It can be advantageous and cost effective to mount a thyristor unit in the primary circuit of a transformer. The following article describes precautions to be taken.
Transformers are frequently used in electrical energy applications for one of three purposes:
i. To achieve galvanic isolation
ii. To achieve a reduction of the supply voltage at the load
iii. To achieve an increase in the supply voltage at the load
The use of thyristors with transformers needs some care as the ‘inrush’ current on connecting the supply to the transformer can easily exceed the maximum current rating of the thyristor unit and cause failure of the semiconductor fuse. The inrush current is of the order of 20 times the normal running current, as a rule of thumb.
Consider first the single-phase case:
Where thyristors are connected between transformer and load, the thyristor unit characters are determined by the load only and no attention needs to be paid to the transformer. In many cases, though, thyristors are connected in the transformer primary – i.e. between the supply and the transformer – for reasons of economy; generally one pays more for current than for voltage capacity. In these cases, it is important to observe some straightforward precautions.
If the B-H curve of Figure 1 is considered, it is a matter of chance where on the curve, thus the magnetic state of the transformer, the equipment is left on disconnection. Even if, as is usual with thyristors, the state is left
corresponding to H=0, there will be residual magnetism in the transformer core. The problem is what whilst the circuit is switched off as current falls to zero, it is switched on as voltage rises from zero. Unity power is rare!
If no precautions are taken then, one time in two, switching on will drive the transformer into saturation in the same direction, allowing excess current to flow. In practice, the fuse will fail about one time in four.
On switching on, if the rms current is increased carefully from zero, the inrush problem will be overcome. This technique is referred to as ‘soft start’ and in most cases is necessary over only a few cycles. Where current limiting also is required on starting, this same ‘soft start’ is usually satisfactory for both requirements. Once running, the magnetic state of the transformer can be easily inferred, so the same soft-starting precaution is not always necessary every time the load is re-energised, having been ‘off’ for a period – as is common in temperature control applications. Even so, the problem of ‘off with current, on with voltage zeros’ remains. To cope with this, a technique of burst firing with delayed first cycle firing has been developed.
Usually a delay of half a half-cycle (as shown in figure 2) is satisfactory for most applications, though the enhancement to be able to adjust the delay start from zero to 90° as a commissioning feature – to cope with individual Q factors – is almost always beneficial.
Thus inrush currents can be minimised whilst the benefits of burst firing control are realised.
Designers are still required to exercise considerable caution when working with loads which are other than purely resistive, however, as experience of control using burst firing techniques in the primary of transformers is still limited. Further, where the technique is used with confidence with elements which run at relatively low temperature, there may be difficulty – either mechanical or thermal – in the case of high-temperature elements. The guidance of the element manufacturer should always be sought when in doubt.
The well-established ‘phase angle’ method of control is now thought to be suitable for all applications.
If in difficulty, the designer would be well advised to seek expert assistance from manufacturers with considerable applications experience.
All the precautions necessary when using single-phase transformers are necessary in the case of multi-phase installations. There are one or two additional considerations occasioned by the interaction of three phases.
The principal consideration is the need to keep the three limb fluxes equal by keeping the three driving voltages equal unless the system has been specially designed otherwise. This means that the three phase currents will be unequal to the extent that the load impedances are unequal. In particular, it is difficult without taking special precautions to equalise the power in three zones – even if this were possible it would be unlikely to be achieved except at one specific element temperature – unless the furnace or load is designed with this need in mind.
This highlights the fact that a three-phase transformer cannot be treated in any way as three separate units – its economy lies in the interaction of the phases.
A further complication lies in the property of the three sine waves of the three-phase system to ‘cancel each other out’ enabling a three-phase system the ability to operate without a neutral wire.
In a similar way, a three-phase transformer is designed often with three limbs only, so that the three fluxes – which follow the voltages – also ‘cancel each other out’.
Unfortunately, this cancelling is true only at the fundamental frequency of the supply (50Hz in the UK). Most transformers are controlled by thyristors in the ‘phase angle’ mode, which means that the current in the windings consists of a range of odd harmonics; dominant amongst which is the third harmonic, (setting aside the fundamental). Reference to Figure 3 will demonstrate that, far from ‘cancelling out’, the third harmonics reinforce each other. This property can have interesting effects on a transformer core, resulting from overheating, since the flux will try to ‘close’ via any metal in the vicinity – frequently the steel case.
There are two easy ways to avoid difficulty. The first is to wind the primary (usually, though the secondary is equally suitable) in ‘delta’ so that the three fluxes will add in a ring round the core, with a resultant of zero. The alternative is to use a five-limb transformer – the outer limbs forming the return paths. These are not the only way to avoid a problem, nor is it guaranteed that they are always satisfactory. However, they are the most usual and fail only in very unusual cases.
There are many installations which have been running without problems for years which do not employ a delta. For whatever reason, whether it is because the transformer is lightly loaded or the third harmonic content is low at the normal working point, or a combination of other factors, these systems emphasise that the rules concentrating the third harmonic are only guidelines.
These are applications, including two-zone furnaces and salt baths where, for supply reasons, energy for only two zones is taken equally (as near as possible) from the three phases of the supply. The tool which is often used for this function is the Scott-wound transformer.
The ‘Scott’ is a three-to-two arrangement where the primary is usually star and the two secondaries are usually separate, as shown in Figure 4.
All the rules and guidelines applying to 3-phase transformers apply also to ‘Scott’. The notable exception is the delta-primary which is not practical. For this reason, Scott transformers are often wound with five limbs when used with thyristors. Additionally, the unusual flux patterns often necessitates a slower ‘soft start’ than with the ‘conventional’ transformer, as because of the harmonic content of load currents, the line currents will not be equal for all phase angle conditions and the power will not necessarily be equal in the two load circuits.
The one really important precaution is the avoidance of parasite voltages in the primaries which can damage thyristor equipment.
This can occur where secondaries are not, in fact, galvanically isolated. A current flowing in one can pass, wholly or partly, through the other and this can in turn generate through ‘backwards’ transformer action, an e.m.f in its appropriate primary.
If the thyristor unit in this circuit is off, the e.m.f. can be large enough to cause damage. Where this problem is likely, expert advice should be sought.
In all transformer applications of thyristors, it should be remembered that the transformer is essentially a 50Hz device – it is indeed a very good 50Hz. The generation of all the harmonics will cause the core to warm up more than in ‘normal’ use and this, coupled with the need to avoid inrush current difficulties means that the core should be run at a relatively low flux density. A value of around 1.2 Tesla is usually used.
Eurotherm offer an on-site advisory service in the choice of suitable thyristors.