difference (LMTD). It is a good idea to use this correlation as we can account for either counter or co-current flow, and it is able to handle non-isothermal systems.
Now when you design any shell and tube heat exchanger, you must decide upon the number of passes. This means, in the context of tubes, the number of loops it creates before leaving. For example, if fluid enters and goes straight out, this is a single pass. If the tube bends twice, this is a double pass.
The correction factor is essential for determining the true LMTD, therefore it is dependent upon the number of passes, both tube and shell. You can use the following formula, or the following graphs, depending on the configuration…
Now we can determine our suggested overall heat transfer coefficient. The duty can be found by three different equations; namely… Q=UA∆T or Q=hA∆T or Q=mCp∆T.
In this simulation we want to determine “U”, however based on the fluid configuration and temperatures we can get our estimate through tables or charts. Here you can use the following chart…
3. Determine the estimated U-value
Now that we know the suggested overall heat transfer coefficient, this will allow us to determine the provisional heat transfer area. This can be achieved using the formula…
Once you have determined your provisional heat transfer area, we can now turn our attention to the number of tubes that will be required to achieve these targets. This will then allow us to determine our estimated overall heat transfer coefficient.
For your tubes, aim to use standard sizes that are widely available. You can get these from almost any suppliers website. The formulas you need to use are as follows…
4. Determine the tubside pitch & bundle diameter
Now it’s time to turn our attention to determining the pitch of our tubes and the overall tube bundle diameter, as this will determine the overall size of the heat exchanger.
These are several different pitch configurations, however, generally, we use either square or triangular pitch. The general formulas for the pitch and bundle diameter are as follows, where “n2 and “K” are pitch constants taken from reference text. A table for reference values has also been included…
5. Determine tube side heat transfer coefficient
Now that the number of tubes is known, we can now begin to understand the type of flow we have through Reynolds number. This will provide us with the knowledge of which Nusselt or direct “hi” correlation we must use for determining the tubeside heat transfer coefficient.
In order to do this however we will also need to determine Prandtl number as this accounts for the thermal properties of the fluid; whereas Reynolds accounts for the flow behaviour.
Once these dimensionless numbers are known, you must then decide upon which “hi” correlation to use. Here are the three general possible correlations as a guide…
6. Determining the shell diameter
The shell diameter (Ds) can be found by taking the tube bundle diameter and accounting for the clearance. This is done using the following chart; whereby depending upon the type of head selected will yield a relative ratio clearance value…
This value is then added to the bundle diameter to give the overall shell diameter (Ds).
From knowing the pitch and bundle diameter, we can now determine the number of tubes at the equator of the shell. This tells us the number of tubes at the widest point, which then allows us to add the additional clearance for the overall shell diameter.
7. Baffle design
Now the overall shell dimensions are known, it’s time to consider the baffles that will be placed inside the shell. These improve the efficiency of heat transfer by inducing turbulence.
As a general rule of thumb, the minimum baffle spacing can be based on the length of the tubes. Generally, we use no more than 40%.
However, a more accurate calculation is taking 20% of the tube length. Then using this value, we can determine the number of chambers that would be created. From here we can determine the actual baffle spacing, using the following formulas…
8. Determine shell side heat transfer coefficient
Just like with the tube side, we will need to determine the flow regime first before we can deduce the shell side heat transfer coefficient. To do this we need to use Reynolds, however, we must use a variant of the standard Reynolds number formula.
We first start off with accounting for the new geometries that will affect the flowrate of the fluid. Therefore, we can determine the crossectional area of the shell; accounting for the baffles and tubes…
Now the shell side velocity flowrate can be found, using the following formula… Alternatively there are several other renditions that can be used as well.
When we consider the shell side Reynolds number, we cannot use the standard diameter. Instead, we use the equivalent diameter as this accounts for the tubes and baffles. Depending upon the pitch used, the shell equivalent diameter will be different.
Reynolds number can now be found using the following formula. Furthermore, Prandtl number can also be found as these will both be required for the corresponding Nusselt number correlation.
It’s worth noting here at this stage there are several different correlations to choose from, here we have shown the most common for square pitch, known as the Grimison Correlation. The constants “C” and “n” are based on empirical data taken from several sources, some can be found in our free online resource library.
From here we can now deduce the shell side heat transfer coefficient by rearrangement of the standard Nusselt number.
9. Determine the actual overall heat transfer coefficient
All the information required to determine the actual overall heat transfer is now available, therefore we can use the overall heat transfer coefficient equation.
It’s also important to determine the accuracy of this value based on the estimated value that was determined in stage 3. Typically the acceptable error is +/- 30% however it is advised to aim for no more than 20%.
What happens if your error is outwith the acceptable range? Well here are several areas which can be adjusted to improve the performance. It is also worth noting, other adjustments can be made, however, these are the most common variables to adjust first…
- Change the tube lengths.
- Change the tube length/diameter ratio.
- Alter the number of tube side passes.
- Alter the number of shell side baffles.
- Look at the resistances for optimisations.
10. Calculate the tube side pressure drop
When calculating the pressure drop of a shell and tube heat exchanger, it is important that we consider the tube side and shell side separately. The reasons for this should be self-explanatory at this stage.
We will first consider the tube side. The formulas are empirical, therefore no derivation is required. You may find variations of these equations, however, you will see that they will yield very similar answers.
In order to use the formula, we must first determine the fanning friction factor for the tube side, which you can see below.
You can see based on the chart, that we need to use the tube side Reynolds number along with the relative roughness. Typically for ideal systems, we can assume a smooth pipe, however, this decreases the accuracy of the final solution.
From here we can now use the tube side pressure drop equation…
10. Calculate the shell side pressure drop
The process is exactly the same for determining the shell side pressure drop. The main distinction is the friction factor chart that is used, as it must account for the baffles, and the overall equation is slightly different.
Congratulations! You Have Made It!
IF you have reached this point then I would like to say a huge congratulations, on completing your Heat Exchanger design. You now have the skills, tools, and equations to design and optimise any shell and tube heat exchanger.
As a huge thank you for taking the time to read this post, I would like to offer you a 30% discount on our extensive Heat Transfer Course. Use code HEAT30 and take your heat transfer knowledge to the next level!
You may also like
The Bitumen Production Process
The Role of Chemical Engineering In Sustainable Development
