Peter Herpertz - Product manager - power, SebaKMT
Good progress is being made in developing the equipment and techniques needed for testing and fault finding on the latest undersea power cables. Nevertheless, in the future, the operators of these cables, if they are to consistently achieve high reliability and minimise the potentially crippling costs that result from downtime, will need to make a much larger investment in test technology.
Renewable energy sources and decentralised power generation are two of the most discussed topics in the power business. Unfortunately, however, the development of test, diagnostic and fault location techniques that will be needed to support decentralised generation is receiving far too little attention. And the first problems associated with the new energy sources are already starting to appear.
The existence of a supply bottleneck in the form of a transmission network that is poorly adapted to working with these new energy sources is universally acknowledged, but no conclusions have yet been reached about how this challenge will be addressed. Let’s look in a little more detail at the financial and technical issues implicit in this challenge particularly, but not exclusively, in the offshore sector.
First let’s be assured that the problems are by no means trivial. Over the next few years, there are plans for the UK to be deriving 18 GW of power from offshore installations and Germany is quite active with 40 GW planned. These are future figures, but offshore wind parks are already starting to come on line. Germany’s first is the Alpha Vertus wind park which, in 2011 contributed 267 GWh of electricity to the country’s national grid and followed this up with a very similar contribution in 2012.
The nominal installed power of the Alpha Vertus wind park is 60 MW, which is equivalent to a maximum power harvest of 1.44 GWh/day. In practice, this figure can never be reached but it is interesting to note that a calculated feed-in tariff of € 0.15 per kWh, the annual revenue from the 267 GWh that is actually generated is over € 40 million or, in very round terms, € 100,000 per day.
Consider the implications if a cable fault puts the wind park out of commission. For every day that the fault remains unrepaired, that € 100,000 in revenue will be completely lost. If the fault takes a month to find and repair, this amounts to a revenue loss in excess of € 3 million! Now remember that Alpha Vertus is only a small “test” wind park. Full scale wind parks planned and under construction typically have a capacity 400 MW compared with the 60 MW of Alpha Vertus, which means that downtime can be expected to cost more than € 500,000 per day!
Time really is money in these situations and, after a cable fault, the highest priority is to locate and repair the fault so that the cable can be brought back into reliable operating condition as quickly as possible, while all the time remembering that each extra hour of delay is costing more than € 20,000. This means that it’s essential to have accurate trustworthy data about the cable to hand, and to be able to quickly obtain and make use of information about the fault.
These are excellent aspirations but we are, unfortunately, a long way from achieving them. In reality it can take days or even weeks to obtain the services of a vessel suitable for investigating a cable fault. Further delays are likely if no suitable fault location technology is available locally, if no technician with offshore approval is available, or if bad weather intervenes. In short, the bald truth is that, in the North Sea and the Baltic at least, there are no systems in place to fulfil the need for speedy location and rectification of faults on undersea power cables.
So what is possible at the present time and what should we be aiming to make possible in the future? To answer these questions, let’s start by looking at the most important criteria that need to be considered in relation to testing and fault-finding on undersea cables.
The first is the position of the cable. Before any work can be done, it is of course essential to know its exact location and, ultimately, the location of the fault. The operating voltage of the cable must also be considered and, for present-day cables, is likely to range from 230 kV AC up to 450 kV DC. The capacitance of the cable, and its consequent ability to store energy, are further important considerations. Cables that are 200 to 700 km long, several of which are planned – for example the link between Norway and UK – can store up to 10 MJoule of energy, which must be handled and discharged safely. Finally, there’s cable length – long cables are a real challenge for existing test technology, which is primarily designed to work with cables up to 50 km in length rather than 500 km.
Turning now to likely fault scenarios, with HVDC cables faults can be expected to become resistive, because of the likelihood of high-energy discharges at the fault location. With HVAC cables, the breakdown voltage can easily be comparable to the operating voltage. Sheath faults will be relatively common. Drifting anchors have already damaged cables, with the largest recent event being the damage to the 60 kV cable linking Bornholm Island in the Baltic with Denmark, which occurred on December 26th, 2012. Movement and motion of the cable can also cause problems even though there are regulations governing cable installation that have been formulated specifically to minimise the risk of damage from this source.
As we have already noted, before any testing or measurements can be carried out on a cable, it is necessary to know its precise location. Despite the regulations that require subsea cables to be laid in such a way that they will not move from their original site of installation, issues such as strong currents, tidal effects and disturbances by ship anchors mean that even carefully installed cables may end up being moved or damaged.
It is essential, therefore, to have reliable and effective means of localising the cable, verifying its actual route in relation to its expected position and evaluating any changes that have occurred. It is also important to be able to determine the installation depth of the cable in the seabed. In on-shore applications involving underground cables, these requirements are addressed using well-established line tracing techniques, which typically involve injecting an audio-frequency test signal into the cable. This signal is radiated by the cable along its length, and a sensitive detector can be used to determine and follow the cable route.
While the same idea is applicable to subsea cables, there are additional challenges. The first is that the seawater strongly attenuates the frequencies that are most used for the test signal in onshore applications. The second is that many types of subsea power cable, especially HVDC cables, produce high levels of electrical interference at the typical test signal frequencies. The final challenge is that, whatever cable tracing technique is adopted, it must be suitable for use on energised cables so that routine confirmation of cable position can be carried out while the cable remains in service.
Electromagnetic signals – in this case the test signals used for cable tracing – propagate very poorly in salt water. In comparison with ordinary soil, which has an average conductivity of around 0.001 S/m (siemens per metre), seawater typically has a conductivity in the region of 4.0 S/m. This means that electromagnetic signals are attenuated by up to 3 dB per metre. This attenuation is frequency dependent; higher frequencies are attenuated more than lower frequencies.
A subsea cable tracing system that takes these factors into account has already been successfully developed. Essentially, this comprises an audio-frequency test signal generator that operates at lower frequencies than those used for land-based cable tracing, complemented by an inductive coupling system for injecting the test signal into the cable sheath without the need for a direct electrical connection. This arrangement meets the requirement for route tracing to be possible while the cable remains in service.
The signal generator and inductive coupler are used in conjunction with a submersible detector that can be mounted on either a remote operated vehicle (ROV) or a remote operated crawler (ROC) to follow the route of the cable. Features of this system include its ability to provide cable location and depth data, vehicle skew angle data, look-ahead information and, in some cases, information about the location of cable faults.
The challenges involved in route tracing for subsea cables are significant, but the challenges involved in testing these cables are even greater and, at the present time, they have only been partially addressed. VLF test systems for 220 kV AC cables with capacitance up to 25 µF (corresponding to a length of somewhere between 80 and 100 km) are technically feasible but have not yet been developed. Safe discharge of these long high voltage cables is, once again, possible but is by no means a standard solution.
Even where testing is possible with present-day equipment, it is often difficult and costly. When testing was recently carried out on a “only” 20 km long 400 kV cable, three resonance test sets, each mounted on an articulated lorry trailer, had to be connected in parallel to energise the cable. The cost of performing this test was more than € 200,000.
Currently available test equipment is able to address some of the requirements associated with subsea power cables. Compact and readily transportable VLF (very low frequency) test sets have, for example, been successfully used to test 35 kV connections between wind turbines and substations at 3U0 with cable lengths up to 33 km and cable capacitance up to 2 µF.
Even when test equipment is available or at least technically feasible, there still remains the question of the most appropriate type of testing to use and this issue generates much discussion. Partial discharge (PD) diagnostic testing cannot be used on cables longer than about 10 km, for example, and even for cables where PD testing is possible, the question is frequently asked whether this should be combined with tan delta testing.
To answer this question, it is worth remembering that tan delta testing is most useful for detected deterioration related to ageing in cables, whereas PD testing is particularly good at detecting incipient faults of the type that are often related to damage that has occurred during the installation of the cable. It is also important to remember that the cables that link turbines and substations in offshore wind parks are designed to last between 20 to 30 years. This is logical because after this time the wind turbines themselves will have reached the end of their lives. When new turbines are installed these are almost certain to have larger generating capacity and they will, therefore, require the installation of new and larger cables.
When these factors are considered it’s easy to see that tan delta testing makes little sense.PD testing is, however, potentially invaluable. As the cables usually have no joints, the problems are likely to be located in the terminations and PD testing is very effective in detecting faults of this kind. It has also been found to be useful in detecting other types of installation problem including delamination and damage caused by excessive bending of the cable.
A future problem about which little is currently known is the effect of salt fog on cable terminations, but field experience is already starting to show that some types of elbow connector are susceptible to surface discharge when exposed to salt-laden air. Since the turbines are usually serviced once or twice a year, it is a good idea to use these opportunities to check the terminations for this phenomenon using a handheld PD scanner or, where the terminations are accessible, by carrying out a visual inspection.
When working on faulty cables, rather than simply assessing the condition of a cable during commissioning, fault pre-location becomes the primary issue. It has proved possible to energise cables up to 400 kV using damped AC techniques, and reflection (TDR) measurements have been performed on very long cables. ARM (arc reflection method) measurements at distances up to 25 km have been successfully achieved.
HV measuring bridges offer an alternative and complementary approach to fault pre-location and, with suitable adaption to cope with the high discharge energies, can presently be used on cables up to 30 km in length. It is worth noting that, while bridges are primarily intended for the pre-location of core-to-core and core-to-screen faults, they can in theory also be used for pre-location of sheath faults and external cable damage.
As an indication of what is possible at the present time in the field of subsea cable testing, it is interesting to examine the trials carried out jointly by Megger and Statnett, the operator of the Norwegian energy system, in September 2013. These involved the 580 km long NorNed HVDC cable that runs between Feda in Norway and Eemshavan in the Netherlands. The cable operates at ± 450 kV, giving it a terminal-to-terminal voltage of 900 kV, which means that the system includes HVDC converters with the highest voltage rating of any in the world.
The NorNed cable has been in commercial use since 2008 and advantage was taken of a scheduled out-of-service period to carry out trials with a Teleflex VX time domain reflectometer from Seba KMT, a Megger company. The results were impressive as the instrument was able to clearly “see” the end of the cable, 579.8 km from the point of connection. At the time of writing, this is a world record.
As we have seen, countries in Europe and further afield are becoming more and more dependent on power transferred via subsea cable connections, yet all too little thought has been given to maintaining and fault finding on these connections. Cable test technology has already risen to many of the challenges involved but much more could and should be done. Progress will require commitment and investment by the cable operators. Understandably the operators find the prospect of spending money without an apparent immediate need unattractive, but they would do well to bear in mind that their investment in test equipment development will be recovered many times over if the new equipment eliminates only a comparatively short period of cable downtime. They should also take into account that, even when the test technology is feasible – and in most cases it is – it is still impossible to develop and deliver innovative test equipment overnight to address an urgent requirement.
In short, cable operators with the foresight to make modest investment in cable test equipment development today will see big dividends tomorrow!