In this section, we aim to introduce the idea of a hydrocyclone, outline its uses and advantages, and take you through our iterative design process .
A hydrocyclone is a filter designed to separate larger particulates from a contaminated solution. Depending on size and shape, a hydrocyclone can be used to filter anything from sand (large particulates) to microalgae (small particulates). The advantage of a hydrocyclone, and the reason that it is already widely used in industry, lies with its relatively simple design structure, and its lack of moving parts. With no moving parts, a hydrocyclone requires little maintenance and is seen as a cost effective, long term solution. For example, hydrocyclones are used in the oil industry to separate oil from water and vice versa. They are also used in the drilling industry to separate sand from the expensive clay that is used for lubrication during the drilling.
A hydrocyclone works by pumping contaminated water through an inlet feed, which is tangential to a cylindrical feed chamber. A large vortex is then created inside the filter, which forces larger particulates to be pushed out to the walls due to centrifugal forces. These larger particulates are then deposited through the spigot at the bottom, while being carried in some waste solution. Meanwhile, a smaller, inner vortex is generated through the centre of the hydrocyclone, which carries the lightest particles (the filtered water) up and out of the vortex finder at the top. The filtered water is now mostly free of sediment and will not block the second stage of filtration.
Hydrocyclones are essential to our project because we cannot afford to have our metal binding reactor blocked by unwanted debris in the contaminated waste that is being filtered. We have used an iterative design process to maximise the separation efficiency (percentage of heavy sediment removed) of each hydrocyclone. Our final design boasts a separation efficiency of 97.4%.
For the 3D design of the hydrocyclones, we used the software package Autodesk Fusion 360. We decided to use Autodesk Fusion 360, as it is free for students and is extremely intuitive for beginners. The first design was influenced, quite simply, by the shapes of other hydrocyclones seen on the internet. Unfortunately, the 3D printer printed support structures on the interior of the first hydrocyclone, which were impossible to remove. From this model, we did learn that the inlet was marginally too small for the piping and we were afraid of leakages, so we went away and made the design slightly larger in order to ensure a tight fit. To prevent the support structures from affecting the interior of the cyclone, we split the design into four separate components which we glued together using Loctite Ultra Control Gel. We chose to use Loctite Ultra Control Gel following some brief research into the most effective adhesives for PLA plastic. Ideally, we would have used a 3D printer that can print dissolvable, PVA support structures so we could have printed the hydrocyclone as a single piece.
After gluing together the hydrocyclone, we realised that to achieve the desired vortex we needed a more powerful pump than the peristaltic pump we have in the labs. Our initial idea was to source a pond pump from an aquatics supplier, however to achieve the desired 800ml/s (approximately), we would need a pump that would cost somewhere in the region of £120. Unfortunately, this meant that the hydrocyclone was pretty much useless. The only testing we were feasibly able to accomplish was running a tap through the cyclone to explore whether the volume of underflow still vastly exceeded the volume of overflow. Literature suggests that the ratio of overflow to underflow should be approximately 80:20 (Arterburn, R.A., 1982). However, the overflow:underflow ratio was closer to 20:80 with this cyclone. Because of this minor set back, we went back to the drawing board to work on Hydrocyclone 2.
A paper titled The Sizing and Selection of Hydrocyclones (Arterburn, R.A., 1982), enabled us to design the hydrocyclone with much clearer direction. For example, we shortened the cyclindrical feed chamber to promote the development of the inner cyclone. To futher promote this development, we also extended the length of the vortex finder. In order to solve the flow rate problem, we designed hydrocyclone 2 to be much smaller; the total volume is now 20cm3 as opposed to the volume of hydrocyclone 1, which had a total volume of 100cm3. However, we still wanted this design to be able to separate slightly larger particulate contaminants from water, such as sand, so we had to ensure that the inlet and outlets were large enough to prevent clogging.
After the initial testing of Hydrocyclone 2 revealed an overflow:underflow volume ratio of 42:100 (an improvement on Hydrocyclone 1, but still not sufficient), we returned to the literature, seeking instruction on how to adapt the design to increase overflow output.
To increase overflow output, the paper titled: Hydrocyclones for Particle Size Separation (Cilliers, J.J., 2000) recommended increasing the angle of the conical chamber from the cylindrical feed chamber from 20° to 30°. It also suggested that we change the diameter of the vortex finder, as the diameter of the vortex finder should not equal the diameter of the spigot (underflow outlet). It is suggested that the size of the spigot should be within the range 0.1Dc - 0.2Dc, where Dc is the diameter of the cylindrical feed chamber, while the size of the vortex finder should be within the range 0.13Dc - 0.43Dc. We also adapted the inlet feed so that it was rectangular as opposed to circular. In the image it looks square, however that opening tapers down to a rectangle with a height to width ratio of 2:1 (still tangential to the cylindrical feed chamber).
Initial testing of the Hydrocyclone was to discover whether the perfect overflow to underflow ratio of 80:20 could be achieved at a particular flow rate. After some initial optimisation of flow rates, we discovered that the perfect ratio could be achieved with a flow rate of 143ml/s. The experiment we conducted was to connect the cyclone to a pump (a simple DC pump from an aquatics store, capable of 1200L/h), and run the experiment until the overflow outlet had been filled to 1L. Our results then showed that at 143ml/s, for every litre of overflow, we had 250ml (±5ml) of underflow. The experiment was conducted three times and an average was taken. The next round of experimentation followed the protocol as detailed below. We tested the hydrocyclone's ability to actually filter larger particulates from the water.
Below is the lab protocol for the next round of experimentation:
A downloadable copy of this protocol can be found here.
You will need:
- Hydrocyclone 3.
- DC pump, capable of flow rates of up to 1200L/H.
- Magnetic Stirrer
- Necessary tubing.
- Aragonite sand solution with a sand to water ratio of 10:90.
- 1 x 5-litre beakers.
- 2 x 2-litre beakers.
- heavy duty scales + a set of more sensitive scales for measuring the volume of sand.
Set up of the experiment:
- While turned off, and unplugged, stick the DC pump to the wall of the 5L beaker, with the inlet nozzle facing the base, 1 inch inch from the bottom.
- Weigh out 500g of aragonite sand and pour into the sample solution beaker. Then fill the beaker to the 5L marker with water.
- Set up a clamp stand over the 'Underflow beaker' and secure the hydrocyclone vertically.
- Connect the necessary tubing; pump to inlet feed, vortex finder to overflow beaker, spigot to underflow beaker. Use tie wraps to secure.
- Place the sample solution beaker (5L) onto a magnetic stirrer and initiate the stirrer to keep the sand in suspension.
- Before conducting the experiment, ensure that the stopwatch is to hand.
- Instantaneously turn on the DC pump and initialise the stopwatch.
- (Before the water level falls below the pump inlet) Turn off the pump and stop the stopwatch. Record the time.
- Leave the underflow and overflow solutions to settle overnight.
- Carefully pour the excess fluid from both beakers into a measuring cylinder, being careful not to disturb the settled sediment. Note the volume of water.
- Dry the remaining sediment in an oven.
- Using the scales, measure the weight of the underflow sediment. Note the ratio of underflow sand:water.
- Repeat this step for the overflow.
Following the protocol above, we conducted three experiments on the third hydrocyclone. We found that our hydrocyclone separated aragonite sand from contaminated water with 97.4% efficiency, leaving only 2.6% of sediment still in solution. Our design also met the industry standard of an underflow:overflow ratio of 20:80.
To calculate the separation efficiency, we simply divided the underflow sediment weight by the total sediment weight. It is very important that the sand is completely dry before measurements are taken.
By following the design, build, test and learn cycle we were able to successfully produce a hydrocyclone that met our team's specific requirements. In doing so, we have demonstrated that this form of early-stage filtration is a cheap, long lasting, and widely applicable solution to a number of filtration problems.
Arterburn, R.A., 1982. The sizing and selection of hydrocyclones. Design and Installation of Comminution Circuits, 1, pp.597-607.
Cilliers, J.J., 2000. Hydrocyclones for particle size separation. Particle Size Separation, pp.1819-1825.