At The University of Manchester we have established an academic research group working on engine particle separator systems, and more widely on separation and filter technology for aero engines (inlet barrier filters, inertial particle separators, vortex tubes separators).Through physics-based predictive modelling, we aim to provide performance analysis of auxiliary systems of gas turbine engines to enable improved cycle scheduling and better estimate component through-costs.


We combine physics-based models and computational fluid dynamics solutions of air cleaning systems for gas turbine engines, in order to predict their performance during operational lifetime. This capability affords the operator the opportunity to justify procurement and installation of such devices, via holistic assessment of through-life costs.

We also assist in the development of dust-sand particle separators for air breathing engines, providing technical advice and modelling tools to comprehensively optimise device efficiency for the chosen application. We specialise in analysis of the two main retro-fit air particle separators for helicopter engine: inlet barrier filters (IBF) and vortex tube separators (VTS) .

Our capability includes gas turbine engine off-design performance simulation, which refers to the behaviour of the chosen engine when operating in conditions that do not match the engine settings. We can provide engine data that would otherwise be inaccessible, but can be used to update existing performance charts to account for auxiliary engine systems such as particle separators and engine bleed.

Engine Installation Losses

When an engine is tested in the factory, it delivers the power for which it is designed. When it is installed into the airframe, however, this level of power is often not attained. This is down to what is known as installation losses – sources of power loss such as insufficient mass flow or obstructions to the ingested air stream. Typical causes of such losses are given in Figure 1. Our modelling capability allows us to predict the shortfall in engine performance when one or more of these causes are present. The diagram indicates the shortfall in the physical quantity that is associated to each cause.


Figure 1 Installed engine power loss sources
Figure 1: Diagrammatic summary of typical sources of typical sources of installed engine power loss


Engine Air Particle Separators

Our specialism is in particle separators for helicopters, in particular those operating in harsh environments such as the desert. The two retro-fit technologies are known as inlet barrier filters (IBF) shown in Figures 2 and 3, and vortex tube separators (VTS) shown in Figure 4 and 5.

Figure 2 Inlet barrier filter for the MD502
Figure 2: Inlet barrier filter for the MD502


Figure 3 Sikorsky Blackhawk with Inlet Barrier Filter fitted
Figure 3: Sikorsky Blackhawk with Inlet Barrier Filter fitted


 Figure 4 Vortex tube separator for the Eurocopter Super Puma
Figure 4: Vortex tube separator for the Eurocopter Super Puma



Figure 5 Vortex tube separator for the Eurocopter Super Puma
Figure 5: Vortex tube separator for the Eurocopter Super Puma


Their purpose is to remove dust and sand from engine-bound air in order to prevent damage to key components and prolong engine life. Their cost is a multifarious performance penalty, which must be endured to realise the device’s facility. Their use can cause:

  1. A loss of total pressure
  2. An inhomogeneity of total pressure
  3. A loss of engine air mass flow rate

A combination of these causes a twofold effect on the engine, both in performance and operability as shown in Figure 6. Performance effects relate to the longer term efficiency of the engine – the reduction in static pressure requires the engine to work harder to produce the same power, which translates to more fuel consumed and a higher working temperature for the turbine. A secondary effect of the higher mean gas temperature is a reduction in turbine blade life, and a more visible infra-red signature. Engine operability refers to the real-time ability to operate the engine safely, and attain the requested power without compromising performance. Use of a particle separator limits the amount of bleed flow available, can cause an imbalance of aerodynamic loads through uneven pressure distribution, and can increase the risk of compressor stall or surge.

Figure 6 Diagrammatic summary of the performance pitfalls of employing a particle separator
Figure 6: Diagrammatic summary of the performance pitfalls of employing a particle separator



Inlet Barrier Filters

Our research on Inlet Barrier Filters ranges covers all parameter length scales. For example, IBF performance is as much affected by micrometer-sized adjustments to the filter fibre, as centimetre-sized modifications to the intake geometry. We use state-of-the-art computational fluid dynamics programs to simulate the flow through filters, beginning with a computer-aided design (CAD) model of the specific filter, as shown in Figure 7. The filter media is pleated in order to increase the filtration area. This reduces the overall pressure drop and improves the holding capacity of the medium.

Figure 7 Computer-aided Design model of a pleat section
Figure 7: Computer-aided Design model of a pleat section


There is an optimum pleat density (number of folds per unit length) for a given flow condition. Our research reveals this optimum design point based on the required demands of the helicopter engine and the type of dust to be filtered. The optimum design point arises from the dual nature of the source of pressure loss across the pleated filter. Pressure is lost both as the flow passes through the filter medium, and as it contracts within the pleat channels. Within the media (the solid region in Figure 7), pressure is lost to friction with the constituent fibres and walls of the pore channels. In the pleat channels the flow contraction causes the air to accelerate (shown by red areas in Figure 9), causing local flow shearing in areas of high velocity gradient. As the number of folds increases, the contribution of the loss due to the channels increases in dominance, while the loss contribution through the media decreases as a proportion of the total pressure lost across the whole filter. What results is an optimum design curve, shown in Figure 9.


Figure 8 Contours of velocity magnitude and flow streamlines through half-pleat section
Figure 8: Contours of velocity magnitude and flow streamlines through half-pleat section


Figure 9 Pressure loss as a function of pleat angle with breakdown of loss sources
Figure 9: Pressure loss as a function of pleat angle, with breakdown of loss sources


The occurrence of the optimum design point allows the filter design to be tailored to specific cases. The flow influent to helicopter engines varies from one platform to the next depending on the location and mass flow requirements of the engine. Furthermore, the design of the internal structure of the filter media, critical to predicting pressure loss, is highly dependent on the target particle size. What results is a lack of general solution; each IBF design is case specific. To deal with this, we have developed a performance assessment tool that considers the following:

Vortex Tube Separators

We are developing a similar capability for VTS technology. A vortex tube assembly consists of several hundred tubes arrange on a panel, each of which provides the engine with a share of the required mass flow. Within each vortex tube is fixed a four-bladed helical vane, which spins the influent air causing particles to centrifuge to the periphery of the tube. Further downstream, a small diameter tube bifurcates the flow, sending cleaner air onwards to the engine while the particulate heavy peripheral flow is scavenged overboard. A simplified diagram of a vortex tube is shown in Figure 10.


Figure 10 Simplified drawing of a single vortex tube separator
Figure 10: Simplified drawing of a single vortex tube separator


Our physics-based models indicate that the performance of the vortex tube is based on several design compromises. The first relates to the proportion of flow that is scavenged: drawing a larger percentage of the tube flow allows more particles to be captured at the expense of a shortfall of engine mass flow, which must be met by additional, power-consuming tubes. The second relates to the geometry of the helix: a helix of smaller pitch turns the air in a shorter distance, imparting a larger centrifugal force to the particles, assisting their separation from the core flow. However, the cost is an increase in pressure loss due to friction with the surfaces of the turning vane. Such compromises provide an opportunity for optimisation for a given flow.

The existence of an optimum design also ensures that there are times at which the vortex tube will not be performing to its best performance, during off-design conditions. We are currently developing a capability that allows the performance of a vortex tube separator array to be predicted across the envelope of operating conditions. This will afford a more comprehensive comparison of the system with inlet barrier filters, the alternative particle separation system.