**2. Materials and Methods**

Section 3.1 is a review of the literature that shows the historical development of SCIMs, focusing on the main technological innovations, material improvements, and various projects. The sources of information are technical documents from SCIM manufacturers or scientific articles that described the processes and the main events that caused the changes in the mass/power ratio between 1890 and 1990, contributing to answering question I.

In Section 3.2, a literature review is presented discussing the variations in performance between 1935 and 2012, contributing to answering questions I and II.

In Section 3.3, the primary data collected in the Technical Test Reports of the Laboratory of Electrical Machines of the Institute of Energy and Environment (IEE) of the University of São Paulo (USP) are presented and discussed.

Between 1945 and 1996, the Technical Test Reports were only available in printed form. Thus, it was necessary to digitize the data and collect them into a spreadsheet. Between 1997 and 2020, the Technical Test Reports were already available in digital format for processing and analysis.

The Laboratory of Electrical Machines of the IEE-USP has a technical collection of approximately 21,000 technical reports. For this analysis, reports with the following characteristics were considered:


Using the conditions expressed in a–g, 359 technical reports of tested SCIMs with speeds corresponding to 2, 4, 6, or 8 poles, with a motor rated output power of 3.7, 37, or 150 kW, were collected for the evaluation of the change in performance between 1945 and 2020. The assessment seeks to answer the questions (I and II) that motivated this research, based on the data collected.

The results are organized into three different output power (kW) categories. The chosen groups include low power (3.7 kW), medium power (37 kW), and high power (150 kW). As the groups chosen to represent SCIMs are of significantly different dimensions, the production processes used in the manufacturing process and the standards of precision/quality of the materials are also different, even when dealing with the same equipment.

The number of poles of the electric motor determines the rotation speed, due to the arrangement and distribution of the electrical conductors of the windings located in the stator slots. In the SCIM market, historically, four speeds have been the most used. Between 80% and 90% of all the SCIMs sold have between 2 and 8 poles; therefore, this research evaluates them in this speed range. In fact, 4-pole SCIMs are dominant, representing between 45 and 70% of SCIMs [21,47].

Using the conditions expressed in a–g, 28 SCIMs with speeds corresponding to 2 poles and motor rated output powers of 3.7, and 4.4 kW were used to evaluate the change in the mass/power ratio between 2000 and 2020, seeking to answer the questions (I and III) that motivated this research, based on the data collected.

The National Institute of Metrology, Quality and Technology (INMETRO) accredits the Laboratory of Electrical Machines at IEE-USP, following ABNT NBR ISO/IEC 17025:2017 [48] under No. CRL 0011. INMETRO periodically carries out audits in accredited laboratories, aiming to guarantee the quality of the measurement results. INMETRO is a signatory to the mutual recognition agreements of the International Laboratory Accreditation Cooperation (ILAC) and the Inter-American Accreditation Cooperation (IAAC), thus following a world standard of quality and reliability.

This research, therefore, used data from standardized performance tests. This is because there may be differences between values measured in neutral laboratories and values reported by manufacturers [49], and when using the measured data, errors and uncertainties are reduced.

#### **3. Results and Discussion**

#### *3.1. The Improvements in SCIMs*

All the technological and theoretical bases for electric motor development were already advanced by the end of the 19th century. Direct-current motors were on the market, and alternating-current motors were in the full developmental stage, with research ongoing in Europe and the United States. The first patent for the electric motor with asynchronous technology was filed by the engineer Nikola Tesla [50] in 1888 and accepted in 1889 [51] in New York. The asynchronous motor became known as an induction motor, based on its working principle. However, Tesla's proposal was similar to the current single-phase auxiliary winding motors, operating with a wound rotor. The text that explained the working principle of the new electric induction motor was published by Nikola Tesla in 1988 with the title "A new system of alternate current motors and transformers" [52].

Parallel to Nikola Tesla's experiments in the USA were those of Galileo Ferraris in Italy. In 1885, Ferraris developed the idea that two out-of-phase currents could be used to produce two magnetic fields that could be combined to produce a rotating field, without the need for switching or moving parts, opening the door to AC electric motors [53–55].

The three-phase squirrel-cage rotor induction motor (SCIM) closest to the type we have today was developed by a German company AEG (Allgemeine Elektricitäts-Gesellschaft), headed by the Russian engineer Mikhail Dolivo-Dobrovolsky between 1888 and 1890 [56]. The electric motor developed by the Dobrovolsky team had very favourable characteristics such as high starting torque, more straightforward construction features, robustness in construction, and low maintenance needs. However, it also had the inconvenience of needing to be powered by a three-phase alternating-current system, which was not yet commercial. Until then, the available electrical systems were single-phase and two-phase systems. This type of supply does not provide efficient starting of the Tesla-mounted motor (starting torque practically non-existent), in addition to imposing some degree of vibration during operation. The SCIM has a high starting torque and does not need auxiliary windings and accessories such as a capacitor and a centrifugal starter, in addition to having a lower operating current compared to a single-phase motor. However, three-phase electric power generation, transmission, and distribution systems were quickly implemented with the objective of feeding the attractive SCIMs [57–59].

Dobrovolsky and the AEG company gained fame for the great invention. The artist Irene Ahrens created the illustration in Figure 3, which was exhibited in Berlin. The engineer appears in the sky, entering the Hall of Fame with his SCIM shown near his feet.

**Figure 3.** Mikhail Dolivo-Dobrovolsky entering the Hall of Fame with his SCIM. Source: [57,60].

In 1891, at AEG, Dobrovolsky coordinated the first serial production of SCIMs with shaft powers between 0.4 and 7.5 kW. The first SCIMs assembled had a performance of approximately 80% for the power range produced and very high mass by today's standards. The first commercial two-pole SCIM with a shaft power of 4.4 kW was marketed in 1891. These SCIMs had a mass/power ratio of 86 kg/kW, as shown in Figure 4.

**Figure 4.** Improvements in SCIM mass/power ratio between 1891 and 1984. Source: [61–63].

The company AEG published the famous image represented in Figure 4, which shows the mass/power ratio from the first SCIMs manufactured by the company in 1891 to the SCIMs manufactured in 1984. The optimization of materials for electrical, magnetic, and mechanical purposes, combined with solid technological innovations, made it possible to reach a ratio of 6.8 kg/kW in 1984, representing only 8% of the total mass of the two-pole SCIMs with an axle power of 4.4 kW produced in 1891, as AEG's first commercial units.

The concept of the SCIM has not changed since the beginning of its commercialization; however, the volume has changed considerably (Figure 4).

The technological progress of SCIMs has been remarkable, stimulated by strong competition and by processes, technological innovations, and improvements in materials. According to Browning (1997) [64], the changes in the mass/power ratio resulted in better operational characteristics, even more excellent reliability, versatility, and longer life.

Browning (1997) [64] identified the following improvements in SCIMs:


The adoption of industry standards has played a significant role in the progress of SCIMs [28,64]. An example is the thermal classification of insulating materials, which first appeared in 1898. In 1911, standardization by the AIEE Standards (now IEEE—Institute of Electrical and Electronics Engineers) established temperature limits for SCIMs. The 1915 edition of the AIEE Standards included definitions of insulation classes A, B, and C and the materials assigned to those classes. In 1929, the first SCIMs built to NEMA standards were made available on the market, setting standard dimensions and operating characteristics for specific ratings for the first time. Users were given the ability to directly replace SCIMs via the concept of stock electric motors for quick replacement in case of failure [64].

It is possible to observe in Figure 5 the tremendous technological innovations that were decisive in reducing the mass and volume of SCIMs.

**Figure 5.** Chronology of 0.75 kW SCIM mass reduction between 1900 and 1990. Source: [65–68].

At the beginning of the 20th century, the first major technological innovation was the development of ball bearings, replacing the traditional plain bearings that were bulky, heavy, and required lubrication with oil. With the new bearings and the reduction in friction losses, the mass and volume of the SCIMs decreased considerably.

Between 1913 and 1940, there were gains in the quality of materials, improving compaction and making it possible to reduce the volume of copper and iron used in SCIMs and to reduce losses. In the 1940s, rotors previously built using iron sheets began to be developed using cast aluminium, adding more mass reduction, as shown in Figure 5. In addition, in the 1940s, with successive advances in metallurgy, SCIM housings could be built in an increasingly closed way and could maintain the cooling of the windings located in the stator.

In the early 1960s, a series of advances in insulation systems were instrumental in reducing the volume of SCIMs. Between 1960 and 1970, SCIMs went through five generations of materials used to construct insulation for electrical conductors. In the first SCIMs, the insulation was composed of paper, and later cotton. Then, insulation with varnish predominated until the present day. Figure 6 shows in white the area necessary to accommodate electrical conductors of the same metallic volume inside the stator magnetic package slot for different insulation technologies [63].

The first significant innovation in SCIM insulation systems was the replacement of the double layers of cotton between the conductors and the sheets with two layers of silk, allowing a reduction of approximately 59% of the groove area in the metal sheets (ferromagnetic material) of the stator. The second major innovation was the introduction of varnish used in conjunction with silk, giving an area reduction of over 2%, as shown in Figure 6. Subsequently, improvements in the quality of silk and varnish allowed an area equivalent to be reached of only 22% of the space required for the same electrical conductor using cotton as an insulator.

**Figure 6.** Space used by different insulation technologies for the same SCIM output power. Source: adapted from [63].

Successive technological innovations and improvements in electrical, magnetic, and mechanical materials achieved significant volume compaction in SCIMs between 1903 and 1974 [28], as illustrated in Figure 7a. Figure 7b shows the changes in appearance and frame dimensions of SCIMs of different powers from the open construction of 1904 to those used in the 1970s, similar to today's drip-proof and fully fan-cooled SCIMs.

**Figure 7.** Dimension trends and housing changes of 11 kW 4-pole SCIMs between 1903 and 1974. (**a**,**b**): the changes in appearance and frame dimensions of SCIMs of different powers from the open construction of 1904 to those used in the 1970s. Source adapted from: [28,53].

Figure 7a presents SCIMs designed for operation at 220 volts and 11 kW, built by General Electric (GE). In Figure 7a, it is possible to observe the changes in the NEMA 404 housing over the years, and two significant innovations are evident in the images of SCIMs between the years 1920 and 1954: axial extension of the rotor at the rear and the closed housing, seen from 1954 and made possible by the improvement in insulation systems, enabling the transfer of heat from the windings to the outside.

Figure 7b shows a small SCIM (1904) built without a fan and considered to be "selfventilated" by the semi-open housing. As early as 1918, SCIMs used a fan attached to the shaft for cooling. In 1930, the 15 kW SCIM already had a more efficient fan and could

adopt a more closed design. In 1972, the engines were already drip-proof. Figure 7b shows 18.5 kW and 45 kW SCIMs. They could be fully enclosed (45 kW), allowing a reduction in SCIM dimensions. As a result of the improved insulation between the conductors and between the conductors and the ferromagnetic material of the sheets, the temperature of the winding wires and the groove walls became more homogeneous, as they were closer together with a thinner insulating layer. The temperature of the set decreased, and for this reason it was possible to increase the power considerably for the same housing. The stator slot was significantly reduced for the same power, and the magnetic section between the slots could be increased. There was also an improvement in the ferromagnetic material, an increase in the magnetic flux, and a consequent decrease in the number of turns per stator coil for the same electrical voltage.

According to Alger and Arnold (1979) [28], to avoid hot spots in the centres of long cores, radial ducts were introduced in the stator and impellers in the rotor operating as fans, creating the airflow through the stator channels. Therefore, the rating given to the NEMA 404 frame with an axle height and length of 25.4 cm and 31.1 cm, respectively, was increased with respect to mechanical power from 5.5 kW in 1897 to 75 kW in 1974, as shown in Table 2.

**Table 2.** Mechanical power increments in the same frame from 1898 to 1974. Source: [28].


The reduction in the volume of SCIMs also made it possible to reduce their costs, intensifying the electrification of industrial plants. For example, in 1890, a 3.7 kW SCIM weighed approximately 450 kg and cost about USD 900, and in 1957, a SCIM of the same power weighed around 50 kg and cost USD 110 [64]. Thus, the relationship between value and mass remained practically the same. However, as mass reduced significantly, the price of the SCIM reduced considerably, since the cost of an SCIM is fundamentally a function of the quantity and quality of materials used.

The company Hitachi produced three SCIMs of 3.7 kW in 1910, and in 2010 the total production was already 40 million SCIMs. The company recorded the significant advances that SCIMs have made over more than 100 years in this period. Hitachi divides advances in electric motors into three distinct periods. Between 1830 and 1890 is the period of inventions, from 1930 to approximately 1950 is the period of scientific initiatives, and between the 1950s and the present day is the period of industrial initiatives [69].

Various technical and technological developments have made Hitachi SCIMs smaller and lighter over the 100 years from 1910 to 2010. Figure 8 presents the leading technologies used by Hitachi that made it possible to reduce the mass of the first SCIM, with a power of 3.7 kW (four poles) manufactured by the company in 1910 with a mass of 150 kg, to approximately 20% of the mass in 2010 (30 kg).

**Figure 8.** Hitachi SCIM mass changes for 3.7 kW (4 poles) SCIMs. Adapted from [69].

The main changes recorded were the use of aluminium in the rotor in the late 1940s. Later, in the 1950s, bearings improved, moving from sliding technology to ball bearings. In the late 1960s, improvements were made with the application of new insulation classes of varnishes on the wires. In the mid-1970s, cast iron frames gave way to lighter sheet steel frames. In the 1990s, aluminium structures closed the cycle of major technological innovations in Hitachi's first 100 years (1910–2010).

Reduction of Volume and Losses of Ferromagnetic Materials in SCIMs

The first electrical devices to use ferromagnetic materials were developed in the second half of the 19th century. Knowledge about these materials, such as their structure, was absent; as a result, the development of the projects was based on trial and error [70].

For SCIMs to thrive, they needed to advance in generating, transmitting, and distributing electrical energy via alternating current (AC) [71]. Charles Proteus Steinmetz was hired by General Electric (GE) (by Thomas Alva Edison) to improve the AC distribution system. He developed the complex representation of variables sinusoidally in time, which is still in use today [72]. Steinmetz deepened his studies of ferromagnetic materials to better compete with Westinghouse, which manufactured the induction motors invented by Tesla.

The first concepts regarding losses in ferromagnetic materials, traditionally known as iron losses, were developed by Steinmetz [73]. Via understanding how the losses behaved with changes in the intensity of the magnetic field, the General Electric induction motors became competitive, due to the reduction in the volume of material used [74].

Steinmetz's secret was to use increasingly thin sheets. The eddy current losses depend on the square of the sheet thickness, the hysteresis losses, and the square of the magnetic field strength [75]. Steinmetz's discoveries led to more efficient rolling mills that produced thinner and thinner sheets.

Understanding the ferromagnetic losses (hysteresis and eddy current) was decisive for selecting increasingly thin sheets to assemble the stator and rotor magnetic package. Thus, it was possible to impose a greater magnetic flux density in the package of sheets, approaching the limit of the magnetic saturation of the plate. This knowledge contributed to reducing the volume of SCIMs to approximately one third of the initial volume between 1891 and 1901 (Figure 4).

Subsequently, the development of ferromagnetic materials focused on reducing iron losses through heat treatment of the materials and the "doping" of silicon to increase the resistivity of the composite [76], thus enabling the intensification of the magnetic field and consequently reducing the volume of the SCIMs for a defined power.

To better understand the reason for the volume reduction of SCIMs over time, as shown in Figure 4, regarding the reduction provided by the improvement in ferromagnetic materials, it is possible to model the volume of SCIMs from the increase in the intensity of the magnetic field in their structures. This imposition of increasingly intense magnetic fields was one of the main reasons for the reduction in the volume of electric motors since their development.

A mathematical expression that translates the volume/power ratio as a function of the imposed magnetic flux density can be deduced from the electromechanical energy conversion equation, where the phase-induced electromotive force is given by Cardoso et al. [70]:

$$\mathbf{E} = \mathbf{4.44fN}\_{\text{eff}} \boldsymbol{\mathcal{Q}} \tag{1}$$

where f is the frequency (Hz), Neff is the number of adequate turns in series per phase, and ∅ is the magnetic flux per pole (Wb).

Furthermore:

$$\mathcal{Q} = 2 \frac{\text{BLR}}{\text{P}} \tag{2}$$

where B is the density of the magnetic flux in the air gap (T), L is the packet length (m), R is the radius of the air gap (m), and p is the number of pole pairs.

The electric current expressed as a function of the magnetic field in the motor air gap was expanded from the classical magnetomotive force equation FMM = NI = φ, and can be expressed as [70]:

$$\mathbf{I} = \frac{\pi \mathbf{p} \mathbf{l}\_{\mathbf{g}} \mathbf{B}}{3\sqrt{2}\mu\_0 \mathbf{N}\_{\mathbf{eff}}} \tag{3}$$

where μ<sup>0</sup> is the magnetic air permeability (H/m), p is the number of pole pairs, and lg is the thickness of the air gap (m).

Ignoring any type of losses, the motor power will be given by P = mEI, where m is the number of motor phases.

Substituting E and I by their values expressed in Equations (1) and (3) results in:

$$\mathbf{P} = \frac{\pi \mathbf{m} \mathbf{p}^2 \mathbf{l}\_{\mathfrak{E}} \mathbf{n}}{3\mu\_0} \mathbf{B}^2 \text{Vol} \tag{4}$$

where n = <sup>f</sup> <sup>p</sup> represents the synchronous motor rotation in rps and Vol <sup>=</sup> <sup>π</sup>R2L represents the motor volume.

Reorganizing Equation (4), it is possible to mathematically verify the volume/power ratio of SCIMs and other equipment that uses ferromagnetic materials, in proportion to the intensity of the internal magnetic field in Equation (5).

$$\frac{\text{Vol}}{\text{P}} = \frac{1}{\text{K}\_{\text{m}}\text{B}^{2}}\tag{5}$$

with:

$$\mathbf{K}\_{\mathbf{m}} = \frac{\pi \mathbf{m} \mathbf{p}^2 \mathbf{l}\_{\mathbf{g}} \mathbf{n}}{3\mu\_0} \tag{6}$$

Equation (5) is inversely proportional to the magnetic flux density characteristic square, expressing a curve similar to that of Figure 4. It suggests that one of the primary explanations for the reduction in the volume (or mass) of SCIMs over the years was the

more significant imposition of the magnetic field on its magnetic structure, as shown in Figure 9.

**Figure 9.** Density of the magnetic flux in the air gap and the volume of SCIMs.

Figure 9 represents the relationship between the improvement in the quality of the ferromagnetic material and the reduction in SCIM volume, considering the same output power. As a result of Equation (5), the curve is theoretical since it is impossible to design an electric machine with unlimited magnetic flux density (B) or a value of B very close to zero. Hence, the curve represents one of the essential reasons for the reduction in the volume of SCIMs from the first units to the present day.

It is estimated that for 110 kW electric motors, the losses in ferromagnetic materials represent, on average, 59% of the total losses [77]. The losses increase with increasing frequency of the electric voltage. These materials have also undergone improvements from the first electric motors to the current ones.

Since Michael Faraday demonstrated electromagnetic induction in 1831 [78], soft magnetic (ferromagnetic) materials have continued to evolve. When iron was the only soft magnetic material available, metallurgists and materials scientists experimented by introducing other elements to improve the efficiency of iron.

The main known losses in ferromagnetic materials are hysteresis and eddy current losses. Hysteresis losses occur through the coercivity of a magnetic material. Each time a material with magnetic characteristics completes an entire cycle of its magnetization curve, the area within this curve measures the energy lost in the magnetization process.

The second primary loss mechanism in soft magnetic materials is eddy currents. Eddy currents are closed electric current paths generated in a conductor whose source is a timevarying magnetic field. These current loops create a magnetic field in opposition to the change in magnetic flux (according to Faraday's law of induction). The energy losses caused by eddy currents scale approximately with the square of the operating frequency and are thus a significant cause of losses in alternating-current machines.

The development of silicon (electrical) steel in about 1900 was a notable event in the advances of soft magnetic materials [79]. Silicon steel still dominates the global soft magnet market and is the material of choice for large-scale transformers and electrical machines such as SCIMs. In 1900, Robert Hadfield, a metallurgist from England, and his team developed unoriented silicon steel by adding up to 3% of silicon to iron and increasing its electrical resistivity (ρ) [80].

The team led by the American metallurgist Norman Goss developed grain-oriented silicon steel in 1933, promoting grain growth along a crystalline direction. The most common applications for silicon steel are large-scale transformers (grain-oriented silicon steel) and electrical machines (unoriented isotropic silicon steel is preferred for rotating machines), for which the economical price is a great benefit [80].

Improvements in magnetic properties were also achieved, from the treatment of iron to minimize chemical impurities to the techniques of slicing the iron into thin sheets. Subsequently, silicon was used to increase the electrical resistance of iron and control the crystal orientation. Figure 10 presents the reduction in losses in the core of electrical machines in watts for each kilogram of ferromagnetic material, highlighting the predominant technological advances of each period.

**Figure 10.** Changes in losses in the core of electrical machines (ferromagnetic material). Source: adapted from [7,76,81].

It is possible to observe in Figure 10 that between 1884 and 1970, the losses in the core of alternating-current electrical equipment reduced from 8.16 W/kg to 0.44 W/kg, which represents an approximately 95% reduction.

Figure 10 shows low frequencies (50 or 60 Hz) and a constant B (T) value, as both directly influence losses in ferromagnetic materials.

Today's primary soft ferromagnetic materials in electric motors are iron and ferrosilicon alloys (2022). However, materials with lower eddy current and hysteresis losses have been developed since the 1970s.

After the energy crisis of the 1970s, the first attempts to use amorphous materials for electric motors were recorded (1981). Mischler et al., demonstrated the low-loss potential of the amorphous stator in a laboratory environment [82].

In 1967, a new class of materials, amorphous alloys, was introduced [83]. In the mid-1970s, interest in amorphous alloys based on iron and cobalt increased, and these materials began to find applications [84]. However, only in 1988 did Hitachi researchers investigate Nb and Cu additives. They added an annealing step to amorphous alloys to produce small-spaced crystallites of iron or cobalt within an amorphous matrix material. The formation of isolated crystallites of transition metals reduced the eddy current losses of these materials compared to traditional amorphous alloys. Despite a higher initial cost than silicon steel, these advanced alloys can reduce the total lifetime costs of electric motors due to reduced losses.

Currently (2022), unique treatments involving thermal manipulation, laser bombardment, and other technologies continue to produce high-performance magnetic materials.
