Control of Cyclone Flow Instabilities for Increased Efficiency, Reduced Pressure Loss and Lower Noise

Lead Research Organisation: University of Cambridge
Department Name: Engineering


The precessing vortex core (PVC) in a cyclone separator is detrimental to separation performance, accounts for 80% of velocity fluctuations, increases aerodynamic tonal noise, and dissipates mechanical energy thereby increasing the pressure loss. The PVC arises because the swirling flow in the cyclone is hydrodynamically unstable. Recent studies at Cambridge (Grimble PhD, Mehl MPhil, Tsailouski & Foster MEng, Inigo PostDoc) have shown that the two regions of the flow that cause this instability (the wavemaker regions) lie (i) in the neck between the cyclone and the dust collector and (ii) at entry to the vortex finder. In region (i), the instability is driven by the strong shear in this region and is not convected out of this region because the same mass of air flows in both directions. These studies have also shown that it may be possible to control this instability by altering the flow, although they have not investigated the consequent influence on separation efficiency.
The overall aim of this project is to reduce the amplitude of PVC oscillations or to eliminate them entirely, without adversely affecting separation performance. Intermediate milestones are:
1 to find the wavemaker region of the instability using local and global stability analysis;
2 to perform an adjoint base state sensitivity analysis to identify how the flow should be changed to reduce the growth rate of the instability;
3 to alter the flow in an existing CFD simulation to test the results of the stability analyses;
4 to perform an adjoint double-decker sensitivity analysis to identify changes to the shape of the boundary that will reduce the growth rate of the instability;
5 to develop a numerical model for separation efficiency of the cyclone, building on work by Dyson and its collaborators, and to validate it against experimental data by Dyson and in the open literature;
6 to work with engineers at Dyson to incorporate this knowledge


10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509103/1 01/10/2015 31/03/2022
1611004 Studentship EP/N509103/1 01/10/2015 30/09/2019 Jack Brewster
Description In this award, we developed adjoint-based sensitivity analysis to the unstable oscillations in a generic flow and then applied this to Dyson's cyclonic separators. This analysis shows how to change the shape of the cyclone in order to stabilize the flow. We then wrapped this inside an optimization algorithm and showed how to minimize the growth rate of oscillations by changing the cyclone shape, while retaining centrifugal separation in the cyclone.

The award objectives were met almost entirely. The only objective that was not totally met was to include sophisticated metrics of particle separation, but this will be a relatively simple addition once the metrics have been agreed.
Exploitation Route The findings will be taken forward by Dyson's research team.
Sectors Other

Description Previous work has linked an unwanted noise, 'cyclone hum', to the precessing vortex core (PVC) in a cyclone separator. In addition to this unwanted noise, the PVC also causes reduced performance, increasing the pressure loss over the cyclone. This PVC arises from a hydrodynamic instability of the swirling flow. We have developed shape optimisation techniques that can be used to modify the growth rate and frequency of this instability. These techniques have been applied to produce the geometry of a cyclone separator with an adjusted frequency. This new design is being tested by Dyson to validate the process.
First Year Of Impact 2017
Sector Other
Impact Types Economic

Description Tech Transfer to Dyson 
Organisation Dyson
Country United Kingdom 
Sector Private 
PI Contribution The University of Cambridge, in this research project, developed software that can devise the shape change that has most influence on reducing the sound from a cyclonic separator. The code developed has been transferred to Dyson and the student spent time at Dyson showing Dyson's engineers how to use the code.
Collaborator Contribution The partners (Dyson) provided the initial motivation for the project, ongoing technical advice, ongoing motivation to re-direct the project as necessary, and contact with other academic researchers. They also provided money (through a CASE studentship).
Impact Single discipline (Fluid Mechanics) Publication: 10.1016/j.euromechflu.2019.11.007
Start Year 2016
Title Optimal Shape Deformations 
Description Python software that automatically identifies the most influential ways of deforming a geometry in order to change a flow-field. These allow the most import parts of the geometry to be identified before performing shape optimisation and so help influence which regions of the geometry should be included in any parametrisation. This provides engineers with insight into how the geometry could be changed without the cost of performing the whole shape optimsation process. 
Type Of Technology Software 
Year Produced 2018 
Impact This software was used to identify the most influential parts of the cyclone and so help decide which parts of the cyclone should be allowed to change during the shape optimisation process. It has been installed at Dyson in order to be used for other flow problems.