No description available.

Lau, Tak Kit. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2009. / Includes bibliographical references (leaves 168-176). / Abstract also in Chinese. / Chapter 1 --- Introduction --- p.18 / Chapter 1.1 --- Motivation and Problem Statement --- p.18 / Chapter 1.2 --- Literature Review --- p.19 / Chapter 1.2.1 --- Avionics Design --- p.19 / Chapter 1.2.2 --- Controller Design --- p.21 / Chapter 1.2.3 --- Dynamics Analysis --- p.23 / Chapter 1.2.4 --- State Estimation for GNSS Outage --- p.24 / Chapter 1.3 --- Outline --- p.25 / Chapter 2 --- Actuation Dynamics --- p.26 / Chapter 2.1 --- Mcchanism Of The Rotor --- p.27 / Chapter 2.2 --- Mcchanism Of Swashplate And Rotor --- p.28 / Chapter 2.3 --- Numerical Analysis Of Cyclic Pitch Angle --- p.31 / Chapter 2.4 --- Helicopter Dynamics --- p.33 / Chapter 2.4.1 --- Aerodynamic Forces And Moments --- p.35 / Chapter 2.4.2 --- Aerodynamic Drag --- p.36 / Chapter 2.4.3 --- Incremental Lift --- p.36 / Chapter 2.4.4 --- Tail Rotor Thrust And Moment --- p.36 / Chapter 2.4.5 --- Deadweight And Moment --- p.37 / Chapter 2.5 --- The Conventional Inadequacy Of Adding A 90° Phase-Lag --- p.38 / Chapter 2.6 --- The Gyroscopic Effect In Helicopter Dynamics --- p.39 / Chapter 2.6.1 --- How Precession Works --- p.43 / Chapter 2.6.2 --- The Analytical Form --- p.45 / Chapter 2.6.3 --- Numerical Analysis Of The Gyroscopic Effect --- p.48 / Chapter 3 --- State Estimation For GNSS Outage --- p.52 / Chapter 3.1 --- GNSS Error And UAV Failure --- p.52 / Chapter 3.2 --- "Kalman Filter, And The Extended Kalman Filter" --- p.53 / Chapter 3.3 --- Unscented Kalman Filter --- p.54 / Chapter 3.4 --- Process And Measurement Model --- p.55 / Chapter 3.4.1 --- The IMU Driven Model And Sensor Error --- p.57 / Chapter 3.5 --- Modifications To The Model And UKF Algorithm --- p.62 / Chapter 3.5.1 --- Acceleration White Noise Bias (AWNB) --- p.62 / Chapter 3.5.2 --- Acceleration Scale (AS) --- p.64 / Chapter 3.5.3 --- Prioritized Propagation Of States (PPS) --- p.64 / Chapter 3.5.4 --- Performance Of The Proposed Enhancements --- p.66 / Chapter 3.5.5 --- Tripled Percentage Reduction Of Position RMSE When Using PPS With AWNB --- p.73 / Chapter 4 --- Autopilot For Attitude Stabilization --- p.84 / Chapter 4.1 --- Oil Test Bondi --- p.85 / Chapter 4.2 --- On Unconstrained Flight --- p.88 / Chapter 4.2.1 --- Tracking Reference Problem --- p.88 / Chapter 4.2.2 --- An Alternative To PID Attitude Control --- p.91 / Chapter 4.2.3 --- The Proposed Hierarchical PD Controller --- p.91 / Chapter 4.2.4 --- Stability Analysis --- p.92 / Chapter 4.2.5 --- Hierarchy Of The Varying Tracking Reference --- p.95 / Chapter 4.2.6 --- Asymptotical Stability And Robustness --- p.99 / Chapter 4.2.7 --- Experiment and Performance Of The Proposed Controller --- p.101 / Chapter 5 --- Avionics And Test Bench Design --- p.105 / Chapter 5.1 --- Avionics Design --- p.105 / Chapter 5.1.1 --- Design Essentials --- p.107 / Chapter 5.1.2 --- Synchronization Of Commands --- p.107 / Chapter 5.1.3 --- Normalization Of Servomechanism Commands --- p.110 / Chapter 5.2 --- Test Bench Design --- p.110 / Chapter 5.2.1 --- The Idea --- p.111 / Chapter 5.2.2 --- Concern --- p.111 / Chapter 5.2.3 --- Test Bench Design Options --- p.112 / Chapter 5.2.4 --- Building The Test Bench --- p.113 / Chapter 5.2.5 --- Disturbance In IMU Data --- p.113 / Chapter 5.2.6 --- The Solution To IMU Saturation --- p.115 / Chapter 6 --- Conclusion --- p.118 / Chapter 6.1 --- Actuation Dynamics --- p.118 / Chapter 6.2 --- State Estimation for GNSS Outage --- p.119 / Chapter 6.3 --- Hierarchical PD Controller --- p.121 / Chapter A --- Appendix - Derivation From Recursive Least Square Estimation To Kalman Filter --- p.122 / Chapter A.1 --- Recursive Least Square --- p.122 / Chapter A.1.1 --- Alternate Estimator form for RLS --- p.134 / Chapter A.1.2 --- Propagation of States and Covariance --- p.137 / Chapter A.1.3 --- Kalman Filter --- p.139 / Chapter B --- Appendix - Actuation by Gyroscopic Effect --- p.144 / Chapter B.1 --- Expression of The Induced Moment Due to Gyroscopic Effect In The Total External Moment --- p.150 / Chapter B.2 --- An Illustrated Example --- p.153 / Chapter B.3 --- Another Derivation By Using A Different Orientation Definition --- p.156 / Chapter B.4 --- Dimensions of the helicopter for experiments --- p.166 / Bibliography --- p.167

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2009. / This thesis describes the successful development of an autopilot for an unmanned radio controlled helicopter. It presents a non-linear helicopter model. An adaptive linearised model is derived and used to design a controller. The adaptive full state controller is tested in various ways, including two aerobatic manoeuvres. A number of analyses are performed on the controller, including its robustness to parameter changes, noisy estimates, wind and processing power. The controller is compared with a non-adaptive counterpart, which leads to the conception, design and analysis of a much improved control structure. Practical flight test results are presented and analysed. In some instances available literature was reworked and re-derived to produce a genericmodel-controller package that can easily be adapted for helicopters of any make, model and size.

Thesis (MScEng)--University of Stellenbosch, 2004. / ENGLISH ABSTRACT: A need exists for the development of an unmanned rotorcraft capable of autonomous flight, as would be required for the survey of high voltage electricity supply lines. A program was initiated at the University of Stellenbosch in December 2002 in order to develop such an aircraft. The first goal of this thesis was the development of software that could calculate the stability and control derivatives of a model helicopter. These derivatives could then be used in the formulation of an appropriate helicopter control strategy. The second goal of the thesis was an investigation of the stability and control characteristics of model helicopters. The trim settings of the helicopter were required in the calculation of the stability and control derivatives. A computer program was developed to determine the trim settings of a helicopter in forward flight. Another program was developed to calculate the stability and control derivatives, using the results of the trim analysis. The trim analysis was based on the assumption of negligible coupling between the longitudinal and lateral modes of motion. The method proposed by Bramwell (1976) was used to perform the trim analysis. The stability and control derivatives were calculated by obtaining the trim settings from the trim analysis. These derivatives were then used to solve the roots of the characteristic equations of the longitudinal and lateral modes of motion. The stability of the helicopters were investigated firstly by examining the stability derivatives and secondly through root-loci analyses. The most important results were the following: • The root-loci analyses indicated that a helicopter without a horizontal stabiliser suffered from instability of the phugoid mode. It was also found that the short-period motion of these helicopters was heavily damped. Fitting a horizontal stabiliser to these helicopters caused the phugoid motion to become stable even at low speeds. This was achieved at the cost of a reduction in short-period motion damping. • The periods of the lateral and longitudinal motions were smaller than those found on full-scale helicopters. This was attributed to the small mass and inertia properties of the model helicopters. An increase in speed is generally accompanied by an increase in the stability of the helicopters. This could be attributed to the effective operation of the tail surfaces at higher speeds. • The axial climbing speed of a helicopter is influenced by the rotor speed. A low rotor speed allows higher climbing velocities at a given power setting. This was due to lower induced power losses at low rotor speed, assuming that no blade stall occurs. • The rotor speed does not influence the incremental amount of power (M:,) required to achieve a certain climbing velocity, due to the fact that the profile power losses are constant for a certain rotor speed. • The simplified horseshoe-vortex theory can be used to analyse the downwash angle at the horizontal stabiliser if the helicopter is in high-speed forward flight. / AFRIKAANSE OPSOMMING: Daar is tans 'n vraag na die ontwikkeling van onbemande rotor-vlerk vliegtuie wat die vermoë beskik om hulself te beheer. Hierdie tipe vliegtuie sal gebruik word om byvoorbeeld hoë-spannings elektrisiteitverskaffingsdrade na te gaan. 'n Program is in Desember 2002 by die Universiteit van Stellenbosch begin om sulke vliegtuie te ontwikkel. Die eerste doel van hierdie tesis was om sagteware te ontwikkel wat die stabiliteit- en beheerafgeleides van 'n model helikopter kon bereken. Hierdie afgeleides kan dan gebruik word om 'n gepaste helikopter beheerstrategie saam te stel. Die tweede doel van die tesis was om die stabiliteit- en beheerseienskappe van model helikopters te ondersoek. Die berekening van die stabiliteit- en beheerafgeleides van die helikopter berus op die beheerinsette benodig om die helikopter in ewewig te hou (trim). 'n Rekenaarprogram is ontwikkelom hierdie beheerinsette vir 'n helikopter in voorwaartse vlug te bereken. 'n Ander program is ontwikkelom die stabiliteit- en beheerafgeleides te bereken met behulp van die ewewig beheerinsette. Die analise van die helikopter in ewewig berus op die aanname dat die grootte van die koppeling tussen die longitudinale en laterale beweginsmodusse weglaatbaar klein is. Die beheerinsette van die helikopter in ewewig tydens voorwaartse vlug is bereken deur van Bramwell (1976) se metode te gebruik. Die stabiliteit- en beheerafgeleides is bereken deur van hierdie beheerinsette gebruik te maak. Die afgeleides is gebruik om die wortels van die karakteristieke vergelykings van die longitudinale en laterale bewegingsmodusse te bereken. Die stabiliteit van die helikopters is eerstens beoordeel deur die stabiliteitsafgeleides te ondersoek en tweedens deur middel van 'n wortel-lokus analise. Die belangrikste resultate is as volg: • Die wortel-lokus analise toon dat 'n helikopter sonder 'n horisontale stabiliseerder phugoid-onstabiliteit (Iangperiode onstabiliteit) het. Die kort-periode beweging van hierdie helikopters het verder groot hoveelhede demping aangetoon. Die phugoid-beweging kon selfs teen lae snelhede gestabiliseer word deur 'n horisontale stabiliseerder aan te heg. Hierdie stabiliteit is egter bereik ten koste van die demping van die kort-periode beweging wat verminder is. • Die periodes van die longitudinale en laterale bewegings is kleiner gewees as vir volskaal helikopters. Dit kan toegeskryf word aan die klein massa en inersie van die model helikopters. • Die stabiliteit van die helikopter is in die algemeen verbeter soos die snelheid verhoog. Dit kan toegeskryf word aan die beter werking van die stert teen die verhoogde snelhede. • Die klimtempo van die helikopter word beïnvloed deur die hoofrotor snelheid. 'n Lae hoofrotor snelheid laat 'n hoër klimptempo toe teen 'n spesifieke drywinginset. Dit is as gevolg van die laer geïndusseerde drywingsverliese teen die laer hoofrotor snelheid. Daar word aanvaar dat die lugvloei oor die lem nie staak nie. • Die hoofrotor snelheid beïnvloed nie die inkrimentele drywing (M,,) wat benodig word om 'n sekere klimtempo te bereik nie. Dit is as gevolg van die konstante drywings verliese teen 'n sekere hoofrotor snelheid. • Die vereenvoudigde perdeskoenwerwel teorie kan gebruik word om die afspoel hoek by die horisontale stabiliseerder te bereken indien die helikopter in hoë-spoed voorwaartse vlug is.

Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2005. / This project started a new research area in rotary-wing °ight control in the Computer and Control group at the University of Stellenbosch. Initial attempts to build a quad-rotor vehicle exposed di±culties which motivated changing to a standard model helicopter as a test vehicle. A JR Voyager E electrically powered model helicopter was instrumented with low-cost, low-weight sensors and a data communication RF link. The total cost of the sensor, communication and microcontroller hardware used is approximately US$ 1000 and the added onboard hardware weighs less than 0:4 kg. The sensors used to control the helicopter include a non-di®erential u-Blox GPS receiver, Analog Devices ADXRS150 rate gyroscopes, Analog Devices ADXL202 accelerometers, a Polaroid ultrasonic range sensor and a Honeywell HMC2003 magnetometer. Successful yaw, height and longitudinal position control was demonstrated. Signi¯cant further work is proposed, based on the literature study performed and the insights and achievements of the ¯rst rotary-wing unmanned aerial vehicle project in the group.

Helicopters suffer from severe vibration levels compared to fixed-wing aircraft. The main source of vibration in a helicopter is the main rotor which operates in a highly unsteady aerodynamic environment. Active vibration control methods are effective in helicopter vibration suppression since they can adapt to various flight conditions and often involve low weight penalty. One such method is the actively controlled flap (ACF) approach. In the ACF approach, a trailing-edge flap (TEF) located in each rotor blade is deflected at higher harmonics of rotor frequency to reduce vibratory loads at the rotor hub. The ACF approach is attractive because of its simplicity in practical implementation, low actuation power and enhanced airworthiness, since the flap control is independent of the primary control system. Multiple-flaps are better suited to modify the aerodynamic loading over the rotor blade and hence offer more flexibility compared to a single flap. They also provide the advantage of redundancy over single-flap configuration. However, issues like the number, location and size of these individual flaps need to be addressed based on logic and a suitable performance criteria. Preliminary studies on a 4-bladed hingeless rotor using simple aerodynamic and wake models predict that multiple-flaps are capable of 70-75 percent reduction in hub vibration levels. Numerical studies confirm that multiple-flaps require significantly less control effort as compared to single-flap configuration for obtaining similar reductions in hub vibration levels. Detailed studies include more accurate aerodynamic and wake models for the rotor with TEF’s. A simple and efficient flap control algorithm is chosen from literature and modified for use in multiple-flap configuration to actuate every flap near complete authority. The flap algorithm is computationally efficient and performs creditably at both high and low forward speeds. This algorithm works reasonably well in the presence of zero-mean Gaussian noise in hub load data. It is also fairly insensitive to small changes in plant parameters, such as, blade mass and stiffness properties. The optimal locations of multiple TEF’s for maximum reduction in hub vibration are determined using Response Surface methodology. Piezoelectric stack actuators are the most promising candidates for actuation of full-scale TEF’s on helicopter rotors. A major limitation of piezoelectric actuators is their lack of accuracy due to nonlinearity and hysteresis. The hysteresis in the actuators is modeled using the classical Preisach model (CPM). Experimental data from literature is used to estimate the Preisach distribution function. The hub vibration in this case is reduced by about 81-86 percent from baseline conditions. The performance of the ACF mechanism can be further improved by using an accurate hysteresis compensation scheme. However, using a linear model for the piezoelectric actuator or an inaccurate compensation scheme can lead to deterioration in ACF performance. Finally, bench-top experiments are conducted on a commercially available piezostack actuator (APA500L from CEDRAT Technologies) to study its dynamic hysteresis characteristics. A rate-dependent dynamic hysteresis model based on CPM is used to model the actuator. The unknown coefficients in the model are identified using experiments and validated. Numerical simulations show the importance of modeling actuator hysteresis in helicopter vibration control using TEF’s. A final configuration of multiple flaps is then proposed by including the effects of actuator hysteresis and using the response surface approach to determine the optimal flap locations. It is found that dynamic hysteresis not only affects the vibration reduction levels but also the optimal location of the TEF's.