ELECTRICAL INTEGRITY ISSUES OF SUBSEA DISTRIBUTION EQUIPMENT AND UMBILICALS

Transcription

1 ELECTRICAL INTEGRITY ISSUES OF SUBSEA DISTRIBUTION EQUIPMENT AND UMBILICALS Author: Neil Douglas, Managing Director, Viper Subsea LLP, Portishead, UK. Date: 2013 Presented at: SUBSEA AUSTRALASIA CONFERENCE 2013 Abstract: Subsea electro-hydraulic and all electric control systems rely on the integrity of the electrical elements within the main umbilical, the electrical jumpers, and electrical connectors throughout the subsea distribution system. This paper looks at a range of examples of failures, the causes, the consequences, and also reviews different design solutions to avoid such failures. In addition, a range of methods for monitoring the electrical integrity of the distribution system, locating points of failure and predicting rates of degradation are considered.

2 1 Introduction Electrical and electro-hydraulic control systems offer significant performance benefits when compared to direct hydraulic control systems, but they do suffer from failures within their electrical distribution components typically umbilicals, umbilical termination assemblies (UTAs), electrical distribution units (EDUs), electrical connectors, and electrical flying leads (EFLs). Historically, the main control system vendors have taken responsibility for the delivery and performance of all the distribution hardware with the exception of the umbilical, but ultimately, the large financial consequences associated with in-service failures are to the oil/gas field Operator s account. Insulation Resistance (IR) measurement is commonly used to determine the integrity of subsea electrical cables and connectors. A low IR effectively means that there is a breakdown in the integrity of the insulating material that surrounds electrical conductors and more practically, indicates that sea water is entering the electrical distribution system. Once water has started penetrating insulating materials, there is a high probability that ultimately electrical breakdown and failure will occur. Subsea electrical failures are experienced in the majority of subsea production systems at some point during the field life cycle. Such failures can be seen during system commissioning, almost immediately after the equipment is immersed in seawater, whilst other failures manifest themselves later in field life and are time related. There are a number of failure modes that result in water ingress which lead to ultimate system failure. The cost of subsea electrical failures in subsea umbilicals and control and distribution systems is significant in terms of marine intervention costs, replacement hardware, and lost production, the costs are typically many millions of dollars per field over its life cycle. The majority of operators work on the basis of when not if an electrical IR failure will occur. With more complex subsea field developments there are typically hundreds of electrical jumpers and thousands of electrical connectors installed all expected to perform in a harsh environment for decades. Unless the industry adopts best practice for avoiding IR failures and continuous monitoring of the subsea distribution system, high operational costs will continue to be incurred. Across the industry there is significant experience in the causes of these electrical integrity failures but also a large number of lessons that can be learned and shared. 2 System Architectures Subsea electrical distribution networks are designed to support a specific field layout. This layout may change during the field s life as new discoveries are made and more wells are required at new locations to maintain production. Generally the electrical distribution systems are bespoke to the project with designs receiving input and demands from the design houses, the Operators, the Control System vendors and also the umbilical, equipment, and component suppliers. Unsurprisingly, with such disparate designs, the industry continues to learn and re-learn the same lessons; a number of failures that we are experiencing in current projects have been seen in previous years and decades. That is not to say that improvements are not being made; lessons learnt by the component suppliers are generally incorporated into product designs and associated processes, but such components benefit from the fact that they are typically used across multiple projects and across all global regions. 3 Failures in Electrical Distribution Systems 3.1 Failure Types Each electrical jumper, connector, and penetrator acts as a potential source of failure with three failure modes: open circuit; short circuit; line insulation failure (current leakage to earth). If there is no fault protection in the distribution network, a single fault (either earth leakage or short-circuit/over-current) can take down the electrical supply and everything that is connected to it. Even with dual redundant supplies a single fault could take out the redundancy and two faults could take out the complete system. Subsea electrical distribution systems historically have suffered from poor reliability. The reasons for this are varied but in general can be broken down into three main headings: System Design Installation and Operating Environment Component Manufacture and Assembly From a systems design perspective, the following characteristics, either fully or partially, can be applied to most electrical distribution network implementations. From these characteristics, it can easily be seen how the system design can have such a large impact on the reliability and availability of the electrical distribution network. Large numbers of series and branch connected umbilicals & jumpers forming a series of networks. Design approach relies on high reliability of components and intervention as the means of repair. Large numbers of interconnected electrical connectors. Minimal condition monitoring. No means of isolating faults except by intervention.

3 Once a failure has occurred and has been identified, it can be categorized into one of two types, namely Intrinsic, (where the faults relate to design, materials or hardware assembly) or Extrinsic (where the faults relate to handling, installation, environment, mechanical stressing and misapplication). Such failures can cause slow degradation of integrity over time or can result in an immediate catastrophic failure. Intrinsic Failures that have been seen in the past include: Cracks in epoxy in connectors (Figure 1) Elastomeric failures (including bellows) Mechanical splice failures Polyethylene moulding failures (cathodic disbondment) Compression gland failures due to ambient pressure on cable Insufficient compensation volume in connectors Shuttle pin assembly failure (Figure 2) Inadequate material selection / material compatibility Inadequate adhesion to connector parts during moulding processes Fracturing of electrical pin inserts Poor soldering workmanship (flux, excessive quantity and sharp edges) (Figure 3) Incorrect boot seals applied (wrong type and size) Damaged/split boot seals Use of incorrect crimping tools (Figure 4) Non-waterproof designs used as secondary barrier (Figure 5) Examples of previously experienced extrinsic failures include Calcareous and/or marine deposit around compensation and pin shrouds (Figure 7) Lateral mechanical stresses placed on connectors when cables are installed on structures ROV impact damage Connectors left exposed subsea for too long (not properly parked or protected) (Figure 8) Cable sheath damage Nascent hydrogen penetration into conductor insulation material Fishing/trawl board impact Harsh seabed currents that can fatigue components 3.2 Installation and Operating Environment Failures Installing subsea umbilicals in deep-water is a challenging task requiring a high level of expertise to avoid stressing the umbilicals. In addition to electrical lines, subsea umbilicals also contain hydraulic, chemical, and in some case optical lines, all terminated to a seabed structure that can weigh greater than 10 tonnes. Installation of this equipment is dependent upon having the right weather conditions and that all the stresses imparted into the umbilical are adequately controlled. There have been a number of instances where equipment has been significantly degraded during installation. After installation the electrical system is subject to external damage such as fishing gear impact and harsh seabed currents that can fatigue components that are not secured to the seabed. The dynamic section of the umbilical from the host facility to the seabed is subject to additional fatigue loads caused by vessel movement (for floating installations) and the various tidal currents that occur throughout the water column. 3.3 UMBILICAL Failures A previously published paper (Rossiter), reported that: i. 50% of all umbilical failures arise from incorrect installation and handling ii. 20% of all failures arise from electrical faults iii. 10% of failures arise from incorrect operation iv. 20% of failures arise either singly or in combination, as a result of material/component defect, incorrect repair, marine life or extreme environment. Experience suggests that there will be a significant decline in the umbilical IR after installation, generally this will be a substantial reduction over time. Such declines in the umbilical IR may well be masked by the more localised faults encountered elsewhere in jumpers, connectors and Subsea Control Modules. Despite the best endeavours of the umbilical manufacturers, moisture penetration is a continuing problem. Many of the problems leading to a reduction in IR have been identified (Lofaro & Villaran, 2010) in a table listing possible failure mechanisms in electrical cables. Whilst the report and table were compiled for the nuclear industry in the USA, the table gives a comprehensive listing of the most likely water intrusion failure mechanisms, causes and effects. It is recognised that some moisture will eventually penetrate the plastic covering materials through diffusion and permeation, owing to pressure and material density differentials, leading in the long run to a soggy dielectric. Permeation of molecules (water and hydrogen) through insulation barriers depends on many factors, including: concentration (partial pressure) temperature, differential pressure either side of the barrier, time of exposure and barrier thickness. Permeation of any hydrogen molecules will also depend on presence of anodes and localised cathodic protection potentials.

4 The only effective way of reducing this is to employ thicker insulation materials, which inevitably increase the capital cost and decreases flexibility but this may be the penalty for getting a more reliable cable. A technique employed by the subsea power cable industry, is to use a double layer, tandem extrusion process. This lays down the dielectric insulation materials around the conductor in two half thickness layers. This prevents any flaw affecting the whole layer and the discontinuity between them acts as a barrier to their growth Water Treeing in Umbilicals Published literature (Boggs & Xu) (Hai & Thang) on failures in subsea cables gives much attention to Water Treeing; however, it is not clear if this is a problem for systems operating at low voltage, i.e. less than about 500V. The effect can be summarised as a tree like breakdown of the dielectric insulation material (see Figure 9) due to stressing by the electric field in the cable. To occur, there needs to be contamination or a local defect or inclusion in the insulation material, which, under the effect of the alternating field and in the presence of moisture, grows to form a tree like structure, which can eventually cross the insulator and cause a breakdown. Such trees can also grow from a surface defect. Water trees need 70-80% relative humidity within the insulation to grow, they do not need free water. Techniques have been developed (Boone, Eichhorn, & Sahadlich) (Abib & Daniel) for minimising this effect in power cables but there is a dichotomy of view in the industry. American manufacturers of waterproof cable have tended to use tree retardant, cross-linked polyethylene (TR-XPLE) where an additive is used to retard treeing. While a successful technique, it has some manufacturing difficulties and affects flexibility of the finished cable. In Europe, the trend has been to employ rigorous cleanliness and attention to quality in the manufacture of the insulation materials and during the extrusion process, to minimise contaminants and inclusions and hence the seeds for treeing. Independent reports suggest both approaches, whilst having various pros and cons, generally result in similar performance. Whilst water treeing may not be considered a serious problem at the voltages generally employed for subsea control systems, it is dependent on both voltage and time and therefore it may be considered that over a long enough period, such effects could become an issue and it is suggested that this might be the subject of an independent investigation by, for instance, a university materials department. 3.4 Electrical Connector Failures The mass majority of failures in electrical connectors are attributed to either sub-standard assembly or elastomeric failure. The following list provides a top level summary of the types of failure seen historically: Cable Boot Crimp Adaptor Insulation Sleeve Pressure Compensation Boot Cable Seal Up stand (Flux) Cable Insulation Gel Fill Shuttle pin While some of these failure points have been eliminated or improved with the latest designs, the quality of the connector assembly can be a major contributory cause of failures, and this applies equally to the latest design as to those earlier generation designs. It has to be recognised that these components are operating in a very hostile environment and that only the very highest standard of assembly can be accepted for every connector, termination and assembly. It should be further recognised that just because a component has passed its FAT, there may still be latent manufacturing faults, which only long term emersion will elicit. Without applying the highest quality approach, failures are bound to continue. Failures in jumpers/connectors have also been identified that have been the result of, or exacerbated by, rough handling, which unfortunately is a natural consequence of the way they have to be installed by remote means. The force required to remove connectors also increases when calcareous and marine growth deposits are present. This is not always considered within connector designs and often leads to failures, especially in older installations. 4 Insulation Resistance It is normal for most subsea installations powered from the surface, to be equipped with devices that continuously monitor the Insulation Resistance of each of the power circuits. These Line Insulation Monitoring (LIM) devices generally impress a dc voltage onto the ac supply and measure the associated dc current leaking to ground. The acceptance criteria for subsea umbilicals and jumper assemblies is for an IR value in excess of 10 GOhms, however experience shows that once the equipment is deployed this level falls off and if sound, tends to stabilise at a lower level. However any faults in the cable/jumper connector, resulting in moisture ingress can result in a continuing reduction in IR over a period of months or years. A LIM device will provide continuous monitoring of the IR for each cable, allowing prediction of likely failures arising in a Subsea Control Module (SCM), or group of SCMs, providing advance warning of un-serviceability and the possible loss of associated production capacity. This method works until some form of galvanic isolation is present in the system such as a transformer, in the subsea UTA or EDU, as is the case in many systems. Even without the galvanic isolation, the LIM located on the topsides is unable to distinguish between an IR failure in the bulk insulation within the umbilical or a failure in the subsea network at the end of the umbilical and beyond. Historically LIMs used in these topsides applications have been products that were originally designed for onshore power networks with much shorter cables, and

5 higher levels of IR. This results in the LIM devices not operating satisfactorily when the IR drops to low levels giving both high errors in the IR readings and also instability. Figure 10 shows the output readings of two electrical pairs in an umbilical in the North Sea. The X axis represents time, and in this particular graph, the total time period is two years. The figure illustrates a number of issues discussed in this paper namely, a failing electrical distribution system, and secondly the instability of the topsides LIM or are the IR figures really varying as shown? The currents flowing in an umbilical are complex, and are shown graphically in Figure 11; as well as the leakage current there is a capacitive charging current and an absorption current. These different currents, if known, can be used to extract much more detail on the performance and integrity of the umbilical insulation. In particular, the data can be used to distinguish between a subsea IR problem and a failure in the umbilical bulk insulation. Viper Subsea has developed a LIM, known as the V-LIM (see Figure 12), that is specifically designed to be optimally suited for subsea umbilicals it offers high accuracy and high stability measurements at low IR levels, operates with power-line signalling and importantly also measures a range of other parameters relating to the integrity of the umbilical enabling the differentiation between umbilical failures and those in the other subsea distribution equipment. 4.1 Minimum Acceptable IR Levels On the topsides, the LIMs have the ability to set an alarm and subsequently to de-energise the subsea circuit at predefined IR thresholds. In practice, operators wish to keep the system running as long as possible in the presence of low IR, and would only shut down if a dangerous or unsafe condition occurred and hence the trip threshold tends to be reduced to the minimum. The main safety device in such circumstances is the over-current circuit breaker, which will be triggered if the leakage current plus normal load current together exceeds the safety trip level. In such circumstances, the main concern would be of excessive heating in conductors or connectors due to excessive current flow, which could eventually cause a local failure. The most common question from operators is What is an acceptable IR level for my system? Overload circuit breakers are designed to trip when a large current flows, usually due to a short circuit, which is generally a transient event to which the circuit breaker will respond within seconds. However the increased load current due to subsea IR reduction tends to be a long term, steadily increasing value, which is unlikely to trigger the circuit breaker until significant current flow is registered. It is questionable whether the topside equipment and wiring will have been rated for such a high continuous current flow. The effect of the increasing leakage current can however result in system failure well in advance of any protective circuit breaker operating: the leakage may result in a loss of voltage regulation in the SEMs and also result in the inability of the network to supply power on demand to the SEMs. Viper Subsea has undertaken a number of system analyses in order to determine minimum acceptable IR thresholds for different projects. Such analyses compare results of different failure modes i.e. water ingress along the length of the umbilical versus ingress into discrete points in the distribution network. These analyses are used to predict not only the envelope of operation with respect to water ingress, but also, with live field data, an estimate in time as to how long the system is likely to stay operational. It is generally surprising, just how low the IR level can go before there is any loss in system operability. The power lost in the umbilical through leakage increases quickly for an IR of less than 100 kohms but a considerable reduction in IR can be withstood. Even when the IR reduces to a few kohms, the supply current would typically only double in value. With this in mind a system which is to all intents and purposes operating successfully below the LIM threshold could be only a few kohms from failure without further warning. It has been shown that it is unlikely that the leakage current resulting from such low values of IR, will of itself be sufficient to trip the over-current circuit breaker. A generalised breakdown of cable insulation, resulting in distributed leakage currents sufficient to trip the circuit breakers, although possible, is generally unlikely. When the circuit breaker does operate, in most cases the leakage and therefore low IR can be attributed to localised breakdown leading to short or near short circuit conditions. Heating due to high leakage currents (modest if distributed evenly and submerged) at a localised defect could cause a local rise in temperature, even in the presence of cold water in the vicinity, leading to further breakdown and an inevitable short circuit to ground. 4.2 Failure Protection It is fairly common practice to utilise some form of current protection in the Electrical Distribution Unit (EDU). Typically such protection would take the form of one of the following: Passive current limiters (resistors or inductors) Positive Temperature Coefficient Thermistors (PTC s) Solid state relays (controlled by local electronics) Fuses PTC devices are not ideal: they are typically restricted to a maximum operating voltage of 260V a.c. and have to be mounted in parallel to accommodate full load current. Fuses are typically one shot and require subsea intervention to replace them. (Fuses can also fail with age and under normal current conditions.) Subsea Electronic Modules (SEMs) typically have their own internal protection fuses, PTC s or current fold-back circuitry in the front end Power Supply Units (PSUs) and because of this, the question has to be asked what is the over-current protection in the EDU protecting? In reality it is providing protection against the failure of electrical connectors and cables in the distribution network only, and not failures within the SEMs or connected sensors. We have already seen that the predominant failure mode of connectors and jumpers is water ingress, which results in a reduction in IR before any over-current situation exists. It therefore should be obvious that the preferred technical solution should be to utilise switches in the EDU that are operated from circuitry that is monitoring IR as well as current. Monitoring of IR in each leg of the distribution network

6 will provide the best electrical integrity data and also assist in fault finding. More recent projects have utilised active over-current protection devices in the EDUs. Typically, this active circuitry would monitor the current flow and use solid state relays to isolate any fault. The use of active electronics allows these relays to be switched on demand from the surface (to assist in fault diagnosis) and also to send current and voltage data back to surface monitoring electronics. 5 Pre-deployment testing All equipment will have been subjected to a factory acceptance test, which tests components to their operational limits to elicit possible faults due to inherent defects and workmanship errors. This applies to umbilicals, jumpers, terminations and distribution units. However, many low IR faults are only found after periods of immersion in service and are not found during pre-deployment testing, suggesting that such testing is not sufficiently rigorous. Clearly not all equipment can be subjected to hyperbaric tests before deployment due to its physical size, but consideration should be given to making the FAT and other pre-deployment testing more searching. Testing could be extended in duration, be to more severe test levels and utilise more sophisticated tests. During any test regime, components must not be over stressed, otherwise latent potential faults could be created. As well as FATs, equipment may well be subject to additional testing on the dockside and/or on the installation vessel prior to deployment and after deployment. This represents an opportunity to get base-line performance characteristics, which can be used as a standard against which subsequent measurements can be compared and to form the base-line for future trending of performance. Such additional measurements might include: Umbilical polarisation index (dc test) Dielectric loss factor (ac test) Time domain reflectometry test 6 Condition Monitoring 6.1 Online Condition monitoring and diagnostics On the basis that despite the best testing, some faults will occur, as has been previously discussed, it has been normal to install Line Insulation Monitoring units at the topside power supply. Since all measurements are made from one end, only a limited picture is available of the condition of the distribution system and there is a need to know not only that a fault has occurred but where it is in the network, to guide intervention planning. The system status can be improved somewhat by using the SCMs to monitor their own condition and in some cases to look back up the distribution system. However this latter capability is only possible if a transformer provides galvanic isolation between the topside and subsea LIMs. A more complete picture of the network status can be made if monitoring is located at the distribution point and is able to monitor each leg of the distribution system separately. Viper Subsea has developed an active electrical distribution unit to provide this functionality: Voltage and current measurements are made continuously, while IR and loop resistance measurements are made on an isolated leg, once the switch is opened. Use of such a circuit on each of the outputs of a distribution unit, enables condition monitoring, fault detection and isolation. Fault isolation can be accomplished either automatically, with over current or low IR triggering the circuit breaker to open, or on command by alerting the Master Control Station (MCS) / operator to an out of specification condition The V-SLIM The V-SLIM (see Figure 13) has been developed by Viper Subsea specifically to assist with subsea fault finding and the trending of the electrical integrity performance with time. The V-SLIM can be deployed into the subsea network to provide instant readings or to be left deployed, gathering integrity data for trending and failure prediction. The V-SLIM can be connected to one cable or bulkhead connector or, more usually, plugged between two halves of a connector pair. It measures voltage and current (normal and leakage) and can identify whether earth leakage problem is upstream or downstream of the unit. It also can undertake a loop resistance measurement and provide full transient (waveform) analyses. It is fully compatible with existing control system equipment from all manufacturers and can operate in circuit with the system powered or isolated. 6.2 Offline Condition Monitoring Non Intrusive Techniques Offline condition monitoring requires manual intervention and specialist personnel. To perform this activity the subsea equipment needs to be shut-down and therefore impacts on oil/gas production. As all testing is performed from the host installation there is no requirement for a subsea Intervention vessel to perform operations with equipment on the seabed.

7 6.2.1 Surface Insulation Resistance (Megger) Testing By utilising a higher test voltage to record insulation resistance than that used by the LIM instruments, a higher accuracy of measurement can be obtained due to the leakage currents becoming much larger. Testing with an Insulation Resistance Test unit such as a Megger can be undertaken at 500Vdc. Testing at voltages above this with a fully configured system is normally not possible due to input voltage limitations of the Subsea Control Modules. A benefit of such testing is that the applied voltage is similar to that of the peak applied voltage that the system normally operates at i.e. in the range 200 to 500V and therefore stresses the insulation materials to similar level as would be seen in normal operation. Such IR testing is always performed immediately after umbilical installation and this provides a bench mark for future interventions. To allow comparison of results it is important that the same test meter type is used otherwise results may vary significantly. Temperature has a major bearing on the resistive properties of cable insulation materials such that just a 10oC rise in temperature can cause the IR to decrease by up to 50% of its value. Field readings however, because of the consistency of both the environment and the measuring equipment, can be expected to be repeatable to an estimated accuracy of +/- 5% Time Domain Reflectometer (TDR) Similar to Insulation Resistance Testing, Time Domain Reflectometers (TDR) are typically used as a reactive test tool in response to a failure. The TDR operates by transmitting a pulse of energy down the defective line. When the pulse reaches a discontinuity in the line a percentage of energy is reflected back to the instrument. As the transmission characteristics of the cables are all known, the time taken for the pulse to propagate through the system can be calculated. Each reflection can be correlated with a known location in the system and anomalies identified. Every connection in the distribution network will be identified in the TDR measurement in addition to areas of significant electrical defects such as broken cables, unmated connections or very low impedance faults (<1kΩ). These significant anomalies will show up as a peak or trough on the TDR. Figure 14 shows a simplified example of a TDR output graph when an IR fault occurs. The result given by the TDR is invariably a graph giving peaks and troughs where the impedance has changed. Provided the distances between each connection in the system are known, it should be clear where any anomalies lie. These anomalies are potentially where the electrical faults will be. However, for this to work appropriately the Velocity of Propagation (VOP) of the cable must be known. This is usually given by the manufacturer of the cable, but there are various methods of finding this with a healthy line. As with the insulation test tool, TDR testing utilises skilled personnel and requires the system to be powered down and isolated. 6.3 Offline Condition Monitoring Subsea Intrusive Techniques When is not possible to identify actual fault locations via online or non-intrusive offline condition monitoring techniques, subsea intervention is necessary. This involves the mobilisation of a specialist intervention vessel that is equipped with Remotely Operated Vehicles (ROVs) that can undertake basic activities such as making and breaking of connections on the seabed, fitting test connectors and undertaking electrical tests Subsea Megger Industrial and Marine Engineering Services Ltd (IMES) have developed, in conjunction with BP, a subsea Megger and loop resistance tool. The tool is based upon a commercially available Megger unit packaged within a one atmosphere enclosure to withstand the ambient hydrostatic pressure. It is mounted on a Remotely Operated Vehicle (ROV) and a flying lead is plugged into the subsea network by ROV. The unit is remotely operated via topside test console located on the intervention vessel with power and communications being provided via the ROV s infrastructure. Within the subsea Megger, a number of relays are provided that allow all permutations of core to core and core to ground testing to be performed. A number of relays are also incorporated that are remotely operable to allow short circuits to be applied when performing bridge testing from the host installation. This allows various combinations of tests to be performed in a single ROV dive trip Line Resonance Analysis The Line Resonance Analysis (LIRA) method (Fantoni, 2009) was developed by the Halden Reactor Project five or six years ago. It is based on transmission line theory. The behaviour of a transmission line depends on its length in comparison with the wavelength of the electric signal traveling into it. When the transmission line length is much lower than the wavelength, as happens when the cable is short and the signal frequency is low, the line has no influence on the circuit behaviour and the circuit impedance, as seen from the generator side, is equal to the load impedance at any time. However, if the line length and/or the signal frequency are high enough, the line impedance seen from the generator does not match the load. LIRA includes an algorithm to evaluate an accurate line impedance spectrum from noise measurements. Line impedance estimation is then used as the basis for local and global degradation assessment. Tests performed with LIRA show that thermal degradation of the cable insulation and mechanical damage on the insulation do have an impact on the transmission line capacitance (C) and to a lesser extent on the inductance (L). Direct measurement of C (and L) would not be effective due to the electrical noise normally present in installed cables. LIRA monitors C variations through its impact on the complex line impedance, taking advantage of the strong amplification factor on some properties of the phase and amplitude of the impedance figure.

8 Since the LIRA technique looks at variations in the complex line impedance, the technique is only beneficial if it is used regularly, and results are obtained when the system is initially installed and is known to be good. It is difficult to discriminate between cable faults (degradation) and load faults (changes in load reactance).

9

10

11 7 Conclusion This paper has summarised most of the known causes of subsea electrical IR failures and the subsea electrical fault detection techniques that are currently available. Awareness of the design, manufacture, and test issues as well as the methods available for monitoring integrity, locating faults and predicting time to failure should lead to a higher availability of the electrical distribution network. Throughout the paper, observations and recommendations have been made. Those that are considered to be the most critical to improve the reliability and integrity of the distribution network are summarised below, together with a few additional recommendations that have not been addressed in this paper: Reducing vulnerability to faults by limiting the number of drops and number of series connectors forming any particular network; More focus should be applied to ensuring component reliability and the selection of components based on reliability; Ensure all electrical flying leads and EDUs are field replaceable; More stringent hyperbaric testing as part of FAT for components/assemblies which need to keep out water from electrical circuits; Implement 100% inspection of component parts and workmanship; Ensure dual redundant water barriers are truly independent; Ensure dielectric fluid has no entrained moisture (to get very good IR, it is necessary to drive out any residual moisture by gentle heating of the components and dielectric fluid in an oven before or during assembly. Use of centrifugal drying of the dielectric fluid might also be considered); Maintainability would be significantly improved by incorporating an ability to detect and identify the location of failures: Monitor IR on each output leg of the subsea electrical distribution unit; Provide an option to isolate a leg when an IR or over-current fault is detected; Enable the operator to remotely increase trip threshold; Failure of the monitoring system should result in power always being available to the consumers. i.e. failure of the monitoring devices maintains system power; Electrical systems in warm water environments should protect the electrical system from marine growth: Suitable shrouding of termination assemblies; Ability to inject calcareous removal chemicals such as sulphamic acid;

12 8 References Boggs, S., & Xu, J. (n.d.). Water Treeing-Filled versus Unfilled Cable Insulation. University of Connecticut. Boone, W., Eichhorn, R. M., & Sahadlich, H. (n.d.). Long life cables by use of tree retardant insulation and super clean shields. Jicable, Fantoni, P. (2009). Condition Monitoring of Electrical Cables Using Line Resonance Analysis (LIRA). 17th International Conference on Nuclear Engineering. Brussels. Hai, V. T., & Thang, N. D. (n.d.). Final breakdown on water tree degraded polymer insulation. Hanoi University of Technology. Lofaro, R., & Villaran, M. (2010). Essential Elements of an Electric Cable Conditioning Monitoring Program. Brookhaven: United States Nuclear Regulatory Commission. Rossiter, S. (n.d.). Long Life Dynamic Power Cables for Deep-Water. London: BPP-Tech.

13 Figure 1 Crack in Epoxy Figure 2 Shuttle Pin Failure

14 Figure 3 Poor Soldering Figure 4 Incorrect Crimping Tool

15 Figure 5 Non-Waterproof Secondary Barrier Figure 6 Consequence of Water Ingress on the Male Pins of a Controlled Environment Connector

16 Figure 7 Calcareous Growth around a Connector Body

17 Figure 8 Marine Growth on an Exposed Connector Figure 9 Water Treeing

18 Figure 10 Output LIM Readings from Two Channels of an Umbilical Figure 11 Non-Load Currents Flowing in an Umbilical

19 Figure 12 High Performance Line Insulation Monitoring for Topside Use (2013) Figure 13 The V-SLIM (2013)

20 Figure 14 Graphical Output from a TDR Test Figure 15 Bridge Configuration

21 Figure 16 Holborn Loop Equivalent Circuits Figure 17 Cable Discharge Mapping Set-up

22 Figure 18 Output from a Cable Discharge Test

23 Please contact Viper Innovations for further information on V-LIM PORTISHEAD OFFICE Viper Innovations Ltd Unit 3, Marine View Office Park 45 Martingale Way Portishead Bristol BS20 7AW United Kingdom Tel: +44 (0) ABERDEEN OFFICE Viper Innovations Ltd Unit F10 The Enterprise Centre Exploration Drive Bridge of Don Aberdeen AB 23 8GX United Kingdom Tel: +44 (0)