Power network monitoring avoids blackouts, saves millions
Partial discharges (PD) result from faults in the insulation of high voltage power systems, and produce low-level, electrical signals which can only be detected by sensitive monitoring equipment. If the signals grow in intensity, then this is a sign of imminent failure, and the power system equipment producing the signals should be taken off-line, and repairs instigated before more serious damage, such as a short circuit and consequent blackout, occurs. New techniques to monitor power networks for impending faults due to PD have saved several large organisations from the costs of expensive, embarrassing blackouts. Two companies which have emerged from this research have saved clients, including the London 2012 Olympics, steel manufacturer CSC and UK power companies, from high costs related to unexpected power failure.
For over 20 years several groups in the University and more recently the Power Conversion group and the Electrical Energy and Power Systems group, have researched methods for detecting faults in power networks, knowing that the investment in detection systems is always orders of magnitude cheaper than the costs of failure.
In the 1990s the initial research team showed how PD analysis techniques could identify faults in electricity cables. The research led to the creation of the first spin-out company, IPEC in 1995.
A second spin-out company was created in 1998: High Voltage Partial Discharge Ltd (HVPD). HVPD soon became a leader in the growing global market for on-line PD testing and monitoring, and incipient PD fault location technologies for high voltage electricity cables. HVPD exports its products to partners in 80 countries.
IPEC and HVPD commercialised a software simulation package, developed by the University, which helps to optimise actuators used in medium voltage switchgear and autoreclosers. HVPD, meanwhile, in collaboration with the University has developed a system to monitor partial discharges in high voltage motors operating in remote hazardous conditions, such as petrochemical plant.
During the 2012 London Olympics technology from IPEC continuously analysed 62km of cable and 170 ring main units for PD faults. The systems found five faulty ring main units during the Games; twice daily checks on these units ensured a constant flow of power and avoided embarrassing and costly interruptions to live TV.
Together IPEC and HVPD now employ 59 staff and boast an annual turnover of almost £4 million.
Similarly, UK Power Networks uses a PD monitoring system on more than 1000 cable circuits installed across its network. So far the company has saved an estimated £900,000 following preventative action on six detected faults and the replacement of 12 cables in its network.
Further afield, in Taiwan the CSC steel plant used IPEC systems to check its 33kV cables which were over 25 years old. Unsurprisingly the cable analysis identified a faulty joint, which was quickly replaced. Servicing this fault eliminated the risk of shutting down the plant, saving the company an estimated $1 million.
Related research projects in the Power Systems group developed techniques to test and verify the safety of electrical power components in aerospace applications. The techniques search for faults in cabling, electrical machinery, de-icing systems and a range of other components. Following this work the researchers made a substantial contribution to a key technical report which has become the pre-cursor to a full technical standard on the use of high voltages in aerospace systems.
The Power systems group has also performed ageing tests on electrical equipment in the airbus A350. Airbus and Boeing suppliers, including Moog, GKN, Ultra Electronics, Goodrich, Rolls Royce, and Liebherr – use these tests on their components.
Ongoing research within the School focuses on further developing PD detection and analysis techniques that can discover defects in the insulation systems of power networks.
The original studies found a method of distinguishing between discharges arising from different high voltage sources. The method involves the conversion of acoustic emission signals from the time domain to the frequency domain using fast Fourier transformation and recognition of the signature frequency spectra.
The researchers developed fast Fourier transformation and neural networks to distinguish internal flaw signals from ultrasonic and acoustic background noise. Applied to substations, the methods allowed data to be collected hourly and used for data mining and correlation pattern recognition.
The team also developed a novel software simulation package to optimise actuators in medium voltage switchgear and autoreclosers. The package simulates the actuator electrical, magnetic and mechanical effects, using a ‘lumped reluctance model’ to optimise the device. This approach overcame many of the difficulties which arise from using finite element analysis.
Recent work in the Power systems group has expanded the focus of research to aerospace components because aircraft increasingly rely on electrical and electronic systems rather than mechanical components. Researchers have studied the mechanisms that underlie the susceptibility of components to degradation and from this work have devised techniques to assess the probability of component degradation and failure.
Studies have also looked at how to calculate the safe voltage rating of unscreened insulated wires within an aircraft. The researchers concluded that the optimal operating point for an aircraft power system requires a trade-off between wire weight and power transfer.