Digital hydraulic actuator for flight control

Linköpings universitet

Linköping University | Department of Management and Engineering

Master’s thesis, 30 ECTS | Mechatronics

Spring 2019 | LIU-IEI-TEK-A--19/03340--SE

Digital hydraulic actuator for

flight control

Felix Larsson

Christian Johansson

Supervisor : Dr. Alessandro Dell’Amico at IEI, Linköping University Examiner : Prof. Petter Krus at IEI, Linköping University

Detta dokument hålls tillgängligt på Internet - eller dess framtida ersättare - under 25 år från publicer-ingsdatum under förutsättning att inga extraordinära omständigheter uppstår.

Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner, skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för ickekommersiell forskning och för undervisning. Över-föring av upphovsrätten vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av dokumentet kräver upphovsmannens medgivande. För att garantera äktheten, säkerheten och tillgängligheten finns lösningar av teknisk och administrativ art.

Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i den omfattning som god sed kräver vid användning av dokumentet på ovan beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan form eller i sådant sammanhang som är kränkande för upphovsman-nens litterära eller konstnärliga anseende eller egenart.

För ytterligare information om Linköping University Electronic Press se förlagets hemsida http://www.ep.liu.se/.

Copyright

The publishers will keep this document online on the Internet - or its possible replacement - for a period of 25 years starting from the date of publication barring exceptional circumstances.

The online availability of the document implies permanent permission for anyone to read, to down-load, or to print out single copies for his/hers own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility. According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement.

For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.

Abstract

In aviation industry, one of the most important aspects is weight savings. This since with a lowered weight, the performance of the aircraft can be increased together with increased fuel savings and thus lowered running costs. One way of saving weight is to reduce energy consumption, since with lowered energy consumption, lowered mass of fuel is required etc.

Most aircraft are today maneuvered with hydraulic systems due to its robustness and power density. It is the primary source of power for primary and secondary flight controls. The control of a conventional system which is using proportional valves is done by altering flow by restricting it to the extent where the desired output is achieved, which implies heat losses since the full performance of its supply is wasted through the valve.

In previous research, more energy efficient hydraulic systems called digital hydraulics has been investigated. In difference with conventional hydraulics, digital hydraulics uses low cost, high frequency on/off valves, which either are fully opened, or fully closed, instead of proportional valves to achieve the desired output. With this comes the benefit of no energy losses due to leakage and restriction control.

The downsides with digital hydraulics is the controlabillity. One way of controlling it is by using several pressure sources which outputs different pressure levels. By using the on/off valves in different combinations, different outputs can be achieved in a discrete manner.

In this thesis, the aim was to remove the impact of the discrete force steps which are present in digital hydraulics by creating concepts with hybrid solutions containing both digital hydraulics and restrictive control.

Three concepts were developed and investigated using simulation. The energy

con-sumption and performance was analysed and compared with a reference model, the concepts redundancy compared to conventional systems was discussed and finally the concepts were tested with an aircraft simulation model.

The concepts were found to reduce the energy consumption with different magnitude depending on the load cycle. The performance was found to be almost as good as the ref-erence model. The redundancy compared with conventional systems should be possible to maintain with slight modifications, but further investigation is needed.

It was found that one of the most important aspects regarding energy consumption is which combination of supply pressures is used to supply the system since it influences leakage and flow due to compression.

This master thesis has been performed at Saab Aeronautics, Saab AB. Examiner is Professor Petter Krus. Supervisor is Dr. Alessandro Dell’Amico from both Saab’s point of view as well as from Linköping University. A special thank you to Alessandro for all help and for his great enthusiasm in our work, and to our manager Niklas Fält who has given us this opportunity to perform this thesis. We would also like to thank all co-workers at Saab for the warm welcome and all help during this journey.

Linköping, May 2019

Contents

Abstract iii

Acknowledgments iv

Contents v

List of Figures vii

List of Tables x 1 Introduction 1 1.1 Background . . . 1 1.2 Goals . . . 1 1.3 Research questions . . . 2 1.4 Delimitations . . . 2 2 Literature study 3 2.1 Hydraulics . . . 3 2.1.1 Load pressure . . . 4 2.1.2 Flow continuity . . . 4 2.1.3 Valves . . . 5 2.1.4 Multi-chamber actuator . . . 7 2.1.5 Hydraulic efficiency . . . 7 2.1.6 Digital hydraulics . . . 8 2.2 Dynamic properties . . . 11

2.2.1 Gain and phase margin . . . 11

2.2.2 Bandwidth . . . 11

2.3 Control . . . 11

2.4 Aviation . . . 12

2.4.1 General aircraft hydraulics . . . 12

2.4.2 Control . . . 14

2.4.3 Air load from control surfaces . . . 15

2.4.4 Aircraft simulation model & system parameters . . . 16

2.5 Redundancy and reliability . . . 17

2.5.1 Static redundancy . . . 18 2.5.2 Dynamic redundancy . . . 19 2.5.3 Hybrid redundancy . . . 19 2.6 Software . . . 19 2.6.1 Hopsan . . . 19 2.6.2 Matlab . . . 19 2.6.3 Simulink . . . 19 3 Methodology 20

3.3 Simulations . . . 22 3.4 Results . . . 24 4 Implementation 25 4.1 Concept development . . . 25 4.1.1 List of requirements . . . 25 4.1.2 Concept generation . . . 25 4.2 Simulations . . . 32 4.2.1 System development . . . 32 4.2.2 Control strategies . . . 34

4.2.3 Full system simulation objectives . . . 40

4.2.4 System analysis . . . 41

5 Results 44 5.1 Baseline simulations . . . 44

5.1.1 Concept 1 . . . 44

5.1.2 Concept 3 . . . 45

5.2 Full system simulations . . . 47

5.2.1 Test 1 . . . 47

5.2.2 Reference model test . . . 50

5.2.3 Test 2 . . . 50

5.2.4 Test 3 . . . 52

5.2.5 Test 4 . . . 55

5.2.6 Redundancy test . . . 55

5.2.7 Flight mission simulation . . . 57

6 Discussion 60 6.1 Method . . . 60 6.2 Results . . . 60 6.2.1 Energy consumption . . . 60 6.2.2 Performance . . . 62 6.2.3 Redundancy . . . 62

6.2.4 Flight mission simulations . . . 63

6.3 Research questions . . . 63

6.4 Future research . . . 64

6.5 The work in a wider context . . . 64

7 Conclusion 65

Bibliography 66

A Complementary results from test 2 68

List of Figures

2.1 Simple hydraulic system. . . 3

2.2 A simple two chamber actuator. . . 4

2.3 The figure illustrates a flow into a volume. . . 4

2.4 A proportional valve able to control direction of flow with spool position xv, pres-sure source ps, load pressure pLand flow qL. . . 5

2.5 Flapper nozzle servo valve. . . 6

2.6 Jet pipe servo valve. . . 6

2.7 Tandem actuator. . . 7

2.8 The figure shows a grid of on/off valves (DFCU) connected to a multi-chamber cylinder with varying piston areas supplied with three different pressure lev-els, which together assemblies the digital hydraulic actuator concept (DHA). The model is created with Hopsan. . . 10

2.9 Hydraulic system with two individual circuits . . . 13

2.10 Multi-chamber with float valves. . . 14

2.11 Roll, pitch and yaw axis. Ailerons in green, elevators in blue and rudder in orange. 15 2.12 The figure illustrates a control surface with an actuator. . . 15

2.13 The figure shows a delta canard aircraft. Ailerons in green, elevators in blue, rud-der in orange and canards in yellow. . . 16

3.1 Flowchart of the project methodology . . . 20

3.2 Steps of the preparation phase. . . 21

3.3 Steps of the concept development phase. Horizontal lines shows workflow and vertical lines shows partial activity within the referred step. . . 22

3.4 Main activities within the simulation phase. . . 22

3.5 Steps of the system development activity. Horizontal lines shows workflow and vertical lines shows partial activity within the referred step. . . 23

3.6 Steps of the model testing and refinement activity. . . 23

3.7 Steps of "Results" phase. . . 24

4.1 System layout of concept 1. . . 26

4.2 Base line simulation with fixed volumes. . . 27

4.3 System layout of Concept 2. . . 28

4.4 The figure shows a Hopsan model of concept 3. . . 29

4.5 The figure shows a Hopsan model of concept 3 with 3 pressure sources. . . 30

4.6 Baseline simulation for concept 3 . . . 31

4.7 The figures shows the force step distribution for 2, 3 and 4 pressure sources. . . 32

4.8 The figure shows the Hopsan model of the grid of on/off valves with 3 pressure sources (DFCU). . . 33

4.9 The figure shows how the available forces are chosen to follow a force reference. . 35

4.10 The figure shows how the differences in force steps when restricting the DHQ. . . 35

4.11 Flow chart of concept 1. VM stands for valve matrix, which contains the boolean matrix with information to the DFCU which valves to open/close. . . 36

4.13 Flow chart of concept 2. . . 37

4.14 Flow chart of concept 3. VM stands for valve matrix, which contains the boolean matrix with information to the DFCU which valves to open/close. . . 38

5.1 Step response test to control load pressure. . . 44

5.2 The figure shows the piston position of the baseline simulation of concept 3. . . 45

5.3 The figure shows the pressure following in chamber 1 from the baseline simulation of concept 3. The red line shows the reference while the blue line shows the output pressure. . . 45

5.4 The figure shows the pressure following in chamber 2 from the baseline simulation of concept 3. The red line shows the reference while the blue line shows the output pressure. . . 46

5.5 The figures shows how the total consumed energy varies with the step restriction with and without leakage for concept 1. PS stands for pressure source. . . 47

5.6 The figure shows how the consumed energy varies for different step restrictions when using 2, 3 and 4 pressure sources. . . 48

5.7 The figures shows how the total consumed energy varies with the step restriction with and without leakage for concept 3. . . 49

5.8 Shows the tests of reference model where the test requirements where found. . . . 50

5.9 The figures shows the results from test 2 for concept 1 with 3 pressure sources and force step restriction of 50 kN. . . 50

5.10 The figures shows the results from test 2 for concept 2 with 3 pressure sources and force step restriction of 50 kN. . . 51

5.11 The figures shows the results from test 2 for concept 3 with 2 pressure sources and force step restriction of 20 kN. . . 51

5.12 The figures shows the results from test 3 for concept 1 with 3 pressure sources and force step restriction of 50 kN. . . 52

5.13 The figures shows the results from test 3 for concept 2 with 3 pressure sources and force step restriction of 50 kN. . . 52

5.14 Modified concept 2 with increased valve resonance frequency to 200 Hz . . . 53

5.15 The figures shows the results from test 3 for concept 3 with 2 pressure sources and force step restriction of 20 kN. . . 54

5.16 The figure shows the results from the redundancy test with concept 1. . . 55

5.17 The result shows the result from test 4 with concept 2. . . 56

5.18 The figure shows the results from the redundancy test with concept 3. . . 56

5.19 Energy consumption for concept 1 and the reference model with and without leak-age during the flight mission simulation. . . 57

5.20 The figures show the responses for concept 1 and the reference model during the flight mission simulation. . . 58 5.21 The figures show the deflection of the control surfaces inner and outer left elevon. 59

A.1 The figures shows the results from test 2 for concept 1 with 2 pressure sources and force step restriction of 20 kN. . . 68 A.2 The figures shows the results from test 2 for concept 1 with 4 pressure sources and

force step restriction of 40 kN. . . 69 A.3 The figures shows the results from test 2 for concept 2 with 2 pressure sources and

force step restriction of 60 kN. . . 69 A.4 The figures shows the results from test 2 for concept 2 with 4 pressure sources and

force step restriction of 50 kN. . . 69 A.5 The figures shows the results from test 2 for concept 3 with 3 pressure sources and

force step restriction of 20 kN. . . 70 A.6 The figures shows the results from test 2 for concept 3 with 4 pressure sources and

force step restriction of 40 kN. . . 70 B.1 The figures shows the results from test 3 for concept 1 with 2 pressure sources and

force step restriction of 20 kN. . . 71 B.2 The figures shows the results from test 3 for concept 1 with 4 pressure sources and

force step restriction of 40 kN. . . 72 B.3 The figures shows the results from test 3 for concept 2 with 2 pressure sources and

force step restriction of 50 kN. . . 72 B.4 The figures shows the results from test 3 for concept 2 with 4 pressure sources and

force step restriction of 50 kN. . . 72 B.5 Modified concept 2 with increased valve resonance frequency to 200 Hz . . . 73 B.6 Modified concept 2 with increased valve resonance frequency to 200 Hz . . . 73 B.7 The figures shows the results from test 3 for concept 3 with 3 pressure sources and

force step restriction of 20 kN. . . 73 B.8 The figures shows the results from test 3 for concept 3 with 4 pressure sources and

2.1 List of system parameters. . . 17

4.1 Requirements of the DHA concepts . . . 25

4.2 System parameters for the baseline simulation . . . 30

4.3 System parameters for the simulation models . . . 32

4.4 Test objectives . . . 40

4.5 Test objective for the redundancy test . . . 42

5.1 Shows the most efficient operating points and the energy consumption for each concept and configuration. First part of the table show values without leakage and second part with leakage. E stands for energy consumed, ps for pressure sources, Frs for force step restriction. . . 49

5.2 Shows the energy saving at the concepts best operating points compared to the reference system. First part of the table show values without leakage and second part with leakage. Negative values shows where a system consumed more energy than the reference model. . . 49

5.3 The energy savings for all concepts during test 4. Negative values means that more energy where consumed compared to the reference system. . . 55

C

1

I

NTRODUCTION

1.1

Background

In aviation industry, one of most important aspects is weight savings. This since with a lower weight, the performance of the aircraft can be increased together with increased fuel savings and thus lowered total costs.

Most aircraft are today maneuvered with hydraulic systems due to its robustness and power density. The maneuverings by use of hydraulic systems are done by moving control surfaces of the wings with use of actuators. The actuators are controlled by hydraulic pressure/flow which in turn is controlled by valves. A pressure source is connected to the valve and if more pressure or flow is needed to the actuator, the valve is opened more and vice versa. When doing this, a lot of potential in the hydraulic system is turned to heat, and thus wasted, due to pressure drops over the valve when the valve is not fully opened or closed.

An upcoming and promising technology is called digital hydraulics, which uses several on/off valves instead of regular proportional valves. This means that the valves can only be fully opened or closed, and by that the wasted energy due to flow throttling will be reduced. The control can be made by having several different pressure sources together with an multi-chamber cylinder actuator with several piston areas, which together gives a possibility to choose between a discrete number of forces acting on the actuator rod. By doing this, the energy consumption can be reduced, but the control possibilities are changed from continuous flow control to discrete force control. This means that it is a lot more difficult to achieve good position and control of a system based on pure digital hydraulics where the load cycle is unknown. Hardly ever will one of the discrete forces be equivalent with the desired force which gives a system which is to fast or slow, or is unable to hold a certain position. This control error could be reduced by switching between force steps in order to stay closer to the desired force. Previous studies has shown that excessive switching of force steps will waste energy and cause unpleasant dynamic behaviours. Further improvements need to be done to be able to implement this kind of technology in applications with high demand of precision.

1.2

Goals

The primary goal with this project is to create a conceptual design of a hydraulic system to control the maneuverings of an aircraft. The concept shall be based on the idea of digital hydraulics, but it shall also possess the ability to restrict pressure from pressure sources.

Thus it will not be a system which consists of pure digital hydraulics, instead it shall be developed towards a hybrid solution which have both digital and restrictive control. This in order to save as much energy as possible and reduce the influence of having a discrete number of forces to choose between to enhance the control of the aircraft.

The secondary goal is to learn whether these kind of systems are redundant enough to be used in aircraft.

The general goal with this thesis is to gather a broad knowledge about the possibilities to combine restriction control with digital hydraulics to provide future research by studying different combinations.

1.3

Research questions

The questions to be answered in this project are:

• Is it possible to modify digital hydraulic actuators to have a system based on digital hydraulics, achieve continuous control and still maintain increased efficiency and re-dundancy compared to conventional hydraulic systems?

• What aspect of these kind of systems is most important to consider in order to reduce energy consumption as much as possible?

• How well does a combination of digital hydraulics and restrictive controlled hydraulics perform compared to a traditional system?

• If digital hydraulics are combined with restrictive controlled hydraulics, is it most im-portant to reduce switching losses or losses due to restriction?

• In what way could the control strategy minimize energy consumption?

1.4

Delimitations

The delimitations for this project will be:

• Only the tandem cylinder actuator will be investigated

• Only a number of two to four pressure sources will be investigated • Only force control will be applied for controlling the system output • Ideal supply systems will be used within simulation models

• The choice of components will be based on realistic properties, but not necessarily off the shelf products

• The flight simulations will only be conducted with the Admire aircraft model • System size and weight will not be studied in detail

C

2

L

ITERATURE STUDY

In this chapter the background study as well as all needed theory will be presented and explained.

2.1

Hydraulics

According to Pascals law a pressure change in one point of a fluid inside a fixed volume will transfer to the rest of the fluid. This law apply to a lot of hydraulic applications.

Figure 2.1: Simple hydraulic system.

For the system in figure 2.1 the pressure can be described as equation (2.1) if the pistons are friction free. In hydraulic applications, usually p ąą pa+ρgh. From this, the pressure can be expressed as in equation (2.2). p= F1 A1 +pa+ρgh1= F2 A2 +pa+ρgh2 (2.1) p= F1 A1 = F2 A2 (2.2)

2.1.1

Load pressure

The pressure difference between the chambers in a two chamber actuator counteracting a load force is called load pressure.

Figure 2.2: A simple two chamber actuator.

This because according to the fore equilibrium in equation 2.3

p1˚A1´p2˚A2=FL (2.3)

equation 2.4 can be derived

p1´p2= pL (2.4)

which is the required pressure difference to counteract the load force.

2.1.2

Flow continuity

The flow continuity equations describes the mass balance for a given volume V according to (2.5). ÿ qin= dV dt + V βe dp dt (2.5)

Whereř q_in is the total flow in or out from the volume. The term dV/dt is the volume derivative with respect to time and the last term(V/βe)(dp/dt)describes the compression

flow which is required to achieve a pressure change in the volume. βeis a modulus of

com-pression of the fluid including its surroundings.

2.1. Hydraulics

2.1.3

Valves

Proportional valves

A proportional valve consists of a cylindrical spool and a matching manifold. The spool position can be changed in the axial direction in order to create an opening area, see figure 2.4. With this adjustable opening area it is possible to control the flow through the valve if there is a pressure difference between the pressure before and after the valve. It is possible to control the flow q since it depends on the pressure difference and opening area [18], see equation (2.6). Proportional valves can be designed in different ways to achieve different functions considering the flow. One option is to only restrict the flow through the valve. Another option is to control the direction of the flow meanwhile restricting the flow. [11]

q=Cqwxv

d 2

ρ(p1´p2) (2.6)

Where q is the flow, Cq is the flow coefficient, w is the opening area gradient, xvis the spool

position, ρ is the oil density, p1is the pressure before the valve and p2is the pressure after the

valve.

Figure 2.4: A proportional valve able to control direction of flow with spool position xv,

pressure source ps, load pressure pLand flow qL. [18]

Servo valve

A servo valve have high response and precision. Different kind of feedback makes it possible to set and hold the spool position at a desired value. In a flapper nozzle, which is a common type of servo valve, there is a feedback needle which is attached to both the spool and an electrical torque motor, see figure 2.5. On each side of this flapper part of the needle there is a nozzle. Electrical input to the torque motor displaces the needle which creates a difference in nozzle opening which in turn gives pressure difference between the nozzles and the main valve spool. Pressure difference will move the spool and the feedback needle will at some point counteract the torque from the torque motor and the spool stops. The spool stops since pressure equilibrium have been achieved when the flapper is returned to center of the two nozzles. [18].

Servo valves consumes a lot of energy due to leakage and high throttling losses. Leak-age occurs even when the valve is closed trough the valve and pilot stLeak-age. The flapper nozzle valve is rarely used in aviation since it is sensitive to pollution, where particles easily can disturb the pilot stage which creates unsuspected movement of the hydraulic load. A more

Figure 2.5: Flapper nozzle servo valve. [18]

common servo valve in aviation is the jet pipe solution, since it is less sensitive to pollution and has a lower and more constant leakage flow than the flapper nozzle valve [13]. The jet pipe pilot solution consists of a jet pipe which is directed to one of the two receiver ports by an electrical torque motor. This gives a pressure difference in the control ports which results in a pilot flow and a spool displacement [18]. The position control feedback can also be done with a needle which counteract the torque from the electric motor [13]. It is also possible to electronically measure the spool position and use this signal as feedback to the controller and get a proportional relation between command signal and spool position [18].

Figure 2.6: Jet pipe servo valve. [18]

On/Off valve

On/off valve is a simple and robust valve which can only be fully opened or fully closed. This means that the valve can not throttle the flow or power transferred to the load, it is only possible to transfer zero power or full power. Ideally, when the valve is open there is no restriction area, thus there will be no pressure difference over the valve. When the pressure difference is close to zero there will be almost no loss of energy due to throttling. Compared to a proportional valve or servo valve the leakage is almost non-existent. On/off valves enables

2.1. Hydraulics

at the same precision as conventional valves. To increase the precision, several of these valves can be put together and cooperate to create different power outputs as described in section 2.1.6. [3]

2.1.4

Multi-chamber actuator

One type of a multi chamber actuator is a tandem actuator. Here, the actuator consists of two pistons mounted on the same rod. By doing this, the cylinder will consist of four chambers instead of two. This is a redundant solution which is commonly used in aviation industry. [13]

In figure 2.7, a sketch of a tandem actuator can be seen. The numbers 1-4 represents the chamber numbers that from here on will be referred to. Since this actuator basically consists of two actuators, the added redundancy comes from if any of the supply lines to any chamber fails. If this failure occurs in a regular two chamber cylinder, the actuator can no longer be used. Here, two separated supply systems can be used to supply both chamber pairs (cham-ber 1 and 2 builds one cham(cham-ber pair as well as cham(cham-ber 3 with cham(cham-ber 4), which means that if one of the supply systems fails, the other chamber pair will continue to function.

Figure 2.7: Tandem actuator.

2.1.5

Hydraulic efficiency

Hydromechanical efficiency

In hydraulic systems, friction contribute to the hydromechanical energy losses. It consist of both mechanical and viscous friction. Mechanical friction comes from for example sealing between piston and the cylinder wall in a actuator. Viscous friction is the friction between the fluid and hoses but also the resistance that a throttling valve creates.

Volumetric efficiency

Volumetric losses is related to the loss in volume flow in the system. The flow can be reduced by leakage both internally and externally from a hydraulic component. For example, if the hydromechanical efficiency of a pump is 92 %, the pump will only deliver 92 % of the ideal flow due to leakage of that component. Compressibility of the fluid also contribute to the volumetric losses since compressed fluid reduces the volume which means some increase in flow is required to compensate the change in volume.

Switching losses

On/off valves minimizes the throttling losses since they only operate fully open or fully closed. Ideally, the only throttling losses created is during the time the valve closes or opens [16]. When using these valves to create different force steps, as explained in section 2.1.6,

switching of the valves needs to be performed in order to change the force step. This means that some valves will open and some will close to establish new pressures in the chambers. Since hydraulic oil is compressible the energy efficiency will be affected negatively by switch-ing. This energy loss occur when a valve with supply pressure p1opens to a chamber with a

volume V, initial pressure p0, bulk modulus of the oil β. The energy loss is given by equation

(2.7). [10]

Esw´loss=

V

2β(p1´p0)

2 _(2.7)

Differences in the supplied energy and the stored energy in the volume, Es and EV, gives

equation 2.8. Esw´loss=Es´EV = ż8 0 p1q(t)dt ´ ż8 0 pV(t)q(t)dt (2.8)

Pressure pV in the volume is described by (2.9) and derived from the flow continuity

equa-tion. pV(t) = β V ż q(t)dt+p0 (2.9)

This shows that the switching loss does not depend on the opening area or opening time of the valve. [10]

Throttling losses

It is common to use proportional valves to control flow or pressure in hydraulic systems. This is done by throttling of the flow which create a energy loss in form of heat. The energy comes from the change in hydraulic power since the power decreases through a valve which throttles the flow. Flow, q, is driven by pressure difference (2.6) and the hydraulic power, P, depends on the flow and pressure, p, (2.10). This explains the loss of power, ∆p, (2.11) through the valve and the loss of energy.

P=qp (2.10)

∆Ploss =q∆p (2.11)

2.1.6

Digital hydraulics

Using traditional hydraulics, the pressure or flow is controlled by a proportional valve. If there is a demand for lower pressure/flow, the valve is closed until the demanded output is received. When doing this, potential power is wasted by the restriction implemented. The idea with digital hydraulics is to let the system work as close as possible to the maximal power the supply system can provide in order to minimize the energy loss due to restrictions. Since the on/off valve only can be fully opened or fully closed, nearly the full potential of the power sources will be used. There are different techniques to control the pressure or flow in the system which will be explained below.

Switching

Switching control is one approach for digital hydraulics. The fundamentals of switching con-trol is adopted from PWM (pulse-width modulation) concon-trol in electronics, where a constant supply can be alternated by pulsing it. For example, if the power source outputs a voltage of 6 volts, and one wants to achieve a 3 volt output, the output can be pulsed between 6 and 0 volts where the duty cycle is 50 % (The output is held at 6 volt 50 % of the time and 0 volts 50 % of the time). This gives a mean voltage output of 3 volts. Depending on the application, the frequency of the pulses is chosen so it does not give the experience of having a "jerky"

2.1. Hydraulics

applications are 1 - 200 kHz, which means one cycle happens in 1 ms to 5 µs. [2]

To obtain a behaviour similar to PWM in a hydraulic system, a constant pressure source with a fast on/off valve can be used to perform the switching. One main advantage (besides the increased efficiency) with this type of system is that the control becomes fairly simple ideally compared to other digital hydraulic systems since it is possible to vary the output between 0 and maximum power. Since this type of system requires high speed switching, there is a desire of fast valves. It is shown that switching frequency is influencing the final performance significantly. A high efficiency can be obtained if the switching frequency is below and up to the natural frequency of the system, while a low frequency also gives high fluctuations in pressure and flow which also leads to poor dynamics of the system. [19] It has been shown that a high efficiency can be obtained with high speed switching if an inertance tube is introduced to the system, which basically is used to introduce fluid inertia. The inertance tube is consists of a hydraulic line with a small diameter and high length. It is shown that the properties of the line (diameter and length) is critical for the final efficiency of the system. [15]

DFCU

Digital flow control unit (DFCU) is a concept where a grid of on/off valves is used to either alter the effective flow or pressure to the actuator. To alter the flow, several on/off valves are connected in parallel and the effective output flow is determined by how many valves are opened at the same time. The valves can either have the same opening area (called pulse number modulation (PNM) coded), have different areas (called binary coded) or a mix of the previously mentioned set ups. With PNM-coded DFCU’s there is a need of many valves to have a high resolution of discrete flow-rates since a higher flow only can be achieved by opening more valves. With binary coded DFCU’s, the flow rates of the valves shall be dimensioned such that their flow capacities is 2N˚Q1where Q1is the maximum flow rate of

the smallest valve and N is the total number of valves minus one. This gives 2N+1number of possible discrete flow rates. If four valves is used in the DFCU, the number of flow rates will be 2(4´1)+1=24=16. Therefor, it is not required to have as many valves as in PNM-coded, but since all valves will be different, they will behave different which will lead to uncertain-ties in the achieved effective flow. [12]

To alter the pressure with a DFCU, which in turn is used to control the output force from the actuator, several pressure sources are connected to a grid of on/off valves which are connected in series and/or in parallel which is connected to a multi chamber cylinder with varying piston areas. This type of assembly build’s a concept called digital hydraulic actuator (DHA) which is the type of concept this thesis is based on. One possible set-up can be seen in figure 2.8.

In figure 2.8, one can see that it is possible to apply phigh, pmid and plow to every

cham-ber in the cylinder. One can also see that the piston area in every chamcham-ber differs. This means that it is possible to generate 3 different forces in every chamber depending on which on/off valve is open at the moment. Since there are 4 chambers, the number of discrete forces possible in this case is 34=81. Therefor, the relation between the number of pressures available (np), the number of chambers with unique areas (nc) and the number of discrete

forces (nF) can be written as follows. [21]

Figure 2.8: The figure shows the grid of on/off valves (DFCU) connected to a multi-chamber cylinder with varying piston areas supplied with three different pressure levels, which to-gether assemblies the digital hydraulic actuator concept (DHA). The model is created with Hopsan. [21]

With this type of solution combined with accumulators and constant pressure sources, en-ergy within the system can be reused by recovering the flow and pressure generated from retraction and extension of the actuator. In figure 2.8 it can be seen that if phigh is applied

to all chambers, the resulting movement of the piston leads to oil flowing out from two of the chambers, and into the other two chambers. The flow needed from the pump is reduced since oil is circulated between the different chambers, which results in a lower energy con-sumption. [3]

Previous research has shown that with DFCU it is possible to reduce the losses, com-pared to conventional systems, by 80 %. The remaining losses mainly comes from friction and hydraulic capacitance. The losses which comes from the capacitance appears when the pressure goes from high to low which makes the oil expand since it is compressible. When the oil expands, energy stored in it disappears into flow- and heat energy. But, these kind of losses tend to be lower in aerospace applications since the valve block often is located close to the actuator. [3]

Further losses associated with this kind of digital hydraulics comes from short circuits. It occurs during a force step change when two on/off valves connected to the same cylinder chamber are open simultaneously a short period of time, one valve will close while the second opens in order to change the supply pressure to the chamber. This is often sorted by delaying the opening of the second valve. [4]

A problem is the discrete amount of forces that comes with the force control with use of the DFCU. If a demanded force is outside the force spectra (in between two available discrete forces), there will be two possible outcomes. Either, the output force will be to high or to low, or high speed switching between a too high and to low force has to take place. With the last option, the same problems as with the switching technique are introduced. [21]

Hybrid digital hydraulics

Most recent research in digital hydraulics is aiming for better control possibilities for contin-uous control in order to reduce the problems with having discrete force steps.

2.2. Dynamic properties

In some research, the DFCU has been modified to include proportional valve(s) to give the possibility for "fine tuning" of the pressure/flow by throttling. One example is an exca-vator where the traditional hydraulics has been replaced with a hybrid concept where the valves in the DFCU is of the proportional type in order to be able to control the flow into a multi-chamber cylinder. It has been proven that the energy consumption for the hydraulic system can be reduced with about 40 % compared to a traditional load sensing system. [9] In another research, a semi-binary hydraulic four chamber cylinder, which consists of both fast switching on/off valves and one proportional valve connected to one of the chambers has been investigated. It was found that the control when no switching is performed, is very good. When switching occurs, there is still some force spikes that occurs. The loss in efficiency compared to a traditional DHA system is acceptable if the piston area ratio is chosen in a good way. [8]

2.2

Dynamic properties

2.2.1

Gain and phase margin

To determine stability of a system, gain and phase margin are used. They are individual measures calculated from the open loop system response to get two relative safety margins. These margins indicates how close the system is to start oscillate or become unstable.[7] Gain margin describes the amount of change one can make in the open loop gain before instability occurs in the closed loop system.[7]

Phase margin is measured between the system frequency and -180°where the amplitude is 0 dB. If the phase margin is zero degrees, the difference in phase between reference signal and output signal is 180°. This will make the closed loop feedback positive and therefor unstable.[7]

It is not reasonable to only take one of these margins into account. A system with good phase margin can be unstable due to zero phase margin. Non of these margins are better than the other, both have to be considered.[7]

2.2.2

Bandwidth

For a closed looped system the bandwidth describes up to what frequency the system output signal can follow the reference signal without any disturbance or load. The reference signal is sinusoidal and at the frequency where the gain has decreased 3 dB or when the phase shift is larger than -90°. [17]

2.3

Control

In most systems where a desired action shall be obtained, controllers of some kind are re-quired. To achieve an action from any system, it needs input which instructs the system to perform its processes.

Controllers can be open loop controllers or closed loop controllers. An open loop con-troller completely acts from its given inputs. An closed loop concon-troller acts both from its inputs and outputs by feeding the output back to the input telling the system if it is close or far from achieving its desired action. [7]

One example of the difference between an open and closed loop controller can be a wa-ter tap filling a glass of wawa-ter. To provide wawa-ter to the glass, the wawa-ter tap needs a input (open the tap). If it is left with the tap opened and no feedback is given, the glass would in some time be overfilled. To make the open loop controller work in this case, the time to fill the glass must be known so the input is changed from open to close when the fill time is achieved. A closed loop system could measure the water level in the glass (the output) and automatically close the tap when the glass is full. This gives the controller an error to act on, the difference between a desired output (glass full of water) and the actual output (measured water level in the glass). From this, the actual opening of the tap can be controlled. When the glass is empty, the error will be large, which results in an fully opened tap. When the water level closes in to the desired value, the error will be smaller and thus the tap opening is reduced. When the glass is full, the error is zero and the tap is closed. To achieve this, a pro-portional regulator is required. This type of regulator multiplies the control error with a gain, K, to give a reasonable input signal to the system. This can be explained by the difference in size between the required input signal to the system (in the tap example, mm of open-ing diameter) and the control error (litres of water in the glass). The control error must be scaled with the gain before the input signal is sent to the system to give an realizable input. [7] For more complex systems, fulfillment of the desired value might need more calculations from the regulator. A PI-regulator both has a proportional part to scale the control error and a integrating part which is proportional to the integral of the control error. It can be written as equation 2.13 shows. [5] u(t) =Ke(t) + 1 Ti żt 0 e(τ)dτ (2.13)

Where u(t)is the output signal from the regulator, K is the gain, e(t)is the control error and Tiis the integrator time. The integral part of the regulator looks at the integral of the control

error, which means that at all time there is an control error, the integral part will increase, which will increase the controller output. When a P regulator is used, the controller output becomes small when the control error becomes small. In some cases, small controller outputs can lead to a static control error since small outputs might not affect the system enough to completely eliminate the control error. With the integral part, the small control error will be integrated, and over time, the regulator will increase its output since the integral will grow as long as the control error remains. Therefore, the control error should be completely eliminated with an PI-regulator. [7]

2.4

Aviation

2.4.1

General aircraft hydraulics

Hydraulics are today the primary source of power for primary and secondary flight controls where the system transmit and control the power from the engine to the aircraft control surfaces. There is many advantages to use hydraulics as source of power since it gives high power amplification from small control input, precise control of position, fast response, smooth dynamics and it is able to handle several loads simultaneously.[20]

The Hydraulic system consist in general of pressure sources in form of pumps and pres-surized tanks, servo valves, pistons and pipelines. The pump is the hydraulic power supply and it is propelled by the aircraft engine. Control elements consist of the servo valves which controls the hydraulic power and the direction of flow to the actuators. A linear actuator, piston actuator, delivers the mechanical power output transformed from the hydraulic power to the control surface.[20]

2.4. Aviation

Safety is of high importance in aircraft hydraulic systems and to meet the safety require-ments one approach is to implement several independent components with the same func-tion within the system, in order to increase the redundancy. Figure 2.9 shows a dual channel hydraulic system with two pumps, several parallel actuators, and a four chamber actuator. A system like this is able to continue functioning although one or, in some cases, several com-ponents fails. For a control surface, more than one actuator is used to increase redundancy, and the actuators for that control surface are supplied from different channels to further increase the redundancy. [20]

Figure 2.9: Hydraulic system with two individual circuits

Figure 2.10 shows a multi-chamber actuator with two float valves. These valves is a part of a redundancy solution and are positioned between the servo valves and the multi-chamber. The function for these valves is to set one or both halves in float mode in case of a failure. It could be a fault of a hydraulic line, a servo valve jam in open or closed position, or a pump failure. The float function is activated when one of the channels need to be disconnected from the multi-chamber in order to make sure that a failure does not disturb the rest of the system. When activated, the pair of chambers is connected to each other and will not contribute to the systems output.

Figure 2.10: Multi-chamber with float valves.

2.4.2

Control

An Aircraft has six degrees of freedom which relates to transitional movement and rotation to three axis, see figure 2.11. Transitional motion relates to the travel from one point to another where a conventional aircraft always point in the direction of travel. The rotational motion relates to the maneuvering of the aircraft which consists of pitch, roll and yaw. Pitch will rotate the aircraft around the θ-axis while traveling along the ϕ-axis to increase the altitude. For a properly performed turn, rotation around the z-axis is implemented, which is normal to the ground. A certain roll angle is maintained and at the same the aircraft need some pitch to compensate for the reduced lift, caused by roll angle, to prevent loss in altitude. Yaw is also used to rotate the aircraft in ψ-axis to new direction of heading. [14]

The aircraft wings has variable control surfaces to create the rotating movements for pitch, roll and yaw. In general there is a pair of control surfaces for pitch (elevators), a pair for roll (ailerons) and a single surface for yaw (rudder). While airborne these control surfaces deflects the air to maneuver the aircraft by rotating it around the desired axis. [14]

2.4. Aviation

Figure 2.11: Roll, pitch and yaw axis. Ailerons in green, elevators in blue and rudder in orange.

2.4.3

Air load from control surfaces

Besides from the static mass loads which comes from wings etc., the control surfaces of a moving aircraft will generate a load force which comes from moving air. This load can be approximated as a spring load which will increase with an increased deflection of the control surface. [21]

Figure 2.12: The figure illustrates a control surface with an actuator.

In figure 2.12 a control surface with an actuator is shown. With an increased δ, the piston position xpwill increase. When the control surface is in its middle position (δ = 0, xp = 0), the

air load will be at its minimum. With increased (positive or negative) deflection, the air load will increase since the control surface area affected by the air will increase. The air load can be described with the following relationship.

FL,air= A ˚ sin

(δ)˚ ρ ˚ v2˚Cd

2 (2.14)

Where ρ is the air density, v is the air velocity and A is the surface area of the control surface. [21]

2.4.4

Aircraft simulation model & system parameters

For aircraft simulations in previous research [4], the ADMIRE simulation model has been used. The ADMIRE model was created from a generic aero-data model [1] to be provided for the research community. This generic model consists of a small single seat fighter aircraft. In difference with the aircraft shown in figure 2.11, this aircraft has a delta canard configuration. Here the elevators and ailerons are placed next to each other at the back on a delta wing. There are also canards in this model, located in front of the delta wing (see figure 2.13). [6]

Figure 2.13: The figure shows a delta canard aircraft. Ailerons in green, elevators in blue, rudder in orange and canards in yellow.

In previous research where a digital hydraulic actuator with 81 discrete forces applied to a model of an unstable delta-canard aircraft has been investigated, the parameters presented in table 2.1 were set up to comply with the system requirements.

Dr. Alessandro Dell’Amico also used a tandem actuator controlled with servo valves for simulation of the conventional hydraulic system within his research. This system, similar to the tandem actuator seen in figure 2.9, will be used as a reference model in this thesis. The reference model consists of two constant pressure pumps (280 bar supply), two pressurized tanks (7.5 bar), two servo valves with the same setup as seen in table 2.1 and a implemented leakage over the valves (1 liter/min). One difference from table 2.1 is that a symmetrical actuator was used with cylinder areas 2.75*10´_{4 m}2_{to fulfill the requirements.}

2.5. Redundancy and reliability

Table 2.1: List of system parameters. [4]

Parameter Value

phigh 280 [bar]

pmiddle 72.22 [bar]

plow 7.5 [bar]

A1,2,3,4 {30.09 26.36 24.87 25.11}*10´4[m2]

Servo valve resonance frequency 250 [Hz] On/off valve resonance frequency 300 [Hz] Servo valve maximum opening area 2.2˚10´6[m2] On/off valve maximum opening area 6.3˚10´5[m2]

Control surface rate 50 [0_/s]

Mass load 300 [kg]

Maximum piston speed 0.0873 [m/s]

Cq 0.6

ρ 850 [kg/m3]

Viscous friction (Bp) 10000 [Ns/m]

Damping ratio (on/off and servo valves) 1

Fmax 150 [kN]

Fmin -150 [kN]

Aircraft model parameters [6] Value

Wing area 45 [m2_]

Wing span 10 [m]

Wing mean chord 5.2 [m]

Mass 9100 [kg]

Ix 21000 [kgm2]

Iy 81000 [kgm2]

Iz 101000 [kgm2]

Ixz 2500 [kgm2]

2.5

Redundancy and reliability

In aviation industry it is important that all systems are reliable and redundant since a loss of function in any system can otherwise lead to catastrophic events.

A system can be designed to either be resistant to failure, or a failure must be tolerated in some extent. With resistance to failure it is meant that the system’s components will never fail during its designed lifespan with correct maintenance or/and due to over-dimensioning of its parts. With tolerance to failure, the systems where for some reason resistance failure can not be achieved, the failure must be tolerated and/or managed somehow. [13]

The systems which need to have tolerance to failure can either be fail active- or fail safe systems. In cases where the systems function must be operative in case of system failure (fail active systems), the damage done must either lead to reduced- or totally maintained performance. If reduced performance is acceptable, the system is said to be fail-functional. If full performance must be held, the system is said to be fail-operative. It can in some cases be tolerated that the system is not functioning for a short while in order to let it change its structure for maintaining full or partial function.

In fail-safe systems, the loss of function can be tolerated as long it is not affecting the performance of any other function. In the event of a failure, the system’s measure to it can be done in three different ways. The first option is that the system stays in its position when it fails and remains it until the damage is fixed, which is when it is called "fail-freeze". The

second option is that the system follows the movement of other systems and "floats", which is called "fail-passive". The third and last option is that when a fail occurs, the system goes back to a predetermined position where it is not affecting any other component’s performance, which is called "fail-natural". [13]

To make sure that fails in actuation are resulting in the previously described manner, different types of redundant solutions can be implemented. Redundancy of systems can be classified in three different aspects: Independence, diversity and segregation. One has to make sure that a system is independent so a failure does not spread to other systems, for instance two actuators can have two different supply lines to prevent contaminated oil to spread. Applying diversity in designs can prevent similar failures in redundant systems by for example choosing two completely different power sources to the same actuator instead of having two similar ones for redundancy reasons. One can also obtain redundancy by imple-menting segregation between different systems in order to further protect the functionality when failures occur. If two different pumps previously driven by the same motor would be segregated and be driven by separate motors, the redundancy would increase. [13]

2.5.1

Static redundancy

Static redundancy represents solutions that contains duplicated systems which works in parallel to obtain the same function in order to have a backup if one of the systems would fail. The requested performance of the function is under normal circumstances obtained by adding the outputs from both elements. If one of the systems would fail, and the needed extra performance requires the still functioning system to increase its performance, the sys-tem is called "shared parallel". If not, then it is a "purely parallel" syssys-tem. One setback of static redundancy is that in practise, when two components that are not completely similar (which no real component is), even if the same supply and the same calculations occur in both systems, they will not generate exactly the same output. This can lead to a phenomena called "force fighting" where the, for example, control surface achieves its desired position with help of two actuators which are not synchronized in their output forces, leading to a higher performance demand than necessary. [13]

The static redundancy can be obtained by simple parallel redundancy, redundancy by averaging, majority voting redundancy and command/monitoring. Simple parallel redun-dancy means that the system contains two similar components to perform a task where their output is summed. If one of two component fails, it must fall back to a fail safe action in order to not disturb the other component’s action. If the function still has to be working, they must be designed in a manner so one has the needed performance to keep the system fully functioning. Redundancy by averaging can be obtained if a system that originally requires at least three components in parallel for full performance, and the components output is limited so that the system can continue to be functional with two thirds of its performance. If one component fails, the system can still be functional with reduced performance. One example is instead of using one large pump, three smaller pumps are chosen that delivers the same flow as the larger pump when combined. If one is broken, the system has two thirds of the original flow instead of no flow at all, which had happen with the larger pump. [13]

In order to reduce the loss of performance from using redundancy by averaging, majority voting redundancy can be used. Here, the output from the two still functioning components are summed, integrated, fed backwards in the loop and subtracted from the input to the com-ponents. This means if the input is, for example, a force, and the output from the components are pressures, the integration of the outputs will give a force that will be subtracted from the input. If the output force is negative and the input force is positive (which practically means

2.6. Software

that the components produces a too low combined force), the resulting input force will be increased which will increase the lowered performance. [13]

2.5.2

Dynamic redundancy

With static redundancy the added reliability comes from duplicated components that ac-tively works in parallel with the original component to achieve the function. The difference between static- and dynamic redundancy is that the added component(s) for redundancy is in "stand-by" mode and thereby does not actively contribute to the function. If a component fail, the idea is to switch from the failed to the added component. [13]

There are two different types of dynamic redundancy. These are called hot- and cold re-dundancy. A system with hot redundancy is always providing the stand-by element with power while a system with cold redundancy does not. If the element is always provided with power, the switch from the failed element will be quicker. Because of this, "hidden faults" such as faulty wiring and damaged hydraulic lines can be detected. The downside is that this consumes power and contributes to wear of the component, even if there is no power generated from it. [13]

One setup of dynamic redundancy is called "detection-correction". Here, the system contains two elements where only one is active and the other is in stand-by mode. The system also contains a simulation model which simulates the element’s function from the generated input signal. The simulation model’s output is then compared with the active element’s output. If there is a too large deviation between the simulated and true output, a signal is sent to a function which switches from the active element to the stand-by element to produce the system output. [13]

2.5.3

Hybrid redundancy

It is also possible to mix static and dynamic redundancy in order to obtain reliability. This solution is called "hybrid redundancy". One example of this could be a duplex actuator where both cylinders are active which is controlled by two computers where only one of them is powered as long as it is proven to be working. If the active computer fails, then the stand-by computer becomes active and perform the desired function.[13]

2.6

Software

Here is the software that will be used in this project.

2.6.1

Hopsan

Hopsan is an open source software focusing on simulation of fluid and mechatronic systems. It is developed at the department of fluid and mechatronic systems at Linköping university.

2.6.2

Matlab

Matlab is a software which performs mathematical and technical calculations.

2.6.3

Simulink

M

ETHODOLOGY

To describe the methodology that was followed in this project, figure 3.1 was created. It de-scribes the general overview of the method for this work.

Preparations

Concept development

Simulations

Results

Figure 3.1: Flowchart of the project methodology

The project was carried out in four different steps - Preparations, Concept development, Sim-ulations and finally Results. The four steps is described further in the following subsections.

3.1. Preparations

3.1

Preparations

In the preparation phase, the project planning was made, the background for the project was studied and relevant literature was reviewed.

Preparations

Project planning Background study

Literature study

Figure 3.2: Steps of the preparation phase.

When this phase was completed, enough information and knowledge to start the practical work were collected.

3.2

Concept development

In the concept development phase, a simple list of requirement for the generation of concepts was created. This was done to conclude the characteristics to be improved and make sure that the concepts created in some way fulfill the research questions. By doing this, it was possible to sort out some ideas by discussion within the concept generation phase, which made sure that the final concepts was worth proceeding with.

Concept development List of requirements Concept generation

Baseline simulations

Figure 3.3: Steps of the concept development phase. Horizontal lines shows workflow and vertical lines shows partial activity within the referred step.

When the list of requirements was completed, the concept generation was started. When a concept was generated, it was presented and discussed to directly decide if simple sim-ulations (baseline simsim-ulations) were needed to determine if the concept was fulfilling the requirements. If simulations were needed, they were conducted. If the concept could be proven to not fulfill the requirements in a satisfying way, it was eliminated. If it was possible to directly say that the concept fulfilled the requirements, it was kept to the next phase for further investigations. When there were a number of concepts which satisfyingly fulfilled the requirements, the concept development phase ended.

3.3

Simulations

After the concept development phase was completed, the simulation phase was entered.

Simulations System Development System validation

Figure 3.4: Main activities within the simulation phase.

The first step of this phase was to build simulation models. This was done in steps. Firstly, the system was divided into smaller parts, which were built and tested separately. When all parts worked, they were assembled to a full system. The full system was tested and developed until it worked satisfyingly.

3.3. Simulations

A list of system parameters with focus on system performance was concluded in order to make it easier to size components in the systems to meet the demands. When all concepts were developed into full system simulation models, they were tested and analyzed with fo-cus on energy efficiency and performance. A reference model was used to find the test cycles for the concepts. The performance test cycles were found by finding the limits of the reference model at different load levels. The limits were found when the amplitude of the output had dropped 3 dB (29 %), or the phase lag became more than 90°compared to the sine wave input due to a too high input frequency. The energy consumption test cycles were found by using a frequency of the sine wave input where the reference model and the concepts performed well. The performance and energy consumption could then be compared with the reference model.

System Development Partial system sim.models Full system simulations

List of system parameters

Component sizing

Figure 3.5: Steps of the system development activity. Horizontal lines shows workflow and vertical lines shows partial activity within the referred step.

When the system development was concluded, one setup of each concept were further inves-tigated and validated. They were used to perform a flight test with an airplane simulation model where the results will be compared to a conventional system. The systems were also further investigated in this step with regards to the required redundancy for aviation indus-try.

System validation Redundancy analysis

Flight mission simulations

3.4

Results

In the last phase, called Results, the results from the simulations were analyzed to see how well the concepts performed, how realistic the results were and what could be improved further. Comparisons between the concepts and systems used nowadays in aviation industry were conducted in order to see how much improvement the concepts implied.

Results Analysis of results

Comparison with previous system

C

4

I

MPLEMENTATION

4.1

Concept development

In this section, the concept generation and development is presented as well as the concepts that are further investigated.

4.1.1

List of requirements

A list of requirements for the concepts was created (see table 4.1) in order to restrict what type of concepts will be generated.

Table 4.1: Requirements of the DHA concepts

Requirements Demand/Desirable

Reduces the impact of having a discrete number of forces Demand Efficiency shall significantly improve

compared to conventional systems Demand

Energy recuperation should be possible Desirable The concepts shall be based on digital hydraulics Demand

Total weight shall be considered Demand

System size shall be considered Demand

System complexity should be considered Desirable

Redundancy shall be considered Demand

Should be controlled with force control Demand

4.1.2

Concept generation

Here the generated concepts are explained. In common all the concepts are based on the DFCU idea to establish force steps and they all include a multi-chamber actuator. The con-cepts can have two, three or four pressure sources. Thus there is three configurations within each concept.

Concept 1

The first of three concepts consist of a multi-chamber piston, two 4/3 servo valves, and a DFCU. The system can, which can be seen in figure 4.1, be divided into two sections where the first part is the DFCU and the second is the servo valves together with the multi-chamber.

p

p

p

Figure 4.1: System layout of concept 1.

The DFCU can consist of different number of pressure sources to be able to investigate the systems energy consumption and controlability. Regardless of the number of pressure sources, the DFCU function is used to establish the required force step.

While the DFCU delivers the discrete force steps, the servo valves create the possibility to continuously control the force output. To do this the servo valves will control the load pressure of each side of the multi-chamber by throttling of the pressures. When the desired load pressures is established, the correct force output is delivered from the actuator. The servo valves will deliver the pressures into the chambers smoothly up to the maximum force step if that is required. If it is needed to reduce the force output without changing force step, the servo valve will switch the direction of the pressures, which results in a reduced force output. It is important to keep track of which pressure source is delivered to each chamber, since if the same pressure source is delivered to all chambers it will be impossible to achieve continuous control.

4.1. Concept development

In this system it is also possible to filter some of the force spikes caused by on/ off valve switching with the servo valves. When the servo valves are closed or only partial open and the force step the is changed, the pressure oscillations will occur before the servo valve instead of inside the actuator chambers.

Baseline simulations

To prove and identify if it is possible to control the load pressure with a 4/3 proportional servo valve, a simple simulation was made in Hopsan. This simulation was made on a sys-tem with the proportional valve connected to two fixed volumes and two constant pressure sources. The volumes where pre-loaded with a pressure of 10 MPa and the pressure sources had a pressure of 5 and 20 MPa. The reference control value was set to 12 MPa. This means that the pressure must increase in one volume and decrease in the other volume. The valve is controlled with a PI-controller where the load pressure is fed back to achieve the desired pressure difference.

Figure 4.2: Base line simulation with fixed volumes.

A transfer function (4.1) was made of the system with valve position (xv) as input and load

pressure (pL) as output. This was realized using the flow equation (2.6), flow continuity

equation (2.5) and equation for load pressure (2.4), which was linearized and Laplace trans-formed. The transfer function was made for a system with a symmetrical piston and the piston velocity is in this case a disturbance. By using a simulation model with fixed volumes, the disturbance is removed. The regulator was therefor created without taking the distur-bance into account. The closed loop system characteristics was determined as a first order system with a determined time constant. The method Lambda tuning was used to find a regulator that would achieve the closed loop characteristics and it showed that it should be possible with a PI-regulator.

PL =2 Kqxv´xpAs V βes+Kc (4.1) Gc= 1 1+0.05s (4.2)

To test this, the regulator was implemented in the Hopsan model together with the hydraulic system and a reference value was tried to be followed. The result of this test can be seen in figure 5.1.

Concept 2

p

p

p

Figure 4.3: System layout of Concept 2.

Concept 2 is a DFCU with all on/off valves replaced with 2/2 way proportional seat valves. This in order to be able to throttle the pressure to each chamber individually which is ex-pected to give a larger freedom in the control strategy. The valves are almost leak-free and directly actuated. In comparison with a servo valve the response frequency is lower as a trade off for reduced leakage. With the possibility to throttle the pressures it is expected that the force step will not have to be changed as often compared to a DFCU with on/off valves.

4.1. Concept development

Concept 3

The third concept builds on proportional directional valves to control the pressure in each chamber. The idea here is to keep the number of valves lower in order to keep total system weight low. Due to the valve design it is possible to choose the supply pressure in use indi-vidually for each chamber. This makes it possible to directly increase or decrease pressure in the chamber for small adjustments. This also makes it impossible to achieve "short circuits" in this concept unlike the other concepts.

p

p

Figure 4.4: The figure shows a Hopsan model of concept 3.

If more than two supply pressures shall be used, the third supply pressure can be connected to the volume between the pressure sources and the valves with an directional on/off valve (see figure 4.5).

If the added pressure is lower than the previously highest pressure, the control signal to the servo valve can be kept since the pressure still will be increased/decreased if the valve moves in the same direction as before. A fourth pressure source can be added in the same way, where the two highest pressures shall be connected to the "left" port of the servo valve and the two lowest to the "right" port to keep the same control signal.

By adding the extra pressure sources in this way, the benefit with not having the possi-bility to short circuit the pressure sources is maintained.

p

p

p

Figure 4.5: The figure shows a Hopsan model of concept 3 with 3 pressure sources.

Baseline simulation

A simple simulation was carried out to see how one can control a two chamber cylinder with two different pressure sources and two directional valves (see figure 4.6).

The objective was to dimension a spring load in such way that a change in pressure in one chamber will result in a movement of the piston. The relevant parameters of the system were set to the values found in table 4.2.

Table 4.2: System parameters for the baseline simulation Parameter Value ps,1 100 [bar] ps,2 10 bar A1 0.001 [m2] A2 0.001 [m2] m 0.0001 [kg] sl 1 [m]