The operation and structure of a typical axial magnetic bearing is pre- sented in Schweitzer G., Bleuer H. and Traxler A., Active Magnetic Bearings, Zürich 1994,244 p., ISBN 37281 2132 0. This book presents a basic structure according to Fig. 12, whereby the force between the stator 1 and the rotor 2 in the axial magnet bearing structure is pro- duced by controlling the current in a coil 3, generating a magnetic flux 4 effective in the magnetic circuit. The magnetic flux 4 effective via the stator 1 and the rotor 2 generates a force which, using the markings of Fig. 13, has the quantity of: <BR> <BR> <BR> <BR> <BR> <BR> F= 2 (B2Aa +Bb2Ab), (1)<BR> <BR> <BR> <BR> Z, ut in which Ba = magnetic flux density at a point 4a of the magnetic flux, Bb = magnetic flux density at a point 4b of the magnetic flux, Aa = the area of the surface perpendicular to the magnetic flux 4 at the point 4a, Ab = the area of the surface perpendicular to the magnetic flux 4 at the point 4b, and pO = permeability of a vacuum.

In the basic structure of Fig. 13, the stator 1 of the magnetic circuit structure is parallel to the shaft 2.

In view of prior art, reference is further made to the publication WO <BR> <BR> <BR> 97/02641 (Lindgren et al.) presenting a structure which is illustrated in Fig. 14. The stator 1 of the magnetic circuit is in the radial direction with respect to the rotation symmetrical shaft, and the magnetic flux4 is excited from the stator 1 to the rotor 2 via a step 2d formed on the outer surface of the rotor. With the radial stator structure 1, the extension of the stator 1 of the axial magnetic bearing can be made shorter. The step 2d applied on the outer surface of the rotor 2 increases the area of incidence of the magnetic flux 4 in the rotor 2, thereby generating a greater force with respect to the basic solution in Fig. 13.

Further, with reference to the prior art, Fig. 15 shows a winding struc- ture of an axial magnetic bearing according to US Patent 4,353,602 which makes it possible to split the stator 1. In the presented solution, the partial coils 3a, 3b of the winding structure 3 are placed to encircle the central axis of the rotor 2, and the magnetic flux 4 is divided in the winding structure 3 into two parallel magnetic fluxes 4h and 4j. In the solution according to the patent 4,353,602, magnetic fluxes 4h and 4j encircle the projection surface about 2/3 times.

Furthermore, the prior art publication WO 95/34763 presents an axial magnetic bearing structure which is a structure biased with a perma- nent magnet. In addition to the permanent magnet, the magnetic bear- ing comprises a winding to generate a force effective on a flange fixed to the rotor, to stabilize the rotor in the axial direction. Part of the mag- netic flux generated by the permanent magnet and the winding is also excited through the end of the rotor, wherein it is possible to compen- sate for an external force effective on the rotor in the axial direction.

To determine the projection area, the concepts of outer diameter da and inner diameter db are first defined (examples shown in Figs. 6,13 and 14). The outer diameter da is the greatest common diameter of the adjacent parts of the rotor and the stator in the axial direction of the rotor. In a corresponding manner, the inner diameter db is the smallest common diameter of the adjacent parts of the rotor and the stator in the

axial direction of the rotor. The projection area corresponding to these diameters, perpendicular to the magnetic flux, is <BR> <BR> <BR> <BR> <BR> <BR> =-),(2) In conclusion to the solutions of prior art, the following can thus be stated by comparing the forces generated by the magnetic flux between the stator 1 and the rotor 2 in the applications according to Figs. 13 and 14. To provide an objective comparison, it is assumed that in the cases of both the Figs. 13 and 14, the diameters of the rotor 2 are equal at the point 2a and are equal at the point 2b.

We shall look at the total area Atot (= Aa + Ab) of incidence of the two magnetic flux parts 4a (partial area Aa) and 4b (partial area Ab) of the magnetic flux 4 excited in the air gap. In the solution of Fig. 13, the total area is typically: <BR> <BR> <BR> <BR> <BR> <BR> A, o"~x Ap (3) In a corresponding manner, in the solution of Fig. 14, the total area is typically When it is desired to obtain great forces with the magnetic circuit structures presently in use, a problem is the relative reduction of the diameter of the part 2b of the rotor 2 with respect to the part 2a with a greater diameter, as well as a great stray flux 4c produced in the stator 1 of the magnetic circuit.

The force is directly proportional to the areas of incidence of the mag- netic flux 4 at the points 4a and 4b in the rotor 2, and inversely propor- tional to the square of the air gap.

When it is desired to increase the force, particularly in the solution of Fig. 13, the ratio between the outer diameter da and the inner diameter db of the rotor2 is increased, wherein the areas are increased at the points 4a and 4b in relation to the difference between the squares of the diameters. In the solution of Fig. 14, in comparison with the solution shown in Fig. 13, it has been possible to increase the area of incidence of the magnetic flux 4 by about 50 %. It is obvious that the height of the step structure (the difference between the outer diameter and the inner diameter) should be minimized in view of the flexural strength of the rotor. Thus, in connection with the use of the step structure 2d, there is a correlation between the quantity of the maximum axial force gener- ated by the magnetic circuit and the flexural strength of the rotor.

Particularly in the solutions of Figs. 13 and 14, a problem is also a stray flux 4c excited in the stator 1 of the magnetic circuit, not through the rotor 2. The quantity of the stray flux 4c depends on the shape of the stator 1 and rotor 2 as well as the coil 3 of the magnetic circuit. The stray flux 4c does not generate a force between the stator and the rotor.

The stray flux 4c is caused by the current in the coil 3. The stray flux 4c causes a current loss, for which reasons an additional current must be supplied to the coil 3, which, in turn, increases the cross-sectional area of the coil 3 or increases the current density. The increased cross- sectional area of the coil 3 will, in turn, increase the size of the bearing, and increased current density will increase losses in the coil 3, thereby increasing the need of cooling the bearing. The increased need of cur- rent caused by the stray flux 4c is a problem particularly in the solution of Fig. 13, because the axial length of the structure is increased. In the solutions of Figs. 13 and 14, the structure of the magnetic circuit is rotationally symmetrical and thus cannot be split in the axial direction of the rotor without cutting the winding. This structural property would be very useful in solutions in which an actuator, such as the rotor of a turbo machine whose outer diameter is greater than the diameter of the part 2b of the rotor 2 with the smaller diameter, is fixed to the end fo the part 2b of the rotor 2. Splitting of the bearing structure has the advan- tage that it is not necessary to remove the actuator installe at the end of the rotor 2 in situations of disassembling the apparatus, for example in connection with maintenance.

It is an aim of the present invention to provide a magnetic circuit which can be used to generate a greater magnetic force than conventional solutions when the solution according to the invention is compared with conventional solutions of the corresponding size range. To achieve this aim, the magnetic circuit structure according to the invention is primarily characterized in that the separation between the parts of the magnetic circuit effective in the stator on one hand and in the rotor on the other hand is arranged in that part of the magnetic circuit in which the direc- tion of the magnetic flux deviates from the axial direction of the rotor, and that the different parts of the magnetic circuit are partly adjacent to each other in the axial direction of the rotor, to direct the magnetic flux effective in the magnetic circuit at least partly in the axial direction of the rotor.

The advantage achieved with the above solution can be illustrated as follows. It is assumed that three rotors are formed, two of which apply magnetic circuit structures of Figs. 13 and 14, and one solution apply- ing a magnetic circuit structure of the invention, constructed in the above described manner. Thus, the diameters of the rotor in that part of the magnetic circuit structure in which the force-generating magnetic flux is excited between the rotor and the stator are, e. g. with reference to Fig. 13,2a and 2b, of which 2a is the greatest diameter and 2b is the smallest diameter, wherein they can be indicated with da and db. The projection area corresponding to these diameters, perpendicular to the magnetic flux, is <BR> <BR> <BR> <BR> <BR> <BR> <BR> =--)(5) In the conventional solutions of Figs. 13 and 14, the magnetic flux is excited through the stator and the rotor through two air gaps, passing only once through the projection surface Ap. The propagation of the magnetic flux between the stator and the rotor takes place substantially in the portion in the axial direction of the magnetic flux. In the solution according to the present invention, one magnetic flux is excited through two air gaps, passing through the projection area Ap more than once,

advantageously more than 1.3 times, preferably more than 1.5 times, but a maximum of about 2 times, wherein it encounters the total area Atot3, that is <BR> <BR> <BR> <BR> <BR> <BR> l < A, ot3 < 2Ap (6) Since the magnetic force generated by the magnetic flux is directly pro- portional to the cross-sectional area, it can thus be found that using the magnetic circuit structure of the invention, a force is generated which is even about three times greater than in the solution of Fig. 13 (cf. equa- tion 3) and, correspondingly, about two times greater than in the solu- tion of Fig. 14 (cf. equation 4). Naturally, a basic requirement for this is that the projection area Ap is equal in all the cases.

Furthermore, with respect to the advantages achieved with the mag- netic circuit structure of the invention, the following comparison can be made. It is assumed that the areas met by the magnetic flux and the lengths of the air gaps are equal. Thus, the magnetic intensity required by the magnetic flux density is <BR> <BR> <BR> <BR> <BR> <BR> Fm = NI = B, uo 280, (7) in which N = number of laps of the coil, I = current in the coil [A], B = air gap flux density at point 4a and 4b of the magnetic flux [T], go = permeability of a vacuum, and 80 = length of the air gap at point 4a and 4b of the magnetic flux [m].

It follows from the above equation that the magnetic circuit structure of the invention requires as much current in the coil as those of conven- tional solutions (Fig. 13 and Fig. 14), although the magnetic flux passes through an additional disc related to the construction of the magnetic circuit of the invention. Using the solution of the invention, the stray flux

in the stator can be made very small, thanks to the more advantageous magnetic circuit.

It must naturally be noted that the magnetic circuit structure of the invention makes it possible, firstly, to increase the force effective in the magnetic circuit with respect to the conventional solutions, by keeping the dimensions of the construction equal. On the other hand, the strength of the construction, for the part of the rotor, can be increased by maintaining the force generated by the magnetic circuit to corre- spond to that of conventional solutions by increasing the diameter of the part of the rotor with the smaller diameter in such a way that the projection area Ap is reduced as required. Yet another advantage is that the losses occurring in the coil in the magnetic circuit structure of the invention are substantially reduced when compared with conven- tional solutions, because less current is needed in the coil to generate the magnetic force.

The advantages of the present invention are disclosed particularly in demanding bearing constructions in which it is desired to maximize the force, to minimize the losses occurring in the coil, to maximize the rigid- ity of the rotor, and to minimize unnecessary magnetization caused by a stray flux. The presented magnetic circuit structure makes it possible to dimension the bearing structure in an optimal way, which facilitates par- ticularly the mechanical design and implementation.

The appended dependent claims present some advantageous embodi- ments of the magnetic circuit structure according to the invention.

In the following description, the magnetic circuit structure according to the invention will be illustrated in more detail with reference to some advantageous embodiments.

Figure 1 shows, in a sectional view in the longitudinal direction, a first embodiment of the magnetic circuit structure according to the invention, utilizing cylinder symmetrical winding, Fig. 2 shows a cross-section at point 1-1 in Fig. 1,

Fig. 3 shows, in a sectional view in the longitudinal direction, a second embodiment of the magnetic circuit structure according to the invention, utilizing a winding placed on the periphery of the stator and consisting of two or more partial windings, Fig. 4 shows a cross-section at the point IV-IV of Fig. 3, Fig. 5 shows, in a sectional view in the longitudinal direction, a third embodiment of the magnetic circuit structure according to the invention, Fig. 6 shows, in a sectional view in the longitudinal direction, a fourth embodiment of the magnetic circuit structure accord- ing to the invention, Fig. 7 shows, in a sectional view in the longitudinal direction, a fifth embodiment of the magnetic circuit structure according to the invention, Fig. 8 shows, in a sectional view in the longitudinal direction, a sixth embodiment of the magnetic structure according to the invention, Figs. 9 and 10 show, in schematical views, a first method of forming the magnetic circuit structure according to the invention, Figs. 11 and 12 show, in schematical views, a second method of form- ing the magnetic circuit structure of the invention, Fig. 13 shows, in a sectional view in the longitudinal direction, the magnetic circuit structure according to the reference Schweitzer et at.,

Fig. 14 shows, in a sectional view in the longitudinal direction, the magnetic circuit structure according to the reference WO 97/02641 (Lindgren et al.), Fig. 15 shows, in a sectional view in the longitudinal direction, a modification of the structure according to US Patent 4,353,602 (Habermann), Fig. 16 shows, in a sectional view in the longitudinal direction, a seventh embodiment of the magnetic circuit structure according to the invention, Fig. 17 shows the electrical coupling corresponding to the magnetic circuit structure according to Fig. 1 in a reduced view, Fig. 18 shows the electrical coupling corresponding to the magnetic circuit structure according to Fig. 8 in a reduced view, and Fig. 19 shows the electrical coupling corresponding to the magnetic circuit structure according to Fig. 5 in a reduced view.

The parts shown in the drawings are listed: -stator 1, -rotor 2, -part 2a of the rotor directly outside the magnetic bearing (greater diameter da), -part 2b of the rotor in the stator area of the magnetic bearing (smaller diameter db), -auxiliary disc 2c placed adjacent to the magnetic bearing, -part 2d of the rotor, at the step between the magnetic fluxes 4a and 4b excited in the air gap in Fig. 13 (wherein diameter 2a > diameter 2d > diameter 2c), -coi) 3, -cylinder symmetrical winding 3a and separate winding 3b placed on the periphery of the stator and consisting of parts, -magnetic flux 4, -magnetic fluxes 4a, 4b, 4c and 4d in air gap,

stray flux 4c, as well as magnetic fluxes 4f, 4g, 4h and 4i in magnetic circuits, -rotor part 2a of the structure according to the invention can be formed at the end parts of the rotor 2 or independently on the rotor shaft 2, these solutions being distinguished in the rotor part with a broken line 5.

Figures 1 and 2 show the basic solution of the invention, in which the stator 1 is formed of one ring equipped with a U-profile form directed to the central axis of the rotor 2, inside which U-form the winding 3a is placed to encircle the rotor 2 as a uniform closed ring. The magnetic flux 4 is placed to encircle the rotor 2 as a uniform closed ring. The magnetic flux 4 is achieved by means of the front surface of the part 2a of the rotor 2 (the projection surface Ap) and an auxiliary disc 2c fitted inside the U-form. The winding 3a can be formed of several partial coils (not shown in Figs. 1 and 2), wherein the magnetic intensity Fm can be calculated as a sum of the magnetic intensities of the partial coils in the following way: <BR> <BR> <BR> <BR> <BR> F. = N I + N2, 2+... +NI,, (8) in which N1 = number of laps in the first partial coil, N2 = number of laps of the second partial coil, Ni = number of laps of the last partial coil, li = current in the first partial coil [A], 12 = current in the second partial coil [A], and Ij = current in the last partial coil [A].

If the magnetic circuit structure also comprises a permanent magnet, it <BR> <BR> will induce a magnetic intensity Fk whose effect in the total magnetic intensity Fm is obtained by adding it to the magnetic intensities of the partial coils.

Figures 3 and 4 show a modification alternative to place the winding 3b which makes it possible to split the stator structure 1 in the axial direc-

tion. Normally, the coil 3a is placed in a cylinder symmetrical way, as in Fig. 1, wherein it is not possible to cleave the stator structure. In the solutions of Figs. 3 and 4, typically two to six windings 3b are set, wherein the coil 3 is placed in the stator 1 to be formed of at least two partial coils 3b which are placed as closed rings at determined intervals on the periphery of the stator 1, wherein the stator 1 is equipped with a perforation in the peripheral direction, through which each partial coil 3b extends, being supported to support elements between the perforations inside the closed ring form of the partial coil 3b.

In the magnetic circuits of Figs. 1 and 3, the coil 3 excites a magnetic flux 4 which passes through two air gaps 4a, 4b between the stator 1 and the rotor 2. Thus, the magnetic flux encounters the partial areas Aa and Ab and generates, in the axial direction of the rotor, a force which is substantially parallel in both air gaps, thus intensifying each other. In this case, the total area encountered by the magnetic flux is the sum of these partial areas, that is <BR> <BR> <BR> <BR> <BR> <BR> A. = Aa + Ab In this case, each partial area Aa and Ab is substantially the same as the projection area Ap, wherein the total area is about twice the projec- tion area Ap. Next, the ratio KA of the partial areas Aa and Ab and the projection area Ap is defined: <BR> <BR> <BR> <BR> <BR> <BR> =/ (10) Thus, in the examples of Figs. 1 and 3, this ratio KA of the partial areas is approximately 2.

We shall next discuss the situation according to e. g. Fig. 8, in which the magnetic flux generates forces in different air gaps, some of which are effective on the rotor in one axial direction and some are effective on the rotor in the opposite axial direction. Thus, when calculating the total force effective in the magnetic circuit, the directions of the forces must be taken into account. As already stated above in this description, a

directly proportional factor which substantially affects the force gener- ated by the magnetic flux is e. g. the area encountered by the flux in the air gap. The direction of the forces can be taken into account in the above-presented formula 10 in such a way that when calculating the total area, the areas in which the magnetic flux generates a force in a first direction are counted as positive and, correspondingly, the areas in which the magnetic flux generates a force in the other, opposite direc- tion, are counted as negative. In this way, the weighted total area can be defined in the magnetic circuit with m air gaps. This weighted total area can be calculated with the formula: Atotw = ((A1 + A2 + ... + Am) + |Ai+ - Ai-|)/2 (11) in which <BR> <BR> <BR> <BR> <BR> <BR> A,, A2,..., A are the areas of the air gaps,<BR> <BR> <BR> <BR> <BR> Ai+ ils the sum of the areas effective in the positive direction, and A ; is the sum of the areas effective in the negative direc- tion.

As examples of Formula 11, the magnetic circuit structures of Figs. 6 and 8 can be discussed. Figure 6 has 4 air gaps 4a, 4b, 4c, 4d, of which it is assumed that the areas encountered by the magnetic flux are substantially equal; consequently, using the markings of Formula 11, A, = A2 = A3 = A4 = Ap, wherein the resulting total area is A, o, W ~ 4Ap.

In practical applications, a small clearance must be left between the rotor 2 and the stator 1, wherein the minimum diameter db at point 2b is smaller than the corresponding minimum diameter of the stator 2. Thus, calculated with the formula 5, the resulting projection area is a greater value than the areas encountered by the magnetic flux in said air gaps <BR> <BR> 4a, 4b, 4c, 4d, wherein the weighted total area Atorw iS slightly smaller<BR> <BR> <BR> <BR> <BR> than 4Ap. The ratio of the areas in this case is Kw = 4.

Correspondingly, in the situation of Fig. 8, there are three air gaps, the force in one air gap 4b being opposite with respect to the other two air gaps 4a, 4c, wherein the total area obtained from the Formula 11 is Atotw = 2Ap and the ratio of the areas is Kw = 2.

In comparison, the weighted total area is also calculated in a magnetic circuit structure according to prior art, as presented in Fig. 14. It is assumed that the rotor diameter at point 2a = da and at point 2d = (da db)/2, and the stator diameter at point 2b = db. Thus, the corresponding partial areas are A, =A, <1Ap. By placing these values in the For- 2 mula 11, the weighted total area obtained is AtoW = ((A, + A2) + |A, + A2 |)/2 < Ap, and ratio ratio the areas is Kw <l.

On the basis of what has been presented above, the ratio Kw between <BR> <BR> the weighted area Atotw calculated for the magnetic circuit structure of<BR> <BR> <BR> <BR> the invention and the projection area Ap is<BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> advantageously 1 < Kw < m, but preferably 2 < Kw < m, (12a) in which m is the number of effective air gaps encountered by the mag- netic flux in the magnetic circuit.

The magnetic flux is typically excited by a coil 3 which can be one coil or it can consist of several partial coils. As is obvious for anyone skilled in the art, the partial coils can be, for example, coils arranged in parallel to encircle the stator as in the embodiment of Fig. 1, coils placed on different sides of the stator as in the embodiment of Fig. 3, or combina- tions of these winding arrangements.

Furthermore, it is conventional in practical applications that the num- ber n of the coils 3 is one less than the number of air gaps. However, the magnetic flux formed by the coil 3 is divided in such a way that a part of this magnetic flux passes through each air gap of the magnetic

circuit structure. Thus, on the basis of the number n of the coils 3, the Formula 12 can also be presented in the format: <BR> <BR> <BR> <BR> <BR> <BR> advantageously l < Kw < n + 1, preferably 2 < Kw < n + 19 (1 2b) In the two-part structure of Fig. 5, the force effective on the rotor can be controlled with currents in separate coils. The stator frames are sepa- rate, and an auxiliary disc 2c is placed therebetween. In other respects, the partial structures correspond to the structures of Figs. 1 and 2 (3 and 4, respectively). Consequently, two completely different magnetic fluxes 4 are formed in this solution.

Figure 16 shows an advantageous modification alternative for the mag- netic circuit structure according to Fig. 5. The first winding 3a therein is arranged to be annular, to encircle the rotor 2, and the second wind- ing 3b is arranged as separate coil rings encircling the stator.

Figure 6 shows an application of the invention, in which the stator 1 consisting of one frame element comprises three coils 3 and three auxiliary discs 2c, wherein four air gaps 4a-4d are formed.

Figure 7 shows a modification alternative, in which the rotor 2 com- prises a total of three discs 2c and the stator 1 consists of two separate stator parts, in which control coils 3c and 3d are formed which generate magnetic fluxes 4f and 4g passing through the auxiliary disc 2c in the middle. The forces generated by the magnetic fluxes 4f and 4g are differently directed in the axial direction of the rotor 2.

Figure 8 shows a modification alternative, in whch the stator 1 consist- ing of one frame element is supplemented with a disc 1a whereby an asymmetric force is generated between the stator and the rotor by means of magnetic fluxes 4h and 4i. The auxiliary disc 2c is placed in the groove of the double-U-shaped profile where the coil 3d is located.

When, for example, an equal but opposite current flows in the coils 3c and 3d, the force generated by the magnetic flux 4i is greater than the force generated by the magnetic flux 4h, wherein the total force has a

different direction than in the case where the current flowing in the coils 3c and 3d is unidirectional. The structure of Fig. 8 is suitable for cases in which an asymmetric force in the axial direction of the rotor 2 is desired.

In certain modification alternatives of the invention, the area 2a of the rotor 2 can be replace with an auxiliary disc according to the part 2c (area separated with a broken line 5).

The magnetic circuit structure according to Figs. 1,3,5,6, and 8 is suitable to be installed symmetrically at both ends of the rotor.

It is obvious that in some solutions, the stator 1 can comprise several discs 1 a as well as the rotor 2 can comprise several discs 2c to arrange various paths and forces of the magnetic flux.

We shall still discuss the magnetic circuits in some examples by using the electrical couplings corresponding to them. As is obvious to anyone skilled in the art, the following analogy is valid between electric circuits and magnetic circuits. The magnetic intensity (Fm) corresponds to the voltage (U) in such a way that U = Fm = Nl, + Nl2+ * +Fpermanent magnei- In a corresponding manner, the following analogy is valid for current (I) and magnetic flux (¢): I = 0 To the resistance (R) of the electric circuit corresponds the reluctance (Rm) of the magnetic circuit. In the magnetic circuit, the reluctance can be defined by the formula '--r , uA, uoA in which

= length of the magnetic circuit,<BR> 6 = length of the air gap, IL, permeability, permeability in a vacuum, and A = area of the magnetic circuit or the air gap.

The analog to current density (J) in the magnetic circuit is magnetic flux density (B) which can be given as the ratio of the flux (¢) and the area (A) : <BR> <BR> #<BR> J = B =<BR> A<BR> On the basis of the analogies presented above, we shall first discuss the electrical coupling corresponding to the magnetic circuit of Fig. 1, which is presented in Fig. 17. The reluctances of the air gaps 4a, 4b are indicated with the resistances Ria, Rib, the stator relucance is indi- cated with the resistance Rs, and the rotor reluctance with the resis- tance Rr. The coil 3a is indicated with the magnetic intensity F. To the magnetic flux ¢ corresponds the current in the electrical circuit.

We shall next discuss the electrical coupling corresponding to the mag- netic circuit of Fig. 8, which is presented in Fig. 18. The magnetic circuit consists of several coils and air gaps, in this example two coils 3c, 3d and three air gaps 4a, 4b, 4c. The reluctances of the air gaps 4a, 4b, 4c are indicated with the resistances Ria, Rib and Ric, the stator reluc- tances with the resistances Rs1, Rs2, and the rotor reluctances with the resistances Rr1, Rr1. The coils 3a, 3b are indicated with the magnetic intensities F1, F2. To the magnetic flux ¢1 and ¢2 correspond the cur- rents in the electrical circuit. On the basis of this, it is possible to draw up circuit formulas for the case of coupled magnetic circuits: from which it is possible to solve the unknown fluxes and further the flux densities in the air gaps 4a, 4b, 4c. The solution obtained for the flux density in each air gap is:

Bu ou<BR> <BR> <BR> A<BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> #1 - #2<BR> <BR> <BR> Bb =<BR> Ab<BR> #2 Bc = Ac and further, the total force is <BR> <BR> <BR> <BR> <BR> <BR> F-1 \BaAa-BbAb+BcAcl<BR> <BR> <BR> 2, uo B, It is observed from the circuit equation that the magnetic circuits are coupled with each other, and there are two magnetic circuits (n) and three air gaps (m = n + 1) in the equations. The direction of the current in the coils 3c, 3d affects the quantity of the flux and thereby also the direction and quantity of the force.

Generally, it can be stated that in a solution containing n number of coil sets, the magnetic flux passes n + 1 = m times between the stator and the rotor; that is, there is one more air gap than there are magnetic cir- cuits.

We shall still discuss the example situation of uncoupled magnetic cir- cuits as presented in Fig. 7. The corresponding electrical coupling is shown in the appended Fig. 19. The first magnetic circuit has a first magnetic intensity F1, a resistance Rr1 representing the first rotor reluctance, a resistance Rs1 representing the first stator reluctance, and two resistances Ria, Rib representing air gaps 4a, 4b. In a corre- sponding manner, the second magnetic circuit has a second magnetic intensity F2, a resistance Rr2 representing the second rotor reluctance, a resistance Rs2 representing the second stator reluctance, and two resistances Ric, Rid representing air gaps 4c, 4d. The partial structures are combined by only one magnetic flux conducting means, in this

example a part of the rotor 2 whose reluctance is represented in the corresponding electrical coupling by the resistance Rc, wherein each partial circuit can be looked at independently from each other. In practi- cal applications, various stray circuits can also be induced between partial structures, but because the force is inversely proportional to the area of the air gap and to the square of the distance, the share of the stray circuits can be disregarded in the review. From the corresponding electrical coupling of Fig. 9, the circuit equations can be derived as follows: from which the unknown fluxes and, further, the flux densities in the air gaps 4a, 4b, 4c and 4d can be solved. As a result, the following flux densities are obtained for each air gap: <BR> <BR> <BR> <BR> <BR> '<BR> <BR> <BR> <BR> <BR> A,,<BR> <BR> <BR> <BR> <BR> <BR> Bb= ,<BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> <BR> Ab<BR> <BR> <BR> Ac<BR> #2<BR> Bd =<BR> Ad and further the total force, which is <BR> <BR> <BR> <BR> <BR> 2Ho<BR> <BR> 2po It can be observed from the circuit equation that the magnetic circuits are not coupled to each other and that the directions of the currents in

the coils 3c, 3d have no effect on the direction and quantity of the force.

Thus, the magnetic circuits can be looked at independently.

Further, Figs. 9 to 12 illustrate the implementation of the magnetic circuit according to the invention.

In the case of the closed magnetic circuit according to Fig. 9, which is the starting point (assuming that the horizontal direction is the axial direction of the rotor), it is implemented by separating the parts of the magnetic circuit substantially in the axial direction of the rotor 2 at a point where the magnetic flux is excited substantially in the radial direc- tion of the rotor 2 (point A, Fig. 9), by directional displacement of the separated parts of the magnetic circuits with respect to each other in the axial direction (point B, Fig. 9), and by displacement of the parts partly adjacent to each other in the axial direction to direct the magnetic flux, on said length adjacent to each other, at least partly to the axial direction of the rotor 2 (point C, Fig. 9). Thus, the result is the situation shown in Fig. 10.

On the other hand, with reference to Figs. 11 and 12, the parts of the magnetic circuits can be separated diagonally in relation to the axial direction of the rotor 2, wherein a rotation clearance is formed at the point of separation between the parts of the magnetic circuits and wherein the magnetic flux is directed at said separation point diagonally in relation to the axial direction of the rotor 2, perpendicularly to the main directions of the mating surfaces at the separation point.

Some advantageous applications of the magnetic circuit structure according to the invention, to be mentioned in this context, include electric motors, such as squirrel-cage motors, and various compres- sors, such as turbocompressors. In a squirrel-cage motor, the rotor of the magnetic circuit structure belongs to the rotor shaft of the squirre- cage motor. In such a motor, the magnetic circuit structure is preferably fixed at both ends of the rotor shaft.

In conclusion of what has been presented above, it can still be stated that using the magnetic circuit structure of the invention, it is possible to

accomplish better magnetic bearing than with magnetic bearings of prior art without increasing the current strength to be fed to the wind- ings. The better efficiency also has the advantage that the minimum diameter of the rotor at the point of the magnetic circuit can be greater than in magnetic bearings of prior art, wherein the mechanical strength and stability of the rotor are also better.

It is obvious that the mechanical features of the parts of the magnetic circuit can vary within the scope of the invention. For example, those parts of the rotor and the stator which constitute the air gap, can com- prise various gradations which do not, however, substantially affect the size of the projection surface. Also in such an air gap, the magnetic flux will flow once between the rotor and the stator and will generate a force in the axial direction of the rotor.

It is obvious that the present invention is not limited solely to the embodiments presented above but it can be modified within the scope of the appended claims.