CORDIS - EU research results

Valve hIgh PERformances for flow control separation in aircraft

Final Report Summary - VIPER (Valve hIgh PERformances for flow control separation in aircraft)

Executive Summary:
The objective of the VIPER European Cleansky project is to develop new technologies for future aircraft enabling a 20-30% fuel burn reduction and related CO2 emissions and a similar reduction in noise levels compared to current aircraft. One of the ways to reach this goal is to improve the aerodynamic performances of current high lift devices.
Active flow control is unanimously seen as the best mean to reach this objective. By suppressing flow separation and/or delaying stall, active flow control will increase wing aerodynamic performances.
The partnership between CEDRAT TECHNOLOGIES (CTEC) and ONERA in the framework of the VIPER project has led to the design, manufacturing and test of an innovative pulsed jet actuator based on a CTEC amplified piezo-actuator (APA). Its aim is to provide a pulsed sonic jet up to 500Hz with a mass flow around 34 g/s through a slot 1mm wide and 80mm long. Coupled with CTEC SA75D switching power amplifier this actuator produces the expected sonic jet with an electrical consumption around 40W thanks to energy recovery.
As results of the actuator characterisation (mechanical, fluidic), the follwing targeted performances of the valve (ejection speed and mass flow) have been achieved and are beyond the state of the art:
• Air exit velocity : 340m/s
• Air linear mass flow : 425 g/s/m
• Actuation frequency range : 0 - 500Hz

Project Context and Objectives:
VIPER : Very hIgh PERformance valve for flow control separation
State of the art – Background
The VIPER (Valve hIgh PERformances for flow control separation in aircraft) Cleansky project aims at developing, manufacturing and testing piezoelectric-based high speed valves to accurately control the air flow (Active Flow Control Actuators) on transport and business aircrafts. The objective is to delay the separation of the boundary layer in flows over airfoils which results in drag reduction.
The project is coordinated by Cedrat Technologies and performed with the help of Onera DMS, acting as subcontractor and Fraunhofer ENAS, acting as topic manager. CEDRAT TECHNOLOGIES (CTEC) & ONERA DMS have been working together for over 4 years in a close partnership over high-performance pulsed jets valves based on APA® (Amplified piezoelectric Actuators). The main advantages of using piezoelectric technology for this type of valves lie in the controllability and fast-response time.
Increasing concern on the part of the government and industry resulted in creation of the ambitious Clean Sky research partnership to develop technologies that reduce the environmental impact of air transport. Wings are an excellent target for such programmes because smooth laminar flow contributes to drag reduction that in turn reduces noise and vibration as well as fuel consumption and emissions related to efforts to steady the plane.
Flow separation refers to a sort of detachment of the air flow from the aircraft and it creates vortices and eddies that can increase drag. The EU is funding the project 'Valve high performances for flow control separation in aircraft' (VIPER) to minimise flow separation with electrical systems. This will maintain more favourable aerodynamic conditions over wing surfaces and will support more-electric aircraft with fewer hydraulic and pneumatic systems.
The targeted performances of the valve (ejection speed and mass flow) are beyond the state of the art:
• Air exit velocity : 340m/s
• Air linear mass flow : 425 g/s/m
• Actuation frequency range : 0 - 500Hz

Description of work
VIPER is developing an active flow control device to produce a quasi-sonic (500 Hz) pulsed jet of air to increase lift and decrease drag. It is the first time that such high frequencies of pulsed air will be investigated. High-lift capability will reduce required thrust during take-off and landing, reducing noise and emissions in the vicinity of airports. Increased lift and decreased drag translate into lower fuel consumption and fewer emissions.
The technology exploits amplified piezoelectric actuators. Piezoelectric materials convert a pressure into a voltage or vice versa. Amplified piezoelectric actuators are specialised piezoelectric actuators that essentially amplify the displacement relative to the voltage of traditional piezoelectric actuators.
In the first year, the team considered several configurations to meet the specifications and chose the most promising valve configuration. Some composite actuators have been manufactured and tested, with results comprising the subject of a publication for the Actuator 2014 International Conference and Exhibition on New Actuators and Drive Systems in Bremen, Germany.
The following years have been used to develop an engineering model of VIPER valve with its electronics for driving and control, and to test it as regard electromechanical and flow performances.
Finally the results are in line with the objectives.
VIPER's active flow control device will significantly decrease noise and emissions associated with air travel, providing important relief from the pressures faced by the public and the planet. Customers have demonstrated great enthusiasm for the actuator technology, which could have widespread application in addition to the current one.

Project Results:
a) Timeline & main milestones
The duration of the VIPER project is 48 months. It stated in November 2012 and ends in April 2016.

b) Results

The main objective of European research in aeronautics is to reduce the fuel burn and environmental impact of current aircrafts by improving their aerodynamic performances. One of the means to fulfil this target is to control the airflow around the wings, rudders, etc. At high angle of attack, the apparition of flow separation decreases the aerodynamic performances on two aspects: the lift is reduced and the drag is increased. Therefore, the main objective of active flow control is to keep the flow attached on the largest part of the wing.

Several techniques are investigated to control the flow. In the framework of the VIPER project, a pulsed blowing jet is studied (see Fig. 1). This kind of actuator uses the engine bleed air and blows it along the span which allows a delay or a suppression of flow separation. The air which is blown through the actuators adds energy to the boundary layer and prevents it from separating from the wing.

Figure 1 – Concept of blowing to suppress flow separation.

VIPER actuator features
Many pulsed jets actuators have been designed and tested in the past.
VIPER actuator is innovative because:
• The mass flow rate reached is the highest one obtained on an actuator within this volume (more than 420g/s per meter span at 500Hz);

• The improved efficiency when the actuator is coupled with CTEC switching amplifier SA75D;

• The high power density of the actuator;

• The enduring lifetime.

Figure 2 – VIPER pulsed jet fluidic actuator and its driving electronic SA75D (lab version).
The fact that the motion is created with piezo ceramics makes it very interesting in terms of bandwidth, compacity and weight.
In parallel to the actuator development, CTEC has designed a specific switching power amplifier called SA75D. This power amplifier is very efficient. Its electrical consumption is around 40W while producing 3.4kVA on the actuator side.

Determination of APA specifications
The fluidic actuator was designed by ONERA so that its characteristics are compatible with the bandwidth and flow rate specified. It led to determine the stroke and the blocked force needed for the APA to be used to drive the valves.
The jet exit area depends on the length of the slot that can be driven by the fluidic actuator. Basically this value is directly linked to the dimension of the APA.
In order to produce a sonic jet through the slot the area for flow circulation through the valves driven by the APA must be larger than the slot area itself. It leads to a minimum value for the stroke of the actuator.
Furthermore the fluidic actuator is designed in the ways that without electric power supply the valve is closed. The piezoelectric actuator must be designed so that the opening of the valve is possible in this configuration. The difference of pressure between the inside of the actuator and the volume downstream the valve which depends on the pressure loss in the fluidic actuator allows one to estimate the requested blocked force of the APA.
This study was based on fluidic actuators using the same principle that were designed and manufactured by ONERA within previous studies for wind tunnel models.

Numerical simulation of the flow inside the actuator
The numerical simulation has been performed with ONERA's unstructured solver CEDRE. The Navier-Stokes solver of the CEDRE code is a fully unstructured solver developed at ONERA with main applications in the energetics and propulsion fields where it can be coupled with other solvers to perform multiphysics simulations [1]. Figures 3 and 4 show different views of the unstructured grid

Figure 3 - 3D view of the mesh.
The work has been carried out as follows. A first geometry has been designed by the ONERA model shop. Then, a first numerical simulation of the flow inside this actuator has been performed. Shape modifications have been proposed to suppress the recirculation zones in order to decrease the pressure loss and improve the velocity homogeneity at the slot exit. These modifications took place mainly in the diffuser and the slot. Finally, a second computation has been done to check the effect of these shape modifications.

Figure 4 - Details of the mesh in the diffuser region (top) and in the slot region (bottom).
The slot geometry has also been modified as shown in Figure 5 to suppress the recirculation zones and so decrease the pressure losses and improve the flow homogeneity at the slot exit.

Figure 5 - Mach number field in the slot (spanwise plane at a quarter of the slot width) before (top) and after shape modification (bottom).
The actuator bandwidth is limited by the flow volume between the valve and the exit slot. A simple model has been developed to estimate this bandwidth. It consists in an isentropic model of the flow inside the actuator cavity that treats the volume filling as a series of isentropic and adiabatic compressions and expansions and flow through the orifice as inviscid has been used.
This model enables to compute the velocity signal at the orifice exit and in particular to find the actuator bandwidth in terms of time response of cavity. It also enables to quickly optimize the volume cavity and the slot width in order to fulfil the actuator bandwidth requirements.
Figure 6 shows the estimated response of the actuator for two frequencies: 375Hz and 500Hz.
Up to 375Hz, the expected peak exit velocity is constant equal to Mach = 1. Beyond this frequency, the peak velocity starts to decrease. At 500Hz the exit peak velocity is close to Mach = 0.97. Nevertheless the closing of the actuator is still observable.

Figure 6 – Mach number at the slot exit as function of time.
The resonance frequency of the APA 1000L being highly greater than 500Hz (1320Hz), one can assume that the actuator bandwidth will be about 375Hz and that a functioning at 500Hz with a slightly lower peak velocity is possible.
Actuator mechanical characterisation
The VIPER prototype was tested at CTEC to measure the piezo actuator capabilities with and without the airflow. The comparison is made between the two types of embedded sensors which are eddy current (ECS) and strain gages (SG) technologies.
First the APA static behaviour is characterized. On the following figures, the stroke measurement is given with and without air flow. On each graph the measurement of each sensor type is given.

Figure 7 – Stroke without (top) and with (bottom) airflow seen for each sensor type.
The sensors are not placed at the same location on the actuator. ECS sensor measures a direct output displacement while SG sensor measures the strain on the ceramic stack. It can be noticed on the ECS curve that the airflow adds some oscillation that is modifying the controllability properties
VIPER control and inputs signals
The input signal is optimized to reduce the oscillations. These oscillations have two sources which are, on one hand, the mechanical ones due to the step response and on the other hand, the ones due to the airflow.
VIPER valve is working with square signals. However, open loop square signal is worst case in terms of ringing excitation. Therefore, 2 techniques have been tested. First, a notched filtered input signal is proposed, in order to cancel contribution of resonance frequency within the input signal. This leads to reduction in ringing but is not compatible with short repeating periods, due to duration of the filter input signal. The second solution is using pseudo period. In that case, square signal is replaced by a controlled ramp corresponding to the exact natural period of the system. This leads to strong ringing reduction, with no need of closed-loop technique.

Figure 8 – Actuator position as function of time.
The test and optimization campaign at CTEC has allowed defining the VIPER actuator capabilities: min/max duty cycle, stroke, operating inlet pressures etc. Following sections summarizes the actuator performances.

Figure 9 – Actuator capabilities in terms of duty cycle.
Actuator fluidic characterisation
The tests performed by ONERA aimed at characterizing the air flow response of the VIPER actuator. The characterisation is first performed in static mode (continuously blowing jet) in order to observe the velocity distribution along the slot length, then in pulsed blowing mode in order to qualify the frequency response in the range [10-500 Hz]. The input pressure is set between 1 and 4 bars (relative pressure) and a square-type input electric command has been used, with duty cycles 25%, 50% and 75%.
First, a Pitot probe is used to assess the spanwise velocity distribution along the slot span. An unsteady pressure sensor installed inside the slot is then calibrated versus the input mass flow, so that the dynamic characterisation of the actuator is performed in terms of instantaneous mass flow rate.
The actuator is attached to a console on the ONERA actuator characterisation bench [2], using a clamping system. From the CTEC amplifier, various cables allow the actuator to be electrically powered as well as the actuator sensors feedback signals.
During the characterisation tests, an additional unsteady pressure sensor has been installed on the actuator cover in order to measure the absolute stagnation pressure inside the air supply chamber of the valves.

Figure 10 – View of the VIPER actuator on the ONERA test bench.
The acquisition equipment employed on the actuator characterisation bench is comprised of a PC dedicated to the measurements and National Instruments cards allowing the simultaneous generation of a command signal for the actuator and measurement of the various sensors.
A software developed under the LABVIEW environment is used to manage the tests. Once the air supply pressure is reached in the feeding tank, the required settings for the actuator are entered through the user interface and an acquisition sequence of 2 seconds is launched. Note that at the beginning of the tests, only one air inlet channel was installed on the actuator; this has been modified and for the final tests, the actuator has been fitted with three air inlet channels to reduce the pressure loss inside the actuator chamber.
The actuation settings are:
- the feeding tank pressure,
- the actuation frequency,
- the duty cycle for the actuation frequency.

Figure 11 – General view of the bench with the feeding tank.
Static characterisation results
The total pressure probe is positioned in front of the slot, where the blockage effect is not significant. Measurements are carried out each 2 mm along the slot. The maximum velocity is first identified (through small probe displacement in the transverse direction), then the probe pressure is recorded and finally the velocity and Mach number are derived from St-Venant equations.
The velocity profile obtained allows the identification of possible sonic blockage for an input relative pressure of 3 bars upstream of the actuator. However, this pressure level includes an important pressure loss at the unique actuator air inlet; this pressure loss will be drastically reduced by the addition of two air inlet on the actuator and the inlet relative pressure level will be brought to 2 bars (instead of 3 initially).
The flow homogeneity is very good (purple curve in Fig. 12). A very small velocity deficit is observed close to the partitions (blue curve in Fig. 12) or flow rectifiers existing in the diffuser upstream of the slot. The absolute pressure recorded on the sensor located in the slot is around 1.4 bar (0.4 bar relative pressure); it can be concluded that for equivalent atmospheric and air supply pressures, such a pressure level will signify sonic conditions (Mach=1) at the slot exit.
Then, a specific device is used to obtain the transfer function between the mass flow rate and the pressure recorded in the slot. It can be noticed that for data of Fig. 12, where the absolute pressure in the slot is 1.4 bar, the mass flow rate is around 37 gr/s.

Figure 12 – Velocity and Mach number response along the slot. The probe is located 2 mm above the slot.
In the following dynamic measurements, it will therefore be considered that, if the instantaneous mass flow rate of 37gr/s is reached, conditions at the slot exit are sonic.
Dynamic characterisation results
The main goal of the dynamic characterisation is to set an upstream pressure value so that the pulsed flow at the orifice slot is sonic (in amplitude) for the maximum instantaneous flow rate, and at all frequencies between 10 Hz and 500 Hz with a 50% duty cycle, and if possible with 25% and 75% duty cycles as well.
Starting from a relative pressure of 1 bar at the feeding pressure tank outlet to a value of 4 bars by 0.5 bar steps, the frequency range is explored for the 3 duty cycles specified.
For each set of input pressure, the frequency response of the maximum instantaneous flow rate is plotted and the results are stored when the mass flow rate equal to 37 g/s is obtained, whatever the frequency is.
Two campaigns of measurements have been performed: the first one for the acquisition of the first four measurement channels (input command signal and pressure measurements); the second one with the addition of the remaining channel out of the 8 available (strain gage and displacements). These two campaigns serve to test the measurements repeatability.
As shown in Fig. 13, the level of 37 g/s (bold dashed red line) is reached for an input relative pressure of 2 bars, especially during the 2nd test campaign (continuous lines) compared to the 1st campaign (dash lines). However, one can observe a significant drop on the curves beyond 250Hz in the 25% duty cycle case. On the opposite, for the 50% and 75% duty cycles, the curves are rather flat, which is close to the ideal case that was targeted during the design phase of the actuator. For the two campaigns, one can notice that these two curves at 50% and 75% duty cycles are always above the theoretical threshold value of 34 g/s (bold continuous red line).

Figure 13 – Mass flow rate as function of frequency for different duty cycles. Inlet pressure is 3 bar (absolute pressure).
With a lower input relative pressure of 1.5 bar (Fig. 14), the curves are within 30 and 35 g/s, which is acceptable with respect to the theoretical mass flow of 34 g/s that was aimed during the actuator design process.

Figure 14 – Mass flow rate frequency response for different duty cycles. Inlet pressure is 2.5 bar (absolute pressure).
Finally, for a greater inlet relative pressure of 2.5 bar, the curves are more scattered and non-reproducible between the two campaigns (Fig. 15).
This phenomenon may be due to the incoming pressure that is perturbing the valve openings, which therefore are not functioning properly anymore.

Figure 15 – Mass flow rate frequency response for different duty cycles. Inlet pressure is 3.5 bar (absolute pressure).

About the time response, an example is given for a 2 bar relative input pressure. Fig. 16 shows the filtered temporal pressure signal coming from the slot exit sensor. The maximum mass flow is directly deduced from the pic amplitudes of such signal for the different frequencies tested.
Similarly, the time response of the strain gage is provided in Fig. 17, and the time response of the APA displacement from ECS 1 sensor in Fig. 18

Figure 16 – Time response of the slot pressure for all frequencies. Inlet pressure is 3 bar (absolute pressure).

Figure 17 – Time response of the strain gage signal for all frequencies. Inlet pressure is 3 bar (absolute pressure).

Figure 18 – Time response of the ECS1 position sensor signal for all frequencies. Inlet pressure is 3 bar (absolute pressure).
In the continuously blowing case first, the homogeneity of the flow at the slot exit has been assessed and showed to be very good. Moreover, it was shown that for a mass flow rate around 37g/s, it is possible to reach a sonic regime almost all along the slot. Then in the pulsed blowing case, an inlet relative pressure equal to 2 bars upstream of the valve openings enables to get the same mass flow rate (in amplitude), therefore insuring a sonic exit flow for this pressure setting value. It was also shown that in these conditions, the frequency response is quite flat over the all bandwidth tested. After testing, the experimental performances of VIPER can be summarized in the following table.

Specification Value Unit
Slot dimensions 1*80 mm²
Pitch angle of the slot exit < 30° °
Exit peak velocities 1 Mach
Exit peak mass flow 462 g/s/m
Actuation Max Frequency 500 Hz
Duty Cycle 50-75 %
Volume 45.4*79.6*208.9 mm3
Efficiency (Flow + Actuator) 37.5 %
Table 1 - VIPER experimental performances synthesis.

Potential Impact:
• How the foreground might be exploited, when and by whom
At first this new know-how would allow to pursue R&D works on pulsed piezo valves for flow control separation in aircraft, as a step further to VIPER.
Secondly, thanks to this know-how and to dissemination on VIPER, CTEC has gained a large national project called ASPIC, related to Flow Control Actuators of the Synthetic Jet Actuator (SJA) type. This project will be realised with ONERA will be go further as it includes a series of SJA on a wing and tests in wind tunnel. A large aircraft end-user is also involved. It has given the support to this new project after having assessed the good work performed by ONERA and CTEC in VIPER.
At last, CTEC will use its know-how to develop customised valves for space, cars, industries, medical, etc. In these domains there are demands for always faster, large flow and/or even more compact valves.
Side results related to improvement of components such as APA with composite shell (APA1000L-CRFP) and Switching Amplifiers (SA75D) will be exploited in various domains via products sales.

• IPR exploitable measures taken or intended
In VIPER framework, a new valve structure has been identified and experimented. This patentable feature is not be published yet. So it would be possible to apply for a patent.
The present strategy is to wait for a new project using this valve leading to higher TRL before to apply for a patent application.

• Further research necessary, if any
To achieve a TRL 6, more R&D works is required to mature the VIPER technology:
- Flow separation efficiency, to be tested in wind tunnel
- Life time test of the valves
- Electronics miniaturisation and compliance with aircrafts
- Integration of a set of valves in a wing segment
- Performance reproducibility
- Environment test (Thermal, dust, ice...)
To achieve a TRL 7, some flight tests are mandatory.

• Potential/expected impact (quantify where possible)
The expected impacts of VIPER valves are a reduction of apparition of flow separation on aircraft wings, an increase of lift and a reduction of drag. The achieved speed (mach 1) and flow rate (420gr/s/m) is enough to generate significant effects on aircraft aerodynamics and fuel consumption.

List of Websites: