design and manufacture of structure and hull for large auv1247567/...page 6 of 53 abstract the main...

of 53

DEGREE PROJECT IN MECHANICAL ENGINEERING, SECOND CYCLE, 30 CREDITS STOCKHOLM, SWEDEN 2018

Design and manufacture of

structure and hull for large AUV

M.S Thesis in Naval Architecture

BILLY VENDEL OLOFSSON

KTH ROYAL INSTITUTE OF TECHNOLOGY SCHOOL OF ENGINEERING SCIENCES Stockholm, Sweden 2018

of 53

Design and manufacture of

structure and hull for large

AUV

B ILLY VENDEL OLOFSSON

M a s t e r ’ s T h e s i s i n N a v a l A r c h i t e c t u r e ( 3 0 E C T S c r e d i t s ) a t t h e S c h o o l o f

E n g i n e e r i n g S c i e n c e s R o y a l I n s t i t u t e o f T e c h n o l o g y

S u p e r v i s o r : S e b a s t i a n T h u n é E x a m i n e r : J a k o b K u t t e n k e u l e r

TRITA-SCI-GRU 2018:048

Royal Institute of Technology School of Engineering Sciences

KTH SCI

Teknikringen 8 SE-100 44 Stockholm, Sweden

URL: www.kth.se/sci

Date: 2018-04-02 [email protected]

mailto:[email protected]

of 53

Sammanfattning Masterarbetet bestod av att designa och bygga en struktur och skrov till en stor AUV som KTH ämnar använda inom forskning. Målet med ubåten, även kallad Maribot LoLo (long endurance long range), är att agera som en experimentiell plattform inom en bred aktivitet av forskning, här alla fyra förmågor inom SMaRC (autonomi, uthållighet, kommunikation och iaktagelse) kan bli testade, förbättrade och demonstrerade.

Strukturen designades och och ska byggas i aluminiumprofiler från Bosch Rexroth med specialtillverkade vinkeljärn som är tillverkade i ett korritionsbeständigt stål. När designen av strukturen var fastställd togs lastfallen och dess tillhörande elementarfall fram. Inga mekaniska beräkningar har genomförts då strukturen, när monterad, planeras bli testad med ovanstående nämda lastfall som pålagd last. Skrovet är byggt i glasfiber med ett laminatupplägg bestående av 0/90 och ±45 mattor. Den totala fiberhalten per kvadratmeter blev 4800 g/m2. Skrovet var tillverkat genom att vakuminjisera glasfiber i en hon-form, metod och resultat finns beskrivet i appendix 1. Den totala kostnaden för tillverkning av struktur och skrov uppskattades till 129 000 SEK, här de största utgifterna var tillverkning av ”han-form” och personal kostnad i samband med tillverkning av skrov. Tillverkning av ett nytt glasfiber skrov uppskattas kosta 16 000 kronor, om det i framtiden tillverkas av studenter på KTH. Motsvarande kostnad för att tillverka skrovet i kolfiber uppskattas till 60 000 kronor. Innan AUVn är sjöduglig krävs följande arbete; montera ihop ramen, kapa skrovet till rätt dimensioner, designa klart och tillverka sammanlänkningarna mellan skrov och ram, samt testa strukturen så den uppfyller de ställda kraven. Eventuella optimeringsarbeten som kan genomföras är; tillämpa laminat teori och beräkna skrovets mekaniska egenskaper. Tillsammans med dessa egenskaper och implementering av balk-teori, uppskappta skrovets utböjning när det evakuerar vattnet. Mekaniska beräkningar på struktur kan eventuellt leda till en optimering av ramens vikt, här kan troligen mindre dimensioner på profiler användass.

of 53

Abstract The main goal with the master thesis was to design and manufacture a frame and hull to a large AUV which KTH will use in research. The general aim for the Maribot LoLo (Long Endurance, Long Range AUV) is to act as an experimental platform and tool for a wide range of research activities, where all four capabilities within the SMaRC area (Autonomy, Perception, Endurance and Communication) can be tested, improved and demonstrated. The frame was designed and shall be built in aluminum profiles from Bosch Rexroth with custom made brackets, manufactured in a corrosion resistant steal. When the frame was designed, the associated load case and beam elementary case was implemented. No mechanical calculations have been executed since the frame is planned to be tested against the above mentioned load case. The hull was manufactured in glass fiber with a laminate built from 0/90 and ±45 fabrics. The laminates total weight per square meter was 4800 g/m2. The hull was manufactured by performing vacuum infusion onto a female-mould. The method used and the result is presented in appendix 1. The total cost for manufacturing the frame and hull was about 129 000 SEK, where the largest expenses were due to the manufacturing cost of the male-mould, and the labor cost when manufacturing the hull. Manufacturing a new glass fiber hull would in the future cost approximately 16 000 SEK, given it’s manufactured by students at KTH and in the same female-mould. Manufacturing the hull in carbon fiber would cost approximately 60 000 SEK. The future work required before the AUV is seaworthy are; assemble the frame, cut the hull to its correct dimensions, design and manufacture the attachment solutions between the hull and frame, and to test the frame against the applied loads. Future optimization work would be applying laminate theory and extract the mechanical properties of the hull. Combine the laminate properties with beam theory in order to estimate the deflection of the laminate when evacuating the water. Mechanical calculations on the frame could as well lead to optimizing the weight of the frame, where hopefully smaller dimensions on the profiles could be used.

of 53

Acknowledgement The work was performed at KTH and Vaxholm Komposit, supported by the examiner Jacob Kuttenkeuler, supervisor Sebastian Thuné and Ulf Brändström CEO of Vaxholm Komposit. I would like to start by thanking Ulf Brändström and Filip at Vaxholm Komposit for giving me practical experience in different lamination techniques. They also had a key-role when the hull was manufactured. I would like to continue by thanking Sebastian Thuné for his enormous support as a mentor through the whole project. He also played a vital role when designing the CAD designs of the mould and hull. He also put down a tremendous amount of work when updating the entire LoLo assembly and creating the bracket drafts. Thank you Jakob Kuttenkeuler for your great guidance and support, and the vital inputs you had through the whole project. Finally I would like to thank Stefan Hallström for contributing with both theoretical knowledge and guidance when the load path of the structure and hull was set.

of 53

Contents

1. Introduction ........................................................................................................................... 10

2. Background ........................................................................................................................... 11

3. Structural design ................................................................................................................... 13

3.1 Load cases ....................................................................................................................................... 14

3.2 Manufacture of frame ..................................................................................................................... 17

3.2.1 Brackets .................................................................................................................................... 18

3.2.1.1 Profiles .................................................................................................................................. 18

3.3 Result ................................................................................................................................................ 19

4. Evacuation of water .............................................................................................................. 20

4.1. Aft-section hatch ............................................................................................................................. 20

4.2. Mid-section hatch ............................................................................................................................ 21

4.3. Flexible hull ...................................................................................................................................... 21

4.4. Scale test .......................................................................................................................................... 22

5. Hull ......................................................................................................................................... 23

5.1 Load case ......................................................................................................................................... 23

5.2 Creation of mould ............................................................................................................................ 25

5.2.1 Creation of male-mould .......................................................................................................... 25

5.2.2 Creation of female-mould ...................................................................................................... 27

5.3 Test pieces ....................................................................................................................................... 28

6. Attachment between structure and hull .............................................................................. 29

7. Result ..................................................................................................................................... 31

8. Conclusion ............................................................................................................................ 33

9. Future work ........................................................................................................................... 34

10. References ............................................................................................................................. 35

11. Appendix ................................................................................................................................ 36

Appendix 1: Mould preparations and method for vacuum infusion ........................................ 36

Introduction .................................................................................................................................................. 36

Step by step guide ...................................................................................................................................... 36

Method .......................................................................................................................................................... 40

Mould preparation ................................................................................................................................... 40

Applying gelcoat ...................................................................................................................................... 40

Packing fabrics and mesh...................................................................................................................... 41

Spiral tubes, in- and outlets ................................................................................................................... 43

Vacuum bag and tubes .......................................................................................................................... 45

of 53

Infusion and chemicals ........................................................................................................................... 47

Result ............................................................................................................................................................ 48

References ................................................................................................................................................... 51

Appendix 2. Technical data sheet DION 9500-501 ..................................................................... 52

of 53

1. Introduction Minimize emissions and increase energy efficiency has always played a key-role when developing competitive products which fulfills the common goal in decreasing the environmental impact. The maritime industry is expanding now more than ever and invests huge amount of resources in research and development of new solutions in order to fulfill the goal. The Royal Institute of Technology’s (KTH) goal is to be in the forefront and act as a role model when it comes to the field of research and never less within the maritime industry. KTH is therefore investing resources every year within the naval architecture department, which lead to the creation of Swedish Maritime Robotics Centre (SMaRC). SMaRC is a national cross-disciplinary industrial research Centre for maritime robotics. The main task is to perform research on, and demonstrate solutions that can contribute to the transition to autonomous intelligent underwater systems. The Centre will focus on four disciplines- autonomy, endurance, perception and communication- with the general aim to develop next-generations maritime robotics for ocean production, safeguarding society and environmental sensing. One of the projects within SMaRC is to develop a demonstrator for

long-range and long endurance (LoLo) missions [1].

The general aim for the Maribot LoLo (Long Endurance, Long Range AUV) is to act as an experimental platform and tool for a wide range of research activities, where all four capabilities within the SMaRC area (Autonomy, Perception, Endurance and Communication) can be tested, improved and demonstrated. There are currently 12 different PhD programs that require an experimental platform in order to test the theories, which is the main reason LoLo is developed. Figure 1 illustrates the core purpose of SMaRC.

Figure 1. The main goal of SMaRC

[1].

The LoLo will be designed based on requirements from primarily the SMaRC partners, but also with the aim to serve and support for further research. The LoLo will be based on open architecture, which will promote easy implementation of novel methods and techniques, thereby make it a versatile platform for testing of new innovations.

The main goal of the thesis was then to design and manufacture the hull of the LoLo and its load bearing structure. It was also desired to design the attachments between structure and hull, as well as designing a water evacuation system for the events when the AUV is recovered at sea from a mothership. The content of this report will describe the method when designing the hull, structure, attachments and evacuation of water, as well the manufacturing process of these.

of 53

2. Background The core purpose with the LoLo AUV is to act as a demonstrator which is not optimized for one specific mission, but several and complete different ones. An example of such a mission could be to transport and drop lotus [2] buoys which measures and collects data at the bottom of the ocean. The AUV is required to achieve the following requirements in order to become a flexible platform;

Have enough capacity for operating autonomously for several days

Work in a range of different environments

Run different missions

Have an operating depth down to 1000m.

The AUV has a “flooded hull”, meaning that it will consist of multiple smaller pressure vessels with a composite hull around these. This will increase modularity that is needed for supporting the many varied missions. The electronics and other equipment are sealed in waterproof vessels. If the AUV would be completely dry (no water inside) it would require much thicker hull and higher requirements when sealed. With this comes the risk of leakage, which could lead to the extreme case in losing the AUV. Another drawback would be that the AUV becomes heavier and by so making the handling more difficult. The desires to have an AUV that is fast and simple while maintaining a high flexibility and modularity were the main reasons why a wet AUV was chosen.

The AUV is composed to three sections, front-, mid-, and aft-section. The front-section where the payload is located. The definition of payload is the research equipment used when performing the missions (lotus buoy, sonar etc.) The batteries and VBS (variable buoyancy system) are placed in the mid-sections together with servo boxes for the control surfaces. The main electronics, antennas and propulsion are located in the aft-section. Figure 2 illustrates the AUV and the different sections in an exploded view.

Figure 2. Exploded view of the AUV

of 53

The purpose is to recover the AUV from a mothership once it has performed the desired mission/s. The AUV shall then find its way back to the mothership once performed the missions. It is beneficial that the AUV is fully drained from water when recovered (by crane from mothership). The added water load would otherwise stress the complete vehicle, especially when exposed to higher accelerations during the recovery procedure (a possible scenario when higher accelerations would occur is when recovered in sea with high waves.) A passive self-draining system would be a possible solution for draining the AUV. An example could be a flexible hull which deflects and releasing the water once lifted above the water line. It is crucial that the AUV can handle the severe loads from the water if the draining system fails to evacuate the water. It is then very important that the load bearing structure, hull and the junctions between these two can manage the applied loads. The assumed wet weight of the vehicle is assumed to 1500kg, compared to a 600 kg dry weight. The load bearing structure is assumed to have a safety factor of 3, which would be similar to 45kN (total mass of the AUV, exposed to 3g). The wet weight is the assumed mass when the AUV is recovered with water inside.

of 53

3. Structural design The most essential part of the AUV is the structure, since it takes all the loads from water, equipment and hull. It is then crucial that the load bearing structure is stiff enough and can manage the resulting torques, buckling, bending and shear stresses which arise from the loads. The loads are transferred through the joints which are keeping the load bearing structure together. The joints are created by assembling profiles with steal brackets, mounted together with M8 screws going through the profiles and creating screw joints. The load is then successfully transferred between the different members of the load bearing structure. Figure 3a and 3b illustrates how the brackets were mounted to the profiles. In this chapter the following points are presented and discussed; different load cases, the manufacture of the frame and the final result.

Figure 3a. Design of the frame

Figure 3b. Zoomed view, Illustrating how the brackets were mounted.

of 53

3.1 Load cases There were different load cases depending on how the hull and equipment was mounted to the structure. In this project mainly three different water evacuation solutions were investigated. In all cases the equipment was assumed to be mounted in both lower and middle longitudinal profiles. The evacuation of water was then assumed to either be from a hatch in the aft-section, hatch in the mid-section or from a flexible hull. The flexible hull solution was chosen and only its load case is presented in the report. The main reason why the hatch solutions were rejected was due to the slow draining procedure, which would stress the load bearing structure even more when recovered. More information about the different water evacuation solutions are described in chapter 4, Evacuation of water. The hull is assumed to be attached to the structure at enough points so that a distributed load from the water can be assumed. The equipment is assumed to be mounted on the load bearing structure and act as point loads. The load case for the flexible hull is described in Figure 2 where the blue distributed loads (Qi) are from the water and the point loads (Fj) are from the equipment and where it was assumed to be mounted. The water mass and volume was then simplified to boxes, which gave arise to the simplified distributed loads in all sections. The black boxes in Figure 4 represent the assumed water mass and volume for each section.

Figure 4. Load case in the longitudinal view

The distributed load and points load are estimated with the following equations:

𝑄(𝑖) =𝑚(𝑖)𝐻2𝑂∗𝑎

𝑙(𝑖) (1)

𝐹(𝑗) =𝑚(𝑗)𝑒𝑞∗𝑎

𝑛 (2)

Where 𝑚(𝑖)𝐻2𝑂 is the total mass of water for each section, 𝑎 the current acceleration

exposing the AUV, 𝑙(𝑖) the length of each section, 𝑚(𝑗)𝑒𝑞 the total mass of equipment for

each number of point loads, 𝑛 the amount of attachment points and for 𝑖, 𝑗 equal to;

𝑖 = 1,2,3 , 𝑗 = 1,2,3,4. In the front-section the point loads are due to the payload, in the mid-section due to the

batteries and the buoyance system, and in the aft-section for 𝑗 = 3 the electronics,

communication etc. and for 𝑗 = 4 the propulsion. There amount of point loads may change in the future depending on how the real systems will be attached. The lower half of the hull is assumed to take the entire water mass that is contained within the vehicle. In turn, the longitudinal mid profiles that the hull is attached to are assumed to take the distributed load.

of 53

The point loads are assumed to act in both the mid and the lower longitudinal profiles in each section, as presented in Figure 4. The load case was then divided into two different cases, one where the evacuation of water was functioning and one where it was not. In the first case the distributed load in the mid-section is equal to zero, since all water is assumed to be evacuated, while it is not equal to zero in the second case. The dimensioning aspect is then towards the second scenario, since this is regarded as the most extreme case. The real case would probably be a combination between these two cases. The frame is described in 3 dimensions with connections linked to each other and therefore making it too complex to be solved in 3 dimensions. The problem was therefore decomposed to several 2D cases. The first 2D case was in the longitudinal view as described in Figure 4 with the assumed load path. Figure 5 presents the 2D case in the longitudinal view with the assumed models and the different and applied boundary conditions.

Figure 5. 2D case and models in the longitudinal view

The boundary condition for each section was changed and laid ground for two different underdetermined systems for each section. The purpose was to convert the clamped system (overdetermined) into two similar but underdetermined systems with different boundary conditions. The two different models for each section are compared against each other and the model with the most conservative solution is the dimensioning one. The truth is probably somewhere in-between the two solutions, why the most conservative model is regarded the safest one. Given the models above the following points are the most critical ones, which in the future would require further investigation (Calculate and compare the resulting stress in each joint due to the applied loads, for each model and section):

of 53

Section 1: B, C and D are critical points due to the discontinuity in the structure. The

torque and shear forces have to be checked and dimensioned for in this case.

Section 2: Symmetry in the mid-section makes this construction very similar to a space

truss, which means that the contribution from bending moments are assumed to be very

small, if not equal to zero. However, due to the combined loads (distributed load and

point loads in the mid profiles); further investigations in point B and E would be very

interesting.

Section 3: B, C and D are critical points due to the discontinuity in the structure. The

torque and shear forces have to be checked and dimensioned for in this case.

The second 2D case was the mid-section in the cross section view. The construction is as mentioned very similar to a space truss. Due to this similarity and the symmetry the contributions from the torque are assumed to be very small. The joints were then assumed to act like torque free links [3]. The advantage with this simplification is that the mechanical calculation of the structure becomes much easier and can be solved with simple equilibrium equations. The brackets are assumed to be very stiff and successfully transfer the loads from the longitudinal profiles to the cross section. The brackets are mounted on the profiles with M8 screws going through the profiles, assembled with the other bracket by a bolt and by so creating a screw joint. Figure 6 present the cross-section view and the external loads. The equations (3)-(5) presents the external forces in the cross-section, followed by equation (6) which is the sum of the all the external forces in the cross-section view.

Figure 6. Load case in the cross-section view

of 53

Equation (3) presents the external force when lifting the AUV:

𝐹1 =𝑚𝑡𝑜𝑡∗𝑎

2 (3)

Where 𝑚𝑡𝑜𝑡 is the total mass of the AUV and 𝑎 the acceleration. The AUV is assumed to be lifted in two points (one in each end of the mid-section), why the total mass in equation (3) is divided by 2. Equation (4) presents the external force from the water.

𝐹2 =𝑚𝐻2𝑂∗𝑎

4 (4)

Where 𝑚𝐻2𝑂 is the total mass of water contained within the mid-section. The distributed load from the water mass is assumed to be transferred into 4 points (2 on each end of the mid-section, see Figure 4), why the total mass in equation (4) was divided by 4. Equation (5) presents the external force due to the equipment.

𝐹3 =𝑚𝑒𝑞∗𝑎

8 (5)

Where 𝑚𝐸𝑄 is the total mass of the equipment. The point loads are assumed to be transferred

into 8 points (4 on each side of the mid-section, see Figure 4), why the total mass of the equipment in equation (5) was divided by 8. Following relationship was received when assuming equilibrium.

𝐹1 = 2𝐹2 + 4𝐹3 (6)

The top beam has to be checked against the combined buckling and torque load when the AUV is recovered (The AUV is assumed to be lifted in eye-bolts mounted on the top-beam). The size of the buckling load is depending on what angle (relatively to the longitudinal center line of the AUV) the ropes haves lifting the AUV. Solving the combined buckling and torque requires non-linear methods, which was too complex to solve within this project. The frame shall then be tested once assembled, in order to prove that it can withstand the applied loads.

3.2 Manufacture of frame The first iteration of the frame was building a mid-section. The purpose was to get a feeling of the real size and what equipment would fit inside it. The aim was also to test the frame so it could manage the applied loads with a safety margin of 3, i.e. the total weight of LoLo exposed to 3g, which could be a possible scenario when recovered in sea by crane. All the parts have been manufactured by hand, i.e. cutting the profiles, brackets and drilling holes. The frame was despite the handmade parts successfully assembled as illustrated in Figure 7.

of 53

Figure 7. First version of the mid-section frame (mock-up).

The load case was developed after the creation of the frame, and it was by then obvious that performing tests on it would not make a proper representation of the real scenario. The tests have then been postponed until the next version of the frame has been assembled, where all sections are included (front, mid and aft-sections). The loads from aft and front-section will affect the whole structure and would be a more accurate representation with higher external validity than the first version, since it will present how the whole system will behaves.

3.2.1 Brackets

As mentioned the prior brackets were handmade and designed in a CAD software. They were cut from a large aluminum plate into smaller pieces with a circular saw. The final design was then created with the same circular saw. The holes were drilled with a drill press. The material used was aluminum 6082 T6 with a thickness of 5mm. This aluminum alloy has a very high yield stress and hardness, which required a special blade in order to cut the pieces without destroying the machine or the material. It was in total 10 brackets that were manufactured for the frame and it took approximately two days to prepare. Figure 5 presents the resulting frame and brackets. The updated frame has 44 brackets with 10 different configurations, where approximately a third of them are bent brackets. Manufacturing these brackets by hand would require much more time compared to the previous ones. The bent brackets would also be much more difficult to create. Because of these reasons and a better tolerance was required, the work was therefore done by professionals. The brackets and holes were then water cut and bent to required dimensions. The new versions of the brackets are created in 4 mm stainless steel, since the aluminum 6082 T6 is too brittle and would crack once bent.

3.2.1.1 Profiles

The profiles for the first frame version were measured by hand and later cut with a semi-automatic saw. The tolerance achieved was ± 0.5mm. In the profiles cross-section the holes were threaded to make it possible to assemble the profiles with the brackets. In the updated version the majority of the profiles were ordered precut. The remaining ones will be collected from the previous frame and then measured and cut to its right dimensions. The reason

of 53

behind this is to minimize the material waste and cost. The profile dimensions were set to 45x45mm, this in order to ensure that the structure can withstand the severe loads. As mentioned, the top beam is exposed to extreme buckling loads and was therefore chosen to have twice the profile height, in order to prevent it from buckling. However, further analyzes and tests are required in order to get proof if the load bearing structure can withstand the loads or not.

3.3 Result The new design of the frame has not yet been assembled since all parts have not arrived. The new design has been created in CAD and is shown in Figure 8. The front section has no beams in the bottom, which depending on missions makes it possible to drop payload from a hatch in the bottom. A towing point in the front-section was added in order to make it possible to tow the AUV behind a boat, which as well can be seen in Figure 8. The mid-section is very similar to the previous version, but with the slightly difference that all longitudinal profiles have been made longer. The reason was to increase the continuity in the structure. The first 40 cm of the hull in the front section has no change in the curvature, it was therefore possible to extend the top beam this far into the front-section. Due to this change fewer joints were required and the structural strength was drastically increased. In the aft-section no top profile was used. The great advantage with this design was the increased accessibly to components placed in this section (electronics, motors, etc.). The lower profiles make it possible to place and attach the equipment on top of them. The aft-section structure was also extended, which minimized the needs of secondary structure for attaching control surfaces and propulsion in.

Figure 8. Latest version of the frame

of 53

4. Evacuation of water It is beneficial that the water is evacuated efficiently once the AUV is recovered from the water. The loads will drastically increase in the events the AUV is recovered and exposed for vast changes in the acceleration, which will increase the overall stress in the structure. Stress concentrations will most certainly occur in points where the hull is constrained to deflection (in those areas the water is pushing against the hull which is attached to the frame and constrained to deflect). There is still uncertainly what effect the stress concentrations will have on the hull and frame. The attachment points would in this case need to become more robust. There were three different solutions evaluated: flexible hull and two different hatch solutions, one in the mid-section or one in the aft-section. The three different solutions are presented and discussed in the following sub-chapters. The three concepts are working as passive systems, i.e. there is no remote control when the system evacuates the water, since the water with its force is opening the hatches or by flexing the hull. The main advantage with a passive system is, its decreases the systems complexity, no electrical wiring is required and it’s more cost effective and takes less space. However, a passive system can’t evacuate the water when the operator requires it to. An example of a passive system is a hatch mounted with hinges and a spring which extends when exposed to loads higher than the spring constant. When the water is successfully drained, the spring is contracted to its original position. The concept chosen for this project was in the end the flexible hull.

4.1. Aft-section hatch The principle with a hatch in the aft-section was to open and release water once lifted above the waterline. The AUV would be lifted with the tail pitching down and, with the help from gravity, transfer the entire volume of water to the aft-section where it’s evacuated. Lifting the AUV with a pitch angle would successfully drain the whole AUV and minimize the risk for creating accumulations of water in the front- and aft-section. The downsides with having a hatch in the aft-section was that all water has to travel through the whole vehicle before reaching the hatch, which requires more time before all water has been evacuated. Shifting the mass center of the whole AUV (which would be the case in this solution) would stress the structure much more than necessary, which would require a more robust load bearing structure that would weigh more. It also puts requirements on the recovery system to recover the AUV with a pitch angle. This will induce more shear into the load bearing structure, which is not designed for high shear loads! Due to these downsides, this solution was not investigated further. Figure 9 presents the hatch solution in the aft-section.

Figure 9. Hatch solution in the aft-section

of 53

4.2. Mid-section hatch The second concept was to place the hatch in the mid-section. Water would then not travel through the whole vehicle but would be drained evenly through the mid-section. However, there will still be stresses transferred from the hull to the frame since the water is not evacuated at once. There would be a slightly risk that accumulation of water would arise in the front- and aft-section, since the hatch is placed in the mid-section and the lower cross-section beams prevents the last remaining of water to be transported to the hatch. A solution would be to create small draining holes in each section where this risk would arise, which would slowly drain the remaining water (the load from this water is not severe since the volume of the remaining water is assumed to be small.) Figure 10 presents the mid-section hatch concept, illustrated in the cross-section view.

Figure 10. Hatch solution in the mid-section

4.3. Flexible hull

The principle with the flexible hull was as the vehicle is lifted from the water the increased water weight would load the lower hull section to an extent where the composite hull would be designed to flex, creating an opening for the water to escape, as shown in Figure 11. In order to achieve this the hull was only attached to the mid profiles in the mid-section, cut and separated from the front-and aft-section and finally cut in the symmetry line in the cross-sections view and by so creating two symmetrical halves. The result is a flexible hull that drains the water when its pressure forces the halves to separate once lifted above the waterline. In the cases when the acceleration is drastically changing and by so creating huge forces affecting the water and equipment, the water is released much faster by forcing the hull to flex even more. The load from the water is decreased and won’t affect the load bearing structure as much, since the hull is free to deflect.

In the worst case if the system fails to evacuate the water, the load case is same as in the second load case as described in Chapter 4.2. The AUV is however dimensioned for the worst case scenario, as mentioned in the previous chapter. Since the water inside the AUV will adjust to the waterline, it would in the best case be fully drained before it leaves the waterline (assumed there is no vast changes in the acceleration). This would probably be the case in the other concepts as well. But in the worst case when the AUV is raptured from the water, the water drainage is not fast enough which increases the load and stress on the structure.

of 53

Figure 11. Flexible hull solution in the mid-section

4.4. Scale test Tests have been performed on water bottles in order to examine how the different systems would behave. The aim was to estimate and compare the draining time for the different solutions. When an air inlet was added in the top of the pet-bottles the draining time was almost decreased to half. The main cause was the change in the outflow, going from turbulent to steady flow due to faster pressure equalization. The air-inlet also made it possible to fill the bottle when submerged, which would not be possible if accumulated air pockets would be created inside. The conclusion was that the flexible hull was the most efficient one when it came to drain and fill water. Adding an air inlet made the system about twice as efficient. The optimal size of the air inlet was not investigated but would be of great importance for further work. Figure 12 presents when performing tests on the flexible hull concept.

Figure 12. Test on the flexible hull concept

of 53

5. Hull It is important to manufacture a hull that can demonstrate the desired feasibilities as mentioned in the previous chapter. In this chapter the load case for the hull is presented, the procedure when creating the moulds (male- and female-mould), and when manufacturing the test pieces which lead to the laminate layup for the first version of the hull. How the hull was manufactured is presented in Appendix 1.

5.1 Load case From the previous chapter the principle of a flexible hull was mentioned and by assuming a flexible hull it was possible to neglect the water load affecting the mid-sections lower profiles. However, in the case when the evacuation system fails to drain the water, the load from the water is transferred to the longitudinal mid profiles. The load case for the whole hull in the cross section view is presented in Figure 13.

Figure 13. Load case of the hull in the cross-section view

𝑄1 is the static pressure where the height of the AUV and acceleration is the major variables, equation (7) presents the static pressure affecting the side of the AUV;

𝑄1 = 𝜌𝐻2𝑂 ∗ 𝑎 ∗ ℎ (7)

Where 𝜌𝐻2𝑂 is the water density, 𝑎 the acceleration and ℎ the height of the AUV. 𝑄2 is the distributed load from the water which affects the “floor” of the AUV and is described in equation (8);

𝑄2 = 2 ∗𝜌𝐻2𝑂∗𝐴∗𝑏

𝐵∗ 𝑎 (8)

Where 𝐴 = ℎ ∗ 𝐵/2 is the half cross-section area, b the width of the beam latter used in the equations (9) - (11), and B the width of the AUV, which is illustrated in Figure 13. Given this load case the hull was assumed to act as two beams. With this simplification the load case was split into two elementary cases, where classical beam theory could be implemented. Figure 14 presents the load case for one half of the lower hull and the applied elementary cases for beams.

of 53

Figure 14. Load case and elementary cases.

The deflection and rotation are estimated from elementary cases [5]. Due to the distributed

load 𝑄1 in the first elementary case, the free end (the point 𝜉 = 0) will deflect and rotate, leading to an offset angle and position for the second elementary case. The rotation for the first elementary case is calculated according to equation (9);

𝜃(𝜉) =𝑄1∗(

ℎ

2)

2

12𝐸𝐼(𝜉4 − 4𝜉3 + 3) (9)

Where 𝑄1 is the static pressure calculated from (5), 𝐸 the elastic modulus for the laminate, 𝐼

the moment of inertia, and 𝜉 the dimensionless constant determine the position in the beam,

which is defined between 0 ≤ 𝜉 ≤ 1. The displacement for the first elementary case is calculated according to equation (10);

𝛿(𝜉) =𝑄1∗(

ℎ

2)

3

60𝐸𝐼(−𝜉5 + 5𝜉4 − 15𝜉 + 11) (10)

The results from equations (9) and (10) are added to the second elementary case where they

act as an offset in the point 𝜉 = 1. Given the offsets from elementary case 1, the deflection for elementary case 2 is calculated according to equation (11);

𝛿(𝜉) =𝑄2∗(

𝐵

2)

3

24𝐸𝐼(𝜉4 − 4𝜉 + 3) 11)

of 53

Where 𝑄2 is the distributed load from equation (8), 𝐵 the width of the AUV (divided by two

since half the hull is calculated), 𝜉 defined within the interval 0 ≤ 𝜉 ≤ 1, and b the width of

the beam (required once estimating the moment of inertia 𝐼). In order to solve the elementary

cases the laminate properties are required (𝐸-modulus and the moment of inertia 𝐼), which at the time of writing this report haven’t been calculated for this laminate, but remains to be done in the future work. The load cases for the front- and aft-sections have been neglected in this thesis since they are overdetermined systems where the beam theory is not a valid simplification. Solving these cases would require plate theory. The loads are however smaller in these sections compared to the mid-section since there is less accumulations of water. Dimension for the mid-section loads would therefore be the most conservative approach.

5.2 Creation of mould There are two different methods for creating composites in a mould, by a male-mould or by a female-mould. A male-mould is basically a piece of block (PVC, foam, etc.…) which has been milled to the same surface as the composite one is attempted to create. It is very important to prepare the surface of the male-mould so it receives a very good “finish”. There would otherwise be a great risk that the laminate get lodged to the mould and may ruin it once separated. Once the surface is good enough, the infusion is performed by laying the fabrics on top of the milled and prepared surface. When a smooth finish on the outside is required, regardless using gelcoat or plain fabrics, one has to grind the surface after the performed infusion. If gelcoat is used, one has to apply this after the executed infusion.

The female-mould can be created in two different ways, by directly mill it out from a block or by creating a single-skin female mould from a male-mould. In the later method the male-mould is prepared as above, a “form-coat” applied (To give the female-mould a very nice inside surface) and perform one infusion on it. The advantage with a single-skin laminate except its weight is the flexibility, allowing the infused laminate to release easier compared to a milled female-mould which is rigid. The inside surface of the female-mould is then what the outside surface of the composites will get. The huge advantage with a female-mould (instead of using a male-mould) is that the preoperational and post work are only needed to be done once, despite the amount of performed infusions. It’s however very important to have a good finish on the surface; otherwise the main advantage with using a female-mould is lost. The hull was decided to be created in a single-skin laminated female-mould. Since the method of creating the hull from a male mould is very time consuming and required much pre- and post-work, therefore this method was ruled out. Creating the female-mould by directly mill it would be the fastest one, but would also result in the heaviest mould. Since the handling was of great importance, the female-mould was created in a single-skin laminate. A male-mould was still required since the female-mould was created as a single-skin laminate, which was created from the male mould. The creations of these two moulds are described in the following two sub-chapters.

5.2.1 Creation of male-mould

The male mould was designed in CAD. The design included half spherical points creating cut lines on the surface. The advantage with cut lines is to easy have them as reference points when cutting the different hull sections to the right dimensions. There were also added a release angle in the top of the design which made it easier to release the female-mould once created. The female-mould receives the same release angle as the male-mould, which in the next step makes it easier to release the laminates once infused in the female-mould. The CAD design was sent to Macromould [8] who manufactured the real male-mould by milling styrofoam to the desired surface. Epoxy based content was added to the male-mould and milled to the final surface and dimension. A tolerance of ±0.01 mm was achieved by using milled epoxy as final surface. The epoxy is not reacting with other chemicals which are applied during the infusion phases, which would otherwise ruin the male-mould and the attempt in creating a female-mould. The purpose behind the flanges on the male mould was to create space for the infusion materials. The manufacturing cost is presented in chapter 7.

of 53

Figure 15 presents the CAD design of the male-mould and Figure 16 presents the milled male-mould.

Figure 15. CAD design of male-mould

Figure 16. Milled male-mould

of 53

5.2.2 Creation of female-mould

The single skin female-mould was created by applying green mould-coat on the male-mould and hand laminating chopped glass fabrics with vinyl ester based resin. Figure 17 presents when the hand lamination was performed and Figure 18 presents the resulting single-skin female-mould. The creation of the hull and choice of method is presented in Appendix 1.

Figure 17. Creating the female-mould by hand lamination

Figure 18. Inside of the female-mould

of 53

5.3 Test pieces Since no laminate theory was performed, test pieces were required to be created in order to achieve the desired laminate properties of the real hull. The first test pieces were created as simple plates. The purpose with this test was to gain some basic experience in performing vacuum infusion. The first test pieces were too weak, which would make a poor representation of the real hull. New test pieces were then required to be created. The next iteration of test pieces was created in the same shape as the mid-section hull and with different laminate layups. The purpose was to test the flexural righty for the different pieces and by so decide the laminate layup for the real hull. Four different test pieces with different mechanical properties were created, which resulted in a hull with a hybrid laminate between test piece A and B. Following laminate layup was chosen for the hull; [±45, ±45, 0, 90]S. The manufacturing process of the hull is presented in Appendix 1. Figure 19 presents the second iteration of test pieces which shapes were the same as the mid-section hull.

Figure 19. Four different test pieces of the mid-section, created for deciding the hull laminate

of 53

6. Attachment between structure and hull The purpose with the attachment points is to both mount the hull to the frame and transfer the hull loads into the load bearing structure. It is of great importance that these can manage the water load and successfully transfer it to the frame without collapsing. Because of the complex curvature in the nose there were, in a few points, large distances between the hull and the frame. The frame would otherwise need to be cut in smaller pieces and fit to the curvature of the hull, which would result in a much weaker frame. The solution was to mill distance blocks that were mounted on the frame and had the same surface curvature as the hull. Figure 20 presents the distance block mounted on the front-section.

Figure 20. Designed attachment blocks

The blocks are able to freely move on the profiles, allowing them to move against and away from the hull and along the profiles. Creating them with this flexibility makes it easier to align them to the point where the hull is mounted. The slots in the front, makes it possible to adjust the vertical position of the hull. The upper hull can also be adjusted in the longitudinal position due to the milled holes. This flexibility is necessary in order to prevent misalignment in the transition between the front- and mid-section of the hull. Since the distance between the hull and the frame is less in the mid- and aft-section, the attachment points were designed as simple brackets. The brackets can move transversally and longitudinally along the profiles, which makes it possible to exactly align its position with the attachment point of the hull. The first version was to use bent brackets with fixed position for the lower half of the mid-section and vertical adjustment in the upper mid-section and for the whole aft-section. The fixed position of the lower mid-section would act as a reference point for which the rest of the hull would align towards. Figure 21 and 22 presents the first version of the bent brackets used in both the mid- and aft-section. The further design and creation of the blocks and bent brackets belongs to the future work and is not further discussed within this thesis.

of 53

Figure 21. Bracket solution for mid- and aft-section

Figure 22. Bracket solution for mid- and aft-section, exploded view

of 53

7. Result The results achieved within the thesis were:

Design and build the first version of the mid-section frame.

Design the new version of the complete inner structure with new brackets.

Set up the load path for the structure and hull.

Performed tests on pet-bottles and compared different water evacuation solutions.

Design attachment blocks and solution for towing.

Manufacture the hull to the AUV.

By achieving these points, a complete solution for frame, hull, integration between hull and frame and a water evacuation system were presented. The price for manufacturing the hull was about 46 200 SEK where approximately 66% was labor expenses, the remaining was the material cost for creating two hull halves. The material used was two different glass reinforced fabrics. The laminate layup and material properties are presented in appendix 1. The cost for manufacturing the milled male-mould (done by Macromould) was 55234 SEK. The cost for all the aluminum profiles and connections were approximately 12 000 SEK. The new brackets costed almost 13 000 SEK. Other expenses as transportations etc. was about 2 600 SEK. The total cost for manufacturing the frame and hull was about 129 000 SEK. Figure 23 presents the 3D designed AUV and Figure 24 the manufactured hull.

Figure 23. 3D designed AUV with frame, hull and attachment points.

of 53

Figure 24. The manufactured hull

of 53

8. Conclusion The tasks that have been done consist of high quality where it was a mix between research, theoretical- and physical work and implementations. The frame needs to be validated through tests so it can withstand the applied loads from water and equipment exposed to 3g. Due to the discontinue load bearing structure, the most critical points are the joints in the front- and aft-section. If the frame fails the tests, the first solution is to reinforce these points. If this does not help, one must perform mechanical calculations and redesign the load bearing structure. No structural analyze of the attachment blocks in the front-section have been done and is required if the milled versions fails the test. This can probably be done with relative high accuracy in a FEM-software (ANSYS or Abacus). The flexural tests that were conducted on the manufactured test pieces could have been performed with higher accuracy by using professional test rigs. The results from the tests gave poor mechanical properties of the tested laminates. However, the new test pieces with same shape as the hull, presented in Figure 19, was of great help when deciding the laminate layup used in the hull. There would in future be very interesting to estimate the mechanical properties for the hull. The total price was as mentioned 129 000 SEK. The major two contributions to this were the labor cost for manufacturing the hull and the cost for manufacturing the male-mould. If the work was performed at the university rather than at a professional company, with a poor quality of mould and no hired labor, the cost would be drastically reduced, consisting of only material cost and a cheap mould. The reason why the expensive procedure was chosen, was the lacking experience in performing infusion and time limitations. The material cost for the hull would probably become higher since the lack of experience would result in more waste of materials and failed infusions. A cheaper mould would require much more preoperational work for achieving a decent result, which would not be possible due to the time limit. Creating new hull pars would in the future only cost the use of materials, since the experience is within the university and the work would most likely be performed by students. It would cost about 16 000 SEK for creating a new hull in glass fiber, compared to 60 000 SEK if created in carbon fiber.

of 53

9. Future work There are two levels of future work; what is required to be done before operating the AUV, and future work for optimization. The tasks required in order to operate the AUV is the following:

Cut the hull to the right dimensions. Exact measurements and cuts are required,

which is challenging due to the complex curvatures in the hull. A possible solution

would be to 3D print a block that acts as a guide rail when measuring and cutting the

hull.

Assemble the frame when the ordered parts have arrived. The ordered brackets will

arrive in the beginning of May, making it possible to assemble the whole frame.

The attachment points have to be 3D printed and assembled with the hull and frame.

The aim is to mill the blocks in a more robust material once the 3D printed parts are

assembled with success.

Attachment solutions for the mid- and aft-section have to be developed and

manufactured.

The Frame has to be tested in order to prove it can manage the external loads. This

can be done by fixating the frame to the floor in a way that would represent a

distributed load. The frame is then lifted in the same way the real scenario would be

once recovered at sea by crane (lifted in eyebolts mounted on the top beam). The aim

is to reach the required load of 45kN, which would be the whole AUV when lifted

wet and exposed for 3g.

The evacuation system is not a crucial task for being able to operate the AUV. It is

however important that it work properly. Testing the system and check its function is

a task that can be tested once the AUV is assembled.

The possible optimizing tasks are then;

Mechanical calculations. The mechanical calculations are not crucial (given that the

load bearing structure pass the test) but just a way to optimize the structure.

However, if the structure fails the test, is it crucial to check the weak points and

perform mechanical calculations.

Perform laminate theory and calculate the laminate properties would be of great

interest when optimizing the hull. There would be possible to calculate the required

bending stiffness for a desired deflection (how much the laminate shall deflect once

evacuating the water for a given distributed load Q).

Calculate laminate for the three sections and make layup for each according to the

study of loads each section is introduced to.

of 53

10. References

[1] Smarc.se [2] Bottom lander (lotus buoys), KTH cluster for underwater technology,

https://www.ave.kth.se/se/avd/naval/cute/projects/bottom-lander-for-long-term-underwater-sensing-lotus-1.560613, accessed Marsh 2018, last reviewed 2015-10-30

[3] Fackverk, inledande ingenjörskurs, Umeå universitet,

http://www8.tfe.umu.se/courses/byggteknik/kurser/Inledande-Ingkurs/Boken/3Fackverk.pdf, accessed December 2017.

[4] Hans Lundh, Grundläggande hållfasthetslära KTH, printed 2011, chapters 2,3,4 and 7. [5] Institutionen för hållfasthetslära KTH, handbok och formelsamling i hållfastighetslära, editor

Bengt Sundström April 1999 KTH, Stockholm, Sweden, chapter 6, 17 and 32 [6] Dan Zenkert, an introduction to Sandwich Structures, 2:nd edition December 2005,

Stockholm, Sweden, chapters 2, 4 and 5 [7] Konstruktion och skjuvning, Skjuvningsboken,

http://pl.fredrika.se/Konstruktion_Filer/Skjuvningsbok/Skjuvning_bok/Skjuvningsboken.htm, accessed December 2017.

[8] https://macromould.se/

https://www.ave.kth.se/se/avd/naval/cute/projects/bottom-lander-for-long-term-underwater-sensing-lotus-1.560613

https://www.ave.kth.se/se/avd/naval/cute/projects/bottom-lander-for-long-term-underwater-sensing-lotus-1.560613

http://www8.tfe.umu.se/courses/byggteknik/kurser/Inledande-Ingkurs/Boken/3Fackverk.pdf

http://www8.tfe.umu.se/courses/byggteknik/kurser/Inledande-Ingkurs/Boken/3Fackverk.pdf

http://pl.fredrika.se/Konstruktion_Filer/Skjuvningsbok/Skjuvning_bok/Skjuvningsboken.htm

http://pl.fredrika.se/Konstruktion_Filer/Skjuvningsbok/Skjuvning_bok/Skjuvningsboken.htm

https://macromould.se/

of 53

11. Appendix

Appendix 1: Mould preparations and method for vacuum

infusion

Introduction The aim of this appendix is to describe the method for making a complete infusion of the LoLo-hull. The appendix will include required material and consumables as well discussing two different solutions for resin outlets. The appendix is divided into two major chapters, a step to step guide for preparing and performing the infusion, and a chapter explaining the theory and function behind the different materials and equipment that were used.

Step by step guide The aim of the step to step guide is to explain the different tasks required to make the infusion of the hull repeatable. Once the female-mould has been created or cleaned from previous performed infusion, the following procedures are required:

1. Apply new layers of wax. If the mould is used only a few times, more layers of wax is

required. For this mould, use Norpol wax w70 and a soft cotton cloth for

application. It is usually a description on how to apply the wax on its canister.

2. Apply release agent. For this mould, use Marbocote 220 or Norslipp 9860. The

release agent is applied with a soft cotton cloth or a sponge. The release agent is only

applied once, since the mould by now is well smeared.

3. Mix the gelcoat GN K072 HA and the accelerator Peroxide 5, the weight percent and

procedure for mixing is described in the datasheet of the gelcoat. Apply the gelcoat

immediately after mixing it with the accelerator.

4. Brush the gelcoat onto the mould. When a thicker layer of gelcoat is desired, brush at

regular intervals new layers on top of the old ones. Not before the last layer of

gelcoat is “sticky”.

5. Measure and cut the fabrics.

6. Apply the fabrics once the gelcoat has become sticky. For fixating the different layers

of fabrics, use an adhesive spray (not necessary between the first of fabrics layer,

since the gelcoat works great as an adhesive).

7. In order to avoid bad alignment between mould and fabrics, cut streaks in the fabrics

and twist the fabrics so they fit the curvature/mould. Cut all superfluous fabrics until

only a few centimeters of overlaps are achieved. Figure 28 presents the result after a

successful packing of fabrics.

8. Add peel-ply once all fabrics have been packed (not necessary if infuply2 is used).

9. Add infusion mesh. Infuply2 or regular infusion mesh can be used. Use adhesive

spray for fixating the infuply2 to the fabrics. Use tacky tape if regular infusion mesh

is used. In order to align it with the fabrics, cut the infusion mesh in the same way as

described in point 7, Figure 29 presents the result after packing the infusion mesh.

10. Add inlet spiral tubes. The inlet spiral tubes are folded and placed in the middle,

longitudinally to the mould. Use tacky tape when attaching the spiral tube to the

infusion mesh, see Figure 29 for the result.

11. The outlet spiral tubes are placed on the moulds flanges. Leave some extra space for

attaching the enclosing vacuum bag, as described in point 29. Use tacky tape when

attaching the spiral tubes, see Figure 29 for the result.

12. Wrap small pieces of infusion mesh around the ends of all spiral tubes (in order to

prevent sharp edges piercing the vacuum bag).

of 53

13. Add t-tubes to the outlet spiral tubes in order to connect the vacuum lines in a later

process, see Figure 30.

14. Add “plastic pucks” to the inlet spiral tubes, in order to connect the resin inlet in a

later process, see Figure 25 and 29.

15. Cut small pieces of fabrics for creating the brakes.

16. Fold the brakes around the spiral outlet tubes and place the remaining between the

infusion mesh and peel-ply (if no peel-ply is used, place it directly on top of the

fabrics), see figure 25 and 30.

17. Measure and cut the VMS2.

18. Check for any imperfections in the VMS2.

19. Cut a few holes in the first layer of plastic (not on the side where the perforated

plastic membrane is placed). The distant between each hole shall be approximately 1

meter. The size of the hole shall be approximately 1 square centimeter.

20. Seal the ends of the VMS2 with tacky tape.

21. Attach the VMS2 on the infusion mesh, see Figure 30. Use tacky tape when attaching

the VMS2.

22. Put plastic pucks over the created holes in the VMS2. Wrap tacky tape around the

plastic puck in order to attach it on top of the VMS2.

23. Connect vacuum lines to the t-tubs that were added to the spiral outlet tubes in point

13, see Figure 25.

24. Connect the other ends of the vacuum lines (point 23) to the catch pots, see Figure

25. If there are limited amount of holes in the catch pots, join several vacuum lines

with t-tubes in order to decrease the amount of lines going to the catch pots.

Remember to leave one hole in each catch pot for connecting it with the vacuum

pump.

25. Connect the VMS2s vacuum lines to the catch pots, (do not connect the other end to

the VMS2). Join several vacuum lines with t-tubes if there is not enough holes in the

catch pots, see Figure 25.

26. Place tacky tape on the outside of the outlet spiral tubes, see Figure 25 and 29. Do

not remove the protecting sealing on the tacky tape before applying the vacuum bag.

27. Cut the vacuum bag. It shall be at least one meter extra bagging around the whole

mould. This in order to ensure there is enough bag.

28. Start to attach the bag on one side and work the way around the mould. Create folds

on regular distances along the mould and place tacky tape in-between. This is

necessary in order to prevent the bag from becoming stretched due to the height

differences in the mould, once vacuum is introduced to the system. The folds shall be

approximately 10-15 centimeter and be placed where there are great changes in

curvature, for example in the ends of the spiral tubes and where its curvatures, see

Figure 31.

29. Connect vacuum lines between the catch pots and the vacuum pump.

30. Slowly introduce vacuum to the system. The air is drained successfully once the

system is completely sealed. If necessary, adjust materials, equipment and vacuum

bag before full vacuum in the system is achieved.

31. Optional: Pierce a few holes in the middle of the laminate and place pressure

instruments on top and seal with tacky tape. This in order to check against local

fluctuation in pressure. See figure 31.

32. Turn off the pump when full vacuum is achieved and check against leakage.

33. Maintain the vacuum in the system for 10-20 minutes and check if the pressure

changes (the pump remains off).

34. Connect the vacuum line for the VMS2 by piercing through the vacuum bag. Seal

with tacky tape.

of 53

35. Turn on the pump and reach full vacuum.

36. Remove the pressure indicators and seal the small holes.

37. Shut the secondary outlet tubes.

38. Pour up the right quantity of resin. The required amount of resin can be estimated

according to the fabrics weight fraction. A rule of thumb is to use approximately 70-

100% of the total weight of fabrics.

39. Check the resin viscosity at the right temperature. (Specified in the datasheet.)

40. If necessary, add some weight percent of styrene in order to lower the viscosity.

41. Prepare the resin inlet tubes (cutting the ends so the get sharp edges and add tacky

tape to the ends which shall be connected to the spiral inlet tubes).

42. Place the other end (with no tacky tape on it) in the bucket where the resin is poured

into. Secure the ends so they cannot move during the infusion phase.

43. Mix the chemicals (inhibitor and hardener) with the resin once the right viscosity is

achieved. It is recommended to have a minimum gel time of 1 hour.

44. Fill the resin inlet bucket with the mixed resin.

45. Pierce the resin inlets with the inlet tubes prepared in 41. Quickly seal with the tacky

tape that were rolled on the inlet tube (which as well was prepared in point 41).

46. The resin will enter the system and start to impregnate the fabrics. During this

process, check against any leakages or air being introduced to the system

47. Open the secondary vacuum line, when the resin has reached the VMS2. (Those that

were shut in point 38).

48. Close the secondary vacuum lines once the resin has fully wet all breaks.

49. Pour water into the remaining resin once its texture becomes similar to gelatin (to

prevent the resin to ignite due to the high temperature during the curing phase).

50. Turn of the pump 30-60 minutes after the resin is fully cured. Maintain the vacuum

in the system until the next day.

51. Clean the laminate from all consumables. Do not release the hull from the mould.

The vinyl ester is shrinking during the curing phase, which in extreme case may

deform the hull. Keep the hull in the mould at least 24 hours after the post cure

process is done, which may decrease the overall deformations in the hull.

52. Post cure the laminate for 24 hours at 40-50 degrees Celsius. Use a normal heating

fan when post curing.

53. Release the laminate from the mould.

54. Clean the mould.

Using two independent vacuum lines creates a redundant system, which in case of one failing prevents a total failure. The infusion can be performed successfully with just one vacuum line system, but is not recommended when larger pieces are being manufactured. There is in general always one failsafe system included when larger pieces is manufactured. If only one system is used, the steps for the other one can be skipped. Figure 25 illustrates the overall setup when performing an infusion on the hull.

of 53

Figure 25. Overall setup when perfoming an infusion on the hull

of 53

Method Following subchapters will describe how the preoperational work was done (preparing mould, packing all materials required), which chemicals that was used and discusses the choice of method.

Mould preparation

A necessary step before using the mould is to apply wax and release agent, this in order to prevent the laminate to lodge into the mould and making it almost impossible to separate them without destroying the mould in the process. There are a variety of different waxes and release agents, the ones which may be used for this mould are “Norpol wax w70” and Marbocote 220. The recommendation from Norpol is to apply 4-5 layers of wax on moulds being used for the first time. The more times the mould is being used, the less layers of wax need to be applied, since the mould becomes well smeared. It is recommended to apply a release agent before every infusion. The amount of layers required is usually described on respective release agent’s container. Norpol have several different release agents, the most common one is “Norslipp 9860” which can be used instead of Marbocote 220. Marbocote 220 was used when the first version of the hull was created, since the company where the hull was created only had this release agent. Figure 26 present some different release agents and wax that may be used

Figure 26. Wax and release agents

Applying gelcoat

The gelcoat need to be mixed with the accelerator Peroxide 5 before being applied. The correct mixture between gelcoat and accelerator are specified in the gelcoats datasheet. The gelcoat requires a few hours of pre-cure when applied to the mould, before imposing the fabrics. It is okay to start packing (put on the fabrics) when the gelcoat is slightly sticky but not wet. A good test is to press a finger against the gelcoat, if it leaves a fingerprint rather than wetting the finger, it’s okay to start the lamination process. The gelcoat has a cure time between two to three days if not mixed with resin (perform infusion). It is very importance to have a long cure time on the gelcoat when larger pieces are being manufactured, since it may take several days before the infusion process is initiated. Another way of solving the risk for premature curing is to apply gelcoat after the laminate is created. The gelcoat and the accelerator used for this hull were GN K072 HA and Peroxide 5. Figure 27 present the gelcoat that was used when creating the first version of the hull.

of 53

Figure 27. Gelcoat

Packing fabrics and mesh

When the gelcoat is ready, the fabrics are imposed on top of it. The fabrics are rolled across the short side of the mould. It is recommended to use an adhesive spray for fixating the fabrics between each layer. The adhesive spray used when fixating the fabrics was “G’ Fix 2”. It is of great importance to have an overlap (approximately 2-3 cm overlap) in the joints between the fabrics, this in order to prevent the fabrics of parting and cause a failure in the laminate! It’s not recommended to have all the overlaps at the same position since it creates vast local thickness in the laminate and increases the risk of local delamination. It was not possible to avoid the joints since the fabrics used in this mould were neither as wide as the mould was long, nor as wide as the mould was wide, therefore it made no difference how the fabrics were rolled. However, it is recommended to roll the fabrics across the mould, since this will make the handling much easier. The longer pieces created if rolling it the other way, will make the packing and handling much more complicated due to the moulds complex curvature. It is impossible to avoid local cuts in the fabrics due to the double curvature in the front and aft section of the mould. This is necessary in order to fit the fabrics against the mould and avoid stretched fabrics. The consequence with stretched fabrics is the risk for creating local accumulations of resin once infused, which would drastically lower the mechanical properties of the laminate! The local cuts are created by cutting streaks in the fabrics where there are superfluous amount of it. The streaks are then overlapped in the same way as described above (2-3 cm overlap), the remaining fabrics are removed. Figure 28 presents how the fabrics were packed for the first version of the hull. The fabrics used for this hull were bidirectional fabrics [0/90] 578 g/m2 and biaxial fabrics [+45/-45] 910 g/m2 (where about 100g/m2 consisted of chopped fabrics). The following laminate layup was used: [+45/-45/+45/-45/0/90]S.

of 53

Figure 28. Packed fabrics

The next step is to apply the external infusion mesh. It is recommended to use the infusion mesh “Infuply2” from AMT Composites [1]. This particular mesh is a combination between a perforated release film and a knitted infusion mesh. The advantage with this specific mesh is that no peel-ply is needed. This saves a lot of time and is easier to fixate to the fabrics [1] (use adhesive spray when fixating it to the fabrics). The purpose with an infusion mesh is to aid the matrix flow and distribute it across the surface of the laminate [2]. A normal infusion mesh is also possible to use, however, it is then necessary to use peel-ply between the fabrics and the infusion mesh. It is despite choice of mesh necessary to cut local streaks in it, in order to be able to align the mesh with the fabrics. A stretched mesh will created accumulations of resin once infused, which as mentioned above, will drastically lower the mechanical properties of the laminate. Figure 29 presents how the mesh is applied and how it was cut in the nose.

of 53

Figure 29. Spiral tubes and mesh

Spiral tubes, in- and outlets

The spiral tubes for distributing the resin (resin in- and outlet) are placed according to Figure 29. The distributing spiral tube for the resin inlet is placed in the middle of the laminate. It is recommended to use “plastic pucks” for connecting the resin inlet tubes with the spiral tubes, which can be seen in Figure 29. It is recommended to have a folded spiral tube when the resin is distributed in both directions, as illustrated in Figure 29. The purpose of using double spiral tubes as inlets is to decrease the risk of vacuum bag coming in-between and prevent the resin to enter (would create local dry spots). The resin outlets are placed on top of the flange, outside of the laminate, as illustrated in Figure 29. Several T-tubes are mounted in the outlet spiral tubes in order to connect these with the vacuum tubes in a later process. It is possible to reverse the layup, i.e. have the inlet in the top and the outlet in the middle. However, it would then not be possible to use the “brakes” as complement. The brakes and the function behind them are described later in the appendix. A conventional resin in- and outlet would look similar to the one illustrated in Figure 29. However, in this project, a vacuum membrane system (VMS2) and brakes were also added, which can be seen in figure 30.

of 53

Figure 30. Installed VMS2 and brakes

VMS2 is a unique product that enables vacuum flow on top of wet impregnated zones throughout the curing process. VSM2 is a multilayer strip comprising of a micro-porous membrane and plastic breather mesh enveloped in a flat tube of vacuum bagging film. The lower layer of film is perforated, which protects the micro porous membrane whilst maintaining high vacuum levels. The micro-porous membrane and breather mesh enable air to be evacuated from the laminate without resin being drawn into the vacuum flow [3]. The main advantages using VMS2 are:

Reduces the risk of dry spots or incomplete wet out, by evacuating air from the

laminate without drawing any resin into the vacuum line.

Reduces the risk of dry spots or incomplete wet out, by enabling vacuum flow on top

of the wet, impregnated areas of the laminate.

Simplifies the infusion network by reducing the amount of vacuum connectors.

Easy to remove from the composite after cured

Reduces resin waste, since there are no resin going to neither the vacuum line nor the

catch pot.

Distributes the resin evenly over the whole laminate, even a few centimeters after the

VMS2, allowing it to be placed a few centimeters in to the laminate (see Figure 30).

However, the VMS2 loses its function if there is any imperfection in it, (holes in the plastic which leads the resin to the vacuum lines) which can ruin the whole laminate. It is very important to check the VMS2 for any imperfections before applying it. The VMS2 is placed a few centimeters in on the mesh, sealed in the ends with tacky tape. A small hole is then pierced through the first layer of plastic in order to create an outlet (The hole shall be approximately 1 square centimeter and is located under the plastic “puck”, which becomes the vacuum line outlet). Another outlet system is used as well in order to create a redundant system, which if one system fails, prevent a total failure. The other system is a conventional outlet through spiral tubes, but with added “brakes”. The function behind the brakes is to lead the resin through itself and not directly through the spiral tubes, which will delay the resin going to the vacuum line and catch pot.

of 53

The main advantages with the brakes are:

Minimizes the risk of resin rush

Decreases the waste of resin (possible to create a smaller batch since less amount

goes to catch pot)

Longer cure time on resin can be used, allowing possible accumulations of air to have

time to be sucked out from the laminate.

System can be turned on longer before the resin reaches the vacuum line and catch

pot.

The “brakes” are small pieces of fabrics rolled around the spiral tubes and goes 3-4 centimeters into the laminate. If peel-ply is used, the brakes are placed between the peel-ply and the mesh. If no peel-ply is used, the brakes are placed under the mesh and on top of the fabrics. Placing them directly on the fabrics will make them a part of the laminate.

Vacuum bag and tubes

After the out-and inlet spiral tubes, VMS2 and brakes have been placed, the vacuum bag and vacuum lines are applied. The VMS2 is the primary outlet and the conventional one which is modified with brakes, is the secondary outlet. The vacuum lines for the secondary outlet are connected between the catch pots and the t-tubes mounted in the spiral tubes placed on the flange. The Vacuum bag is then cut. A larger piece of vacuum bag is required to be cut, since the vast curvatures in the mould and its height differences requires extra bag. It is recommended to have approximately 1 meter of extra bag on each side of the mould (all four sides of the mould). The extra meters of bag is required in order to be ensured that there is enough bag for creating folds. The purpose with the folds is to prevent the vacuum bag to become stretched due to the moulds appearance (vast curvatures and large difference in height). Any kind of height differences might cause a stretch bag, even from the spiral tubes for resin in-and outlet. It is therefore important to create many folds in the bag where there are height differences which may risk stretching the bag. A stretched bag will create pockets of accumulated resin, which will result in drastically decreased laminate properties and increase the laminate weight. It is very important to check the bag and fabrics against any stretched areas before the infusion is performed. The vacuum bag is placed outside all infusion equipment (spiral tubes, fabrics etc.), and sealed with tacky tape on the flange. Figure 31 illustrates how the vacuum bag is enclosed around the entire system and how the folds are created.

of 53

Figure 31. Vacuum bag, folds, vacuum line system and pressure indicators.

The vacuum pump is installed to the catch pots. The vacuum lines for the primary system are temporarily plugged in the ends which later will be attached to the VMS2. The other ends of the primary system’s vacuum lines are connected to the catch pots. The pump is then turned on which will decrease the pressure within the enclosed area. Before reaching vacuum, the bag is adjusted so it fits well on top of the surrounding equipment. The pressure is then decreased to vacuum. It is optional to put a few pressure indicators on top of the vacuum bag, placed in the middle of the laminate. This is done by piercing small holes in the vacuum bag, place pressure indicators over the holes and seal with tacky tape. The purpose with the extra pressure indicators is to check against local pressure fluctuations, which may be the case when larger pieces are manufactured. The system is then tested against leakage by turning off the pump and checked for any inlet of air. The undesired inlet of air shall be fixed before performing the infusion. The system is decreased to complete vacuum once again when all undesired inlets of air have been sealed (the pressure might have been increased due to air being introduced to the system). The pump is once again turned off, if the system succeeds in maintaining the vacuum for at least 10-20 minutes, it’s assumed to be complete sealed. The primary vacuum lines are now connected to the VMS2 by piercing thorough the vacuum bag, inserted in the plastic puck, and sealed with tacky tape. The vacuum lines for the secondary system are temporarily sealed for preventing resin to reach the secondary outlets vacuum lines and to the connected catch pots. The reason is to use the primary system first, and reopen the secondary system once the resin reaches the VMS2.

of 53

Infusion and chemicals

The resin used for the hull was Dion 9500-501 which is a pre-accelerated vinyl ester resin. This resin possesses a unique combination of properties as high tensile elongation, good toughness, low shrinkage, good fatigue resistance and excellent adhesion properties [4]. The downside with this resin is that it ages faster than other resins, which increases the viscosity over time due to its pre-accelerated content. Other disadvantage is the low storage stability, which is only 3 months from date of manufacture, compared to 6 months for other resins [4]. It is necessary to check the viscosity for each batch being created, since it changes over the time. One way of checking the viscosity is to use a Ford viscosity cup (with a 4mm outlet). The viscosity shall be between 300-330 mPa*s at 23 degree Celsius [4]. It is possible to add a few weight percent of styrene if the viscosity is too high. However, addition of styrene will increase the shrinkage in the resin once cured, which will affect the laminate result. The different resin gel times for a given mixture of chemicals are presented in Appendix 2. The different chemicals are weighted before poured in to the resin. It is important to check that the viscosity and temperature are correct before pouring in the hardener into the resin. The resin is poured into 10 liter buckets and made ready by mixing the chemicals together. The infusion phase is initiated once the resin inlet tubes are pierced through the vacuum bag and connected to the spiral tubes in the middle of the laminate. It is during the whole infusion phase important to check against leakages, which have to be fixed before the resin is cured, which otherwise would lower the mechanical properties of the laminate. The closer the resin comes to cure, the higher temperature and lower viscosity it gets. It is therefore very important to check against any leakages or failures during this transition phase.

The impregnation time is approximately 30 minutes, but may change drastically if the viscosity is higher than the specifications, or if air is introduced to the system. It is therefore recommended to have a long cure time when creating larger pieces, especially with large height differences, in order to have a margin if failures occur during the infusion process. The vacuum pump is turned on until the laminate is fully cured, which is possible if VMS2 is used as primary outlet (long difference between gel time and fully wet fabrics would result in enormous waste of resin if a conventional system would be used instead). It is possible to close the resin inlet tubes for a conventional and allowing the pump to be turned on until the resin is fully cured, but only if the laminate is fully wet and there are no accumulations of air inside the system. This will decrease the amount of resin going to the catch pot.

of 53

Result Figure 32 present a successful infusion of the laminate, where the bag, tubes and mesh have been removed. The overlaps from the fabrics can as well be clearly seen in the figure below. Figure 33 presents the outside finish of the laminate with the applied gelcoat. The laminates were post cure for 24 hours in 40-50 degrees Celsius before released from the mould. The post cure process is important in order to get good bondage between the resin and fabrics, which will increase the mechanical properties of the laminate. Table 1 presents the material required and the overall cost for two infusions of the hull.

Figure 32. Result after infusion

Figure 33. Outside finish with applied gelcoat

of 53

Table 1: Required material and total cost for two infusion of the LoLo-hull

Description Unit First infusion of

hull

Second infusion of

hull

Total amount of

units

Price per unit [SEK]

Total cost of units [SEK]

Fabrics

Bidirectional Fabrics, [0/90°], 600

g/m2

[kg] 10,7 9,6 20,3 56,7 1151,0

Biaxial Fabrics, +45°/-45°, 900g/m2

[kg] 27,5 24,1 51,6 56,7 2925,7

Consumables

VMS2 [m] 10 10 20 41,4 828

Vacuum bag [m2] 24 24 48 30,6 1468,8

Structural flow media (Mesh)

[m] 6 6 12 50,4 604,8

Peel-ply [m2] - - - 37,8 0

Tacky-Tape 10mm [Rolls] 2,5 2,5 5 48,6 243

Tacky-tape 30mm [m] 1 1 2 48,6 97,2

T-tube [Piece] 18[1] 3 21 19,8 415,8

Tap [Piece] 8[2] 2 10 20,7 207

Spiral tube 10mm [m] 18 16 34 14,4 489,6

Regular tube 10m [m] 45[3] 30 75 14,4 1080

G’ Fix 2 (adhesion) [Can] 3 3 6 181,8 1090,8

Bucket, 10L [Piece] 4 3 7 19,8 138,6

Dosing bucket, 1L [Piece] 4 3 7 3,8 26,6

Chemicals

Vinyl ester resin 9500-501

[kg] 32[4] 22 54 88,2 4762,8

Hardener, Peroxide 24

[kg] 0,640 0,440 1,080 84,6 91,4

Styrene [kg] 0,640 0,440 1,080 42,3 45,7

Inhibitor, acetylacetone

[kg] 0,048 0,033 0,081 78,3 6,3

Total cost for materials excl. VAT

- - - - - 15673,1

Labor cost excl. VAT

[Hour] 42 16[5] 58 525 30450,0

Total Cost for one hull, excl. VAT

- - - - - 46123,1

Total cost incl. 25% VAT

- - - - - 57653,9

[1]: Uncertain amount of recycled t-tubes and more t-tubes were needed in the first

infusion due to more outlets compared to the second infusion where almost all t-tubs

where recycled and much less outlets were required.

[2]: Less taps were used in the second infusion

[3]: Less regular tubes were required in the second infusion since fewer outlets were used.

[4]: Miscalculation in required resin for the laminate, lead to a waste of 11kg resin in the

first infusion.

[5]: Less amount of hired labor were required in the second infusion since the major part

was done by me. The hours required for this infusion was due to my absence when the

last packing and infusion phase were performed.

The fabrics used for the second infusion were less since remains from the first infusion

were used to the second one.

of 53

As can be seen in Table 1 the labor expenses were approximately 2/3 of the total cost for creating the hull. The main reason was lack of experience of lamination, why extra guidance was required for the first infusion. In the second one, no guidance was required, but due to personal issues the last phase of the packing and infusion was done by Vaxholm Komposit. In the future when new hull parts are being created, the aim is to reduce the labor cost to almost zero, since all the work will be done at KTH by students or by employed personnel.

of 53

References [1] AMT Composites, INFUPLY2,

https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/infusion-mesh/infuply2, accessed February 2018.

[2] AMT Composites, INFUSION MESH, https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/infusion-mesh, accessed February 2018.

[3] AMT Composites, VMS2, https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/vms2/vms2, accessed February 2018

[4] Appendix 2, Technical data sheet DION 9500-501 Vinyl Ester Resin [5] Fibre Glast, Vacuum Infusion guide, https://www.fibreglast.com/product/vacuum-

infusion-Guide, accessed December 2017. [6] Epotex, general fabric properties,

http://www.epotex.se/industri/komposit/komposit, accessed January 2018. [7] Fibre Glast, Gel Coat Troubleshooting Guide,

https://www.fibreglast.com/product/gel-coat-troubleshooting-guide/Learning_Center, accessed January 2018.

[8] Ben-Gurion University of the negev, Materials Science and Engineering introduction, Composites, http://in.bgu.ac.il/engn/mater/Documents/LaboratoryBriefings/4/Materials%20Science%20and%20Engineering%20introduction%20Chapter%2015%20Composites%207th%20ed.pdf, accessed February 2018.

[9] Quest, The effect of post-curing temperature of vinyl ester/e-glass, https://www.quest-global.com/wp-content/uploads/2015/07/Effect-of-Post-curing-Temperature-of-Vinylester-E-Glass-composites.pdf, accessed January 2018.

https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/infusion-mesh/infuply2

https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/infusion-mesh/infuply2

https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/infusion-mesh

https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/infusion-mesh

https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/vms2/vms2

https://www.amtcomposites.co.za/products/vacuum-bag-infusion-consumables/vms2/vms2

https://www.fibreglast.com/product/vacuum-infusion-Guide

https://www.fibreglast.com/product/vacuum-infusion-Guide

http://www.epotex.se/industri/komposit/komposit

https://www.fibreglast.com/product/gel-coat-troubleshooting-guide/Learning_Center

https://www.fibreglast.com/product/gel-coat-troubleshooting-guide/Learning_Center

http://in.bgu.ac.il/engn/mater/Documents/LaboratoryBriefings/4/Materials%20Science%20and%20Engineering%20introduction%20Chapter%2015%20Composites%207th%20ed.pdf



https://www.quest-global.com/wp-content/uploads/2015/07/Effect-of-Post-curing-Temperature-of-Vinylester-E-Glass-composites.pdf

https://www.quest-global.com/wp-content/uploads/2015/07/Effect-of-Post-curing-Temperature-of-Vinylester-E-Glass-composites.pdf

of 53

Appendix 2. Technical data sheet DION 9500-501