INVESTIGATION OF SHIELDING AS APPLIED TO TURBOGENERATORS

V. G. SHUMAKHER

(Presented by the scientific seminar of the departments of electrical machines and general electrical engineering)

The electromagnetic stray field in the region of the end parts of the stator winding of a turbogenerator, penetrating into the components of the end zone, leads to a concentration of losses and to increased local heating. In some types of large turbogenerators, a copper shield is used to protect the clamping ring and the stator core back from stray fields. The structural components of the end zone also possess shielding properties, in particular the stator clamping ring itself.

An accurate calculation of the stray fields, losses, and heating in the end zone is impossible without taking shielding into account.

Fields and shielding in the end zone are investigated on full-scale models of large turbogenerators [1] and on reduced physical models of generators [2]. In both cases, the researchers measure the resultant stray field without determining the shielding effect.

The investigation of shielding was carried out on a rotating-field mock-up, which consists of two annular laminated packages of electrical steel arranged parallel and coaxially (Fig. 1).

Fig. 1. Sketch of the rotating-field mock-up: 1 — core-back package, 2 — specimen under investigation, 3 — inductor winding, 4 — inductor package.

The field source—the inductor—is sufficiently distant from the second package, the core back, with the specimen under investigation placed on it. The three-phase annular winding of the inductor creates, in the space between the packages (in the working zone of the mock-up), a rotating electromagnetic field whose induction has a sinusoidal distribution around the circumference and a distribution close to uniform along the radius. The annular windings under investigation—

of different mean diameter \(D_{\mathrm{av}}=450\ \mathrm{mm}\), width \(B=80\ \mathrm{mm}\), and thickness \(h=8\ \mathrm{mm}\), were made of copper and nonmagnetic steel.

To determine the shielding, experiments are carried out both in the absence and in the presence of a specimen in the working zone. In order to ensure constancy of the exciting-field magnitude, the inductor current and the gap between the inductor stacks and the back were kept the same in all experiments. The copper specimen was investigated at frequencies of 14, 30, 50, 100, 200, and 400 Hz; the nonmagnetic-steel specimen at 50, 100, 200, and 400 Hz. The induction of the exciting field at the surface of the back in all experiments was about 160 G.

In the experiments, the amplitude of the magnetic induction normal to the plane of the back and its phase are measured, both at individual points along the radius of the working zone and averaged over the width of the specimen. The sensors are multi-turn coils, which are located at a distance of 3 mm from the surface of the back stack and from the installed specimen. The induction magnitude is determined from the readings of a V3-13 tube voltmeter, and the phase by means of a BIP-5 instrument. The rms error of the relative induction does not exceed 3%, and that of the phase is \(\pm 8^\circ\). The measurement error in the central part of the specimen may be somewhat greater.

The shielding of specimens is characterized by the following parameters [3]:

\(W\) — back-action coefficient;
\(S\) — shielding coefficient, averaged over the width of the screen;
\(S_0\) — shielding coefficient at the center of the screen.

The back-action coefficient \(W\) characterizes the increase of the field at the boundaries of the screen. The smaller the shielding coefficients \(S\) and \(S_0\), the more effectively the screen reduces the exciting field. The coefficients \(W\) and \(S_0\) are determined as the ratio of the inductions at the corresponding points of the working zone in experiments with and without the specimen; the coefficient \(S\) is the analogous ratio of the mean inductions over the width of the specimen.

The model imitates the end part of the stator back with a nonmagnetic screen or retaining ring located near it. The results of research on the model can be used for a turbogenerator by applying the theory of modeling. It is known [4] that the condition for modeling quantities characterizing the electromagnetic field in geometrically similar systems is a definite relationship between the similarity scales of the geometric dimensions \(m_l\), frequency \(m_f\), electrical conductivity \(m_\lambda\), and magnetic permeability \(m_\mu\). In the case of nonmagnetic specimens \((m_\mu=1)\),

\[ m_l^2 = m_f \cdot m_\lambda . \tag{1} \]

The modeling condition (1) can be modified by using the concept of field penetration depth

\[ \frac{1}{m_l}= \sqrt{\frac{f_{\mathrm{mod}}\cdot \lambda_{\mathrm{mod}}}{f_{\mathrm{orig}}\cdot \lambda_{\mathrm{orig}}}} = \frac{\delta_{\mathrm{orig}}}{\delta_{\mathrm{mod}}}, \tag{2} \]

where the subscripts mod refer to the model and orig to the original. Consequently, for identical geometric dimensions of specimens made of different materials, the shielding parameters depend only on the depth of field penetration.

A copper screen or the retaining ring of a turbogenerator can be modeled on the model by a specimen of the same material by increasing the frequency. In addition, it is possible to model a nonmagnetic retaining ring by using a material with better specific conductivity, for example copper. In this case the need for a high-frequency power source is eliminated. The field frequency in modeling is found from the expression

\[ f_{\text{mod}}=\frac{f_{\text{orig}}}{m_1^2}\cdot \frac{\lambda_{\text{orig}}}{\lambda_{\text{mod}}}. \]

Investigations by means of modeling on a rotating-field mock-up are distinguished by the simplicity and convenience of carrying out experiments. It should be noted that the stray field in a turbogenerator decreases along the end surface of the core in the radial direction, whereas in the mock-up the field along the radius of the working zone is close to uniform. In this respect, the shielding data obtained on the mock-up are suitable only for a qualitative evaluation of shielding in turbogenerators.

Results of the investigations

1. The characteristic dependences of induction along the radius of shielding specimens (Fig. 2) have a dip in the central part of the specimen, then, at the specimen boundaries, a value exceeding the exciting field, and a drop beyond the specimen. With decreasing \(\delta\), the depth and region of the dip, as well as

Fig. 2. Distribution of induction along the radius of a copper specimen at different frequencies.

the field peak at the specimen boundaries, increase. The phase of the field along the radius of the shielding specimen changes substantially, reaching a value on the order of \(100^\circ\) at the center of the copper specimen at \(f=50\) Hz. The better the shielding, the greater the phase change, which is a more sensitive indicator of shielding than the change in induction. The phase of the field beyond the specimen boundaries remains practically unchanged.

In a turbogenerator, the field peak at the inner edge of the shield falls in the region of the end part of the core and the teeth at the bottom of the slot, which even without this has a high concentration of local losses; this must be taken into account when cooling the end parts.

1. The shielding of the specimens under investigation (of the order of 1%) begins to appear at \(\delta \cong 3\ \text{cm}\) (Fig. 3); in the region \(\delta = 2 \div 0.65\ \text{cm}\), the shielding and the field peak at the specimen boundaries increase significantly, and then less so. From this point of view, for constant dimensions \(D_{\text{av}}\) and \(b\) of the turbogenerator shield, beginning with a certain thickness, any further increase in it becomes only slightly effective.

Fig. 3. Shielding parameters of the investigated specimens at different frequencies: ○ — specimen of nonmagnetic steel; ● — specimen of copper.

1. The experimental dependence \(W, S, S_0 = f(\delta^{-1})\) can be used to obtain data on the shielding of the original object—clamping rings and turbogenerator shields made of nonmagnetic materials, geometrically similar to the specimens investigated. The shielding parameters at \(f = 50\ \text{Hz}\) for a copper shield are found from (2) at the point \(\delta^{-1} = m^{-1}\), since \(\delta_{\text{orig}} = 1\ \text{cm}\) for the original object made of stainless steel grade 1Х18Н9Т—at the point \(\delta^{-1} = (m_1 \cdot 6.5)^{-1}\ \text{cm}^{-1}\). Thus, a copper shield with dimensions \(b = 226\ \text{mm}\), \(h = 22.6\ \text{mm}\), \(D_{\text{av}} = 1270\ \text{mm}\) \((m^{-1} = 2.83)\), located \(17\ \text{mm}\) from the core, corresponds to \(S = 0.22\) p.u., \(W = 1.5\) p.u., \(S_0 = 0.02\) p.u.; that is, the field in the central part of the shield is practically completely shielded (Fig. 2, \(f = 400\ \text{Hz}\)). A clamping ring made of stainless steel with dimensions \(b = 400\ \text{mm}\), \(h = 40\ \text{mm}\), \(D_{\text{av}} = 2250\ \text{mm}\) \((m^{-1} = 5)\), located \(30\ \text{mm}\) from the core, corresponds to \(S = 0.65\) p.u., \(W = 1.22\) p.u., \(S_0 = 0.57\) p.u. The dimensions of such a clamping ring correspond sufficiently accurately to the dimensions of the clamping ring of a TVV-320 turbogenerator with a capacity of \(300\ \text{MW}\); it is modeled by the copper specimen at \(f = 30\ \text{Hz}\) (Fig. 2).

1. At frequencies up to 400 Hz, no shielding was found for a back packet laminated from annular sheets of electrical steel grade E31

\[ \left(\rho = 0.5 \frac{\Omega \cdot \mathrm{mm}^2}{\mathrm{m}}\right) \]

0.35 mm thick, which indicates the absence of shielding of a stator back made of this steel with dimensions \(b = 283\) mm, \(D_{\mathrm{av}} = 1275\) mm \((m^{-1} = 2.83)\) at \(f = 50\) Hz.

Conclusion

The results of the investigation of specimens with ratios \(b:h = 8\) and \(D_{\mathrm{av}}:b = 5.6\) can be used for a qualitative assessment of the shielding of geometrically similar parts of turbogenerators, in particular, nonmagnetic pressure rings and shields in the end zone. Similar data can be obtained on a rotating-field model for parts with other geometric proportions. On the model it is also possible to compare various variants of the use of shields in a turbogenerator.

LITERATURE

1. V. I. Kosachevskii, G. M. Khutoretskii, P. V. Oshurkov. On the shielding of end packets of the stator core of a turbogenerator. — In: Investigation of Electromagnetic Fields, Parameters, and Losses in High-Power Electrical Machines. “Nauka,” 1966.

2. G. Schastlivyi, V. Zentsev. Investigation of the magnetic field in the end zone of a turbogenerator on a physical model. Elektrotekhnicheskaya promyshlennost’, No. 334, 1969.

3. G. Kaden. Electromagnetic Shields. Gosenergoizdat, 1957.

4. A. V. Ivanov-Smolenskii. Electromagnetic Fields and Processes in Electrical Machines and Their Physical Modeling. “Energiya,” 1969.