Hurricane Sandy and the documented explosion in a New York substation have reminded us all of the dangers associated with failures of the distribution network. Fortunately, on this occasion, there were no reported injuries as a result of the blast. But it is well known that faults on the network, whether caused by man or natural phenomena, can result in the flow of very large currents until the supply can be interrupted by the safety mechanisms in place.
Although a vital task, determining the fault currents that can flow on an electrical network in the event of a fault has always posed a challenge to utilities. Accurate knowledge of the maximum current that could possibly flow, known as the fault level, is essential to maintaining safety on the network, to ensure optimal operation and to optimise investment in infrastructure.
The safety of the transmission and distribution networks depends on circuit breakers to interrupt and manage safely the electricity supply to affected sections of the grid in the case of a fault. These circuit breakers must be of a high enough capacity to make or break the flow of current even under the worst possible fault conditions. The network infrastructure must also be able to withstand the currents that flow during the period before the circuit breakers isolate the affected network to prevent destructive failure.
A circuit breaker is designed to interrupt or connect to a certain maximum current known as its fault level rating. If the actual fault level should be found to exceed this rating or more capacity needs to be added, the network operator must reinforce or reconfigure the network to ensure continued safe operation. However, upgrading of switchgear, cables or transformers, for example, are all costly and potentially disruptive activities. It is essential to know the fault level to maintain efficient, safe and predictable network operation.
This requirement to consider fault level is becoming increasingly relevant to today's network operators. Since the amount of distributed generation (DG) on the network is on the rise, the network is being pushed closer to its existing fault level limits.
Existing fault calculations fall short
These limits are calculated by utilities in two ways or through a combination of both. First, by applying 'rules of thumb' calculations based upon the known contribution from the network's major elements, such as transformers. And second, by creating a detailed, analytical model for the relevant section of the network – which provides an accurate value for the fault current but is complex and time consuming so has been used mainly at higher network voltages, such as for 33 kV and above in the UK.
|Outram FLM's output can be sent to portable devices, such as tablets, via Bluetooth|
Network operators face rising pressure to deliver a continuous, low-carbon supply of clean, stable and value-for-money electricity. Unfortunately, as this pressure mounts the job gets harder, as Infrastructure is often ageing and becoming increasing unsuitable.
Not only are non-linear loads increasingly being connected to the electricity network, producing harmonics as well as other undesirable network effects, but so is DG such as combined heat and power plants, solar panels and wind turbines.
Adding DG increases uncertainty by not only taking power from the grid but also adding it back. Fault levels are continually changing, which means that rules of thumb go out of the window and models cannot with conviction take all this changing activity into account.
Search for a new solution
Following a discussion with the distribution network operator (DNO) Scottish Power Energy Networks (SPEN) on the merits of our power quality measuring products, SPEN asked whether we could design a product that could measure fault level from natural disturbances.
|The Outram FLM is a highly versatile instrument because it can be deployed at any voltage level|
For the several years we have been working on a portable instrument that can predict fault current by measuring natural disturbances on a radial network during normal operation. The Outram fault level monitor (FLM) is a compact, portable device that takes measurements of natural voltage and current disturbances at a specific point on the network and predicts a fault level value for that point. This involved developing a completely original set of algorithms and implementing them on a hardware platform, the 600V Category IV, Outram PM7000 power quality analyser.
It works using a novel, patent-pending algorithm to determine worst case fault current, separating out peak 'make' current, RMS 'break' current and contributions from downstream motors. This is achieved by extracting the tiny characteristics of interest from voltage and current signals at the arbitrary times at which natural disturbances occur, such as tap changes and load switching.
The instrument distinguishes the various components capable of delivering the fault level results, processing these mathematically to optimise signal to noise, then filtering and computing the desired fault level values from these data.
This means that high-quality results can be obtained from voltage disturbances of as little as 0.15 per cent. No previous knowledge of the network is required and the FLM can be deployed at any voltage level, assuming the necessary CT/VTs are in place.
A succesful testing programme
Lab trials produced both accurate and encouraging results. With partial funding from Ofgem's Innovation Funding Incentive, field trials were conducted with SPEN to compare results generated by the FLM with fault currents calculated by analytical modelling.
The results compared very positively with SPEN's detailed IPSA and DigSILENT models. They were consistently within 3–5 per cent of the modelled values showing that the Outram FLM can accurately calculate fault current in real-world applications.
To produce these results, the FLM must rely on voltage disturbances seen on the network during normal operation, so it stands to reason that if there are no disturbances then no fault level prediction can be given. The results generated for the user are therefore ascribed a confidence level dependent on the magnitude of the disturbance seen. Whilst voltage disturbances of <0.15 per cent are given little significance, the larger the disturbance the higher the weighting given for the instantaneous or composite results.
The relative accuracy of the Outram FLM result versus the model result ultimately depends on sensor accuracy and noise for the FLM, and the fidelity of the network characterisation for the computer model.
Clearly, where network modelling is poor or non-existent and sizeable disturbances are plentiful, the FLM will produce superior results. Conversely where networks are interconnected and there are multiple current paths, or suitable disturbances are infrequent, a single FLM operating only on natural disturbances is likely to be inferior to a well characterised model.
But while interconnected network complexity increases the difficulties for both the model and FLM (the latter requiring multiple synchronised devices), the introduction of artificial disturbances on a known single current path is expected to cut through the uncertainty and embrace all current sources, enabling the FLM to produce correct results in the most complicated of interconnected arrangements.
Network disturbance on demand
Another UK DNO, Western Power Distribution (WPD), is planning to take the Outram FLM one step further. With funding from a Tier 1 Low Carbon Networks Fund project, WPD has a vision for an active network management solution. With the benefit of further communications developments on the FLM, WPD aims to introduce a means of supplying it with a network disturbance on demand, so that a fault level value with a high confidence rating can be obtained in real time. WPD has identified a real network environment in which it aims to trial active fault level-based control of a DG connection.
Stage 1 of this project, 11 kV lab testing, has already been completed. WPD carried out trials with the Outram FLM and an IntelliRupter PulseCloser, a unique alternative to conventional automatic circuit reclosers, manufactured by S&C Electric Company, at its Advanced Technology Center in Chicago, US – by comparison the IntelliRupter significantly reduces the amount of energy the network is exposed to when testing the circuit to see if a fault is still present.
In the lab, non-customer affecting quarter to half-cycle current and voltage disturbances, generated during the IntelliRupter's pulseclosing operations, were used to enable the FLM to predict the fault currents produced by separate, deliberately instigated and independently measured bolted faults produced under identical source conditions. During the trials it was therefore possible to compare the predicted fault level with the actual measured peak current seen in response to the bolted fault.
These trials were successful, the results showing that the FLM predictions were between 2.5 per cent and 5 per cent of the current measured by lab instrumentation – an error margin seen as very hard to improve upon because noise on the network and sensor deficiencies will always produce some degree of uncertainty.
WPD's ultimate aim is to further develop its ability to actively manage its distribution network by optimising the network arrangement through predicted fault level values.
Satisfying a growing requirement
Following the successful trials, two clear applications for the Outram FLM have emerged for both DNOs and other network operators, large or small. The first is to validate and refine existing models by obtaining greater visibility of actual system fault levels and the contributions from customer equipment. The second is identifying the fault level in areas of the network where models do not exist or it is difficult to build them, such as in sections of the 11 kV or low-voltage network.
The validation or prediction of unknown fault level values will deliver many benefits and therefore ensure a considerable return on any investment. Network operators will be able to reduce health and safety risks by identifying and subsequently managing fault level issues previously unknown due to inaccurate or non-existent models.
|The Outram FLM is able to accurately determine worst-case fault currents by separating out peak 'make' and RMS 'break' currents|
Financial rewards may be achieved through optimal operation or interconnections on the network, which previously could have been restricted because of perceived fault level issues. The DNOs should also be able to improve their regulatory performance through faster rectification of fault level issues following more accurate monitoring.
There are not just financial and operational but also environmental benefits to be realised. Measurement of the real day-to-day impacts on fault level could enable faster connection of renewable generation, which is important with today's emphasis on low-carbon generation, and can also delay the premature decommissioning of equipment through unnecessary fault level upgrades.
Fault level is a parameter much needed by network operators across the globe. Being able to predict it on any network will bring many benefits, not least enabling the presently assumed safety margins to be realistically assessed. The Outram FLM provides a unique solution to this need, which has been largely unfulfilled.
We worked very closely with SPEN on the requirements and delivered a solution with all the functionality required by operators worldwide.We are currently in talks with other DNOs to deploy the Outram FLM in a number of further sites.
Outram Research limited would like to thank the following for their help and support on the projects mentioned in this article: SPEN, WPD, S&C Electric Europe Limited and Parsons Brinckerhoff.
John Outram is managing director and founder of Outram Research Limited. For more information, visit www.outramresearch.co.uk.