B 32 -The Brown Boveri Review- BBC House Journal and Transformers- Part 3 Special Issue on Transformers

The Special Issues on Transformers were published from 1942 to 1976 in BBC Review as listed below.

- 1942 Vol 29, No. 11/12 November/December, Pages 319-335

- 1945, Vol 32, No 3, March, New transformers and Choke coils designs, Pages 91-100.

- 1956 Vol 43, No.6 June 1956 Pages 187-230

- 1965 Vol 52, No. 11/12 November/December, Pages 799-92

- 1972 Vol 59, No.8 August, Pages 363-426

- 1976 Vol 63, No. 7 July, Pages 419-470

www.dhirubhai.net/pulse/b-31-brown-boveri-review-bbc-house-journal-part-2-p-ramachandran-1zduc/?trackingId=G9mqHd79QV6%2B8U8b0XXe1w%3D%3D

As I noted in my scrap notebook in February 1971, the following are some knowledge nuggets from the 1965 special issue on Transformers. As revealed in this summary, transformer engineering information may still be helpful for young engineers practicing in transformers. These are relevant even half a century later.

1) Dr A Goldstein, Technical Director, in his introduction

In 1956, the maximum size of transformers was 200 MVA 380 kV with a weight-to-power ratio of 1 kg/kVA and 0.5 % losses. By 1965, the ratings reached 660 MVA 400 kV and service voltage 735 kV with a weight of 0.65 kg/kVA and 0.3 % losses. Behind these dry figures is a tremendous amount of development work performed on a comprehensive basis. In addition to the technological, construction and manufacturing progress in producing iron cores, windings and insulation systems, much has been done to locate and control the electric and magnetic fields in the winding space.

2) M Itschener- Progress in Transformer Design

The manufacture of transformers is an example of how an industrial product, without any revolutionary innovations, can be repeatedly improved by perseverant development work on all components. Though not spectacular, the achievements are nevertheless surprising.

Nowadays, the purchaser of a transformer seldom considers the purchase price alone but weighs the annual costs incurred by the iron and copper losses under the respective operating conditions. We adapt our designs to these requirements by selecting the most favourable solution and optimising calculations with an electronic computer.

The primary task was to handle the steady increase in the directional sensitivity with the progressive improvement in the loss factor of the core laminations. This development helped to mitre the laminations, though initially viewed with scepticism, to supersede 90° interleaving, which had hitherto been standard practice for the core columns and yokes. With the radially laminated design, which has now come to be regarded as almost classical, though it proved possible to modify the interleaving of the return yokes in this manner, the butted joint between the laminations of the column and those of the yokes is inevitable; moreover, appreciable transfer paths occur at right-angles to the direction of rolling of the broadest limb laminations. The low losses of the high-grade grain-oriented silicon steel laminations can then not be utilised to the full, so now the radially laminated design is only employed when the iron losses are not as vitally important as the indisputable constructional advantages of this design. For years, the development of the iron core has been characterised by grain-oriented sheets, with all the questions arising from these. The primary task was to consider the steady increase in the directional sensitivity with the progressive improvement in the sheet's loss factor (W/kg). Joints in cores were changed from 90 degrees to 45 degrees mitred joints, though initially, engineers viewed this with scepticism.

Due to the higher sensitivity of the better-grade grain-oriented sheet for the direction of the flux, bolt holes in the core column and yokes are unsuitable from a loss angle. They reduce the value of almost the entire strip width between them. We, therefore, decided to eliminate the bolt holes almost entirely. This was facilitated by the fact the cold-rolled sheets are extraordinarily smooth and flat and can be pressed at a low pressure (which is desirable from the point of view of the loss factor), yielding a good space factor. Core bolts were replaced by core bandages, for which new material was adopted. In some cases, using modern adhesives, it is possible to give the cores a very high mechanical strength without using any bolts, both during assembly and especially after wedging the windings. Cores built up in this manner with homogenous cross-sections have low losses and a low magnetisation current and are very favourable regarding noise.

……… The radially laminated core represents a most valuable component for EHV shunt reactors. The stray fluxes in the surrounding air gaps can enter the radially laminated stub unhindered at any point. Such reactors have been built with one column (limb) for 220/(root3) kV and two columns for up to 735/(root3) kV, the latter for a single-phase rating of 110 MVA, 60 c/s. Our design produces little noise or vibration and hardly any harmonics due to the core iron's low percentage of ampere-turns (i.e. flux density).

BBC French subsidiary, Cie Electro-Mecanique (CEM), developed a revolutionary core for small transformers in the 200-10,000 kVA range. This core was made from three wound core strips acting as limbs, with triangular wound strips forming the yoke.

With increased copper prices, the adoption of aluminium was occasionally considered. Still, the price relationship is not so favourable that far-reaching modifications would have been justified. Aluminium conductors are only considered where extreme weight requirements are specified.

The main development lines for the windings aim to improve insulation and cooling and reduce the effect of leakage fields; the short-circuit strength also had to be adapted to meet the new requirements.

No fundamental change has been made to the transformer insulation from the material aspect. Preparation and drying have been progressively improved. Adding oxidation inhibitors in oil has proved advantageous for slowing down oil oxidation.

External water cooling always uses circulating oil pumps to cool the oil. This method reduces the size of the external cooling equipment and helps to improve the cooling process in the winding with cooling ducts in windings.

The leakage fields cause eddy-current losses in windings. This, with the heavy currents and high field strengths of large transformers, imposes severe restrictions on the dimensions of the conductors and reduces the winding space to very low levels, especially at high voltages. The use of Roebel conductors (CTC), whose varnished strands are continuously transposed and covered by paper insulation, allows the eddy-current losses to be considerably lowered, not only in the main axial leakage field but also in the radial transverse field at the ends of the windings. Special Computer programs simplify the quantitative determination of these losses……. The eddy losses in the tank from the leakage field from the winding can be eliminated to a certain extent by applying laminated magnetic screens.

When there is an imbalance in the ampere-turn distribution between the primary and secondary windings, the axial forces act on the winding supports. This occurs with body taps, thicker conductor insulation at the line end of the winding, or inaccurate winding assembly with axial asymmetry. Wherever possible, these problem sources should be avoided. Their effect must be calculated whenever they are present, and countermeasures like suitable extra bracing must be taken. For some years, we have been using a computer program to calculate these axial forces, which has enabled us to reduce them, often by simple countermeasures.

3) W. Heiniger – The effect of loss capitalisation on the design of large transformers.

The mere purchase price does not decide the cost of a transformer because once it is in service, its losses impose a steady burden on the user in the form of energy losses. It is, therefore, common practice to capitalise on the losses, add their value to the purchase price, and arrive at the cheapest offer.

The capitalised value of each kW of loss (= price of energy loss incurred during future N service years in today’s money value) is estimated based on

? The price of energy

? Operating hours per year

? Average load (kVA) on transformer

? Number of years in service(N)

? Rate of interest for money (inflation rate)

If the unit cost of energy losses is Rs 1 per annum, the value of, say, ten years of payment in today’s rupee will be Rs 7.7 at an interest rate of 5 % (Table 1).

Suppose a transformer is in service for an average of 8,000 hours per year for 25 years. In that case, if the energy price per kWH is Rs 4, and the interest rate is 5%. Then, the cost of 1 kW of no-load losses for the entire transformer life in today’s money will be MFe = 8000x 4x14.1 = Rs 4,51,200 per kW, which is the loss capitalisation per kW of quoted iron loss.

Copper loss capitalisation for the above transformer with an average load of 58 % (root mean square of the load in kVA) MFc = 8000x4x14.1 x square of 0.58= Rs.1,50,400. The cooler loss (energy consumption in fans and pumps) is also capitalised at the rate of copper loss. Then, the capitalised price of losses is added to the quoted price to evaluate various bids.

The total loss capitalisation amount can be more than the quoted price. The lowest bid is the minimum sum of the quoted price plus the loss capitalisation amount. Some countries compare the annual costs instead of the total costs. The yearly cost of energy loss + annual interest, depreciation, and maintenance charge is considered for the comparison. In the figure below, annual costs are plotted against the load. The investment and maintenance costs Mo and the cost of the iron losses MFe do not vary with the load and are therefore represented by a horizontal straight line. The copper losses, on the other hand, increase as the square of the load, beginning with zero at no load and yielding a parabola, which may be denoted by kP2, where k is a constant, and P is the load. The annual cost at any operating point is M = Mo + M Fe + kP2. If now a straight line is drawn from the origin through any working point, and the angle made by this line is referred to as ’a’, we may write tan a = (Mo + MFe+kP2) / P. Minimum annual cost is when the line is tangential to the parabola. At this point, the yearly cost of the copper losses equals the cost of interest, amortization and maintenance plus the iron losses, i.e. square of kP = Mcu = Mo + MFe

In Fig 2, annual costs are compared when transformers of different rated MVA (20-60 MVA) are considered to meet an average load of 25 MVA. The purchase cost increases with a higher rating as more material is involved. The cost of copper losses decreases as the higher-rated transformer is partly loaded. The total annual cost of smaller rated units ( that will be overloaded to meet 25 MVA) is very high and comes down with a higher MVA rating. Surprisingly, the annual cost is almost the same in the optimum region, between 35 to 60 MVA. This means the yearly cost is the same for the MVA bandwidth of 1:1.7, with a deviation of barely 3 %. Hence, to meet a load of 25 MVA, the total annual cost is the same as that of transformer ratings from 35 to 60 MVA. A higher MVA rating gives the margin to meet future load demand increases and acts as an operational reserve. The negative side is the higher initial investment and need for an assured life span.

The above knowledge led to the standardising of the transformer ratings with a ratio of 1:1.6. See the table below.

Another conclusion from Fig 2 is that it is not worthwhile in most cases to adopt a higher current density or overload the transformer to meet a particular load requirement.

A long-standing "rule" of transformer design, also mentioned in Prof. Milan Vidmar's Standard book, is that the manufacturing cost for the iron core should equal that for the winding for optimum design with minimum cost. However, this rule is difficult to follow due to fluctuations in copper and core steel prices and transport and factory infrastructure limitations. When the above rule for optimum design is achieved, the ratio of iron losses to copper losses is around 5. With current prices and technology, it is 5-6 for two winding transformers and 6-10 for auto-transformers.