Fullscreen HTML5Med Res (??MB)HTML4

PT.10: ELECTRIFICATION IN CENTRAL EUROPE THE EVOLUTION OF ELECTRIC RAILWAYS In this chapter we look at the electrification of Central European railways, ponder on some interesting synchronization problems and discuss the Germon development of cycloconverters. By BRYAN MAHER Switzerland, lacking indigenous coal supplies, had experienced great difficulty supplying her steam locomotives during the 1914-1918 war. Consequently, as soon as peace reigned again in Europe, the Swiss Federal Railways (SBB) set about electrifying all their main lines. Taking the highly successful example of the Bern-LotschbergSimplon railway system and the current work then developing on the Lappland Railway (see previous chapters), the Swiss Federal Railways also opted for a 15kV, 16.6Hz AC system, using series AC motors. Even though such motors do have a commutator and brushes, necessitating regular maintenance, the very high starting torque and variable speed characteristics were predominant advantages. Beginning, as was the custom of the time, with rod-drive locomotives, the Swiss Federal Railways built particularly powerful locos. This was by necessity, as the main line around their country has only two choices: go over every mountain in its path or go through them, steep gradients of 1 in 40 or more being common. Because one of their most successful steam locomotives in the pre-electric era was the 1913 type 2-10-0 design, the Swiss elected to build electric locomotives which also had large numbers of driving wheels. Their first example was an articulated electric locomotive having a two-piece hinged mainframe with one cab riding over all, the wheel arrangement being equivalent to 2-6-6-2; ie, twelve driving wheels. In electric locomotives this arrangement is called a "1-C-C-1" type, the translation being that "C" stands for three driving axles, the "1" meaning one non-driving axle as used in leading or trailing bogies having smaller diameter wheels. Smaller diameter leading bogie wheels help locomotives to follow curved track at high speed, as the smaller diameter wheels do not tend to ride up over the outside rail on curves. Being long and articulated, the 1-C-C-1 types earned the nickname "crocodile". Swiss loco classification A problem with electrifying existing tunnels for high voltages is the clearance needed by high-voltage insulators above the train. In some cases, this has required modification to the tunnel roof. (Photo SJ). 76 SILICON CHIP The Swiss invented their own classification of locomotive wheel arrangements. They give the 1-CC-1 type the classification "Ce6/8". The first upper-case letter is a maximum speed rating, the small "e" following means "electric" and "6/8" means six driving axles out of a total of eight axles. The speed rating letters originally chosen Concrete sleepers are now used by the Swiss Federal Railways in place of the older wooden sleepers. The overhead wires carry 15kV 16.6Hz AC. (Photo SJ). were "A" for highest speed, with successive letters meaning lower speed ratings, the slowest being "E". Thus the Ee3/3 class built in 1928 was limited to a top speed of 40km/h and was used for shunting service. The 3/3 means that three axles out of a total of three were driven. The "C" classification usually meant a speed rating of around 65km/h, "B" meant 70 to 80 km/h and "A" was applied to locos rated between 90 and 12 5 km/h. The Ae4/7 of 1927 was therefore an example of one of their top speed locomotives of the day, being rated at 100km/h. It also had a maximum starting tractive effort of 196 kilonewtons supplied by a traction motor-to-driving wheel gear ratio of 1:2.57. Their classification system presented a small problem in 1964 when a locomotive rated at 149km/h was built, as there is no earlier letter than "A" in the alphabet. This problem was overcome by using the letter "R" to denote high-speed locomotives. High speed locomotives Since 1946 the Swiss have built bogie type electric locomotives of both eight wheel and twelve wheel types. The former we would call a "Bo-Bo" type; that is two bogies each with four independent driving wheels. The Swiss call the eight wheel loco an "Re4/4" meaning high speed electric and four driving axles out of a total of four. The modern Re4/4 locos built in 1982 are rated for a maximum speed of 160km/h and have a maximum starting drawbar pull of 300 kilonewtons. Their total power is 4.960MW (6650hp) and their high speed traction motors are geared 1:2.77 to their bogie driving wheels. For all this power they weigh only 80 tonnes - wonderful engineering design! In later years the Swiss were the first in the world to produce locomotives with more than 1000HP per axle, a remarkable achievement. Larger still are the Swiss 12-wheel bogie Re6/6 locomotives built from 1972 to 1980. These are powered by six traction motors which are each rated at 1.308MW (1753hp), giving a total of 7.850MW (10523hp) - all this in a locomotive only 19.31 metres long and weighing 120 tonnes. This enormous power per axle in such a small locomotive was made possible by the latest technology in traction motor design and wheel slip control which we will investigate in a later chapter. For the moment we must say that the Swiss railway engineers, together with the research and manufacturing companies Brown Boveri, Swiss Locomotive Manufacturing Co, Verkehrshaus der Schweiz, and the Swedish company ASEA have been in the forefront of new developments. Now that ASEA and Brown Boveri have merged, we may expect to see still more startling advances in European locomotive design. Synchronous motoralternators An interesting point arises (in any country) with the starting up and placing on-line of the synchronous motor-alternator frequency conversion sets. Commonly, such sets are started by an auxiliary induction "pony" motor having two poles ·less than the main motor. Once running and on load all sets are synchronised with all other alternators and synchronous motors, both on the 50Hz side and on the 16.6Hz side as Fig.1 shows. You would expect that the procedure would be to start the 50Hz synchronous motor and, when it reached full speed, synchronise it with the 50Hz mains. The low frequency alternator would then surely be generating 15kV 16.6Hz. However there is only a 33 % probability that it is in synchronism AUGUST 1988 77 Series AC traction motors require regular maintenance. In this photo, the slots in the motor's armature are being cleaned, ready to accept a new winding. (Photo SJ). SYNCHROSCOPE 6kV 50Hz THREE-PHASE BUSBARS "' THREE-PHASE 50Hz SYNCHRONOUS MOTORS SINGLE PHASE 15kV 16.6Hz ✓BARS SINGLE PHASE 16.6Hz ALTERNATORS Synchronising two substations 'INCOMING MACHINE CLOSED r INCOMING 6kV 50Hz THREE-PHASE SUPPLY FEEDERS I CLOSED CLOSED THREE-PHASE CIRCUIT BREAKERS ALL CLOSED SINGLE PHASE CIRCUIT BREAKERS No. 4 t RUNNING MACHINES ON LINE / 15kV 16.6Hz ELECTRIC RAILW AY OVERHEAD CONTACT WIRE Fig.1: sketch of an electrical system with four synchronous motoralternators. Machines 2, 3 & 4 are running and on-line , supplying train lines. Machine No.1 is running at full speed but needs to be synchronised before it can be connected to the other machines. with the 16.6Hz supply being generated by the other motoralternator sets. Therefore, it cannot simply be placed on line without further thought. The reason for this problem is that the 50Hz motor has three times as many poles as the 16.6Hz alternator, so the 50Hz motor could have 78 SILICON CHIP started up and synchronised from the 16.6Hz side, it follows that the 50Hz synchronous motor (now acting as a generator) must also be in synchronism with the 50Hz supply. (2). If started and synchronised on the 50Hz side and the 16.6Hz side turns out to be not in synchronism, a procedure called " pole-slipping" can be adopted. The operator simply retards the rotor one or two thirds of a rotating electrical circle to find the synchronised position. Pole-slipping consists simply of removing the DC field supply for a very short time, during which time the rotor " slips back" a little in rotating angle. The DC field supply is then restored and chances are that the rotor has slipped back exactly one pole pitch and the correct rotating angular position has been found. Synchronisation would then be possible. If the first attempt at pole-slipping is not successful, the procedure must be repeated. However, another large problem arises in a complete electric railway system operated at 16.6Hz derived from the national 50Hz grid' system. its rotor in any of three angular positions and still be in synchronism with the 50Hz supply. But only one of these three angular positions will give synchronism for both the 50Hz and 16.6Hz supplies. Two methods are available to fix this vital problem: (1). If the motor-alternator set is Consider two motor alternator substations A and B situated 30km apart, each substation supplied by the same 50Hz power grid system and each supplying 15kV 16.6Hz power to its own section of the overhead contact wire system. Normally each section is kept separate from neighbouring sections so that faults will not affect all trains in all sections. The motor side of all motor-alternator sets in both substations are automatically in phase ; ie, in synchronism (because they are on a common 50Hz system). Also we have seen how the 16.6Hz sides of all motoralternators in any one substation are brought into synchronism. Say some fault , perhaps a heavy short circuit, causes all machines in substation B to trip off, following which they are immediately restarted and synchronised again on their 50Hz sides and all their 16.6Hz alternators brought into synchronism with each other. We now have a problem: there is no guarantee that the 16.6Hz supp- THREE-PHASE 50Hz COMMON SUPPLY TO ALL SUBSTATIONS SUBSTATION o· SUBSTATION C SUBSTATION SECTION 2 SECTION 3 B SUBSTATION A EMERGENCY CIRCUIT BREAKERS NORMALLY KEPT OPEN SECTION 1 RAIL RAIL Fig.2: sketch showing four sections of overhead contact wire. The sections are normally kept isolated so that a fault in one section will have no affect on other sections. Note that a locomotive with two pantographs will bridge two sections. If section 2 is not in phase with section 3, the locomotove will cause a short circuit. ly generated by substation B is in phase (ie, in synchronism) with that generated by substation A. In fact, there is only a 33% chance that both substations will be in synchronism. Of course different trains running in each separated contact-wire section would never know the difference. At the meeting of two sections, the overhead contact wires are usually kept separated by an insulator and trains running across the join may simply jump the gap with a momentary but unnoticed power interruption. Multiple unit passenger trains in which each power car has its own pantograph in contact with the overhead wire, or electric locomotives using one pantograph, have no problems in this situation. Even if substation A and substation B ,were out of phase, the motors in trains running over the join feel no ill since they are not synchronous motors, but series motors with commutators which run the same direction no matter what polarity or phase current is applied to them. So where is the problem? It becomes very apparent when the first large electric locomotive comes along with both pantographs up. Common practice is for very highpowered locomotives to raise both pantographs (connected in parallel) to share the current when heavy train loads and mountain line star- ting conditions cause the loco traction motor currents to be high. When twin pantographs, electrically connected directly in parallel, mounted atop one European 15kV AC locomotive, approach a junction of two overhead contact wire sections which happen to be out-ofphase, watch out! For you are about to see fireworks. This would cause huge short circuit currents to flow from substation A, via the overhead contact wire section A, through both parallel pantographs, through over head contact wire section B, through substation B and back to substation A via the running rails and return conductors. Such a short circuit would cause a violent explosion at the front pantograph of the locomotive at the moment of contact. Probably circuit breakers in both substations would trip on over-current and such a fault might even stop some of the machines. It is imperative that such a short circuit situation is never allowed to happen between two remote substations. The remedy is that a feedwire is run from substation A to substation B so that they can be synchonised (using the pole-slipping method) before they are connected together. German electrification Germany can lay claim to having the first electric railway carrying fare-paying passengers: the 1879 demonstration DC electric railway built by the Siemens brothers in Berlin. This was soon followed by a 2.5km electric line in 1881 from which a suburban electric system grew. Mainline German electrification from 1922 has used the 15kV 16.6Hz AC system as pioneered by the Swiss BLS. Today much of West Germany is electrified , allowing international travel by electric train. For example one can travel behind electric locomotives from Italy, through Switzerland and West Germany to Holland in through coaches. Static AC-DC rectifiers The Siemens company of Germany has been active in the development of static frequency conversion methods since the 1930s and are credited with the invention of a static 50Hz-to-16.6Hz frequency converter using mercury-arc rectifiers. These were used in a ' 'cyclo-converter' ' configuration which simply divides the frequency by a factor of three. Long before the invention of semiconductor diodes, thyristors and GTOs (gate turn-off thyristors), the mercury-arc rectifier had been used as a high power rectifier. For example, in Sydney's outer suburban railway DC substations, 6-phase steel-case water cooled mercury-arc rectifiers supplied AUGUST 1988 79 SIX PHASE 50Hz SUPPLY, STAR POINT GROUNDED STAR SIX PHASE TRANSFORMER FEED-THROUGH INSULATORS Static frequency conversion STAR POINT i---,,---ANODES FLASH-OVER SHIELDS CIRCULAR STEEL TANK, EVACUATED AND CONTAINING MERCURY MERCURY -VAPOUR ARC FROM MOST POSITIVE ELECTRODE DC NEGATIVE Fig.3: basic sketch for a 6-phase mercury-arc rectifier. An ionised mercury vapour arc is struck between the most positive AC anode and the common mercury cathode pool on the bottom of the tank. 1.5kV DC for trains. A 6-phase rectifier is effectively six separate mercury-arc anodes in one evacuated steel tank having a pool of liquid mercury at the bottom. The tank and the pool of mercury becomes the common cathode of the multiple diode. Our sketch (Fig.3) shows the essence of the system which operates by an arc of dense ionized mercury vapour being struck between the most positive AC anode and the common mercury cathode pool at the bottom. The mercury liquid is boiled to a vapour and ionized by the electric field into heavy positive mercury ions and much lighter negative electrons. When the anode is on the positive half of the AC cycle, the light-weight negative electrons are attracted to the positive anode, constituting a heavy current flow, experiencing an almost-constant voltage drop across the arc of about 15 volts. High current capability Many thousands of amps may 80 SILICON CHIP Typically, a mercury arc rectifier could cope with a 500% overload for about five or 10 seconds, and lesser overloads for longer times. Many mercury arc rectifiers are still in service throughout the world (there are still a few left in the Sydney electric railway system) but all will eventually be replaced by banks of silicon diodes. easily pass in this direction. When the same anode is on the negative half cycle of the AC supply the negative electrons are rejected but the heavy positive mercury ions are now attracted by the electric field. The comparatively much greater mass of those heavy positive mercury ions prevents any great acceleration towards the negative anode but a few do travel that path, thus giving a small "reverse leakage current", As the forward electron current is thousands of times more than the reverse leakage, the mercury-arc system is an efficient diode of quite low output impedance. They were used extensively for high-current rectification before silicon diodes took over the task. The advantage of the mercury arc rectifier has always been its very large short-duration overload capability, an excellent characteristic for supplying the large currents demanded when, say, five electric trains happen to start up simultaneously, For cyclo-converter (ie, frequency conversion) applications, the mercury-arc rectifier was constructed in single diode format, with many diode units in a ring formation, Cyclo-converter diodes must be switchable; ie, it must be possible to have them in the non-conducting state at times even though the anode is at positive potential. Then on command the diode can be made to conduct. This controlled rectifier action was accomplished by means of a cylindrical control-grid structure mounted between the mercury pool and the anode electrode. If this control grid is held sufficiently negative it strongly rejects the negative mercury ions, so the mercury arc diode cannot conduct even though the anode may be positive. If the negative potential is then removed from the control grid while the anode is still on the positive half cycle, the mercury arc immediately forms from cathode pool to main anode, and the diode conducts. Once the mercury arc is formed and the diode fully conducting, the control cylinder-grid loses control. Applying a negative potential to the grid cylinder now cannot stop the dense arc of negative ions unless the anode is made negative for a short time. "Bouncing" the line voltage This can be accomplished by using a pulse transformer in the anode circuit to produce a negative pulse superimposed on the AC supply to the anode. Such a pulse had to be of long enough duration for the flow of negative mercury ions to come to a halt. After that, a negative grid cylinder could prevent re-ignition of the arc even neither of these methods but applied a small positive potential to the grid to induce ionization, whereupon the main arc would strike as soon as the main anode became positive. If AC phase control is used, a small leading angle of control grid voltage was needed for full output. In this mode, these machines were analogous to the single pole mercury diodes called "Ignitrons", once used to switch large currents, and to the much smaller gridcontrolled gas-filled rectifier tubes known as "thyratrons", valves such as the EN32 series of the 1940s and 1950s. THREE PHASE 50H1 INPUT NOT GROUNDED SUPPLY DELTA CONNECTED TWELVE SINGLE PHASE CONTROLLED MERCURY ARC RECTIAERS IN SIX PULSE THREE PHASE CYCLOCONVERTER CONAGURATION Constructional details SINGLE PHASE 16.6Hz OUTPUT Fig. 4: basic scheme for a cycloconverter using 12 single phase mercury arc rectifier known as lgnitrons. though the anode may be positive, as the circuit of Fig.3 illustrates. The large currents demanded by electric railway service produced considerable heating in the mercury and tank, such heating being approximately equal to the 15V forward voltage drop in the arc multiplied by the thousands of amps of current flowing. This heat was easily removed by circulating water through a jacket outside the steel tank. The steel tanks admitted the anode AC circuit via porcelain insulator bushings, the conducting rod down the centre of the bushing being kept air-tight using liquid mercury as a sealant. The vacuum within the tank kept the mercury in position and if any did spill into the tank no harm was done as it just mixed with the pool at the bottom. Striking the arc At each cycle the mercury arc had to be initially struck. This could be achieved in one of three ways: (1). A heater placed under the surface of the mercury pool, in a large 6-phase or 12-phase rectifier, could vaporize sufficient mercury for the most positive anode to strike an arc. The resultant heating due to the main arc would keep the arc alive when the next anode became more positive and took over conduction. (2). In single anode mercury arc rectifiers, an auxiliary small anode could maintain sufficient arc continuously, such that as soon as the main anode became positive there was enough ionization present for the main arc to immediately strike. (3). Some grid-controlled singleanode mercury arc rectifiers used OA:::!, 0 ,,~ ~a ,:c u~ ~ c Wf' - r oa .~ RCS Radio Pty Ltd is the only company which manufa.ctures and sells .every PCB & front panel published in SILICON CHIP, ETl and EA. 651 Forest Road, Bexley, NSW 2207 Phone (02) 587 3491 for instant prices 4-HOUR TURNAROUND SERVICE Some steel tank mercury-arc rectifiers with many large electrical bushings entering the tank could not maintain vacuum for long periods. Therefore, these were fitted with a vacuum pump running continuously or when needed. The Hewitic Company of England was one organization which developed "pumpless mercury arc rectifiers" in which the seal was sufficiently good so as to not require continuous pumping. The German invention of the mercury arc cycloconverter in the 1930s gave the 16.6Hz AC supply needed by the railways without any need for rotating machinery. This giant step forward was a precursor of similar techniques which would be used in the future, once silicon controlled rectifiers (SCRs) were invented. The steel tank of a mercury-arc rectifier is alive at full DC output potential, so the tank is mounted on porcelain insulators. A limitation on all mercury-arc rectifiers is that they must be used in stationary position with the mercury liquid pool at the bottom. Transport while switched on is prohibited because of the dire consequences of "sloshing" of the mercury pool at the bottom. Short-circuits could easily occur. This prevents any such rectifier being installed in a locomotive. Such schemes had to wait until silicon diodes rated at thousands of volts and many thousands of amps were developed. :It AUGUST 1988 81