RO

NN

a

ARRAA

KBFHPPR

1

rrvdfHTHoHiimFa[

q

0h

Applied Ocean Research 46 (2014) 131–143

Contents lists available at ScienceDirect

Applied Ocean Research

journal homepage: www.elsevier.com/locate/apor

eliability centered modeling for development of deep water Humanccupied Vehicles

. Vedachalam ∗, G.A. Ramadass, M.A. Atmanandational Institute of Ocean Technology, Velachery–Tambaram Main Road, Pallikaranai, Chennai 600100, India

r t i c l e i n f o

rticle history:eceived 3 January 2014eceived in revised form 24 February 2014ccepted 3 March 2014vailable online 3 April 2014

eywords:atteryailure rate

a b s t r a c t

Human Occupied Vehicle operations are required for deep water activities such as high resolutionbathymetry, biological and geological surveys, search activities, salvage operations and engineering sup-port for underwater operations. As this involves direct human presence, the system has to be extremelyreliable. Based on applicable standards, reliability analysis is done on 5 key representative functions withthe assumption that the submersible is utilized for ten deep water missions per year. Analysis is done onthe results obtained to find the influence of the subsystems on the reliability of the overall submersible.Analysis include, influence of battery technologies and reliability centered battery and hydraulic systemconfigurations. Dependence of seal sizes and seal seat surface finish on the leak tight integrity of the per-

uman Occupied Vehicleersonnel sphererobability of failureeliability

sonnel sphere is also discussed. It is found that for submersible housing 75 kWh energy storage batteries,the probability of failure of the hard tank buoyancy ascent function with lead acid batteries configuredfor 300 V terminal voltages and non-redundant hydraulic configuration is 37.74%. The probability of fail-ure can be reduced to 5.24% with lead acid batteries with terminal voltage configured to 120 V and withredundant hydraulic configuration. The results presented shall serve as a model for designers to arriveat the required trade-off between the capital expenditure and the required reliability.

. Introduction

Human Occupied Vehicles (HOVs) are useful tools for theesearches investigating deep sea life and for exploring oceanesources. Even though unmanned vehicles have improved maneu-ering capabilities and excellent vision systems which resembleirect observation, HOVs provide a feel of direct physical presenceor the researches. The successful operation of the first generationOV, Trieste [1,2] at a water depth of 10,906 m in the Marianarench triggered initiatives for the development of more efficientOVs. Further technical developments have greatly expanded theperating range and improved the operational efficiency of theOVs used in scientific research. The second generation HOV which

s centered on the development of a lighter hull for the crew,mprovement of the power supply for propulsion, and establish-

ent of reliable systems, include Alvin [3] of USA, Nautile6000 ofrance, Shinkai6500 of Japan [4], MIR 6000 submersibles of Russiand Jiaolong 7000 of China, developed during the period since 1964

5].

With the experience gained in the development and successfulualification of deep water unmanned Remotely Operable Vehicle

∗ Corresponding author. Tel.: +91 9884869101.E-mail address: veda1973@gmail.com (N. Vedachalam).

141-1187/$ – see front matter © 2014 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.apor.2014.03.001

© 2014 Elsevier Ltd. All rights reserved.

ROSUB 6000 [6–9], the National Institute of Ocean Technology withthe objective of augmenting India’s capability in deep sea research,is planning to develop a HOV capable of operating in deep watersand used for carrying out scientific exploratory activities [8,10].

As HOVs are not electrically powered from the surface, theyrequire self-contained power supply with high energy storage [2]which increases the weight and volume of the submersible, whichin turn limits its operational endurance. When a human is in thesystem he/she must be protected from the hostile deep waterenvironment, and hence, the reliability of the system is of utmostimportance [11]. This calls for man-rating certifications, and is doneby bodies including the American Bureau of Shipping (ABS) [17],Germanischer Lloyd [18], Det Norske Veritas (DNV) [11] and oper-ational guidances by International Marine Contractors Association(IMCA) [19].

This paper reviews the recent technological developments in thesystems required for reliable operation of deep water HOV. Basedon the applicable DNV guidelines RP-203 for qualification proce-dures for new technology [12] reliability analysis are made on thefive identified key representative HOV functions, with the assump-tion that the HOV is utilized for ten deep water dives per year, each

deep water mission clocking 4 h of subsea operation. Studies werecarried on the results obtained to find the subsystems influenceon the overall submersible reliability and it is found that energystorage batteries and hydraulic systems are major contributors in

132 N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143

dia

2

(

(

((

(

(

d

(

penF

tpsausadsfif

c

Table 1Standards followed for FIT estimation of subsystem components.

Components Standards

Batteries, Power electronicconverters, Transformers,Isolators, Motors, Pumps, Lamps,Cameras

MIL and IEEE

CPU, DC-DC Converters, Fuses,Connectors, Ethernet Converters,Input and Output modules forData acquisition cards

FIDES

Subsea sensors and transmitters,Solenoid Valves, O-rings, Gaskets

OREDA, NSWC, ABS

Electronic Pressure cases, ABS, DNV, Germanischer Llyods

comprising of spheres, pumps and compressed air bottles fordescend, traverse, and ascend operations, navigation systems forposition determination and safe vehicle maneuvering, Ship-HOV

Table 2Failure-in-time data used for configuration.

Component FIT (in 109 h)

BatteriesLead acid battery 12 V, 100 Ah 3140Lithium ion battery 12 V, 3.8 Ah 237

Control and network electronicsReal time controller 41Analog input/Output module 12Ethernet switch 70Industrial Computer 527

Vision support systemsCamera 355Light 350

Navigational sensorsPhotonic Inertial Navigation Sensor (INS)with Micro-Electro-Mechanical Systems(MEMS) based INS in redundancy

4650

Doppler Velocity Log (DVL) 300Power conditional systems

DC–DC converter 75DC–AC converter 245

Propulsion systemsThruster motor electronic controller 477Thruster Brush-Less Direct Current (BLDC)motor of 8 kW capacity

455

Hydraulic systemsPump 5400Pressure sensor 60Control valves 341

Fig. 1. Reliability bath tub curve indicating the life cycle [13].

eciding the HOV operational reliability. A detailed analysis on thenfluence of battery chemistry, battery architecture, and hydraulicrchitecture and personnel sphere seals is presented.

. Major standards followed for reliability modeling

The following is the major list of standards followed:

1) FIDES Guide for the estimation of Reliability [13] for electroniccomponents and systems, considering mission and environ-ment specific analysis.

2) MIL HDBK 217F, Military handbook for Reliability Estimation[14] of Electronics Equipment.

3) OREDA Handbook [15] for Offshore Reliability Data.4) IEEE 493 IEEE Recommended practice [16] for Design of reliable

Industrial and Commercial Power Systems.5) DNV RP-A-203-Qualification procedures for new technology

[11].6) Handbook for reliability prediction procedures for mechanical

equipment by Naval Surface Warfare Center (NSWC) [20].

The failure rate determination for the system components wasone by the following methods:

(a) Based on the manufacturers’ data and interpretation accordingto the mission profile.

b) For systems where Failure-in-Time (FIT) data are not avail-able, failure rates are calculated from component failure datafrom the respective standards (FIDES, IEEE, and MIL), takinginto consideration the mission profile, operating conditions andoperating stresses.

FIDES [13] approach is based on physics and failures, sup-orted by the analysis of test data, feedback from operations andxisting models along with statistical interpretations over theormal operating life period of the system and is indicated inig. 1.

The provision in the FIDES standard considering the influence ofhe operating temperature, amplitude and frequency of the tem-erature changes, vibration amplitude, humidity and operatingtresses based on Arrhenius, Norris and Basquin laws [13] waspplied for calculations. The provision given for accounting man-facturing and integrating quality factors are also applied. Thetandards also provide Commercially-Off-The-Shelf (COTS) boardspproach for calculating the failure rates of the systems for theefined mission profile with the functional requirements and mis-ion profile as inputs from the user. The failure rate determinationor the components is done at the circuit component level, and is

ncorporated in the reliability trees [23]. Table 1 gives the standardsollowed for the major systems and components.

Table 2 shows the FIT values of major systems considered foromputations. The values are obtained from indicated standards

Personnel Sphere, Entry Hatches,View Ports

and the values arere-computed based on the actual mission profileand operating conditions adopted.

The TOTAL-SATODEV GRIF tool [21] is used for realizing failuretrees.

3. Overview of HOV systems

Fig. 2 shows the overview of typical deep water HOV [2]. Themajor systems of the HOV include the pressure rated personnelsphere (PS) for human occupancy, equipped with hatch and viewports. They house the systems required for the control of the HOV,communication aids and life safety systems required for the crew.Electro-optical penetrators enable electro optical communicationfrom the external systems. The propulsion system comprises ofbatteries and thrusters to enable the vehicle maneuverability inmultiple degrees of freedom, variable buoyancy enabling systems

OthersShape memory alloy element 92Subsea cables 244Subsea connectors 20

N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143 133

of a h

cEc

Frsp

TD

Fig. 2. View of the major systems

ommunication systems, Emergency Surfacing Systems andntanglement release systems. The details of the subsystemsonsidered for analysis are shown in Table 3.

The control architecture considered for analysis is shown inig. 3 where the real time controller is located inside the pressureated enclosure communicating with the computers inside the per-

onnel sphere with hardware supporting Ethernet communicationrotocol.

able 3etails on the considered typical HOV subsystems.

System Description

Propulsion systemPropulsion batteries Pressure-compensated lead acid type,

75 kWh energy capacity, 300 V terminalvoltage, 4 pods of each 26 kWh, and eachpod with 25 numbers of 12 V 100 Ahbatteries

Auxiliary batteries 24 V–15 Ah Communication battery,24 V–50 Ah emergency batteries

Energy budget 54 kWh for a 4 h subsea mission, 3.5 kW forascent buoyancy engine, 8 kW for PS loadsand control system, 10 kW for contingency

Thrusters/pumps 300 V 8 kW BLDC motor operatedControl systems Real time controllers with Ethernet based

network connectivity with personnelsphere, multi-application industrial gradecomputersand control system componentsdistributed inside exostructure mountedpressure rated enclosures and personnelsphere.

Buoyancy tanks Pressure rated tanks with sea water pumpin/out using pumps of 5 kW capacity. Tankinternal pressure feedback system

Manipulators Propulsion battery powered andhydraulically operated pumps and valves

Navigation sensors DVL guided fiber optic and MEMS basedgyroscope for navigation guidance.Redundant camera and lights for visualnavigation

Personnel sphere seals Pressure rated redundant gaskets for hatch0.5 m diameter and three view ports of0.25 m diameter

uman occupied submersible [2].

4. Functions considered for reliability analysis

Based on the recommendations of the DNV RP-A-203 standardsand procedures, Failure Mode Effective Criticality Analysis (FMECA)studies [14] are carried out on HOV operations, and it is identifiedthat the following key representative operations are studied fromreliability point of view.

(1) Thruster enabled visual survey operations with navigationalaid.

(2) Thruster enabled ascent operation.(3) Drop weight enabled ascent operation.(4) Buoyancy enabled ascent operation.(5) Manipulator operations.

4.1. Function 1: thruster enabled visual survey with navigationalaid

This function involves subsea maneuvering of the HOV withvisual and navigational aids. The propulsion function is realizedusing top and lateral thrusters when the HOV is in a near neutrallybuoyant condition [3] with the operating power provided by the on-board propulsion batteries. Visual support function constitutes a setof cameras and lamps located on the HOV exostructure. The posi-tion information required for the HOV is provided by an InternalNavigation Sensor (INS) [33] aided by Doppler Velocity Log (DVL)and acoustic based Long Base Line (LBL) systems [34,35]. The failurecontributions from the major subsystems toward this function areas tabulated below in Table 4.

Fig. 4 indicates the probability of a 4 h mission failure when 25%battery bank capacity fails during the start of the deep sea opera-tions. Fig. 5 shows portion of the tree forming part of propulsionsystem batteries in Fig. 4. It indicates that the major failure con-tribution (93.36%) is from the battery. As per the energy budgetdesign, four pods are required for a successful 4 h subsea mis-sion. Fig. 5 shows a condition when 1 of 4 pods fail during the

start (T0) of a 4 h deep sea mission. Present day battery banksare equipped with battery management systems [48] which pro-vides in situ battery information such as state of charge and stateof health which helps in advancing the decisions for surfacing

134 N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143

ered

tm

4

sssitoeeai

TP

Fig. 3. Control architecture consid

hereby reducing the probability of failure during the rest of theission.

.2. Function 2: thruster enabled ascent

Vehicle stability, buoyancy, depth keeping and emergencyurfacing requirements of deep water HOV are defined by thetandards from classification societies [17–19]. Present day ascentystems [2,3,31] fall under one or a combination of systems involv-ng soft tank variable buoyancy, hard tank buoyancy or usinghrusters. Alvin and Shinkai6500 use buoyancy engine or thrustersr a combination of both for an ascent and descent operations. Clark

stimated that the neutrally buoyant Alvin HOV requires 8 kWh ofnergy for ascending using thrusters or 3.5 kWh for operating vari-ble buoyancy engine for ascending operation [3]. Operating powers provided by on-board batteries [2]. DC battery power is converted

able 4robability of failure of the function and its subsystem.

Level 1 Thruster enabled visual survey with navigational aid 94%Level 2 Propulsion system energy degradation by

25% of mission time 93.86%Loss of real tim

Level 3 Top thrustersoperation 93.76%

Lateral thrustersoperation 93.76%

Camera Functiofailure 6.8%

for the analysis of HOV functions.

to AC by the inverter housed inside pressure-compensated [23,24]or pressure-rated enclosures [25] mounted on the exostructure.

Fig. 5 shows indicates the probability of thruster based ascentfailure when the demand for ascent function arises at the end ofthe deep sea mission and Table 5 shows the failure contributionto the thruster enabled ascent function by its major subsys-tems.

Fig. 6 shows the probability of failure of thruster enabled ascentoperation in T0 + 4 h instant, in a 1 year period indicating 93.76%.When the ascent operation is demanded at the end of the mis-sion period and if each of the battery pods is considered to havea residual energy of 10%, it would be approximately equivalent to

about 2.5 kWh. As the ascent operation requires 8 kWh of energy,the operation requires all the 4 pods. Table 6 shows the failurerate when the thruster enabled function is demanded at differentinstances of the deep sea mission.

e video 7.42% Navigation data failure 7.71%

nal Lights functionalfailure 7.11%

Position datafailure 6.84%

Pilot PC failure7.24%

N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143 135

Fig. 4. Failure trees indicating probability of failure during 25% battery bank capacity failure.

Fig. 5. Tree showing battery requirements for thruster system.

Table 5Probability of failure of the thruster enabled ascent function and its subsystem.

Level 1 Thruster enabled ascent 93.86%Level 2 Thrusters operational failure 93.76% Switching command of thrusters from PS failure 6.52%Level 3 Thruster-1 operation

failure 93.7%Thruster-2 operationfailure 93.7%

Switching usingJoystick failure 7.39%

Switching using Softpanels failure 7.39%

136 N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143

ilure of thruster enabled ascent operation.

tipd

4

iwdnatWw4a[

tfowSnSldto

Table 6Failure rate when the thruster enabled function is at demand at different instancesof deep sea mission.

Time of ascent No of pods required Probability of failure

T0a 1 9.8%

T0 + 1 h 1 9.8%T0 + 2 h 2 33.55%T0 + 3 h 3 69.58%

TP

Fig. 6. Tree showing probability of fa

As in the case of thruster enabled visual survey with naviga-ional aided mission function, the probable risk of battery failurempact on the ascent function could be reduced by having therovision to monitor the health of the batteries and advancingecisions.

.3. Function 3: drop weight enabled ascent

During the descent phase, based on the required descent veloc-ty and the level of ballast, the HOV is provided with jettisonable

eights, normally 1 ton [3,4]. When the HOV reaches the desiredepth, a portion of the weight is jettisoned, so that the HOV attainsear-neutral buoyancy and the descent stop. Depth keeping ischieved using the top thrusters. As the buoyancy is near-neutral,he energy required for the thrusters shall be maintained minimal.

hen the mission is completed, the HOV jettisons the remainingeights so as to initiate the ascent at a required velocity of around

0 m/min [3]. Jettisoning the weights is done by electro-magneticctuators or pyro actuators or SMA based isolating mechanisms26,27].

Pyro-actuated devices are normally used for releasing opera-ions in space industries [29]. Pyro-actuated pin pullers developedor Viking craft were demonstrated to have a failure probabilityf 0.4% [29] which corresponds to about 96FIT. NASA LaRC studiesere focused on alternatives to pyro actuated devices, where the

hape Memory Alloy (SMA) was well accepted as a potential alter-ative [28,30] as it also has an equivalent level of reliability. TheMA element is highly stable, and is not subject to consider life time

ike pyro-actuated devices, and follow specific rules for explosionevices, and this simplifies the procedure, resulting in a cost advan-age. The SMA based spring loaded nut was tested for 50 continuousperations [26] and found successful. Ti–Ni SMA based frangi-bolt

able 7robability of failure of drop weight enabled ascent function and its subsystem.

Level 1 Drop weight enabled ascent 7.54%Level 2 Hard switching failure 7.6%

Level 3 Joystick failure 0.04% Communication bus toJoystick failure 2.02%

Data fa

T0 + 4 h 4 93.86%

a T0 indicates the start of the deep water survey operation.

recorded half of shock level compared to its pyro-actuated coun-terparts [28].

The function involves dropping the weights by actuating theSMA using the energy from the propulsion batteries with the com-mand by the pilot through the control systems. Table 7 shows thefailure rate of the function when the drop weight operation is atdemand.

When the drop weight enabled ascent operation is demandedat the end of the mission, there is a total residual contingencyenergy of 10 kWh excluding the contingency energy, while, eachpod has energy of 2.5 kWh. The energy requirement for the dropweight function can be met even with 1 pod. Fig. 7 shows that theprobability of failure is 8.7% in a 1 year period.

4.4. Function 4: hard buoyancy enabled ascent

The function is based on a buoyancy engine [5] which com-prises of distributed and inter-connected pressure rated tanks

located in the HOV exostructure. The level of buoyancy can be var-ied by changing the volume of water inside the tanks. The tanksare provided with pumps [32] and remotely actuated valves forpumping the sea water in and out of the pressure rated tanks. The

Switching using soft panels failure 8.4%ilure to PS 7.54% Data failure to PS 7.54% Pilot PC failure 7.24%

N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143 137

Fig. 7. Tree showing probability of failure of drop weight enabled ascent system.

Table 8Probability of failure of Hard buoyancy enabled ascent function and its subsystem.

Level 1 Level 2 Level 3

Hard buoyancy enabled ascent 36.7% Variable ballast pumping system failure 35% Pump system failure 34.95%Switching of pump failure 0.08%

Variable ballast control hydraulic pressure relief valvefailure 0.59%

Flow control valve-1 failure 0.2%Flow control valve-2 failure 0.2%

Seawater inlet pump pressure relief valve failure 0.298% –l valve%

eedba

pctswotd2

fop

4

o[

Seawater flow controInlet Filter failure 0.06Pressure transmitter f

ropulsion battery powered pumps and hydraulic actuators areontrolled by electronic controllers. Water mass flow meters andank internal pressure measurement systems are used to mea-ure the in situ tank buoyancy and therefore the vehicle buoyancy,hich in turn decides the vehicle ascent velocity. When the ascent

peration is demanded at the end of the deep sea mission, and withhe 2.5 kWh of residual energy in each pod, the operation whichemands 3.5 kWh can be carried out successfully with minimum ofpods which results in a probability of failure of 30.84% to battery.

It can be seen from Table 8 and Fig. 8 that the probability ofailure for this function is 36.7% while Fig. 9 showing the probabilityf failure consisting pressure transmitter feedback system formingart of the tree shown in Fig. 8.

.5. Function 5: manipulator operations

The HOV is normally equipped with manipulators for carryingut subsea physical intervention and scientific sampling activities4]. The operation is carried out by the pilot using the joystick

failure 0.298% –Pilot PC failure 7.24%

ck failure 7.85% Data failure to PS 6.62%

located inside the personnel sphere. If the manipulator jaw func-tions fails when holding a task [36], the HOV will be stranded in thesame location, which is an unsafe condition. Forced pull-out leadsto damages to the HOV as well as the task. Table 9 details the proba-bility of failure of manipulator operation along with its subsystems.It can be seen from Fig. 10 that the probability of failure is 11.85% ina 1 year period when the demand for the manipulator operationsoccurs at the end of the deep sea mission with 2.5 kWh of resid-ual energy in each pod. When the manipulator system operationis demanded, which could last for a period of 10mins, which mayrequire 1 kWh (approx.) could be carried out successfully with aminimum of 1 battery pod. This results in a probability of failure of6.11% to battery system.

5. Contribution of subsystems to functional failures

The contribution of battery, hydraulics and electronics in thediscussed functional failures are summarized in Table 10. It can

138 N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143

Fig. 8. Tree showing probability of failure of h

Fp

bw

5

ba

TP

ig. 9. Tree showing probability of failure of pressure transmitting feedback formingart of tree in Fig. 8.

e seen that batteries play a significant role in functional failureshich is followed by hydraulics and electronics systems.

.1. Batteries

From Table 10 it can be seen that the on-board energy storageatteries plays a significant role in deciding the reliability of thenalyzed key representative HOV operations.

able 9robability of failure of the function and its subsystem.

Level 1 Manipulator entanglement release failure 11.85%Level 2 Manipulator switching control failure 6.52%

Level 3 Command throughcommunicationnetwork failure 6.48%

Manipulatorcontrolling joystickfailure 0.04%

ard buoyancy enabled ascent function.

5.1.1. Energy storage technologyThe energy density of a battery is the total amount of energy

it can store for a given volume and weight. The energy densityand reliability of the onboard batteries are critical in terms of HOVoperational reliability. Majority of present day subsea based energystorage solutions [50] are realized using lead acid batteries and Libased batteries [51] operated in a pressure compensated manner[4,25]. Lithium based batteries are superior than lead acid batter-ies as they offer higher energy density, power density and excellentreliability compared to lead acid batteries [39,50]. Lead acid batter-ies have the specific advantage of more number of charge-dischargecycles and deep discharge allowance than Li-based batteries [13].Li-ion batteries have 3–4 times life time of lead acid batteries.However, the Depth of Discharge (DOD) plays a major role in deter-mining the life of the batteries. Both Li-ion and lead acid batterieshave their life time super-linearly extended if they are charged anddischarged with similar DOD, while the life of the Lithium Ion bat-tery within 50% DOD is 5 times than that of lead-acid at 100% DOD[46,47].

Lithium Ion batteries and lead acid batteries have energy den-sities of 150 Wh/kg and 40 Wh/kg respectively. For this energydensity, to store 100 kWh of energy, 2750 kg of lead acid batter-ies are required. Based on the volume–weight ratio [49–53], itrequires approximately 7 m3 of space to house lead acid batter-ies. With the energy density of Li batteries, 670 kg is requiredwith the volume occupancy of around 2 m3 to house the sameenergy of 100 kWh. The calculation of failure rate of a 100 kWh

capacity lead acid battery bank is done based on the failurerates of lead acid batteries collected by the US department ofenergy [39] and is estimated to be around 3140 FIT for a 12 V,100 Ah battery. The reliability of Li-ion batteries for the same

Manipulator operating hydraulic failure 11.77%Pump system failure11.69%

Switching of pumpfailure 0.8%

N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143 139

Table 10Contribution of subsystem in HOV failure.

Function Probability of failure in 1 year Contribution of subsystems in %

Battery Hydraulics Electronics Others

Thruster enabled visual surveywith navigation aid

94% 93.6% 0% 4% 1.3%

Thruster enabled ascent 93.76% 97.6% 0% 1.5% 0%Drop weight enabled ascent 7.54% 80% 0% 1.5% 8%Hard buoyancy enabled ascent 36.8% 78.3% 11% 2.6% 7.2%Manipulator operations 11.85% 82.8% 9.75% 1.7% 7.5%

Fig. 10. Tree showing probability of failure of manipulator entanglement release failure.

Fig. 11. Tree showing probability of failure of battery with 120 V configuration to meet 75 kWh capacity.

140 N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143

nfigur

crc

1pvb1l

55tbtB1muti

5tbao1so

TP

ABCDE

Fig. 12. Probability of failure for the 300 V co

apacities is calculated based on FIDES standards [13] for theequired number of charge discharge cycles and the depth of dis-harge.

It is found that, for achieving energy storage capacity of00 kWh, specific configurations are better from the reliabilityerspective, limited to the commercial availability of power con-ersion and operating voltage levels of propulsion equipment onoard HOV. Two feasible configurations with terminal voltages of20 V and 300 V DC using pressure compensated lead acid and

ithium batteries are considered for analysis.

.1.2. Configuration with lead acid battery chemistry

.1.2.1. For 120 V output. To realize this configuration, it is requiredo have 10 pods with each pod stacking 10 numbers of 12 V 100 Ahatteries in series. Each pod can deliver 10 kWh effectively andhe reliability of each pod is estimated to be around 31,400 FIT.ased on these values, Fig. 11 shows the probability of failure of a00 kWh battery bank to deliver 75 kWh is 44.56% in the definedission period. With the configuration indicated in Fig. 11, the fail-

re of the 3rd pod is treated as a degraded condition correspondingo an effective energy delivery of 70 kWh. This corresponds to thendicated 3 out of 10 option.

.1.2.2. For 300 V output. To realize this configuration, it is requiredo have 4 pods with each pod stacking 25 numbers of 12 V 100 Ahatteries. Each pod can deliver 26 kWh effectively, and the reli-bility of each pod is estimated to be around 78,500 FIT. Based

n these values, Fig. 12 shows the probability of failure for the00 kWh battery bank to deliver 75 kWh is93.6% in the defined mis-ion period. With the configuration indicated in Fig. 12, the failuref the 1st pod is treated as a degraded condition corresponding to an

able 11robability of failure with different battery configurations.

Battery Config./Function Probability of failure of HOV in %

300 V Pb acid, 75 kWh 300 V

A 94% 82.08%B 93.76% 81.34%C 7.54% 74.51%D 36.8% 25.4%E 11.85% 8.96%

: Thruster enabled visual survey with navigation aid.: Thruster enabled ascent.: Drop weight enabled ascent.: Hard buoyancy enabled ascent.: Manipulator operations.

ation battery bank to meet 75 kWh capacity.

effective energy delivery of 72 kWh. This corresponds to the indi-cated 1 out of 4 option.

Table 11 summarizes the probability of failure of HOV withdifferent battery configurations. It can be seen that, even thoughhaving extra energy could reduce the probability of failure, the120 V configuration results in reduced probabilities of failure.

5.1.3. Configuration with 300 V lithium based chemistryTo realize this configuration, it is required to have 11 pods with

each pod stacking 27 numbers of 4 V, 39 Ah batteries. Each pod shalldeliver approximately 9 kWh effectively and the reliability of eachpod is estimated to be around 4125 FIT. Based on these values, theprobability of failure for the battery bank to deliver 75 kWh is esti-mated to be 4.2 × 10−4% in a 1 year period. It can be seen fromFig. 13, that the failure of the 4th pod paralyzes the system.

Even though it can be seen that the failure rates of Lithiumbased batteries is very low, pressure compensated Li ion batter-ies are 10–12 times more expensive than lead acid batteries [47]with lead acid batteries costing 200$/kWh and Li ion batteriescosting 2000$/kWh. With the present technological developmentsin Li Ion technologies, the cost of Li-ion batteries are expectedto come down to 300$/kWh in the year 2015 [47]. Thus a fairtrade-off decision between the reliability and cost is inevitable overtime.

5.2. Hydraulic systems

From Table 10 it can be seen that the hydraulic systems playa reasonable role in deciding the reliability of the key repre-sentative HOV operations. Hydraulic operations are required forhard buoyancy enabled ascent and manipulator operations and

Pb acid, 100 kWh (with 25% extra energy) 120 V Pb acid 75 kWh

47.97% 45.83% 1.52%

8.87% 6.12%

N. Vedachalam et al. / Applied Ocean Research 46 (2014) 131–143 141

ure of lithium ion battery in 300 V configuration.

tTpitsrt

6

vrtsA[avtpst5

rp

TI

Fig. 13. Tree showing the probability of fail

he influence of providing redundancy is tabulated in Table 12.he hydraulic pump is driven by a Brushless DC (BLDC) motorowered from the propulsion battery pods. The hydraulic energy

s used for operating the valve actuators. From the analysis,he probability of pump operational failure is 4.62%. Fig. 14hows the part of the failure tree with redundancy incorpo-ated in variable ballast pumping system and control valves sohat the probability of failure is reduced to 2.45% from 11.85%.

. Water tight intergrity of personnel sphere

The pressure rated water tightness of the PS entry hatch andiew ports are realized by the use of seals and gaskets [38]. Theisk to the personnel could result out of the implosion of the PS dueo structural integrity failure, failure of the acrylic view ports andeal failures in the hatch, windows and the electric penetrators [2].

study of the failures from submarine accidents [22] and OREDA15] indicates no failure occurrences due to sphere implosion andcrylic port damages. Sealing failure rates for the hatch and theiew ports are calculated based on the NSWC guidelines [20] forhe subsea operating environment, with standard hatch and viewort dimensions [2]. Redundant seals are considered with a seal seaturface finish of 0.8 �m. Fig. 15 shows the failure trees indicatinghe failure probability of 15.06% for missions accumulating up to

00 h.

The personnel sphere (PS) in the HOV shall be designed to haveoom for three persons, provided with a hatch for the entry ofersonnel and view ports that enable external vision [2,5]. The PS

able 12nfluence of hydraulic redundancy in HOV functions.

Failure probability in 1 year

With 120 V Pb acid 75 kWh capacitya

Hard buoyancy enabled ascent 8.87%

Manipulator operations 6.37%

a Derived from Table 11.

Fig. 14. Tree showing the probability of failure using redundant hydraulic system.

contributes to around 30% of the HOV weight [40]. Creep properties,cyclic pressurizing and collapse strength studies are to be done forthe PS during the design phase [41]. As the sphere is subjected toexternal hydrostatic pressure, elastic-plastic Finite Element Anal-

ysis studies need to be done to ensure design compliance duringthe pressure variation phases [17]. Numerical computation is themethod to predict the ultimate strength of spherical personnelsphere accepted by various societies. Li et al. [41] have indicated

With 120 V Pb acid 75 kWh capacity & hydraulic systems redundancy

5.24%3.1%

142 N. Vedachalam et al. / Applied Ocean

Fs

seccm1awii

dpKec3in

TF

ig. 15. Fault tree showing failure probability of gaskets and seals used in the per-onnel sphere

tress analysis and stability analysis needs to be performed tonsure conformity to the maximum stress result in the yieldingondition and the yielding zone spread leading to final plasticollapse. A personnel sphere used to accommodate three crewembers shall be of 3T weight with an inner diameter of around

.8 m [2]. Even if the hemispheres are geometrically imperfect to small extent, their ability to resist external hydrostatic pressureill be considerably reduced [41–43]. As a result, overall thickness

s to be increased, so that their safety factors are much larger. Thisncreases the overall cost and weight of the PS and the submersible.

Ross [43] presented the dome design which had a concave cupome which had a 30% weight saving and 70% increased collapseressure over conventional cap domes. Thomson has applied theoiter theory [43] to study the buckling behavior of spheres due toxternal pressure due to imperfect spherical steel shells and con-

luded that imperfections of 0.2%, 0.5%, and 1% could result in a0%, 45%, and 60% reduced buckling pressures respectively. Thus,

t is possible to design a PS with minimum weight. Out of Round-ess (OOR) is an important factor to be considered so as to design a

Fig. 16. Seal FIT as a function of seal siz

able 13ailure probabilities with quality seal seat finish and hatch and view port dimensions.

Parameter

Seal seat contact surface finish (with fixed seal length to width ratio of 3)

Seal length to seal width ratio (for hatch and 3 view ports with seal seatsurface finish 0.38 �m)

Research 46 (2014) 131–143

sphere with minimum thickness and weight [39]. The overall OORis the maximum difference between the nominal and actual shellradius. The local OOR is measured from a separate circular templatehaving nominal inner and outer radius of the PS.

Precise 3D machining is done to achieve the required out ofroundness. Both the overall OOR and the local OOR is to be lessthan 0.5% of the hull inner radius as per the rules laid by the clas-sification societies including ABS, DNV and Lloyds. Even the deepwater submersibles currently in operation have a design safety fac-tor of 1.2–1.55, but the safety factors are yet to be standardized [40].All the deep water submersibles use Titanium alloy 6 Al–4 V whileMir1 and Mir 2 uses Martensite Nickel steel. This Ti alloy has excel-lent post weld performance as the alloy has a low oxygen content[44].

Manufacturing processes [44,45,54] involve hot press formingof titanium spheres and precise 3D machining for sphericity whichis followed by electron beam welding. DNV rules section B300describes the welding standards, heat treatment, NDT while A203section defines the tolerances for Class I pressure vessels whilesection E100 defines the life time and the designed pressurizationcycles. NEMO model 2000 [37,45] gives confidence on acrylic plastichulls tested for implosion resistance during long term submersionsin deep waters and such exposure do not generate cracks duringrapid depressurization.

The reliability of static seals and gaskets to meet the requiredreliability level is a function of the fluid differential pressure,leakage allowance, seal size, contact stresses, seal hardness, sealsurface finish, surrounding fluid viscosity and temperature [20].Fig. 16 shows the FIT values calculated for gaskets with varyingsurface finishes and seal length-to-width ratio, which are used fora differential pressure of 600 bar in sea water, with zero leakage tol-erance and designed for operation at 2 ◦C. Table 13 summarizes thesealing failure probability for a personnel sphere with a hatch of

0.5 m diameter and view port diameter of 0.25 m for an operatingperiod of 500 h. It could be seen that the dimensions of the hatch,view ports and the seal contact surface finish, play a significant rolein deciding the pressure rated integrity. This demands a reasonable

e and seat seating surface finish.

Value Probability of failure in 1 year

0.38 �m 15.06%1.27 �m 48.53%2.54 �m 79.98%3 2.2%15 27.12%30 59.94%

Ocean

to

7

aHrf4fio7b3iahfiniso

A

trc

R

[[[[[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

N. Vedachalam et al. / Applied

rade-off decision in identifying the sealing reliability in the lightf HOV operational reliability and human safety.

. Conclusions

This paper reviews the present day technological developmentsnd reliability centered design in development of a deep wateruman Occupied Vehicle Reliability analysis were done on five rep-

esentative functions with an assumption that the HOV is utilizedor 10 deep water missions per year, with each mission clocking

h of subsea operations. A detailed analysis done on the identi-ed functions to find the influence of subsystems on the reliabilityf the overall submersible reveals that for a submersible housing5 kWh energy storage, the probability of failure of the hard tankuoyancy ascent function with lead acid batteries configured for00 V terminal voltage and non-redundant hydraulic configuration

s 37.74% and the same failure could be reduced to 5.24% with leadcid batteries terminal voltage configured to 120 V and redundantydraulic configuration. The importance of the seal contact surfacenish, the hatch and view port dimensions in deciding the person-el sphere safety and the role of the personnel sphere imperfection

n the weight of the hull is also described. The results presentedhall serve as a model for designers to arrive at the required trade-ff between the capital expenditure and the required reliability.

cknowledgements

The authors gratefully acknowledge the support extended byhe Ministry of Earth Sciences, Government of India, in funding thisesearch. The authors also wish to thank other related staff for theirontribution and support.

eferences

[1] Talkington H. Manned and remotely operated submersible. San Diego, CA:Naval Undersea Centre; 1976.

[2] Frank Busby R. Manned submersibles. Office of the Oceanographer of the Navy;1976.

[3] Clark RP, Brown CP. Selection of descent & ascent method for the WHOI-RHOV.SNAME; 2000.

[4] Iwai Y, Nakanishi T, Takahashi K. Sea trials and supporting technologies ofmanned submersible Shinkai 6500. In: Proceedings IFREMER. France: Acts deCollogues; 1990.

[5] Kohnen W. Manned research: state of technology 2004/2005. Marine TechnolSoc J 2005;39(3):121–6.

[6] Manecius S, Ramesh R, Subramanian AN, Sathianarayanan D, Harikrishnan G,Jayakumar VK, et al. Technology tool for deep ocean exploration: remotelyoperated vehicle. In: Twentieth international offshore (ocean) and polar engi-neering conference. 2010.

[7] Vedachalam N, Ramesh R, Ramesh S, Sathianarayanan D, Subramaniam AN,Harikrishnan G, et al. Challenges in realizing robust systems for deep water sub-mersible ROSUB6000. In: International symposium on underwater technology.2013.

[8] Sathianarayanan D, Ramesh R, Subramaniam AN, Harikrishnan G, Muthuku-maran D, Murugesan M, et al. Deep sea qualification of remotely operablevehicle (ROSUB 6000). In: International symposium on underwater technology.2013.

[9] Ramesh R, Ramadass N, Sathianarayanan D, Vedachalam N, Ramadass GA.Heading control of ROV ROSUB6000 using non-linear model-aided PDapproach. Int J Emerg Technol Adv Eng 2013;3(4).

10] SUT Journal. UnderwaterTechnology UT2; 2013. p. 26–7.11] Rules for certification/classification of submersible, 1988 rules, DNV, Norway.12] DNV. Qualification procedures for new technology; 2001.13] FIDES. Reliability methodology for electronic system; 2010.14] Military Handbook MIL HDBK-217 F. Reliability prediction of electronic equip-

ment. Washington, DC: Department of Defence; 1991.15] OREDA. Offshore reliability data handbook by SINTEF and group of oil and gas

companies; 2003.16] IEEE Standard. IEEE recommended practice for the design of reliable indus-

trial and commercial power systems. IAS/PSE – Power Systems Engineering;

1997.

17] ABS rules for building and classing. Underwater vehicles, system and hyper-baric facilities; 2010.

18] Germanischer Lloyd. Rules for classification and construction for underwatertechnology. Germanischer Lloyd; 2009.

[

[

Research 46 (2014) 131–143 143

19] IMCA, A0DC022 Code of practice of the operation of manned submersible craft.USA: IMCA: International Marine Contractors Association; 1984.

20] Naval Surface Warfare Center. Handbook for reliability prediction of Mechan-ical equipment. Maryland: Naval Surface Warfare Center, Caderock Division;2011.

21] TOTAL’s GRIF (Graphical Representation of Fiability). User manual version;2012.6

22] Christopher Tingle. Design safety submarine accidents: A 60 year statisticalassessment. Professional safety; 2009. p. 31–9.

23] Pettini R, et al. Pressure tolerant power electronics: IGBT gate driver foroperation in a high pressure hydrostatic environments. J Power Energy Eng2012:1500–8.

24] Mehnertm TH. Handbook of fluid-filled, depth/pressure compensated systemsfor deep ocean applications. Deep Ocean Technology Program. Maryland: NavalShip Research and Development Centre, Annapolis Laboratory; 1972.

25] Mehnert TH. Handbook of fluid-filled, depth/pressure compensated systemsfor deep ocean applications. Deep Ocean Technology Program. Maryland: NavalShip Research and Development Center, Annapolis Laboratory; 1972.

26] Bement LJ, Schimmeh ML. A manual for pyrotechnic design, development andqualification. Hampton, VA: NASA technical memorandum; 1995.

27] Brauer KO. Hand book of pyrotechnics. New York: Chemical Publishing Co. Inc.;1974.

28] Lucy MH, Hardy RC, Kist Jr EH, Watson JJ, Wise SA. Report on alternative devicesto pyrotechnics on space craft. In: Proceedings of 10th Annual AIAA/USU con-ference on small satellites. 1996.

29] Falbo MJ. Robert & Robinson, NASA technical note, NASA TN-7141: Apollo expe-rience report. Washington, DC: Spacecraft Pyrotechnic System; 1973.

30] Murray J, Mignon A, Promper C, Stalmaus R, Lafraconi R. A shape memory alloyactuated satellite isolation value. In: Proceedings of 4th international spacecraftpropulsion conference. 2004.

31] Brown C, Clark R. Using a novel vehicle conceptual design utility to evaluate along range, large payload UUV. IEEE 2010;8(10).

32] Wikstrom P, Terenes L, Kobi H. ABB industries AG, CH-5306, Turgi, Switzerland,reliability, availability and maintainability (RAM) of high power VSD systems.IEEE; 2013. PCIC-98-16.

33] Necsulescu DS, Sasiadek JZ, Kim B, Green DN. Fusion of inertial and kinematicnavigation systems for autonomous vehicles. In: Proceedings of IEEE-IEE vehi-cle navigation and information systems conference. 1993.

34] Panish R, Taylor M. Achieving high navigation accuracy using inertial navigationsystems in autonomous underwater vehicles. IEEE; 2011.

35] Fizoptika, failure analysis on fiber optic based gyroscopes. http://www.fizoptika.com/docs/fizoptika eng52.pdf

36] Vedachalam N, et al. Reliability centered development of deep water ROVROSUB 6000. MTS; 2013.

37] Jerry D, Stachi W, ASME board on Pressure Technology codes and standards ofpressure vessels for human occupancy Handbook of acrylics for submersible:Hyperbaric chambers & aquaria. MTS; 2012.

38] Liang XF, Yi H, Zhang YF, Li D. Reliability and safety analysis of an underwaterdry maintenance cabin. Ocean Eng 2009;37:268–76.

39] Pan B, Cui W. An overview of buckling and ultimate strength of spherical pres-sure hull under external pressure. Marine Struct 2010;23:227–40.

40] Pan B, Cui W. Structural optimization for a spherical pressure hull of a deepmanned submersible based on an appropriate design standard. IEEE J OceanEng 2012;37(3):564–71.

41] Dadrasi A. On the effect of imperfection on buckling load of perforated rectan-gular steel plates. Res J Recent Sci 2013;2(3):36–43.

42] Hutchinson JN. Imperfection sensitivity of externally pressurized sphericalshells. J Appl Mech 1967.

43] Ross CTF, Casson AG. Collapse of dome cup ends under external hydrostaticpressure. Ocean Eng 2000;27:933–52.

44] Alcocer A, Fores P, Giuffre GP, Parareda C, Roca A, Roca J. Pressure Hull design &construction of manned submersible ICTINEU. Instrumentation view point 8;2011. p. 35–6.

45] Stachiw JD. Nemo model 2000 Acrylic plastic spherical hill for manned sub-mersible operation at depths to 3000 feet. San Diego, CA: Naval UnderseaCentre; 1974.

46] Linden D, Reddy TB. Handbook of batteries. New York: Tata McGraw-Hill Pub-lications; 2002. p. 1453.

47] Pedram M, Chang N, Kim Y, Wang Y. Hybrid electrical energy storage systems.ISLPED; 2010.

48] SAFT Batteries Battery Management System. Press release. Note no. N21821-2-0412: www.saftbatteries.com

49] Shukla AK, et al. Electro-chemical capacitors: energy storage and beyond bat-teries. Curr Sci 2000;79(12):25.

50] Hadjipaschalis I, Poullikkas A, Efthimiou V. Overview of current and futureenergy storage technologies for electric power applications. Renew Sust EnergyRev 2009;13:1513–22.

51] Doughty D, Peter Roth E. A general discussion of Li ion battery safety. Elec-trochem Soc Interface 2012:37–44.

52] Zhou W, et al. Battery behavior prediction and battery working states analysis ofa hybrid solar-wind power generator system. Renew Energy 2008;33:1413–23.

53] DeAnda MF, Miller J, Moseley P, Butler P. Reliability of valve-regulated lead-acidbatteries for stationary applications. California: Sandia National Laboratories;2004. p. 68.

54] Welding Research Centre. Welding of managing steels. Welding Research Cen-tre; 2011. Bulletin 159.