An alternative concept of docking for the ConeXpress satellite

(1)

MASTER’S THESIS

2006:075 CIV

INGA-LENA HANNUKKA

An Alternative Concept of Docking for the

ConeXpress Satellite

MASTER OF SCIENCE PROGRAMME in Space Engineering

Luleå University of Technology Department of Space Science, Kiruna

(2)

An Alternative Concept of Docking for the ConeXpress Satellite

Inga-Lena Hannukka January 19, 2006

(3)

(4)

Preface and Acknowledgements

This Master’s Thesis Report is part of the final work for the Master of Science degree in Space Engineering at Luleå University of Technology, Sweden. The work has been carried out at Dutch Space, Leiden, the Netherlands, in the Advanced Systems and Engineering Division during July to December in the year 2004.

I wish to express my gratitude to everyone at Dutch Space for giving me the opportunity to learn and develop my knowledge during the work on my Master’s Thesis. Most of all I would like to thank my supervisor and mentor Aad van Swieten who has been most patient and helpful during these months. I also want to thank my former teacher Peter Berlin for helping me get the position in the first place and Paul-Robert Nugteren, Han Sholten and Flip Zijdemans who all made it happen in the end.

I had great help from many of the people at Dutch Space in completing this Master’s Thesis work. Some of them are Wim van Leeuwen, Robert Blommestijn, Lex Meijer, Harm- Jan de Graaf, Erik Schenkeveld, Aad Eggers, Gijs Oomen, Gerard Kester, and Thomas Peters and I wish to thank them all. Also big thanks to my room mate Hans Kollen for keeping me focused at my work during late afternoons and to Seona Candy, Bart Schat, Alastair Wise, Jeroen Heymans, Carlos Dias and Hanjo Hoogendoorn for nice lunch conversations!

Many thanks also to my supervisors and examiners at Luleå University of Technology and at Kiruna Space and Environmental Campus, Björn Graneli and Priya Fernando, for all the time spent on this report and for motivating me in the first place.

And I wish to give all my love to Mattias for always being there for me.

Båstad, December 2005 Inga-Lena Hannukka

(5)

(6)

Abstract

ConeXpress is an affordable multi-mission platform that will be launched as auxiliary payload on the Ariane 5 launcher and as such it exploits the excess launch capacity of Ariane 5 for easy access to space. The satellite can be used for different missions depending on the payload placed in it. The primary objective of ConeXpress is to perform in-orbit servicing.

Two years ago the company Orbital Recovery Ltd started to pursue the market of in-orbit servicing and has now teamed with Dutch Space to develop ConeXpress as an Orbital Recovery System. ConeXpress Orbital Recovery System will prolong the operational life of telecommunication satellites by supplying guidance, navigation and propulsion for station keeping and attitude control for an additional 8-10 years.

This Master’s Thesis is concerned with the docking phase of the ConeXpress project.

The model for the baseline docking phase is currently being developed at Dutch Space and the main mechanical element of the docking system is the Capture Tool, constructed by DLR in Germany. The Capture Tool is inserted into the apogee kick engine nozzle of a client satellite where it expands and locks in the throat of the engine nozzle. This Master’s Thesis will deal with a possible alternative concept to the docking phase. This approach will reduce the mass of the satellite by using only sensors and control and not any mechanical structures that would complicate the docking. Using the alternative concept for docking the Capture Tool could be abandoned and also the high position and orientation control requirements associated with it.

The work started with an investigation of possible alternative docking concepts and after a trade-off the most suitable concept, a 3D Laser Camera System from the Canadian company Neptec, was chosen for further investigation. In order to test the performance a model of the camera system needed to be implemented in the DSF, the Design Simulation Facility (based on MATLAB/Simulink); the simulation environment used in the ConeXpress project. To get acquainted with the program, simulations were first made on the original docking concept and also on the rendezvous phase. After this, simulations for the docking phase using the Neptec 3D Laser Camera System were started. A simplified version of the System, based on the docking camera with range, field of view and resolution adapted for the Neptec system was implemented in the simulator. The results from this simple model clearly showed first of all that the control system of the satellite in the simulator needed to be improved to be able to handle the higher speeds and longer range during the docking phase, but also that the Neptec 3D Laser Camera System handles the procedure well. It is therefore an interesting alternative to the Capture Tool docking approach and should be investigated further.

(7)

(8)

Sammanfattning

ConeXpress är ett prisvärt satellitprojekt med många användningsområden. Det använder sig av outnyttjat utrymme på Ariane 5 raketen och detta medför snabb och billig åtkomst till rymden. Satelliten kan användas för många olika missioner och i olika sammanhang beroende på vilken nyttolast som placeras i den. Det primära användningsområdet är servicetjänster på befintliga satelliter som redan finns i bana runt jorden. För två år sen började företaget Orbital Recovery Ltd att undersöka marknaden och intresset för denna tjänst och de har nu börjat ett samarbete med Dutch Space för att utveckla ConeXpress som ett Orbital Recovery System.

ConeXpress Orbital Recovery System kommer att förlänga livstiden på telekommunikationssatelliter genom att erbjuda attityd- och bankontroll i ytterligare 8-10 år.

Detta examensarbete behandlar dockningsfasen i ConeXpress projektet. Modellen för den nominella dockningsfasen utarbetas i nuläget på Dutch Space och den huvudsakliga mekaniska delen i dockningssystemet är en fångstarm (Capture Tool), utvecklad av DLR i Tyskland. Fångstarmen kommer att skickas in i dysan på målsatellitens apogeummotor och där kommer toppen att expandera och låsas fast i dysans hals. Detta examensarbete kommer att utforska ett möjligt alternativ till den nominella dockningsproceduren. Målet är att minska massan på satelliten och endast använda sensorer och reglering av satelliten och inte några mekaniska delar som kan komma att komplicera dockningen. Genom att använda det alternativa dockningskonceptet kan fångstarmen uteslutas och därmed försvinner även de höga kraven på positions- och orientationskontroll som förknippas med denna.

Arbetet började med en undersökning av möjliga alternativa dockningskoncept och efter en trade-off valdes det mest lämpliga konceptet, ett 3D Laser Camera System från det kanadensiska företaget Neptec, för fortsatta studier. För att testa prestandan hos laserkameran skulle en modell utvecklas och implementeras i DSF, Design Simulation Facility (baserat på MATLAB/Simulink); simuleringsprogrammet som används i ConeXpress projektet. För att lära känna programmet och simulatorn gjordes först simuleringar på den nominella dockningen, och även på delar av rendezvous-fasen. Därefter påbörjades simuleringarna av dockningsfasen då laserkameran från Neptec användes. En förenklad version av systemet som baserades på dockningskameran, men med ändrade värden för räckvidd, synfält och upplösning, implementerades i simulatorn. Resultatet från denna enkla modell visade först och främst att kontrollsystemet för satelliten i simulatorn behövde uppdateras för att klara av de högre hastigheterna och längre avstånden som krävdes under dockningsfasen, men också att laserkameran från Neptec klarade uppgiften väl. Den är därmed ett intressant alternativ till den nominella dockningen och bör utrönas ytterligare.

(9)

(10)

Acronyms and Abbreviations

AOCS Attitude & Orbit Control Subsystem BOL Beginning of Life

CoM Center of Mass

CSF ConeXpress Simulation Facility CT Capture Tool

CX ConeXpress

CX-ORL ConeXpress - Orbital Recovery Ltd

DLR Deutsches zentrum für Luft- und Raumfahrt

DS Dutch Space

DSF Design Simulation Facility EOL End of Life

EOT End of Transfer

EP Electric Propulsion

ESA European Space Agency GEO Geostationary Earth Orbit GTO Geostationary Transfer Orbit GPS Global Positioning System KT/DLR Kaiser Threde / DLR LED Light Emitting Diode LEO Low Earth Orbit

LHLV Local Horizontal Local Vertical frame MEO Medium Earth Orbit

MTR Mid Term Review

ORL Orbital Recovery Limited ORS Orbital Recovery System RVD Rendezvous & Docking SER Systems Engineering Report SSC Swedish Space Corporation

(11)

(12)

1 Introduction

The assignment for this Master’s Thesis was to investigate an alternative concept for the docking phase of the ConeXpress project. ConeXpress is an affordable multi-mission platform that will be launched as auxiliary payload on the Ariane 5 launcher and as such it exploits the excess launch capacity of Ariane 5 for easy access to space. The satellite can be used for different missions depending on the payload placed in it. The primary objective of ConeXpress is to perform in-orbit servicing. ConeXpress Orbital Recovery System will prolong the operational life of telecommunication satellites by supplying guidance, navigation and propulsion for station keeping and attitude control for an additional 8-10 years. [1, 36]

1.1 Thesis objectives

The ConeXpress satellite will need to rendezvous and dock with telecommunication satellites in geostationary orbit in order to perform its mission. The model for the baseline docking phase is currently being developed at Dutch Space. The main mechanical element of the docking system is the Capture Tool, constructed by DLR in Germany. The Capture Tool is a mechanical device, placed at the end of a retractable boom, which will be inserted into the apogee kick engine nozzle of a client satellite. Once in place in the throat of the engine nozzle, the tip of the Capture Tool will expand and lock. The boom is then retracted, the client satellite will follow and a latching mechanism will attach the two satellites together.

The objective of this thesis is to investigate an alternative approach for the docking phase. Instead of using a sensitive and mechanically challenging device such as the Capture Tool it is believed that an alternative approach, based on only light sensors in some configuration and control of the satellite, will be more advantageous. Possible achievements might be reduction of mass of the satellite and less demands on the control system. The work will consist of an investigation of possible concepts and solutions that might be used in the alternative approach, a trade-off between the concepts and finally simulations of the chosen concept.

The underlying idea in the alternative approach is that it should be possible to detect the nozzle of the Client satellite by the sensors of ConeXpress as it approaches. From the information given by the sensors, the control system of ConeXpress should be able to adapt its position with respect to the nozzle so that it ends up in the correct position for the latching mechanism to lock the satellites together. The concern of this thesis is the part of the docking phase from 7 meters up until the moment where the latching mechanism takes over. The function of the latching mechanism itself will not be investigated.

First of all a literature study needs to be done. Information on ConeXpress, the structure and the current docking procedure needs to be understood. After learning about the project the investigation part of the thesis will be started. This includes a discussion regarding if and in which ways the alternative approach will be advantageous, the position and placement of sensors, how the sensors will detect the nozzle of the client satellite and how the ConeXpress can use this information to control its position. Different concepts will be discussed and a trade-off will be made between them in order to determine which concept is the most favorable.

If time is left, the chosen concept will finally be simulated in the Design Simulation Facility (DSF), based on Matlab/Simulink, in order to investigate the function of the concept.

The results from the trade-off and the simulations are to be collected in a System Engineering Report and a Test Report respectively, for Dutch Space records.

(15)

1.2 Dutch Space

Dutch Space is situated in Leiden, the Netherlands, and was originally a member of the Fokker group, Fokker Space. In 1995 it was established as an independent company and has about 350 employees. Due to their involvement in the space industry since the early days, Dutch Space has a strong reputation and a secure position as supplier of space products.

Figure 1-1 Dutch Space, Leiden

Dutch Space is Europe's leading manufacturer of solar arrays for spacecraft, both for telecommunication and scientific applications as well as earth observation applications.

Dutch Space is also an important supplier of launcher structural systems. The company’s expertise in lightweight, rigid and strong structural components have made them a major contributor to the Ariane program, providing launcher structure systems for the new Ariane 5 launch vehicle.

As a specialist in advanced robotics technology, Dutch Space is prime contractor of ERA (European Robotic Arm), an 11 meter long, stand alone space robot. The ERA is developed for the European Space Agency (ESA) by a consortium of companies from 7 countries. The ERA was supposed to be launched in 2002 on the Space Shuttle but due to the explosion of the shuttle Colombia the launch of the arm has been postponed until the shuttle program is in operation again. The ERA will be used on the International Space Station where it will be responsible for the assembly and servicing of the Russian segment.

Other specializations of the company are design, development and testing of crucial systems and subsystems for spacecraft within the field of simulation, payloads and remote sensing, microgravity, and thermal products. [37]

(16)

2 Literature Study

In order to be able to investigate the alternative docking approach a study of the background of the project, the ConeXpress satellite itself and the current docking procedure needed to be done.

2.1 The ConeXpress Project

The ConeXpress project started at Dutch Space in the year 2000 in response to an ESA telecom request. ESA wanted a proposal concerning a phase A level feasibility study of an affordable telecom demonstration solution for satellites in geostationary orbits. The launch is the main cost element for small satellites to GEO or other high orbits and this led to the concept of the ConeXpress mission, which is a new and efficient auxiliary payload system for Ariane 5. Arianespace has been closely cooperating in the development of the concept. They expect to have approximately 10-20% of excess launch capacity (about 1 ton) for their future commercial GTO launches. ConeXpress exploits this excess capacity to realize important GEO missions to very attractive GTO launch prices. [26]

Figure 2-1 The ConeXpress satellite just released by the Ariane 5 launcher

The ConeXpress satellite is a 1200 kg platform designed for smaller GEO or other high orbit missions such as earth observation missions, small telecommunication missions and other research missions. ConeXpress can be launched with either a single or a dual Ariane launch configuration where it replaces the payload adaptor of the lower primary passenger, see Figure 2-2. The cone shaped payload adaptor of the Ariane launcher is always present at every launch and in its current utilization the adaptor unit is empty inside. By using this already available structure as base for the platform, the ConeXpress can be launched as auxiliary payload which gives an easy and affordable access to space. This also allows flight- proven hardware to serve as structure for ConeXpress, and opens regular launch opportunities

(17)

onboard the Ariane 5 launcher. The platform will be equipped with an electrical propulsion system, guidance/flight control electronics and different scientific payloads for different missions. [1, 2]

An auxiliary payload launch costs about 1 million euro for 100kg, which results in about 12 million euro for the ConeXpress satellite. This is a significant reduction in launch costs compared to other launch options. Also, the large adaptor leaves sufficient room for both platform equipment and payloads. The total capacity is about 150 kg, which is more typical for smaller primary launcher satellites than for auxiliary payload satellites. [26]

Figure 2-2 Position of the ConeXpress within the Ariane 5 launcher

In the year 2002 the application of the ConeXpress satellite as an in-orbit servicing vehicle was found. Telecommunication satellites in GEO typically cost about 250 million dollars and they are designed for a useful in-orbit lifetime of 10-15 years. [36] After their fuel is depleted the satellites are boosted into a higher graveyard orbit and are decommissioned even though the income bringing communication relay payloads are still fully functional.

The American company Orbital Recovery Ltd has identified about 100 telecom satellites currently in orbit that are candidates for life extension before the year 2015. [36] The primary targets are telecommunication satellites in GEO with a maximum mass of 2500 kg, 3- axis stabilized and fully operational. ConeXpress will prolong the life of these satellites with up to 10 years under nominal conditions; however in the simulation facility a design lifetime of 12 years will be demonstrated. [1] ConeXpress will perform attitude and orbit control only, so no electrical power or fuel transfer will take place. Other applications can be to rescue satellites that have been placed in a wrong orbit by their launch vehicles or stranded at the wrong location after bad position maneuvers. In these cases ConeXpress will work as a tugboat, pushing the client satellites into useful orbital slots so that their missions can continue.

(18)

Orbital Recovery Ltd has signed Dutch Space as primary contractor of the ConeXpress Orbital Recovery System, together with the sub contractors Swedish Space Corporation, DLR and Kaiser-Threde of Germany, Sener and GMV of Spain, Snecma and Arianespace of France, Contraves Space of Switzerland and ESA. Together they will design, build and test the spacecraft based on the Ariane 5 payload adaptor unit. The first launch is scheduled for 2007/2008 and after this there will be many launches over several years. [36]

2.1.1 The structure, dimensions and coordinate frame The cone shaped structure of ConeXpress has the following dimensions:

• Top diameter: 1194 mm

• Bottom diameter: 2624 mm

• Height: 860 mm

When stowed in the launch configuration the six round solar panels are folded in underneath the satellite, which adds an extra 484 mm to the height. The solar panels are deployed just after the release from the Ariane launcher. [1]

Figure 2-3 The ConeXpress in launch configuration

Another special feature of the ConeXpress is the so called “crater”. The crater is situated at the top of the cone and intrudes 660 mm down into ConeXpress. This space needs to be empty to make room for the apogee kick engine nozzle of the primary payload, which is attached to the adaptor cone at launch.

The Capture Tool will be placed at the bottom of the crater and during the docking phase it will be extended out of the crater and into the apogee kick engine nozzle of the client satellite. Once the satellites are in the mated position the apogee kick engine nozzle of the client satellite will be inside the crater.

(19)

X

Z

Y

Figure 2-4 ConeXpress with the crater;

it can be seen as an opening in the middle of the figure.

The coordinate frame with respect to the ConeXpress is defined in x, y and z coordinates. The coordinates are attached to the ConeXpress and are orientated as shown in Figure 2-5. The x- axis is always pointing out of the crater, the y-axis is orientated along the solar panels and the z-axis completes the right-hand rule. [1]

Figure 2-5 Coordinate frame of ConeXpress

The orientation of ConeXpress can also be defined in a coordinate frame with respect to the client satellite, based on R-bar, V-bar and H-bar. This is especially useful for the rendezvous and docking phase. [5, 6, 7]

The R-bar is pointing radially towards the earth from the client satellite. During the docking the ConeXpress will be approaching from the negative R-bar direction, meaning that it will be situated above the client satellite and closing in towards it and the earth. Since the ConeXpress will be orientated with the crater facing the client satellite during this phase, the R-bar direction will thus be equal to the x-direction.

The V-bar is the direction of the velocity in the orbit, i.e. tangent to the orbit, and it is perpendicular to the R-bar. During the rendezvous ConeXpress will catch up with the client satellite from behind. It will be orientated with the crater facing the travel direction, thus, in this phase the V-bar is equal to the x-direction.

The H-bar concludes the right hand rule.

(20)

2.1.2 The mission phases

After the launch into GTO orbit the ConeXpress mission begins. It consists of 6 different phases.

1 Transfer phase 2 Rendezvous Phase 3 Docking phase 4 Mated phase

5 Transfer Phase to Graveyard orbit 6 Return to GEO and a new client Transfer Phase

The ConeXpress is using Electric Propulsion transfer from GTO orbit to GEO. The Electric Propulsion Subsystem design is based on the one developed by Snecma of France for SMART-1 [1, 5] and the input from SSC on the performance of the units is a valuable source of information. The system has four EP-thrusters in total, but it uses only two of them simultaneously and the challenge is to keep both the transfer time and the propellant mass at a minimum. The transfer time is to be less than 150 days. [1, 5]

The current strategy under consideration is the Koppel trajectory shown in Figure 2-6.

From GTO, first both the apogee and perigee are raised so that ConeXpress spirals to a super- GTO orbit; next the apogee is lowered and the perigee raised so that the orbit is circularized.

The initial apogee raising increases the efficiency of the apogee thrusting, so the total transfer time decreases. In order to minimize stay time in the Van Allen belts, the strategy could be altered. [5]

super-GTO

GTO

GEO

Van Allen belts

Figure 2-6 The Koppel Transfer strategy

Rendezvous Phase

The rendezvous phase, the closing in on the client satellite, starts at a distance of 25 km behind the client satellite in the GEO orbit. It is investigated by GMV of Spain. The sensors to be used are in the current baseline both a GPS sensor and two rendezvous sensors, one for medium range (25-1 km) and one for short (1km -100 m) and close range (100 – 5 m), see Figure 2-7. [7]

(21)

3 m 100 m 1 km 25 km Close Range

(CR)

Short Range (SR)

Medium Range (MR)

0 Docking

Rendez-Vous

Figure 2-7 The Rendezvous range

The rendezvous starting point, range 25 km is determined by absolute orbit determination for ConeXpress with respect to the client satellite.

The rendezvous flight from 25 km behind the client to a distance of about 7 km behind is pre-programmed in the onboard computer in various steps and the use of electric propulsion is preferred over the use of the cold-gas system for mass budget reasons.

At 7 km the first visual contact will be made and a gradual transition from absolute to relative distance measurement will occur. The closing in will be made by “jumps” in the v-bar direction, see Figure 2-8. In the very last part a circular motion is made, under and in front of the client to stop at 50 m above (-50m in R-bar). From here a forced entry towards the client will be performed in order to stop in the right position for the docking phase to begin. This will be at 7 m above the client, with respect to the centers of mass of the two satellites. (The actual distance will be 5 meters). [5, 7]

Figure 2-8 The closing in

The Docking Phase

Since telecommunication satellites have not been intended for docking, this is quite a delicate operation. The docking approach is made from the negative R-bar direction (w.r.t. the client satellite) in order to interfere as little as possible with the transmissions of the telecom satellite down to earth. R-bar docking is special due to the specific orientation of the Client satellite, i.e. earth pointing. The docking frame will be in the ”local horizontal local vertical” (LHLV) frame. This is a frame in which the local horizontal axis is along the V-bar and the local vertical axis is along the negative R-bar. [6]

(22)

The only feature on the satellites that can be considered to always be more or less the same is the nozzle of the apogee kick engine and the adaptor ring where the telecommunication satellite was attached to its launcher. This is where the ConeXpress will be attached. The nozzle of the Client satellite is always pointing in the anti-nadir direction (away from the earth, negative R-bar) and it is selected as capture object for the following reasons:

• universal orientation and position (always pointing along the symmetrical centre line of the client satellite)

• standard shape, although different in size

• small variation in throat diameter allowing a universal attachment device in the form of an expanding tip.

The docking system in the current baseline consists of the docking camera and the Capture Tool and is supplied by DLR and Kayser-Threde of Germany. The docking camera will target the client satellite and the ConeXpress will automatically close in towards the client the final 7 meters. The position will be corrected and co-aligned by the attitude control of ConeXpress, using the cold gas thrusters. The Capture Tool is mounted on a retractable boom, a bi-stem, and will be inserted in the client’s apogee kick engine nozzle where the expanding tip will lock in the throat. After locking, the boom will be retracted until the client satellite is resting closely against the adaptor ring and the latching mechanism can rigidify the connection. [1, 6, 9, 15]

Mated Phase

Once the docking is completed the main mission will begin. ConeXpress will take over north- south and east-west station keeping as well as the attitude control of the client satellite. This will be done by using the electric propulsion system as well as the reaction wheels supplied by Sener of Spain. This will be quite a challenge. ConeXpress will take command of the client satellite; meaning that it will need to be able to handle the extra weight and the change in centre of mass. The EP system and the reaction wheels will have to be placed in such a way that they are able to control both the ConeXpress itself, as well as the mated configuration, with a satisfying result. These new operational procedures are currently investigated and will also be tested. The reaction wheel off-loading will be performed using cold gas thrusters. The ConeXpress will have enough propellant to drive these procedures for up to 10 years. It is assumed that the client is electrically still fully functional and that it is three-axis stabilized.

[1]

Two things are important during this phase. One is that the ConeXpress should not interfere with the clients objectives as a telecom satellite and obstruct the transmissions in any way. The second is that the solar panels of both satellites remain un-shaded to give maximum amount of solar power. The first objective is solved by having ConeXpress docked on the side of the client pointing away from the earth where no transmissions are made. The second objective is solved by having the solar panels positioned at an angle of 90 degrees with respect to the solar panels of the client, see Figure 2-9.

(23)

Figure 2-9 The ConeXpress in mated configuration

Transfer Phase to Graveyard Orbit

After the end of life of the client satellite the ConeXpress will transfer it to a graveyard orbit, 300 km above GEO. There it will undock and leave the dead telecom satellite behind.

Transfer to new client

If the ConeXpress has enough propellant and if there is another client satellite close by that needs the services of the CX-ORS, it will relocate in GEO and dock to the new client.

2.1.3 The client satellites

There are a large number of different telecommunication satellite platforms in the GEO belt.

Due to this a survey has been made with the intention of finding possible candidates for a life extending mission. This was done to limit the number of possible platforms that can be considered. A total of 33 different platforms were included. [1] The restrictions on the platforms are that they should be electrically fully functional, three-axis stabilized and that the apogee kick engine nozzle should be unobstructed for enabling the docking.

For the simulations one client satellite needed to be chosen, the so-called “Straw- man”. It has been selected from an inventory of identified telecommunications satellites currently in orbit. The satellite chosen is the Boeing 601 High Power (BSS-601HP) platform –

“Superbird 4”. It serves as input for the several system and subsystem implementation trade- offs during phase B1 to MTR. [1]

Figure 2-10 The Straw Man satellite and its apogee kick engine nozzle

(24)

2.2 The current docking approach

As mentioned earlier the docking procedure is somewhat challenging. We need to understand the current docking approach before we can start investigating an alternative procedure. And to know the current docking approach, we needed to learn some more about docking in general.

2.2.1 Docking in general

Docking is a fascinating subject that is needed in many occasions, for instance on the Space Shuttle. The shuttle is intended to work as a transport vehicle and the docking feature is necessary to be able to deliver parts, perform reparations and so on.

ISS is one of the places where the shuttle docks. ISS has different docking ports depending on the type of spacecraft that are to be docked. Both American and Russian, as well as European, spacecrafts needs to be able to dock to the space station.

There are two concepts to mention. First of all there is Docking. This is when a spacecraft connects directly to another spacecraft with an airlock or other connection device.

This is what the space shuttle does. And secondly there is Berthing. This is when the chaser spacecraft has a device, like an arm or so, sticking out to capture or grab the target spacecraft.

Then the arm is retracted and the two spacecraft is connected thoroughly. [33]

Usually the docking feature has been thought of already at construction of the spacecraft. There is something to steer towards; markings to get the right position, a camera lens with markings and markings on the target that are to match together.

Automatic docking, using on-board computers only, has been done before in LEO but never in GEO.

2.2.2 Docking on the ConeXpress

In the ConeXpress project the target – the client telecommunication satellite – is not intended to be docked to. There are no markings to steer towards and no default area where the docking device will fit. ConeXpress needs to dock to an uncooperative target.

The only feature that is the same on all telecommunication satellites, as well as on any satellite, is the apogee kick engine nozzle. The apogee engine is used after launch when the satellite is traveling to its orbit. Usually, if the satellite is going to GEO, the launcher drops the satellite in a GTO orbit. This orbit has its apogee at GEO and its perigee at LEO. It is elliptical. In order for it to reach the geostationary orbit it needs to circularize its orbit. This is done by burning the apogee kick engine at apogee. The satellite gains speed and the perigee is raised. After reaching GEO the apogee kick engine is not used any more and the nozzle of the apogee engine is basically just sitting there, not being good for anything. Until ConeXpress! It is this feature, as well as the adaptor ring (once used to attach the satellite to its launcher) that ConeXpress will take advantage of in the docking procedure.

When ConeXpress begins the docking phase at a distance of 7 meters (w.r.t CoM) the stereo docking camera system will be imaging the client satellite. It will see the nozzle and the adaptor ring and during the approach the control system will be able to keep ConeXpress on the right path.

The main mechanical element of the docking system in the current approach is the Capture Tool, manufactured by DLR. The Capture Tool is mounted on a boom, a bi-stem. The bi-stem is made out of two stems. It can be deployed and retracted. (Think of the reflective bracelets that are made of a metallic, curved band. To make it stiff you bend the band and when you strike it over your arm it rolls up around your wrist.) Once ConeXpress is close enough the bi-stem is deployed. At the end of the bi-stem the Capture Tool is attached. It is inserted into the apogee motor nozzle of the client satellite.

(25)

There are two laser systems mounted on the Capture Tool, one at the front and one at the end.

Each system consists of three lasers mounted around the Capture Tool with equal spacing (120°). These lasers will detect the inside of the nozzle. A beam will be sent out from each laser and a small lens will detect the dot on the inside surface of the nozzle. From this the distance to the nozzle surface will be measured and lateral position of the Capture Tool with respect to the nozzle will be calculated. If the Capture Tool is not centered it will be possible for the on-board computer to adapt and correct any position errors. When also the lower laser system starts to detect the inside of the nozzle it is possible to get a reading of the orientation (the angle with which the Capture Tool enters) and correct this if necessary.

At the moment there is no feature added for detecting the depth; how far the Capture Tool has intruded inside the nozzle. This may be added by using the curvature of the nozzle.

There is a connection between the length and the radius of the nozzle. This curve is optimized in order to maximize the effect of the exhaust of the engine. [16] So by knowing the radius at each point it will be possible to calculate the length, i.e. how deep into the nozzle the Capture Tool has reached.

Once the tip of the Capture Tool has reached the throat of the nozzle (the narrowest part) a proximity sensor will give a signal, the tip will expand and lock so that the client satellite is captured. The bi-stem will be retracted all the way back until the latching system will be able to lock down on the adaptor ring of the client satellite. [1, 6, 8, 9, 10, 11, 12, 22]

Figure 2-11 The docking system in the current docking approach

ConeXpress cross-section ConeXpress

cross-section

Client satellite

Nozzle

Latching Mechanism Latching

Mechanism

BI-STEM actuator Capture

Tool

Lower Laser System Upper

Laser System

Camera Camera

(26)

2.2.3 Accuracies and resolution

In the lateral plane, the stereo camera system will detect misalignments of 20 mm radius at a distance of 5 meters and 3 mm at 1 meter. In the depth plane the camera will be operational from 5 meters to 1 meter and will be able to handle the same accuracy as in the lateral plane.

The orientation accuracy will be 1 degree. [12]

For the Capture Tool it will be possible to detect misalignments in lateral movement of about 1-2 mm. Depth measurements and corrections will not be detected at this point and the orientation misalignment will be possible to detect down to <0.1 degree. [10]

2.2.4 Control issues

From a control system point of view the Capture Tool approach will be delicate. The accuracy demands increases when the Capture Tool gets further into the nozzle; there is less and less space to move and less and less room for accidents and misalignments. The excess space is on sub-mm level [6, 15] and this gives the need for very accurate control of the ConeXpress with respect to the client satellite. When the tip is to expand in the throat of the nozzle, it has to be in the exact right position. In order to reach this point-like final area, high demands are posed on the control system of ConeXpress.

Figure 2-12 Capture tool approach from control point of view (Horizontal axis can be considered as the time line for the docking approach and the vertical as a measurement of the accuracy needed)

2.2.5 The Latching System

The mechanical part, the latching mechanism, of the docking system is a challenge for the ConeXpress project.

The structure of ConeXpress, the adaptor cone, is used on the Ariane rockets to mount the main payload and to keep it secure during the launch. At release, 12 powerful spring- loaded pins push away the payload from the launcher. These pins will stick out 66 mm from the adaptor ring of the adaptor cone and the power needed to push them back in will not be achieved by the approach speed of the ConeXpress towards the Client satellite when docking.

[13, 14, 26] Because of this, the adaptor ring of the client satellite cannot rest firmly against the adaptor ring of ConeXpress.

Latching Mechanism is operated

(27)

There are also different sizes of the adaptor rings for different client satellites, and this needs to be taken into account. There are four different possible rings, three circular with different diameters (1666 mm, 1194 mm and 937 mm) and one square (1663 mm). [1]

One of the suggestions for the latching system consists of three tables with 120 degrees spacing that fold out of the crater of ConeXpress to a height of 100 mm above the adaptor ring, in order to avoid the pins. On these tables the adaptor ring of the client will rest.

When in position, the latching mechanism will be operated and the connection will be rigidified. The latching mechanism itself consists of a moveable head that is able to adapt to the different dimensions of adaptor rings of the different client satellites. The actual locking and rigidifying action is made by hooks that move towards the adaptor ring, both from the inside and the outside. When completely closed, the adaptor ring and the client is locked down in the correct position. [26]

Figure 2-13 The latching mechanism in stowed position (left) and operational (right)

Figure 2-14 The latching mechanism in position with respect to ConeXpress and the Client The German company Kaiser Threde is currently also working on a suggestion for the latching mechanism.

The function and size of the latching mechanism is quite important when looking at the docking procedure. The size of the latching mechanism will determine how exact the control of the docking will need to be in order to end up at the right position. Smaller latching mechanism and latching area gives the need for more exact control. However, since the decisions regarding the latching mechanism is not determined at this point, the latching mechanism will not be taken into account any further in this Master’s Thesis Report.

(28)

3 Investigation of Alternative Concepts for Docking

The objective of this thesis report is to investigate the possibility and the advantages of an alternative docking approach for the ConeXpress project. The reasons for doing this investigation are to find a simpler docking process than with the Capture Tool, get less demands on the control system of ConeXpress by using a more robust sensor system and to decrease mass and volume if possible.

It is believed that the docking can be made simpler without the high accuracy and the mechanical demands of the Capture Tool, by using an optical sensing system for detection of the nozzle of the client satellite. As the nozzle enters the crater the sensing system will detect this. The sensing system should be able to detect lateral movements as well as the depth and the orientation. The control system of ConeXpress should then be able to use this information to adjust its position if it is wrong.

This type of automatic docking has never been performed in GEO before. Due to the distance, the communication from earth will have a delay. If something goes wrong in the critical final moment of the docking approach the ConeXpress will not be able to respond in time to the commands. Because of this the docking approach needs to be completely automatic, using an algorithm programmed in the onboard computer.

In this chapter we will discuss the alternative concept by looking at difficulties that may occur, special criteria that has to be implemented and discuss possible concepts. Finally a trade-off will be made between the concepts and one will be chosen.

3.1 The alternative docking concept

In the alternative docking approach the intention is to use a simpler and more robust sensor than the Capture Tool in order to decrease the demands on the agility and the accuracy of the control system. The idea is to detect the presence and shape of the nozzle of the client satellite as it enters into the crater of the ConeXpress. How this can be done will be investigated in the following chapters. The theory so far is to use light diodes and receivers placed in the crater in a certain configuration with several layers. As the nozzle enters some beams will be cut off.

The position and number of beams cut will tell us the position of the nozzle. Either an absolute or a relative sensing concept can be used. A relative measuring concept measures the nozzle relative to itself and then calculates the position relative to the crater (the Capture Tool operates in this way). An absolute measuring concept measures the exact position in the crater.

When the nozzle is detected, its position can be calculated with a certain probability depending on the resolution of the detection sensors. When knowing the position of the nozzle, the position of the adaptor ring of the client can be established. We need to know this since it is on the adaptor ring that the final locking of the latching mechanism will take place.

The camera system will still be available but not the Capture Tool. We will go from a mechanical solution to a strictly control solution depending on the proper configuration, distribution and number of sensors which are to be placed in the crater of ConeXpress.

The following phases can be determined:

• Camera controlled (5- 0.5m)*

• Optical sensor controlled (0.6m – 0.04m)*

• Latching mechanism controlled (0.04m – 0.01m)*

( )*: Distances from ConeXpress (upper) separation plane (= adaptor ring) to Client satellite separation plane (=

adaptor ring)

(29)

a

c b

Mounting plate

Adaptor ring

The alternative approach in steps:

1) Coming out of the Rendezvous phase at a distance of 5 meters from the client satellite, ConeXpress is placed in the right position with respect to the Apogee Nozzle of the client satellite.

2) Illumination system is switched on to provide enough light for the camera system.

3) Client satellite attitude/position is acquired using the cameras.

4) Latching mechanism is deployed.

5) Camera system is used to keep ConeXpress in the right position (by means of cold gas thrusters/reaction wheels) as it is closing in towards the nozzle of the client from 5 to 0.5 meters.

6) The Clients AOCS is in stand-by mode.

7) Nozzle entering the crater of ConeXpress at a distance of approximately 0.6 meters.

Camera system still operational.

8) Optical sensor system detects the nozzle. Lateral position is achieved and corrections are made, if required. Partly operator/on-board control.

9) Lower layers of the optical sensor system detect the nozzle, again lateral position is achieved. Combining this with detection in the first layer gives the orientation/attitude of the nozzle. Deviation from crater centre line is calculated and necessary corrections made. Control completely switched to on-board processing.

10) When the nozzle and adaptor ring are in the right position at 0.04 meters distance, proximity sensors gives a signal and the latching mechanism is activated, after which locking is automatically performed.

11) Latching mechanism locks on Client satellite adaptor ring to rigidify the mated configuration.

3.1.1 AKE nozzle dimensions and mounting issues

The ConeXpress mission needs to be compatible with a wide range of client satellites, all with different kinds of apogee kick engine nozzles. The optical sensing system has to work for all of them, thus an investigation of the different kinds of nozzles needed to be done. [17] The measurements of a standard nozzle are shown in Figure 3-1 below.

Figure 3-1 The AKE Nozzle

(30)

The dimensions for the minimum and maximum nozzle and the straw man can be found in Table 3-1. [17]

Nozzle

Nozzle exit diameter

[mm]

(a)

Estimated distance [mm]

between adaptor ring of client and nozzle exit

(b)

Length [mm]

between mounting plate and exit

nozzle (c)

Minimum values 244 254 (250 with

margin)

504

Straw man (R-4D-11-300) 365 300 676

Maximum values 365 449 (500 with

margin)

676 Table 3-1 Nozzle dimensions

The distance between the adaptor ring (separation plane) and the nozzle exit of the client satellite (measurement b) determines how far the nozzle will reach into the crater of ConeXpress. This distance can be different for different satellites even though they are using the same apogee engine, depending on the mounting of the engine; i.e. how far into the satellite structure the engine is mounted. This is why the measurement can only be estimated.

How big will the impact of the nozzle size be compared to the size of ConeXpress? To answer this, a study of the construction drawings of ConeXpress needed to be done. The result can be seen in Figure 3-2, where the maximum and minimum nozzle sizes compared to the ConeXpress are shown to scale.

Figure 3-2 Dimensions of ConeXpress and the min and max nozzle

As can be seen in the figure, the diameter of the nozzles does not vary that much and will not pose any difficulties. However it becomes clear that the range of lengths is far wider than what was first thought. Due to this the depth sensing has to be given special attention because it has to be able to detect and adapt to this wide range of nozzle lengths.

Min.

nozzle

Max.

nozzle

(31)

Another unexpected issue was found when studying the construction drawings: How can the optical sensing equipment be placed? The drawings showed that the area where it is possible to place the detection devices over the whole circumference of the crater begins at a distance of 379 mm down in the crater, on the inner side of the inner structure, see Figure 3-2. Above this structure there is no firm base for anything to be mounted. The housing cylinders for the push-away springs are the only thing that is present here. They belong to the Ariane adaptor cone structure and cannot be changed. There are twelve springs and they are placed with a spacing of 30 degrees at approximately 555 mm radial distance from the centre line. In between the housings the three arms for the latching system will be placed. As Figure 3-2 shows, this means that the possible detection area will be too far down for having a safe detection (or any detection at all) of the nozzles.

Possible solutions can be

• placing some form of mounting device or bracket on the deck in between the spring housings. The mounting device could be

- point wise brackets, with an even spacing to get the appropriate coverage, on which sensor pairs can be placed.

- a circular structure that goes around the entire rim of the crater, in front of the spring housings, to get a complete coverage. This may however prove to obstruct the working area of the latching arms when they will be folded up upon approach of the client.

• using the latching arms (that already will be there) for mounting sensors upon. This may not give enough coverage of the crater since there will be only three arms.

This issue will however be left un-dealt with so far.

3.1.2 Defining the no-escape zone

During the entire approach there is no contact between ConeXpress and the Client satellite, not until the latching mechanism locks them together. How will we now when to operate the latching mechanism? The client satellite must be in the exact right position, without any contact being made. We invent an area called the no-escape zone. It is the final part of the docking approach, where the adaptor ring of the client satellite is in correct position with respect to the latching mechanism. In this area the client satellite has been captured by the latching mechanism, the adaptor ring is “inside” the latching arms and the soft docking is completed, without any physical contact being made between ConeXpress and the client satellite. The client is captured and cannot get away anymore, i.e. there is “no escape”.

Figure 3-3 Schematic drawing of the no-escape zone

The size of the no-escape zone depends first of all on the range of the latching mechanism, but also on the size of the adaptor ring of the client satellite and the approach speed. But

ConeXpress Client satellite

Latching Mechanism Area in which the

Client cannot escape anymore

(32)

instead of using a very big area like in Figure 3-3, we can use a smaller area positioned in the center of the latch mechanism and separation plane of ConeXpress. In the current latching mechanism the area where the adaptor ring of the client is in the correct place, and where the latching devices are in range to capture it, is 40 mm high and 10 mm lateral. [13] This is completely transferable to the no-escape zone in the center, which will be a small cylinder of height 40 mm and radius of 10 mm situated in the centre of the separation plane of ConeXpress.

Figure 3-4 Real interpretation of the no-escape zone

The latch point is situated in the centre of the adaptor ring of the client, i.e. also in the center of the nozzle. The adaptor ring is in the correct position, with respect to the latching mechanism, when the latch point is somewhere inside the no-escape zone. In the alternative docking approach we detect the position of the nozzle. This means that we need to relate the dimensions and position of the nozzle to the position of the latch point (in the center of the adaptor ring). For instance, if we know that the nozzle extrudes 254 mm from the adaptor ring, we know that we have to detect the rim of the nozzle 254 mm down in the crater to know that the center of the adaptor ring is in correct position with respect to the latching mechanism. We have a margin of 40 mm (the height of the no-escape zone). We also need to detect the lateral position, the position compared to the central axis. Here we only have a margin of 10 mm (the radius of the no-escape zone). This is shown in Figure 3-5.

The rim of the nozzle is a good feature to detect.

This is the first point of the client satellite that the optical sensing system detects and it can easily be followed down into the crater using layers of sensor with appropriate spacing. It is also possible to detect the lateral movement of the nozzle by using sensors in the plane in a certain configuration and appropriate spacing to be able to keep the nozzle centered.

Although, we also have to take into account the orientation aspect, i.e. if ConeXpress approaches the client satellite with a tilt, an angular deviation. The adaptor ring of the client cannot move more than 40 mm in the height direction at the circumference of the adaptor ring, where the latching mechanism will operate. This means that the actual height deviation in the no-escape zone in the center of the adaptor ring is a lot smaller than 40 mm. The actual range depends on the diameter of the adaptor ring; the larger the adaptor ring the smaller the angle can be to still be within the 40 mm range by the latching mechanism. Thus the parallelity between the separation planes should differ no more than 1 degree over the radius of ConeXpress.

[18]

No-escape zone – connected to ConeXpress Adaptor ring of Client Latch point

Figure 3-5 No-escape zone for

Crater

(33)

If it is found that the no-escape zone is too small in order to manage the docking with the envisaged low resolution (= large spacing) of the optical sensor system, an investigation needs to be done regarding the possibility to change the current latching mechanism in order to meet the requirement of the alternative docking concept.

3.1.3 Resolution

The issue of resolution is important for the alternative docking approach. When using an optical sensor system the spacing between the light sources and the detection sensors will determine the resolution. Since the no-escape zone is a larger area in the alternative docking approach than in the Capture Tool approach the need for a very fine control system is no longer necessary. If we want a rougher control system there is no need for a fine sensing system. In fact a rougher sensing system with larger resolution elements is preferable because in combination with ending in a large no-escape zone this will make the approach simpler.

However, having a large sensor resolution leads to the need of having a larger detection area in the crater. For control purposes there need to be a margin of at least three resolution elements. When the nozzle is detected by an element the control system starts to react in order to bring the nozzle to the correct position. But the change will not be noticed instantaneously and in the meantime the nozzle will have had time to pass into the next element. By then the thrusters have reacted to the command from the control system and the nozzle will start to move back to the right position. But it still has some speed and the trajectory will pass into the third resolution element before it goes back. If these resolution elements are larger due to a rougher sensor system, the sensing area will also become larger.

3.1.4 Control issues

Comparing the alternative approach with the Capture Tool approach from the point of view of the control system, the most striking difference will be the ending in a wider zone. It will be in the dimension of tens of millimeters, instead of the tolerances of the Capture Tool approach which are on sub-millimeter levels. The contraction point for the latching mechanism will be the same in both approaches, but for the Capture Tool the transfer to the contraction point of the latching mechanism will be made by retraction of the bi-stem from a point where the first capture already has been made. This first capture point will actually be a risk factor in the Capture Tool approach, due to the enhancing demands on the accuracy in orientation and position, whereas for the alternative docking approach the first capture will be in the wide area of the no-escape zone. From any random point in the no-escape zone the latching mechanism will be able to lock on the adaptor ring of the client until it is firmly captured in the contraction point of the latching mechanism. See Figure 3-6.

It is also shown that the docking camera needs to be operational for a longer time period, since the optical sensing system will not be in operation until the nozzle already enters the crater of the ConeXpress.

(34)

Figure 3-6 Comparison between the two approaches from control point of view

(Horizontal axis can be considered as the time line for the docking approach and the vertical as a measurement for the accuracy needed)

3.1.5 Comparison and advantages

We have so far discussed the alternative docking approach in general without going into details. In order to summarize a bit here will follow a small comparison and the advantages that can be seen so far with the alternative approach.

The alternative approach needs a completely new design of the optical sensor system and of the onboard software processing algorithms compared to the Capture Tool approach. In the latter a partly qualified design and hardware are used, although changes have to be applied to fit the ConeXpress requirements. Also the measuring range of the optical sensor system in the alternative approach needs to be large enough to be able to handle all available sorts and sizes of nozzles. This may be quite challenging with respect to the mounting issues.

The optical sensing system has a higher reliability since the sensors can detect the nozzle even if some of them are not working. Also the redundancy is higher. With the Capture Tool approach there is no alternative. There is only one single device and if it does not work due to some issue or malfunction, the whole mission fails.

The camera system needs to be working for a longer range than in the Capture Tool approach. The sensor system will not be initialized until the nozzle enters the crater, so the camera system needs to be functioning as close as approximately 0.5 meters minimum. This may cause difficulties in case the camera lenses cannot operate at such close distances.

Comparing this with the Capture Tool approach, the bi-stem with the Capture Tool will be deployed and take over control from the camera system at a distance of 1 meter.

In the Capture Tool approach the first capture takes place in the throat of the nozzle.

During the retraction of the bi-stem, when the two satellites are connected together, the possibility exist that unwanted forces, torques and momentums are built up due to the flexibility of the bi-stem. By omitting the bi-stem in the alternative approach the mechanical risks is no longer present. The docking approach will be safer. Also the risk of malfunction in the deployment and retraction mechanism is out, as well as complex electrical cabling in the bi-stem. The early contact between the two satellites may also influence the attitude control

(35)

systems and the risk is that the two AOCS go into a “fight” regarding who is in charge. All these issues are not present in the alternative approach.

The latching mechanism in the alternative approach probably needs a higher complexity since it will perform the one and only capture. In the Capture Tool approach the client satellite is already captured when the latching mechanism are to be operated. It is more accurate and thus the latching mechanism does not need as big a range as in the alternative approach.

In the alternative approach the capturing will take place directly on the adaptor ring of the client satellite. The status of the adaptor ring can easily be seen with the camera system and we can determine the form, fit and function. Comparing this with the Capture Tool approach where the capturing takes place inside the apogee kick engine nozzle, of which the condition after many years in space cannot be fully known, the latter seems more hazardous.

For both approaches possible obstruction of structures sticking out of the client satellite aside from the nozzle, such as heat shields or other objects, must be taken into account.

With the alternative approach it is believed that the mass of the optical sensor system will be less than the mass of the Capture Tool and bi-stem together and mass will most likely be saved or at least exchanged to keep the mass budget neutral.

In the alternative approach the final no-escape zone is larger than in the Capture Tool approach. This reduces the demands on the control system of ConeXpress and the resolution of the sensors can be less accurate.

As can be seen there are several advantages achieved by using the alternative docking approach. It will also be an interesting experiment since this kind of automatic docking never has been investigated or performed in GEO orbit before. Finally, a small summary of the advantages found:

• Mass is most likely saved

• Control can be less accurate

• No mechanical devices to cause malfunction

• Easier latching directly on the adaptor ring

• Less complex electrical cabling needed

• No unwanted force, torque or momentum build-ups

• No single point failure

• Higher reliability

• Simpler redundancy

(36)

3.2 Possible Concepts

The main function of the optical sensing system is to detect and measure the position, orientation and depth of the nozzle when closing-in, as part of the docking operation. In this chapter six measuring concepts will be discussed and the best will be selected after a trade- off.

• Concept 1: Detection bands

• Concept 2: Taking middle lines

• Concept 3: Shading in three directions

• Concept 4: Vertical sensing strips

• Concept 5: Neptec 3D Laser Camera System

• Concept 6: Inversion of the Capture Tool 3.2.1 Concept 1 – Detection bands

This is the very first idea for the alternative docking approach. It has been the base for the discussions and concerns so far. The idea is to use a strip of light sources on one side of the crater and a strip of light sensitive detecting sensors on the other side, much like the diodes and receivers used for automatic door opening systems. There is one light source/LED-diode dedicated for each receiving sensor. The LED’s create a band of light beams across the crater.

As the nozzle enters, some of these beams will be cut, and no signal will be received on the detection strip. By getting information on which beams are cut, we can use this to calculate the position of the nozzle. Depending on the resolution of the LED’s (the spacing of them) we get the position of the nozzle with certain accuracy.

Instead of the LED’s we can also imagine using a laser with a slit and a collector lens to make a parallel bundle of light in one plane. We then have a continuous beam of light without resolution elements and by using a continuous sensor strip as well, we get a better resolution.

The LED’s (or laser device) and the sensor strip can be placed in various configurations. The first idea was to have three bands rotated about 120 degrees. The mounting could then be in connection to the latching arms, but this idea was discarded because of the overlapping of transmitting and receiving areas. The design is instead to put the sensing strips so that the measurement bands cross at straight angles, Figure 3-7.

Max nozzle centered radius: 182mm

Min nozzle with offset of 30mm to the right

radius: 122mm

Figure 3-7 Concept 1: Detection Bands

Min nozzle (R=122mm) with offset of 30mm to the right.

Max nozzle (R=182mm) centred.

(37)

The mounting of the LED’s and sensor strips might however be a bit problematic. See Chapter 3.1.1, where the mounting issues have been discussed.

Calculations

The detections that are received by the sensors need to be transformed into data that is relevant for control purposes. In order to do this we first have to calculate the size needed for the detection bands. The bands should be able to detect all kinds of nozzles, see Table 3-1.

The radial range needs to be covering both the minimum (122 mm) and the maximum nozzle (182 mm). This gives a range of the measuring band of 60mm (182-122 = 60).

The docking camera can get the nozzle centered with an accuracy of 10 mm in radial direction [12] so by adding an extra 10 mm to each side of the range we will get complete coverage. Now we have a range of 80 mm.

For the control there also has to be extra margin. The margin needs to be at least 3 resolution elements in each direction, as discussed in Chapter 3.1.3. Assuming a resolution of 10 mm there has to be an additional 30 mm to each side. In total this give a measuring range of 140 mm, see the gray areas in Figure 3-7 (to scale).

The measuring range can be broken up into a configuration of 14 pairs of LED’s and receiving sensors, each with a spacing of 10 mm. The measuring area will need to start at 82 mm from the centre of the crater (minimum nozzle 122 minus margin of 40 mm) and ends at 222 mm (maximum nozzle 182 plus margin of 40 mm). If a maximum nozzle is measured, the read-out of the sensor is subtracted by 182 mm to get the actual centre line position of the nozzle. For the minimum nozzle 122 mm is subtracted in the on-board control. The same applies for the y-direction. Of course the algorithms will be changed depending on the actual dimensions of the client nozzle.

For depth measuring, multiple layers of this lateral concept can be applied. How many layers that is necessary depends on with what accuracy the depth needs to be known. All nozzle lengths need to be detected. Due to this we need to have layers all the way down to 500 mm counting from the separation plane of ConeXpress. With a required resolution of 10 mm we need a new layer every 10 mm, in total 50 layers to cover the whole depth. This may not be feasible because of the very high amount of sensors needed. In fact, comparing with the current approach of the Capture Tool, where no measurement principle of the depth detection is included, basically all that is needed for doing the same job is three layers; one that detects first entrance, one that detects the final depth of the shortest nozzle and one that detects the final depth of the longest nozzle.

Stray light

Since this concept is based on light sources and light sensitive devices, the issue of stray light needs to be accounted for. Stray light may origin from

• surrounding light sources

• the sun shining into the crater

• the artificial illumination system belonging to the camera

• reflections from the light sources, the sun or the illumination system

The problem can be solved by using pulsed light, either with a certain frequency or a certain array of light being on and off, or by having a certain frequency of the light and a filter that shuts out all other light frequencies so only the correct one is detected.