Safe Forever?  Electrical Protection Methods
CEE Relays Ltd presents a brief history of electrical protection equipment; how electrical fault detection has improved and how it might change in the future.
Fuses
Fused distribution board
Fuses are some of the oldest and most straightforward forms of electrical protection. They remain popular because of their ability to limit fault energy; once enough energy has passed to melt the fuse, the circuit is interrupted.

For fault finding, fuses present a challenge because they really only protect against high-current faults. Other problems such as short-term overloads, frequency problems or short circuits on high-impedance earthed systems will all be overlooked. Locating a fault is very hard because there is no indication of which type of fault has occurred, nor the ambient conditions that prevailed prior to the failure of the fuse. For example, sustained voltage dips may cause equipment to draw extra current and ultimately blow a protective fuse even though the root cause may have nothing to do with the equipment supplied by that fuse. Fuses can also be pre-stressed by fluctuations in ambient temperature, making them more sensitive.
(Left): Fused circuits have been a common means of protecting electrical distribution systems for many years
Electromechanical Relays
Electromechanical relay
The first electromechanical relays were simple (definite-time) attracted armature devices for detecting overcurrent and tripping circuit breakers. Their only real advantage over fuses was that they could be re-set once a fault had been cleared. As networks became more complex, relays were re-designed (circa 1920) to allow selectivity between up-stream and down-stream devices. These induction-disc relays operated more quickly for high currents than for lower currents; their characteristics are the origin of the IDMT curves we use today.
Selectivity benefits fault-finding by tripping only the relay closest to the fault; this not only makes it easier to determine which part of a network experienced a problem but also gives back-up protection by letting up-stream devices operate when a circuit breaker fails on a final circuit. Since only problem circuits are isolated, healthy plant will not be isolated during the fault.
(Right): Electromechanical relay with the induction disc visible in the centre.  Relays like this often had transparent covers so that the mechanism could be inspected easily.
More advanced electromechanical relays were developed to detect vector-shift, voltage and frequency faults; issues which would have previously been undetectable and impossible to trace. Auxiliary contacts could also be added to send remote alarm signals to a control room.  

As more protection relays were added to networks it became necessary to add simple logic to trip circuits to manage the interactions between all the different relays. These can improve fault-finding by, for example, inhibiting a trip unless two protective devices have detected a possible fault. 
Electronic ("Static") Relays
Analogue electronic relays using vacuum tubes have existed for almost as long as electromechanical relays. They were useful because they placed a lower burden on current transformers and contained few moving parts. However, these devices were never popular; perhaps because it was impossible to physically see them operating.
Static relay
Transistor relays first emerged in the 1950s but is was several decades before they became the cheap alternative to electromechanical relays. For fault finding purposes they are useful because they allow finer control of setting ranges and higher accuracy than electromechanical equivalents. Since they are not subject to mechanical wear, they require far less preventive maintenance than their electromechanical predecessors.  

Other innovations to make fault finding easier were made around the same time. Removable, rack-mounted relays could be easily replaced or taken out of a switchboard for bench-testing. Switchboards were also built with test points so that control circuits could be measured with an external meter in the event of a fault. Some later relays also have indicator lamps to allow technicians to easily see when a fault has been detected.  
(Left): One of CEE's many static relays
The logical signals in trip circuits changed very little even after the introduction of static relays. Finding the source of a trip in a complex circuit still relied on checking which relay “flags” had operated, or in their absence, which contacts were open and which were closed; in more extreme cases by physically tracing panel wiring. 

As was noted in the July 2018 newsletter, static relays are still used extensively in offshore, nuclear and military installations. For more information, see our 7000 series page.  
Multi-function Microprocessor Relays
Microprocessor relays were initially introduced to make electrical protection systems more compact by performing several protective functions in a single relay. For motor protection applications they make fault finding easier by closely modelling the thermal characteristic of the motor being protected; therefore minimising the number of spurious trips. Another useful fault-finding feature is the built-in memory of past alarms and trips which can give clues to how a fault emerged. Unlike the mechanical “flags” on static relays, the events recorded by microprocessor relays can also be time-stamped to show not only that a fault has occurred, but when it happened.
(Right): A microprocessor relay
Microprocessor relay
The adaptability of microprocessor relays proved useful since devices could be programmed with more than one setting group for the same protection functions. This is important in multi-incomer systems where the down-stream protection relays have to adjust their settings depending on how many power supplies are in service.
Microprocessor relays with their own built-in logic gates and timers help to minimise the number of external auxiliary devices are needed. However, since they are internal to the relay it is impossible to physically measure the status of logical inputs and outputs which can sometimes make fault finding harder. Relays which can communicate over serial networks can speed up fault finding by sharing real-time information with a central control (SCADA) system. However, these networks have never been considered sufficiently reliable to provide “trip” signals and therefore trip circuits have historically remained hard-wired.
Microprocessor relays have been a staple of electrical protection systems for some time. CEE’s offerings include the Procom series and the NP800 series.
"Smart" Relays
The very latest protection relays such as CEE’s NP900 series are considered “smart” devices since they are self-adjusting, may be programmed remotely and have many advanced protection features contained in a single package. Like the earlier microprocessor relays they have internal logic features and timers; however the status of logic signals can be interrogated in real time. This is very beneficial to the process of fault finding since operators can now see every step of the “trip” and “alarm” logic via a network without needing to be physically present. Signals can also be simulated for testing and commissioning purposes.
 
Smart relays have radically changed the process of detecting a fault on an electrical network. Up until this point, faults could only really be detected after they had occurred; either a relay has tripped or it has not. Now, emerging faults can be detected by remotely monitoring electrical quantities. Examples of emerging problems include monitoring equipment which draws increasingly high currents on start-up or circuit breakers which take increasingly long times to open or close. Relays with disturbance recorders can also keep an electronic record of measured electrical quantities (current, voltage, frequency etc.) before, during and after a trip. These records make the true cause of a trip much easier to identify and can even be re-created on a test set to simulate the same fault for test purposes.  
New systems of sharing information and signals between relays are now sufficiently reliable that they may be used to send “trip” signals instead of the traditional hard-wired trip circuits. Our January 2018 newsletter explained how the (GOOSE) messaging system has self-supervision features and sends regular status updates between devices. Advances in networking hardware such as the use of PRP networks are more resilient against damaged signal cables or optical fibres since they provide two alternate pathways for sending the same signal.
The Future of Protection
There will always be some room for improvement in electrical protection, depending on what the industry needs. These are some of the ways in which future protection relays might develop: 

Increased speed. Keep doing what you are doing, but do it faster. When there is a genuine fault, it must be isolated as quickly as possible to prevent damage to equipment or harm to personnel; this is especially true of arcing faults. Faster relay processors and faster trip signals will help to cut trip times. 

Machine Learning. As far back as the IMM7000 relays, there was some capacity to measure network conditions and adapt relay behaviour. In this example, the IMM relays “learned” the thermal time constant of the motor they were protecting. In the future, it is likely that protection relays will modulate their responses based on conditions locally and elsewhere in the network. For example, settings could be lowered if available generation reduced or the utility fault level dropped.  

Inter-connectivity. Networking of relays for the purpose of remote monitoring and control is now standard practice for new installations; thanks to the IEC61850 / GOOSE protocols, information is also shared between the relays themselves via communication networks. As utilities try to match electricity supply and demand more closely, perhaps sending real-time control and monitoring information from consumers to grid operators will become compulsory? For example, electricity companies might require advanced warning before a consumer starts a large electrical machine.