Microprocessor Protection Devices: the Present and the Future

The paper presents the analysis of the basic constructive disadvantages of the present day microprocessor-based protective devices (MBR) and offers the basic principles for creating a new MBR that can be used in newly constructed devices.


Introduction
Over a long period of time, electromechanical protective relays of the past generation were meeting completely all requirements set for the protection devices of electrical power equipment. In the latest microprocessor-based devices the function of relay protection has been united with functions of other devices: communication and data transmission devices, fault recorders, substation logic units, etc. Such multipurpose complexes are compared now to the individual one-functional electromechanical relays which have been functioning for decades and which are now considered as rather worn out. The advantages of microprocessor relays over electromechanical ones are quite obvious. What has been overlooked, though, is that these devices perform completely different functions which simply cannot be compared with each other [1]. Specialists of the world-leading manufactures have published a considerable number of half-advertising articles stressing only the positive qualities of microprocessor-based protective relays (MBR). The literature provides but a few publications devoted solely to the analysis of problems related to the transition to MBR although there are many more others. Though the problems related to MBR expansion are obvious, the fact that their distribution is increasing and that they are fully replacing electromechanical relays is inevitable owing to the fact that almost all world-leading manufacturers have stopped producing electromechanical protective relays. This is not due to the fact that electromechanical relays have insuperable disadvantages of any great importance (they have just worn out and have not been improved for the past 30 to 40 years), but owing to the enormous profit (gained by MBR producers) which is beyond comparison in relation to the manufacture of electromechanical protective relays [1]. As the future of relay protection is inevitably connected with MBR (at least, for complex protection functions), the forecast of ways of development of this equipment is of definite interest. one say about the MBR's so called 'ability' to constantly control even the serviceability of output electromechanical relays other than to be an advertising gimmick? It would be very interesting for us to learn how is it is possible to control the serviceability of the electromechanical relay all the time without causing it to operate. Our research has shown that the matter is about the control of a relay coil's continuity carried out by letting weak current flow through it, not causing the relay operation thereby. Is it the relay coil but not its contacts that is the most loaded and most unreliable element? Certainly not, but it is much easier to organize a useless control of the coil than the control of the condition of the contacts. What is controlled is apparently absolutely unimportant for a new advertising slogan.

2.2.
The MBR printed circuit board (PCB) is more commonly implemented by using SMD (Surface Mounting Devices) technology; as the density of the SMD micro-components on the board is so high, Fig. 1, it is absolutely hopeless even to discuss troubleshooting and repairing such a PCB. One can only remove it and replace it with a new one. Let the reader imagine the approximate cost of the relay REL-316 PCB with the CPU Intel-486 type, represented on (Fig. 1).
Nowadays there is a tendency to place SMD micro-components together with large power elements, electrically connected with each other on one PCB (Fig. 2).
With regard to the fact that manufacture companies never give the consumer detailed circuit diagrams of the PCBs, searching for malfunctions in them and repairing them turn out also very difficult. One often has to remove it , despite the number of large discrete components usually found on them. The tendency of increasing the density installation is peculiar to PCBs on the basis of usual discrete elements. As for power supplies functioning at relatively high voltages and powers (Fig. 3), installations such as these result in a dangerous approach to printed conductors having different potentials on the board and increase the possibility of electrical breakdown of dusty PCBs with the increase of humidity. In addition, dense installation makes thermal conditions of electronic components work more difficult, which leads to a considerable shortening of their operating life. It is not at all simple and cheap to search for malfunctions and repair PCBs due to dense installation such as these, and one should observe substantial expenditure of time this entails.

2.3.
A serious problem is represented by the electrolytic capacitors widely used in MBR power supplies. After 7 -10 years of operation in switching highfrequency power supplies, even high-quality capacitors produced by well-known Japanese companies, begin to leak, emitting a chemically aggressive electrolyte. Even with this, the power supply functionality continues until adjacent micro components are destroyed or copper tracks of printed circuit dissolved. Even the layer of special firm lacquer called "mask" cannot protect them (Fig. 4). Nobody would dare guarantee the good condition of the unit even after attempts have been made to rectify the damage and to put such crucial device as MBR into operation again after damages of this kind have been made.

2.4.
Only the desire for decreasing the MBR dimensions can explain the fact that almost all the world-leading producers of MBRs use miniature electromagnetic relays for the direct switching of trip coils of high-voltage circuit breakers and to control rather powerful external intermediate relays.
As it is shown in [2][3][4], the technical specifications of a vast majority of such relays fall short of their real operating conditions in the MBR. Naturally, this reduces protection reliability.

2.5.
In some types of MBR produced by world-leading companies, small connectors (slots) are found even in 250V voltage circuits not intended for functioning at such voltages. In order to decrease the risk of electrical breakdown between close outputs of miniature relays and connectors, one sometimes makes special cuts in PCB of MBR between the outputs (Fig. 5).

2.6.
Constantly increasing the density of installation considerably affects electromagnetic compatibility, especially in the atmosphere of the increased risk of deliberate remote impact on the MBR by very powerful directional electromagnetic pulses [5]. Special powerful pulse, ultrahigh-frequency generators ( Fig. 6) allow disabling of the existing MBRs instantly from a distance of several dozens of meters to half a kilometer even from a driving car, let alone military target killers like electromagnetic bombs and warheads which anyone can freely buy at arms market today.

2.7.
It is absolutely obvious by now that constantly increasing the concentration of functions of relay protection in one device along with simultaneous increase of loading factors of electronic components does not at all favour the growth of MBR reliability. In this case, only one component serving any general purpose for the whole MBR unit may fail (central microprocessor, memory, power supply, etc.) leading thus to false switching off of an important power object or serious damage due to the fact that MBR has failed to respond in an emergency situation.

Fig. 5 -Fragment of MBR PCB with cuts made between close outputs of small connectors (1) and miniature output relays (2)
with the view of preventing of electrical breakdown. of an MBR of other type. Moreover, even the substitution of an MBR of one producer for the MBR of another producer at functional substation causes difficulties due to great variety of forms and dimensions of these devices (Fig. 7). The input and output of this circuit through the insulation optrons Opt1 and Opt2 are cut into the MBR electrical circuits connected with microprocessor. A signal coming from optron Opt2 logically simulates (repeats) the presence or lack of input voltage. The functioning of this circuit may be blocked on a software basis. With this, the blocking signal is sent to the inhibiting input of a cell (Opt1 input) from the microprocessor through corresponding electronic circuits. The problem of this logic cell consists in the high level input signal (220 -250 V DC), which should be reduced to the 1.5 -2 V at which the optron, Opt2, functions. The common current consumed by the circuit is divided between resistors R1 and R2 into approximately equal parts, this being the reason why they both, in fact, should have equal power. However two large resistors (like R1), set at a distance sufficient to ensure cooling, cannot be placed together on board. That is why the MBR designers decided to give the functions of an element dissipating surplus power to transistor VT which does not function in a key mode (as it is common in such circuits), but in a booster mode. Certainly, in such case, resistor R2 has low power and small dimensions. It is an original decision, but, in practice, this fall on miniaturization leads to serious problems: often resistor R2 burns out completely and this sometimes results in burning out of all parts of the board as well as the adjacent elements. The problem lies in transistor VT.
Constantly functioning in the booster mode and dissipating surplus power, transistor VT heats up to 70-80ºС. Unlike a normal resistor, in which the resistance grows insignificantly with increasing temperature, the resistance of the direct transition of transistor decreases significantly, as a result of operating point shift at the characteristic and amplification factor increases. This leads to collector current increase, i.e., current through resistor R2. When the simultaneous heating of many R1 resistors and VT transistors in ten-fifteen input circuits occurs, the temperature may rise significantly in part of the MBR case (separated by partitions) in which the board of digital inputs is installed. This leads to a further shift of the transistor operating point and a current increase to its collector completes the burn out of resistor R1. In this situation, diode VD, which according to the scheme should stabilize the transistor operating point, appears to have little effect, as its temperature and the one of the transistor differ by 50-60 degrees. Thus, the producer's desire for reducing the MBR dimensions using this transistor instead of powerful resistor for dissipating surplus power has led to the deterioration in MBR reliability. The problem of insufficient reliability of this scheme connected with miniaturization lies not only in the difficulty described above. One more reason discovered in practice for burning the resistor R2 is poor level of insulation of miniature capacitor C. Partial breakdowns in its inner insulation and an increase of leakage current at a constant voltage 200V also leads to resistor R2 fusing. These breakdowns can also be due to the insufficient efficacy of varistors RU used in the MBR -the smallest (for saving room) and, therefore, the most low-powered of all existent types. Owing to their size they are not always capable of absorbing the energy of surge discharge. Additionally, it is not clear what protects what: does the varistor protect the electronic components or vice versa, as the clamping voltage of the varistors (650V) installed on the board significantly exceeds the voltage rating of electronic components, e.g. 350V for transistor VT. With such parameters ratio, the electronic components will fail until varistors are actuated. 2.10. In order to increase the reliability of the MBR power supply, the direct operational current is used as the rule of a thumb. However, even using direct operational current does not exclude malfunctions in the accumulator circuits. For example, the supply of the direct current bus bars at the substation is provided only by a charger supplied from the direct current mains. In case of emergency conditions in high-voltage power mains protected by an MBR, the supply of the charger disappears and the MBR loses its power supply at the most crucial moment. Our research [6] has shown that under real operating conditions, MBRs of different types have time to generate a signal for switching the circuit breaker off only if they work in mode of momentary cutoff, i.e., without time delay.

3.1.
In our opinion, the basic design principle for future MBRs should be similar to modern PCs: hundreds of varieties of containers, motherboards, storage units, and peripheral devices of many different producers perfectly combined with each other in such manner that they are interchangeable and allow upgrading certain units without affecting the other ones. Why not applying the same principle to new MBRs? We see the MBR of the future as certain functional modules implemented on the printed circuit boards in a single constructive style (standard) provided with plug-in connectors.
We think that one should transfer this principle from the sphere of PCs also into the sphere of software: a basic "relay" prototype of Windows TM and a set of applications realizing functions of concrete types of protection on a general basis is the most perspective line of software development.

3.2.
MBRs should be placed in special metallic boxes (Fig. 9), produced according to a special technology ensuring reliable protection against external electromagnetic radiation over a wide frequency spectrum. Nowadays such boxes are produced by many companies, for example, R.F. Installations, Inc., Universal Shielding Corp., Eldon, Equipto Electronics Corp., European EMC Products Ltd., Amco Engineering, and many others. The boxes should have guides for inserting the printed circuit boards and corresponding connectors for connecting the boards. One such box may contain several different MBRs. Protection relays occupy a very small area of substation or station; therefore the problem of limitation of space occupied by these boxes should not be the prime consideration of new principles of protection devices design. On the contrary we think that the density of installation of PCBs implemented on usual elements should be reduced so as to allow a semiskilled technician to search for malfunc-tions and repair the device without difficulties. Trends of this kind, but not the above-mentioned FUF, should become the criterion of MBR technical level.

3.5.
The main processor unit with memory and all auxiliary elements should be placed in a separate shielding section and be galvanically insulated from all other units by means of optical coupling. 3.6. One need not change the board of the main processor, connection and data transfer board, board of emergency situations meter implemented on microcomponents of surface mounting. The rest of the boards should be implemented using the usual components and the design of the PCB should be such that it can be searched for malfunctions and repaired. It is obvious that this requirement will be urgent until every MBR board is extremely expensive but board repair remains profitable.

3.7.
Though there is a possibility of uniting functions of different types of protection devices placed in one box and to realize them on one general, powerful microprocessor, this possibility, in our opinion, should not be actualized because of the risk of excessive concentration of many protection functions in a single device. Even so, the registration of emergency modes and the activation of all protection devices installed in the box may be implemented on microprocesor and general flash-memory for the whole box. The connection system and data transfer can also be general for all protection devices installed in the box.

3.8.
Input transformers of current and voltage (analog input units) are highly reliable elements of MBR and one has not discovered cases of their damage in process of MBR service. This is why their power can be rated for share usage by all MBR devices installed in the box. Units of analog inputs should have functionally completed construction which allows removing the whole unit from general scheme of the MBR and replacing them with blocks of totally other constructions, for example, with optical inputs designed for receiving optical input analog signals from new generation optical transformers of current and voltage. 3.9. Aluminum electrolytic capacitors of power supplies should be placed in a separate block (PCB). This block should be replaced with a new one every 5 years regardless of the capacitors condition.

3.10.
Power supplies should be general for the whole box. The power supply should contain two separate independent units: basic and reserve (which begins to function automatically as soon as basic unit fails). In addition, the power supplies should contain small hermetic non-serviced accumulators with charger as is done, for example, in fire alarm systems. It is rather instructive in this case of using this type of accumulator instead of such accumulator electrolytic capacitors of large capacity at voltages not less than 450 -500V. In North Caroline University one developed capacitors with accumulated energy which exceeds energy of usual capacitors by 5 -7 times [7]. It is also possible to use super capacitors which ensure power of relay protection box during some period of time sufficient for protection actuation in case of failures in centralized power system [6]. Power circuits of every separate module should be disconnected from power circuits of the next module such that damage in one of them cannot affect the functionality of the other. 3.11. Relay protection boxes should have a board-module with highperformance, high-reliable electromechanical elements implemented, for example, on reed switches or hybrid reed switch-semi conductive elements. Such elements as these will ensure high-performance actuation of output protection relays around microprocessors at current big multiplication factors for protection device input, when the mode is definitely determined as emergency, and no other processing of input signal is not needed to determine mode as an emergency. As is shown in author's works [8,9], simple modules on such elements have high performance (0.8 -1.5 ms), which significantly exceeds the highest performance of MBRs (20 -40 ms).

3.12.
As it is shown above, the analysis of MBR condition only by means of imbedded software, as it is currently done, is definitely not effective. On the other hand, real self-diagnostics, including the special sending of standard test signals automatically to analog and digital inputs of an MBR from an external source, and registration of changes in condition of relay and its outputs circuits is impossible without the complete disconnection of relay. In our opinion, the possibility of periodic automatic disconnection of an MBR with further automatic actuation of diagnostic system on active objects of electric-power industry is rather doubtful. Besides, to check certain functions of relay it is often needed changing (or canceling) some protection settings with further return on test completion. Complexity and costs of the system which provides checks such as these will be similar to cost of MBR itself. If we add the inevitable reduction of reliability to this whole complex because of its complication, and also the necessity of periodic calibration of diagnostics system, the hopelessness of this way of development is obvious, at least, in the nearest future. In our opinion, a much more attractive idea is the creation of a portable, mobile universal automatic diagnostic complex controlled by means of an external portable computer and a set of specialized programs for every type of protection, which operating in a half-automatic mode would take 15 -30 minutes to effectively test the MBR. In general, a good basis for such diagnostic centers already exists: it is devices produced by companies such as Omicron, Doble and others. It is necessary to provide these devices with multi-channel interfaces for the synchronous connection of numerous inputs and outputs of the MBR, and also to develop a set of application programs, which ensure testing different MBR in an automatic mode. We think that with such diagnostic complexes, it makes sense to return to periodic check-ups of MBR with economically motivated intervals.

Conclusion
The principles of MBR construction that have been discussed can ensure, in our opinion, the increase in its reliability, flexibility and usability. This can provide opportunities for modernizing protection with minimum costs. If some MBR units function unsatisfactorily, one can easily replace them with the units of another producer. Having decided to use new generation microprocessor, one should not replace the whole device, including input circuits, power supplies, filters, etc., but only one or two PCBs that need to be changed. Introducing the use of optical transformers of current and voltage will involve changing only one PCB of analog inputs, but not the whole protection device, as it is required today. Likewise, one could change and modernize protection software. All this will lead to the rise in the number of new companies in the market specialized in production of separate functional blocks and separate application programs, as well as in the competitiveness, and, finally, to rash quality improvement of MBR with simultaneous cost reduction.
The author understands that not all aspects of the problem have been addressed and not all principles of construction of future MBRs formulated. Therefore, he invites all interested specialists to discuss this topic.