JP2008280998A - Shaft bending calculation system of turbine rotor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system for calculating shaft bending of a turbine rotor. <P>SOLUTION: This system comprises: an input part for setting measurement points of a rotor disc; a deflection amount detecting part for deriving a radial displacement amount of four or more points along an outer surface of the turbine rotor in the circumferential direction based on a measured value of a displacement meter; a storage part for storing the radial displacement amount and measured angles of the measurement points; and a calculating part for calculating most probable circular core deviation data. The calculating part selects three arbitrary points from the radial displacement amount and the measured angles out of all measured points to calculate a calculated circle value, calculates a difference between the calculated circle value and the radial displacement amount as an error amount to derive an error amount total value, and selects the smallest calculated circle out of the error amount total value as the most probable circle. The calculating part includes a core deviation calculating means for calculating deviation between a center of the most probable circle and the rotational center of the turbine rotor as the most probable circular core deviation data, a determining means for determining whether or not the maximum core deviation amount is within a reference value, and a shaft bending distribution calculating means for calculating the shaft bending distribution from the most probable circular core deviation data. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a system for calculating a shaft bending of a turbine rotor.

  Generally, when a shaft bend or misalignment occurs in the turbine rotor, the turbine rotor becomes unbalanced, which causes shaft vibration during operation. If this shaft vibration becomes excessive, an abnormality occurs in the bearing portion and normal operation cannot be performed. Further, if this excessive vibration further progresses, it may lead to a shaft breakage accident. Accordingly, in order to prevent such an accident, it is important to keep the shaft bending and misalignment of the turbine rotor within allowable values.

  The turbine rotor that is the subject of the present invention includes a steam turbine rotor in addition to a gas turbine rotor.

  Hereinafter, a gas turbine rotor will be specifically described as an example. FIG. 9 shows a general structural diagram of a gas turbine rotor. The gas turbine rotor 1 includes a compressor rotor unit 10, a turbine rotor unit 20, and an intermediate shaft 25 that connects the compressor rotor unit 10 and the turbine rotor unit 20. Both the compressor rotor unit 10 and the turbine rotor unit 20 are implanted with blades 11 radially on the outer periphery. It is composed of an ellipsoidal rotor disk 50. The gas turbine rotor 1 is an integral structure in which these rotor disks 50 are stacked in the rotor axial direction and fastened with spindle bolts 30 and both ends are supported by bearings S1 and S2.

  When shaft bending or misalignment occurs in the gas turbine rotor 1 having such a configuration, shaft vibration is caused. Further, the gap between the tip of the moving blade 11 mounted on the outer periphery of the rotor disk 50 and the outer casing (not shown) is adjusted to be constant in the circumferential direction. When the shaft vibration becomes large, the tip of the rotor blade and the casing may interfere with each other and operation may become impossible. Therefore, it is necessary to adjust at the time of assembling the rotor so that the amount of shaft bending or misalignment falls within the allowable value, and when the shaft bending exceeds the allowable value, it is necessary to correct the shaft bending.

  The following procedure is used to correct shaft bending. In the configuration of the gas turbine rotor 1 shown in FIG. 9, misalignment data including a misalignment amount and a misalignment angle is calculated for each rotor disk 50, and a shaft bending distribution of the gas turbine rotor 1 is calculated. An example of the axial curve distribution of the gas turbine rotor 1 is shown in FIG. The horizontal axis indicates the distance from the bearing S1 of the gas turbine rotor 1, and the vertical axis indicates the misalignment amount of the rotor disk 50.

  One of the factors that cause the shaft bending is that the thickness of the rotor disk 50 is not uniform. Therefore, depending on how the rotor disks 50 are stacked, the amount of misalignment of the rotor disks 50 may exceed an allowable value. In such a case, the rotor disk 50 to be corrected is selected from the axial curve distribution of the gas turbine rotor 1, and the contact surfaces of the rotor disk 50 (surfaces 50b and 51b where the rotor disks 50 and 51 contact each other in FIG. 8) are selected. The axial bending is corrected so that the contact surface angle (α) between the rotor disks 50 and 51 is reduced by cutting (FIG. 8).

  FIG. 8 shows a state in which the axial bending of the gas turbine rotor 1 has occurred. The rotor disk 50, 51, the rotor disk contact surface FG, the contact surface angle (α) between the rotor disks 50, 51, and the rotor This shows the relationship between the radial deflection amount and the misalignment amount of the rotor disks 50 and 51 with respect to the rotation center line RC. The gas turbine rotor 1 is formed by stacking and integrating a plurality of rotor disks 50 in the rotor axial direction, but only a part of the rotor disk is shown in FIG.

  The amount of runout in the radial direction of the rotor disk 50 is determined by selecting a plurality of measurement points at equal intervals in the circumferential direction while rotating the rotor on the outer surface 50a of the rotor disk 50 and reading the displacement meter at each measurement point. It is obtained by measuring the amount of radial displacement at the measurement point. That is, with the measurement start point as a reference (for convenience, the displacement amount at the measurement start point is set to 0 (zero)), the radial displacement amount with respect to the measurement start point at each measurement point is used as the shake amount at each measurement point. Various known sensors are applied as the displacement meter. For example, in addition to a contact sensor such as a dial gauge, a non-contact sensor such as a laser sensor, a capacitance sensor, or an ultrasonic sensor can be used.

Misalignment data is calculated from the measured value of the shake amount at each measurement point. As shown in FIG. 8, the radial deflection amount of the gas turbine rotor 1 is indicated by the fluctuation range of the distance between the outer surface 50 a of the rotor disk 50 and the rotor rotation center line RC. Here, the rotor rotation center line RC refers to a straight line connecting the centers of the bearings S1 and S2. The figure center O ₁ of the cross section of the rotor disk 50 to be measured is calculated from the measured value of the shake amount on the outer surface 50a of the rotor disk 50, and the deviation between the calculated figure center O ₁ and the rotor rotation center line RC is calculated. Misalignment. Quantitative display of the misalignment obtained in this way is misalignment data consisting of the misalignment amount and misalignment angle. By plotting the amount of misalignment of each rotor disk according to the distance from the bearing S1, an axial curve distribution as shown in FIG. 10 is obtained.

Patent Document 1 discloses a rotating body misalignment calculating method and calculating apparatus. Patent Document 2 discloses an example of a vibration amount measuring apparatus for a turbine rotor.
JP 2001-91244 A Japanese Patent Laid-Open No. 11-230733

  In general, when calculating the shaft bending of the turbine rotor, the least square method disclosed in Patent Document 1 is applied to calculate the roundness of the rotor and the misalignment amount of the rotor shaft center, thereby calculating the shaft bending. The method is known.

  However, although the conventional method disclosed in Patent Document 1 is a highly accurate calculation method, there is a problem in that calculation is complicated and enormous calculation is required.

  The present invention has been made to solve such a problem, and an object of the present invention is to provide a system for calculating the shaft bending of a turbine rotor by a simpler method than the conventional method.

  The invention according to claim 1 is a turbine rotor shaft bending calculation system, comprising: an input unit for setting measurement points of a rotor disk constituting the turbine rotor; and at least a circumferential outer surface of the turbine rotor. With respect to four or more measurement points, a shake amount detection unit for deriving a displacement amount in the radial direction of the turbine rotor based on a measurement value measured by a displacement meter, and the radial direction derived by the shake amount detection unit A storage unit that stores a displacement amount and a measurement angle of the measurement point; and a calculation unit that calls the data stored in the storage unit to calculate the most probable center misalignment data of the turbine rotor. A calculation circle is calculated by selecting any three points from the radial displacement amount and the measurement angle at all measurement points stored in the storage unit, and all the measurement points are calculated from the calculation circle. Calculate the calculated circle value, calculate the difference between the calculated circle value and the displacement in the radial direction as the error amount for each measurement point, and sum the error amount to derive the total error amount value for all measurements. Repeat the calculation for the number of combinations of three measurement points for each point to calculate each error amount total value, and find the smallest calculation circle from the total error amount values for all the obtained combinations. Misalignment calculating means for calculating a deviation between the center of the most probable circle and the rotation center of the turbine rotor as the most probable circle misalignment data, and the most probable circle misalignment of all the rotor disks. Misalignment determining means for calculating a maximum misalignment amount with reference to the data, determining whether the maximum misalignment amount is within a reference value, and an axial bend for calculating an axial bend distribution from the most probable circular misalignment data Distribution calculating means.

  According to the configuration of the present invention, at least four or more measurement points are set along the circumferential direction of the rotor disk, and the calculation circle is calculated from the displacement amount and the measurement angle measured at each measurement point. Since the most probable circle is determined and the most probable circle misalignment data is calculated from the most probable circle, the axis bending can be calculated by a simpler method than the conventional method.

  The invention according to claim 2 is characterized in that the calculation unit includes a correction disk selection means for selecting a correction target rotor disk, and a correction amount determination means for determining a correction amount of the correction target rotor disk. .

  According to the configuration of the present invention, when the maximum misalignment amount exceeds the reference value, the correction disk selection unit selects the correction target rotor disk, and the correction amount determination unit determines the correction amount of the rotor disk. It is easy to correct the rotor shaft bending.

  The invention according to claim 3 is characterized in that the correction disk selecting means selects a rotor disk having the maximum misalignment amount as a correction target rotor disk.

  According to the configuration of the present invention, since the rotor disk having the maximum misalignment amount is targeted for correction, the maximum misalignment amount after correcting the rotor disk can be minimized.

  The invention according to claim 4 is characterized in that the correction amount determining means selects the maximum misalignment amount of the correction target rotor disk as a disk correction amount.

  According to the configuration of the present invention, since the disk correction amount is determined based on the maximum misalignment amount, the correction method is the simplest and the shaft bending of the turbine rotor after the rotor disk correction can be minimized.

  According to the configuration of the present invention, since the most probable circle can be easily selected, the shaft bending of the rotor can be accurately calculated by a simple method without performing enormous calculation as disclosed in Patent Document 1.

  Although embodiments of the present invention will be described with reference to the drawings, these are merely examples of the embodiments and are not limited to these embodiments. The constituent elements of this embodiment include those that can be easily replaced by those skilled in the art or those that are substantially the same. In the following description, a gas turbine rotor will be described as an example.

  When calculating the axial curve distribution of the gas turbine rotor 1, the calculation of the misalignment amount of the rotor disk 50 is the basis. The basic concept of the method for calculating the misalignment amount will be described below.

FIG. 1 shows a cross section of the rotor disk 50 of the gas turbine rotor 1 and shows the relationship between the measured value, the reference circle, and the calculated circle in the cross section. Along the circumferential direction of the rotor disk 50, the outer surface of the rotor disk 50 is divided into a plurality (m) at equal intervals to define each measurement point X _i (i = 1 to m), and the rotor disk 50 is outlined. Measured values measured by the displacement meter 2 at each measurement point X _i (i = 1 to m) while rotating once in the direction of the arrow (the distance from the installation position of the displacement meter 2 to the outer surface of the rotor disk and the measurement start point) , The radial displacement amount a _i on the outer surface of the gas turbine rotor is derived. That is, the distance to the outer surface of the displacement meter 2 and the rotor disk 50 at the measurement starting point X ₁ for convenience "0 (zero)", the distance between the measurement starting point X at each measurement point X _{i (i} = 2 to m) The distance at ₁ is compared, and the relative distance difference between them is the displacement amount ai. Note that the measurement points X _i (i = 1 to m) may be selected at equal intervals or not at equal intervals.

The reference circle is not an element directly related to the configuration of the present invention, but is displayed for convenience as a circle whose graphic center coincides with the rotation center O ₂ of the rotor disk 50. In the case of a gas turbine rotor, the reference circle is a perfect circle, and the center of the reference circle is the rotation center O _{2 of the} rotor.

The calculated circle is determined from the measured value P _i (measurement angle θ _i and radial displacement a _i ) of the shake amount at each measurement point X _i (i = 1 to m). Define the position and measuring the angle of each measurement point _X i (i = 1~m) with respect to the circumferential direction of the measurement points (m), the measurement values _P i at each measurement point _X i (i = 1~m) ( If any three points are selected from the measurement angle θ _i and the radial displacement amount a _i ), one circle can be determined without fail. A circle determined by these three points is a calculated circle. The calculation circle is determined by a combination of arbitrary three measurement points among all the measurement points (m), and there are ( _m C ₃ ) combinations in total. Here, ( _m C ₃ ) is the total number of combinations in which all three combinations are selected when any three points are selected for (m) measurement points X _i (i = 1 to m). Means. Therefore, if n = ( _m C ₃ ), there are (n) calculation circles.

In the present invention, arbitrary three points are selected from all the measurement points X _i (i = 1 to m), and one calculation circle is calculated from these three points. FIG. 1 shows a calculation circle calculated from three measurement points X ₁ , X ₂ and X _m as an example.

Next, for each measurement point X _i (i = 1 to m), each measurement value P _i (measurement angle θ _i and radial displacement a _i ) and the radial deviation between the calculated circles, that is, each The difference between the measured value P _i and the calculated circle value (meaning the value on the calculated circle; the significance of the calculated circle value will be described later) is calculated as the error amount Δ _{i, j} . Specifically, a calculated circle value Q _{i, j} at each measurement point X _i (i = 1 to _m) is calculated from the calculated circle _, and an error amount Δ _i is calculated from the measured value P _i and the calculated circle value Q _{i, j. , J} is calculated. Further, an error amount total value ΔS _j for one calculated circle is calculated from each error amount Δ _{i, j} . Next, other calculated circles are sequentially calculated from the other combinations of all three measurement points, and the error amount total value ΔS _j is similarly calculated for each calculated circle.

After calculating the error amount total value ΔS _j (j = 1 to n) for all the calculated circles, the smallest error amount total value ΔS _j (j = 1 to n) is determined as the minimum error amount total value ΔS _j. (J = a) is selected, and the calculated circle corresponding to this minimum error amount total value ΔS _a (j = a) is the most probable circle. Assuming that the most probable circle displays the figure closest to the cross-sectional shape of the gas turbine rotor among all the calculated circles, the center of the most probable circle is considered as the figure center. The misalignment between the center of the most probable circle and the center of rotation is the misalignment. Which quantitatively displays the status of the misalignment is a misalignment data misalignment amount (eccentricity e) consists misalignment angle theta _a. The calculation of the misalignment amount (eccentricity e) the misalignment angle theta _a, the degree of misalignment of the rotor disk 50 can be easily determined, it is easy to validity of the judgment of the data.

For each calculated circle, one error amount Δ _{i, j} is calculated for each measurement point X _i (i = 1 to _m ), and one error amount total value ΔS _j can be calculated for each calculated circle. Further, one most probable circle is determined for all measurement points X _i (i = 1 to m).

The above method will be specifically described with reference to FIG. In FIG. 1, each measurement value at each measurement point X _i (i = 1 to m) is represented by P _i (θ _i , a _i ). Here, the symbol “i” is selected from any one of “1” to “m”, and the measurement point X _i means the “i” -th measurement point from the measurement start point (X ₁ ). To do. The symbol “θ _i ” represents the measurement angle in the clockwise direction from the measurement start point X _{1 at} the measurement point X _i (i = ₁ to m), and the symbol “a _i ” represents the measurement point X _i (i = 1 to 1). The displacement in the radial direction in m) is shown.

  From a combination of arbitrary three measurement points, one calculation circle can be determined by the method described later (Formula 4). In addition, other calculation circles can be calculated from other combinations of any three measurement points by the same method, and (n) calculation circles can be determined at the front.

Next, the significance of the calculated circle value Q _{i, j} will be described with reference to FIG. The calculated circle value Q _{i, j} is a value on one calculated circle selected from a combination of arbitrary three measurement points among all measurement points X _i (i = 1 to m). The calculated circle value Q _{i, j} has a point corresponding to the measurement point X _i (i-th measurement point from the measurement start point), that is, has the same measurement angle θ _i as the measurement point X _i and is calculated from the calculation circle. It is a value on the calculation circle. The calculated circle value Q _{i, j} is indicated by the sign “Q _{i, j} (θ _i , b _{i, j} )”. Here, similarly to the above, the symbol “θ _i ” indicates the clockwise measurement angle from the measurement start point of the measurement point X _i , and the symbol “b _{i, j} ” indicates that the measurement angle is “θ _i ”. Indicates a calculated value on a certain calculation circle. The calculated circle value Q _{i, j} (θ _i , b _{i, j} ) can be calculated from the calculated circle and the measurement angle θ _i if the calculated circle is determined. In the case of (m) measurement points, since there are (n) calculation circles, the codes “i” and “j” displayed below are changed from “1” to “m”. "J" means a point selected from any one of "1" to "n". That is, the symbol "i" indicates a rank number of the measurement points from the measurement starting point X ₁ with respect to (m) number of measurement points, code "j" is symmetrical with respect to (n) pieces of calculated circles The ranking number of the calculated circle.

The difference between each measured value P _i (θ _i , a _i ) and the corresponding calculated circle value Q _{i, j} (θ _i , b _{i, j} ) is calculated at each measurement point X _i (i = 1 to m). if error amount Δ _{i, j,} the error amount delta _{i, j} is represented by equation (1).
(Expression 1) Δ _{i, j} = [P _i (θ _i , a _i ) −Q _{i, j} (θ _i , b _{i, j} )] ²

In Equation 1 _, the difference between the measured value P _i and the calculated circle value Q _{i, j} is squared to eliminate the influence of the plus / minus sign of the difference value between the two, and the measured value includes an abnormal value. This is because the difference between the abnormal value and the normal value is further expanded in consideration of the case where the abnormal value is selected, and the abnormal value is easily selected.

Next, with respect to the target calculation circle, the error amount Δ _{i, j} is calculated by Equation 1 for all measurement points X _i (i = 1 to _m ).

Further, the sum of these error amounts Δ _{i, j} is the total error amount value ΔS _j and is expressed by the following equation (2).
(Expression 2) ΔS _j = Σ (Δ _{i, j} )
For the target calculation circle, the total error amount ΔS _j is obtained by adding the error amounts Δ _{i, j} at each measurement point shown in Equation ₁ from the measurement point X ₁ to the measurement point X _m .

Next, another calculation circle is determined in the same manner from any other combination of three measurement points X _i (i = 1 to m). Further, the error amount Δ _{i, j} and the error amount total value ΔS _j are calculated for each calculation circle using Equation 1 and Equation 2. Since one error amount total value ΔS _j can be calculated for each calculated circle, (n) total error amount values ΔS _j (j = 1 to n) are calculated for (n) calculated circles. It can be calculated.

(N) After calculating the error amount total value ΔS _j (j = 1 to n) for each of the calculated circles, the smallest error amount total value ΔS _j among the error amount total values ΔS _j (j = 1 to n). (J = a) is selected, and the calculated circle having this minimum error amount total value ΔS _a (j = a) is the most probable circle. The most probable circle is regarded as displaying the figure (true circle) closest to the cross-sectional shape of the rotor among all the calculated circles, and the center of the most probable circle is considered as the figure center of the cross section. The misalignment between the center of the most probable circle and the rotation center O ₂ of the rotor disk 50 is referred to as misalignment. In FIG. 1, the deviation length (eccentric distance e) between the rotation center O ₂ of the rotor disk 50 and the most probable circle center is the misalignment amount. The angle theta _a to in the clockwise direction around shows the misalignment direction from the measurement starting point X ₁ is a misalignment angle. Incidentally, the rotation center O _2, as described above, the same meaning as the rotor rotational center O ₂ shown in FIG.

  If the misalignment is determined by such a method, data can be acquired by a simple method as compared with the least square method or the like which is the conventional technique disclosed in Patent Document 1.

  Even if the measured value includes an abnormal value, the abnormal value is surely excluded in the process of calculating the most probable circle. That is, since the calculated circle is a circle determined from arbitrary three measured values, there is always a calculated circle that does not include an abnormal value. Accordingly, the circle with the smallest total error amount is necessarily selected as the most probable circle from the calculated circles that do not include an abnormal value. Further, since the abnormal value can be specifically identified, the abnormal value can be excluded and replaced with the measured value after remeasurement (remeasured value). On the other hand, in the least square method or the like, which is a conventional technique shown in Patent Document 1 and Patent Document 2, an outlier value is taken in and the misalignment is calculated. Moreover, since an abnormal value cannot be specified, it is difficult to exclude an abnormal value and replace it with a remeasured value.

  Next, regarding the method of calculating the error amount by calculating the calculated circle from the measured value, the significance in the plane coordinates will be outlined.

The change in runout when the rotor having misalignment is rotated can be approximated to the runout of the eccentric disc cam. FIG. 3 shows a conceptual diagram of the eccentric disk cam. In FIG. 3, the eccentric disk cam is composed of a rotating disk A and a follower B, and the follower B is composed of a flat plate C and a shaft portion D fixed thereto. The follower B is in contact with the peripheral surface of the rotating disk A through a flat plate C at a contact P. The follower B has a structure in which the shaft portion D can only move in the axial direction (vertical direction on the paper surface of FIG. 3) in the restraining member E, and the follower B as a whole is rotated according to the rotation of the rotating disk A. It has a structure that can move up and down.
Further, the rotating disk A rotates around a point O ₂ eccentric from the figure center O ₁ by a distance e, and the rotation center O ₂ is located on the axis of the shaft portion D of the follower B. In such an eccentric disc cam, when the rotating disc A rotates about the eccentric point O ₂ , the follower B moves up and down with respect to the paper surface as the rotation angle β changes.

FIG. 4 shows how the contact point P varies with the movement of the rotating disk A. FIG. 4 shows that when the rotating disk A rotates about the rotation center O ₂ and the rotation angle β changes clockwise from 0 ° to 360 ° in increments of 45 °, the contact P is the contact P _1. from to the contact point P _9, it shows how the changes in the vertical direction relative to the rotational angle beta.

According to FIG. 4, when the rotation angle β is “0 °”, the disk center O ₁ , the rotation center O ₂ and the axis of the shaft portion D of the follower B coincide with each other in the vertical direction (on the paper surface). This means that the rotation center O ₂ is between the contact P (P ₁ ) and the disk center O ₁ . In this state, the chord Z ₁ Z ₂ forming the diameter of the rotating disk A exists on a vertical line where the axes of the disk center O ₁ , the rotating center O ₂ and the shaft portion D coincide with each other in the vertical direction with respect to the paper surface. To do. The rotation angle β that changes with the rotation of the rotating disk A is the chord Z ₁ Z ₂ and the vertical line (a straight line in the vertical direction with respect to the paper surface that connects the rotation center O ₂ and the axis of the shaft portion D). ) And the clockwise angle.

In FIG. 4, when the relative positional relationship in the vertical direction with respect to the rotation center O ₂ of the contact P is viewed, the position of the contact P moves up and down as the rotation angle β changes. The locus of the contact P draws a sine curve (substantially the same even if called a cosine curve) as will be described later. At the position where the rotation angle β is “0 °”, the height of the contact P (P ₁ ) shows the minimum value (the vertical distance between the flat plate C and the rotation center O ₂ is minimum), and the rotation angle β is “180 °. ", The height of the contact point P (P ₅ ) indicates the maximum value (the vertical distance between the flat plate C and the rotation center O ₂ is maximum).

The difference between the maximum value (P ₅ ) and the minimum value (P ₁ ) of the displacement of the contact P is the maximum swing width. This maximum deflection width is twice the amount of misalignment of the rotating disk A, that is, the eccentric distance e between the graphic center O ₁ and the rotating center O ₂ of the rotating disk A. Further, the rotation angle β in FIG. 4 is synonymous with the measurement angle θ in FIG. 1, and in the following description, the rotation angle β is replaced with the measurement angle θ.

Assuming that the displacement of the follower B of the eccentric disk cam, that is, the displacement in the vertical direction of the contact point P is “y”, the displacement y is expressed by the following equation (3).
(Equation 3) y = e (1-cos θ)

Equation 3 represents the displacement of the follower B when the rotating disk A rotates by the measurement angle θ, that is, the displacement of the contact P, and is represented by a sine curve passing through the origin of the y-θ coordinate. . Further, as described above, the state where the measurement angle “θ” is “0 °” means a state where the axis of the disk center O ₁ , the rotation center O ₂ and the axis D are coincident, and the displacement of the contact point P is minimum. Means the position. The displacement y at this time is “0 (zero)”. It can be considered that the displacement y of the contact point P corresponds to a change in radial deflection of the gas turbine rotor in which misalignment occurs in the present invention.

Measurement of deflection amount in the radial direction of the gas turbine rotor, the displacement at the measurement starting point X ₁ (distance from the displacement meter 2 until the outer surface of the rotor disk 50) to "0" (zero), the measurement starting point X The displacement at other measurement points with reference to ₁ is measured as a change in the reading of the displacement meter. On the other hand, in Equation 3, when the measurement angle θ is “0 °”, the displacement y is “0 (zero)”. Generally, when measuring the roundness of a gas turbine rotor having misalignment, the position where the measurement angle θ is “0 (zero)” (the position where the displacement is minimum) is unknown at the start of measurement. Therefore, actual measurement is started when the measurement angle θ is “θ _a ” and the displacement y is “y _a ”, and this point is set as a measurement start point X ₁ . In addition, coordinate conversion of Formula 3 is performed so that the measurement angle at this time is “0 °” and the displacement Y is “0 (zero)”.

The shake equation after the coordinate conversion is expressed by Equation 4 on the XY coordinates.
(Equation 4) Y = e [1-cos (X + θ _a )] − y _a

This equation is a shake equation that is the basis of the misalignment calculation method according to the present invention. Here, the measurement angle X means a measurement angle (rotation angle) from the measurement start point. The displacement Y means the amount of displacement at the measurement angle X. The angle “θ _a ” is called an initial angle, and the displacement “y _a ” is called an initial displacement. The eccentric distance e between the rotation center O ₂ and the figure center O ₁ of the eccentric disc cam shown in FIG. 3 corresponds to the amplitude of the sine curve of Formula 4 (1/2 of the total deflection width).

Expression 4 variables X, by substituting the measured value of the three measurement points in Y, Sadamari constants _e, θ _a, y _a, wherein the one calculated circle is determined.

  FIG. 2 shows the relationship between the measured values of the rotor shown in FIG. 1, the reference circle, and the calculated circle in plane coordinates. In FIG. 2, the horizontal axis X indicates the measurement angle from the measurement start point, and the vertical axis Y indicates the displacement at the measurement point. The shake equation represented by Equation 4 is represented by a sine curve passing through the origin O on the XY coordinates shown in FIG.

In FIG. 2, the calculated circle is indicated by a solid line. The reference circle is a perfect circle, and all the displacements for all measurement angles are considered to be “0 (zero)”. Therefore, the reference circle coincides with the X axis. Origin O is the measurement starting point _{X 1.} The measurement points in the circumferential direction of the rotor are displayed as measurement points X _i corresponding to the measurement angle θ _i by dividing the X axis from 0 to 360 ° into (m) pieces on the XY coordinate. . The measurement value P _i for each measurement point X _i (i = 1 to m) is an actual measurement value. The relationship between the original coordinates y-θ coordinate and XY coordinate is shifted by the measurement angle “θ _a ” on the X axis and the displacement “y _a ” on the Y axis. The amplitude of this sine curve (1/2 of the total runout width) corresponds to the misalignment amount. The initial angle “θ _a ” corresponds to the misalignment angle.

As described above, the calculation circle determined from the measurement values at the arbitrary three measurement points in FIG. 1 expanded to the XY coordinates corresponds to the locus of the calculation circle indicated by the solid line in FIG. In FIG. 1, a calculation circle determined from three points of measurement values P ₁ , P ₂ , and P _m for the measurement points X ₁ , X ₂ , and X _m is displayed as an example. The calculated circle shown in FIG. 2 is obtained by expanding the calculated circle in FIG. 1 into a XY coordinate as a sine curve. 1 and 2 show only one calculated circle (jth calculated circle), there are actually (n) calculated circles determined from Equation (4).

Further, the error amount Δ _{i, j} displayed by the equation 1 is represented as a difference between the measured value P _i and the calculated circle value Q _{i, j} on the calculated circle in FIG. More specifically, the measurement value of the shake amount is displayed as P _i (θ _i , a _i ) at the measurement point X _i (measurement angle θ _i ) on the X axis. The calculated circle value on the calculated circle is displayed as Q _{i, j} (θ _i , b _{i, j} ). Accordingly, the error amount Δ _{i, j} can be displayed as a difference between the measured value P _i (θ _i , a _i ) and the calculated circle value Q _{i, j} (θ _i , b _{i, j} ). However, as described above, the error amount Δ _{i, j} is expressed by Equation 1 in consideration of the difference in plus / minus sign of the difference between the measured value and the calculated circle value and the ease of selecting the abnormal value. As shown by the equation, the difference between the measured value P _i and the calculated circle value Q _{i, j} is squared.

Next, an error amount Δ _{i, j} is calculated, and an error amount total value ΔS _j (j = 1 to n) is calculated. If the minimum error amount total value ΔS _j (j = a) is selected after calculating the total error amount ΔS _j (j = 1 to n) for each calculated circle, this minimum error amount total value ΔS _a (j = The calculated circle with a) is the most probable circle.

The difference between the center of the most probable circle finally selected and the center of rotation (coincident with the center of the reference circle) is misalignment. That is, the misalignment data is represented by the misalignment amount and the misalignment angle. 2, the misalignment amount is calculated as the amplitude of the sine curve of the most probable circle, the misalignment angle is calculated as the initial angle theta _a. The misalignment amount and misalignment angle thus determined are the most probable circular misalignment data obtained by the present invention.

  Increasing the number of measurement points (m) in the circumferential direction of the rotor increases the calculation accuracy of the most probable circular misalignment data, but increases the calculation amount. On the other hand, if the number of measurement points is reduced, the calculation accuracy of the most probable circle misalignment data becomes worse. However, from the concept of the present invention, the number of measurement points needs to be at least 4 or more. This is because the basic idea of the present invention is not established when the number of measurement points is 3 or less.

  Next, the configuration of the gas turbine rotor shaft bending calculation system based on the rotor disk misalignment calculation method will be described with reference to FIG.

  The system 60 includes an input unit 61, a shake amount detection unit 62, a storage unit 63, a calculation unit 64, and a display unit 65.

  The input unit 61 receives initial values such as the number of rotor disks (DM) and the number of circumferential measurement points (m) for each rotor disk. The number of measurement points (m) may be a different number of measurement points for each rotor disk.

  The shake amount detection unit 62 selects at least four measurement points in the circumferential direction along the outer surface of all rotor disks from the input measurement points (m) in the circumferential direction of the rotor disk, and the rotor disk The displacement meter 2 is installed close to the outer surface. The rotation angle of the rotor disk is measured by a tachometer 3 provided separately or a tachometer 3 provided in the gas turbine. While rotating the gas turbine rotor 1, the measurement values at each measurement point of the rotor disk 50 are read and taken into the storage unit 63. The measurement values to be measured are the shake amount (radial displacement amount) and the measurement angle (rotation angle from the measurement start point). Various known sensors are applied as the displacement meter. For example, in addition to a contact sensor such as a dial gauge, a non-contact sensor such as a laser sensor, a capacitance sensor, or an ultrasonic sensor can be used.

The measurement points of the rotor disk 50 are selected in advance by marking the measurement positions of all measurement points X _i (i = 1 to m) on the outer surface of the rotor disk. The measurement values at each measurement point are automatically taken as the measurement values when the rotor disk 50 is rotated at a low speed and reaches the measurement position at a predetermined measurement point. The position of the measurement point is confirmed by a CCD sensor or the like (not shown) at a preset measurement position. For the measurement per rotor disk, the measurement starting point is determined, and the rotor is rotated once while confirming the position of the measurement point, thereby collecting the measurement values of all the measurement points. As the measurement point, only the measurement start point may be selected on the outer surface of the rotor disk, and the other measurement points may be selected based on the rotation angle from the measurement start point. Further, the position of the measurement start point may be input from the input unit 61 for each rotor disk.

  When the measurement of one rotor disk 50 is completed, the rotor is moved, the position of the measurement start point of the next adjacent rotor disk is determined, and measurement is performed in the same procedure. When the measurement of all the rotor disks is completed, the rotor disk measurement work is completed. In measurement, the displacement meter may be moved without moving the rotor.

  The calculation unit 64 includes various functions including a misalignment calculating means 641, a misalignment determining means 642, an axial bend distribution calculating means 643, a corrected disk selecting means 644, and a correction amount determining means 645.

In misalignment calculating unit 641 calls the measurement values P _i consisting of the measured value of the measured angle and shake amount incorporated in the storage unit 63 (the displacement amount in the radial direction), based on the equation (4), from 3 measurement points Determine the calculation circle. A calculated circle value Q _{i, j} for each measurement point X _i (i = 1 to _m ) is calculated from the determined calculated circle. Next, an error amount Δ _{i, j} is calculated from the measured value P _i and the calculated circle value Q _{i, j at} each measurement point X _i (i = 1 to _m) based on the equation (1). Further, an error amount total value ΔS _j is derived from the error amount Δ _{i, j} by Equation (2). By similar processing, other calculation circles are determined from combinations of the other three measurement points, and an error amount total value ΔS _j (j = 1 to n) for each calculation circle is calculated. The minimum error amount total value ΔS _a (j = a) is determined from the error amount total value ΔS _j (j = 1 to n), and the calculated circle is determined as the most probable circle. The deviation between the center of the most probable circle and the center of rotation is calculated, and misalignment data (center misalignment amount and misalignment angle) is determined. This is the most probable circular misalignment data of the target rotor disk. The calculation is repeatedly performed for all the rotor disks, the most probable circle misalignment data of each rotor disk is calculated, and stored in the storage unit 63.

  The misalignment determination means 642 refers to the misalignment amount from the most probable circular misalignment data of all the rotor disks called from the storage unit 63, and determines the rotor disk having the maximum misalignment amount and the maximum misalignment amount. Next, it is determined whether or not the maximum misalignment amount is within a reference value (reference value 2). When it is determined that the maximum misalignment amount is within the reference value, each calculated most probable circular misalignment data is determined to be appropriate, and a shaft bending distribution of the gas turbine rotor is created and stored in the storage unit 63. As for the most probable circular misalignment data, one most probable circular misalignment data is determined for each rotor disk, but the maximum misalignment amount is one maximum misalignment for the entire turbine rotor assembly. The amount is determined.

  When the maximum misalignment amount exceeds the reference value (reference value 2), it is determined that the turbine rotor shaft bending needs to be corrected. The basic idea of correcting the turbine rotor shaft bend has already been outlined in the background art (FIG. 8). That is, a rotor disk to be corrected may be selected to determine the disk correction amount. A specific method will be described later.

  Based on the most probable circular misalignment data determined to be appropriate, the axial bend distribution calculation unit 643 plots the misalignment amount and misalignment angle for each rotor disk, and creates the axial bend distribution of the gas turbine rotor. An example of the axial curve distribution is shown in FIG.

  In the display unit 65, the measured value (measurement angle and radial displacement amount) of each rotor disk called from the storage unit 63, the most probable circular misalignment data (center misalignment amount and misalignment angle), and the gas turbine rotor Displays the distribution of the axis curvature of the. Further, when an abnormal value occurs in the maximum misalignment amount, the corresponding measurement point and the measurement value and the error amount at the measurement point are displayed. An example of the input / output screen of this system is shown in FIG.

  The correction of the shaft bending is performed by the correction disk selection means 644 and the correction amount determination means 645. Hereinafter, the correction disk selection unit 644 and the correction amount determination unit 645 will be described.

  As described above, the correction disk selection unit 644 selects only one rotor disk to correct the rotor disk, and selects the rotor disk 50 having the maximum misalignment amount as the correction target disk. This is because the maximum misalignment after the disk correction can be minimized as compared with the case where another rotor disk is selected as the correction disk.

  The correction amount determination means 645 will be described with reference to FIG. FIG. 8 shows the state of the shaft bending in which the gas turbine rotor 1 shows the maximum misalignment amount Z in the rotor disk 50 as described above. In FIG. 8, a graphic FGHI shows a cross-sectional shape in the rotor longitudinal direction when the rotor disk is cut along a plane including the rotor axis. The cross-sectional shape of the adjacent rotor disk 51 is indicated by a graphic FGLM. The contact surfaces FG of both the rotor disks 50 and 51 become the contact surfaces 50b and 51b of the rotor disks. In the gas turbine rotor 1, a plurality of rotor disks are stacked via a joint surface, and both ends are supported by a bearing S <b> 1 and a bearing S <b> 2 to constitute an integrated gas turbine rotor 1.

In addition, FIG. 8 exaggeratedly shows the state of the shaft bending of the rotor. In practice, the horizontal distance from the bearing S1 to the rotor disk 50 is sufficiently longer than the rotor misalignment amount Z. Therefore, the outer surface 50a of the rotor disk is a surface substantially parallel to the rotor rotation center line RC. Further, the figure FGHI showing the cross-sectional shape of the rotor disk 50 in the rotor longitudinal direction is expressed as a trapezoidal shape, but is actually a shape close to a rectangle. In this example, the axis O ₁ of the rotor disk 50 is displaced by the maximum misalignment Z with respect to the rotation center O ₂ of the rotor, and the rotor disk 50 (cross section FGHI) is moved to the contact surfaces 50b and 51b (contact surface FG). ), The case where it joins with the adjacent rotor disk 51 (cross-section FGLM) is shown.

The method of correcting the shaft bending is performed by selecting a correction target disk and correcting the thickness distribution of the rotor disk. That is, in FIG. 8, the parallelism of the contact surfaces 50b and 50c, 51b and 51c of the rotor disk 50 to be corrected and the adjacent rotor disk 51 is corrected, and the contact surface angles (α ) To reduce the shaft bending. If the parallelism of the contact surfaces 50b, 50c, 51b, 51c is ensured, the contact surface angle (α) is substantially “0 (zero)”. That is, how to correct axial deformation of the gas turbine rotor, so that the position of the axis _{O 1} of the rotor disk 50 and 51 moves to the position of the rotation center _{O 2} of the rotor, Settemen 50b of the rotor disk 50 and 51, 51b What is necessary is just to cut a part of.

  In FIG. 8, the cutting method will be described in detail by taking the rotor disk 50 (cross-sectional shape FGHI) to be corrected as an example. A perpendicular is dropped from the point G of the figure FGHI to the side FI, the intersection is R, and the length of the side FR is L1. If the cross section FGR indicated by the hatched portion is cut out of the cross section shape FGHI, the parallelism of the contact surfaces 50b and 50c of the rotor disk 50 is maintained. Similarly, with respect to the adjacent rotor disk 51 (cross-sectional shape FGLM), a perpendicular is dropped from the point G to the side FM, the intersection is T, and the length of the side FT is L2. If the cross section FGT indicated by the hatched portion is cut, the parallelism of the contact surfaces 51b and 51c of the adjacent rotor disk 51 can be secured. In this manner, the portions (cross-section FGR, cross-section FGT) indicated by the hatched portions of the two rotor disks 50, 51 are cut across the joint surfaces 50b, 51b of the adjacent rotor disks 50, 51, and the rotor disks are parallel to each other. If the degree is corrected, the shaft bending of the rotor can be corrected.

  A specific method for calculating the cutting amount of the correction target disk will be described below. Here, the distance on the rotor rotation center line RC from the bearing S1 to the joint surface 50b of the rotor disk 50 is LL1, and the angle between the rotor axis CC1 and the rotor rotation center line RC is the rotor axis inclination angle (α1). And Similarly, the distance on the rotor rotation center line RC from the bearing S2 to the contact surface 51b is LL2, and the angle formed between the rotor axis CC2 of the adjacent rotor disk 51 and the rotor rotation center line RC is the rotor axis tilt angle ( α2). The diameter of the rotor disks 50 and 51 is DD.

In the cross-sectional shape FGHI, the triangle FGR of the rotor disk 50 to be cut is similar to the triangle O ₁ O ₂ S1 formed from the bearing S1, the rotor axis CC1, and the rotor rotation center line RC. Therefore, the length L1 of the side FR is expressed by Equation 5.
(Equation 5) L1 = Z × (DD / LL1)

Similarly, the length L2 of the side FT of the adjacent rotor disk 51 is expressed by Equation 6.
(Expression 6) L2 = Z × (DD / LL2)

Therefore, a cross-section FGR corresponding to the length L1 is cut from the joint surface 50b toward the bearing S1 with respect to the longitudinal direction of the rotor on the basis of the outer surface of the rotor disk 50, and on the basis of the outer surface of the adjacent rotor disk 51. If the cross section FGT corresponding to the length L2 is cut from the joint surface 51b toward the bearing S2, the shaft bending of the gas turbine rotor 1 can be eliminated. The cross section FGR and the cross section FGT corresponding to the side FR (length L1) and the side FT (length L2) determined in this way are correction amounts given to the correction target rotor disk.
The contact surface angle (α) is the sum of the rotor axis inclination angles (α1) and (α2).

  In the above correction method, the case where the rotor disks 50 and 51 on both sides are cut across the joint surfaces 50b and 51b of the rotor disks 50 and 51 adjacent to each other is shown. However, the rotor disk 50 having the maximum misalignment amount. You may cut only. In this case, on the basis of the outer surface of the rotor disk 50, the cutting is performed by correcting the length L1 of the side FR and the length L2 of the side FT from the joint surface 50b toward the bearing S1. That is, only one rotor disk 50 is cut using a length obtained by adding up the lengths L1 and L2 of the outer surface standard corresponding to the cutting amount of the rotor disks 50 and 51 on both sides across the joint surfaces 50b and 51b. And you can fix it. This is because, since the distance LL1 from the bearing S1 to the rotor disk 50 is sufficiently larger than the diameter DD of the rotor disk, a large error does not occur.

  The above simple method may be applied to the adjacent rotor disk 51 instead of the rotor disk 50, but it is preferable to apply to the rotor disk 50 having the maximum misalignment amount. This is because the maximum misalignment after correction is the smallest.

  According to the configuration of the above-described gas turbine rotor shaft bending calculation system, it is possible to calculate the gas turbine rotor shaft bending by a simpler method than the conventional method.

  Next, the calculation procedure of the axial bending of the gas turbine rotor according to the present invention will be described with reference to FIG.

  First, the number of rotor disks (DM), rotor disk specifications (disk diameter DD, disk thickness, distance LL1, LL2 from the bearing, etc.), and the number of measurement points (m) in the circumferential direction of the rotor disk at the input unit 61 are as follows. Entered. Based on these input data, the shake amount (radial displacement amount and measurement angle) of each rotor disk is derived by measurement by the shake amount detection unit 62 (step S1).

Next, the number (n) of combinations of three measurement points is determined from the input measurement points (m) in the circumferential direction of the rotor disk (step S2). The number of combinations (n) can be determined by n = ( _m C ₃ ).

  Next, arbitrary three measurement points are selected (step S3). One calculated circle is determined from the three selected points.

  The measured values of the selected three points are substituted into Equation 4 to determine a calculation circle (step S4).

Calculated circle values Q _{i, j} (θ _i , b _{i, j} ) for all measurement points X _i (i = 1 to _m ) are calculated for the determined calculated circle. From each measured value P _i (θ _i , a _i ) and calculated circle value Q _{i, j} (θ _i , b _{i, j} ), at each measurement point X _i (i = 1 to _m ) according to Equation (1). The error amount Δ _{i, j} is calculated (step S5).

When the calculation of the error amount Δ _{i, j} with respect to all the measurement points X _i (i = 1 to _m ) is completed, the error amount total value ΔS _j is calculated for the target calculation circle by Equation (2) (step S6). . By completing this step, calculation of one error amount total value ΔS _j for one calculated circle is completed.

The calculation from step S3 to step S6 is repeated for the number of combinations (n) of three measurement points with respect to all the measurement points (m) (step S7). By repeating the calculation for all combinations (n), one error amount total value ΔS _j can be calculated for each calculated circle for (n) calculated circles.

Next, the minimum error amount total value ΔS _a is selected from among the (n) total error amount values ΔS _j (j = 1 to n), and the most probable circle is determined (step S8). By determining the most probable circle, a calculated circle value corresponding to each measured value for the most probable circle, that is, the most probable circle calculated circle value can be calculated. The error amount between the most probable circle and the measured value, that is, the most probable circle error amount Δ _{i, a} (i = 1 to m, j = a) is determined from the most probable circle calculated circle value and each measured value P _i. .

Then, for each measured value P _i, it is determined whether or not the necessity of determining abnormal value (step S9).

The reason for following this procedure is as follows. Even if the measured value includes an abnormal value, the most probable circle determined is the correct most probable circle. That is, normally, there are very few abnormal values for the total number of measurement points (m) of the measurement points X _i (i = 1 to m). Therefore, even if it contains an abnormal value on the measurement values P _i, in the process of calculating the calculated circle by a combination of any three points, always a combination of three points does not include the abnormal value measured value P _i present To do. That is, since calculated circle always exists that does not include the abnormal value, even if tentatively containing outliers measurement values P _i, the most probable circle to be finally determined, including the abnormal measurement values P _i There will be no correct most probable circle. In other words, it is used to determine the measurement value most probable circle while containing an abnormal value to P _i, there is no difficulty in determining the most probable circle. Therefore, if you want to proceed easily with misalignment calculation, correct misalignment data can be obtained correctly even if you perform misalignment calculation without determining the abnormal value of each measured value, and misalignment calculation work as it is Can be terminated.

When determination is not necessary in the process of determining whether or not abnormal value determination is necessary, the most probable circular misalignment data (center misalignment) is calculated (step S10), and the misalignment calculating operation ends. The most probable circle misalignment data is the misalignment between the center of the most probable circle and the rotation center O _2, and includes the misalignment amount (eccentric distance e) and the misalignment angle. Specifically, the misalignment amount corresponds to the amplitude of the sine curve of the most probable circle, misalignment angle corresponds to the initial angle theta _a. Even if the most probable circle is determined and the most probable circle misalignment data is calculated and the operation is terminated without determining whether or not the abnormal value determination is necessary, the invention is substantially the same as the present invention. It is. The procedure from the calculation of the number of combinations (n) (step S2) to the misalignment calculation (step S10) constitutes the misalignment calculating means 641.

  If it is determined that the abnormal value determination is necessary, the presence / absence of the abnormal value and the determination of the abnormal value are performed in the following steps S11 and S12.

That is, for the determined most probable circle, the error corresponding to the most probable circle among the error amounts Δ _{i, j} for each measurement point X _i (i = 1 to _m ) calculated in step S7 is the most probable circle error amount. Δ _{i, a} (i = 1 to m, j = a) is selected, and the most probable circle error amounts Δ _{i, a} (i = _{1 to} m) for all measurement points X _i (i = 1 to m). , J = a) is determined whether it is within the reference value (reference value 1) (step S11).

When the most probable circle error amount Δ _{i, a} (for example, i = f, j = a) exceeds the reference value, the measurement value P _f (θ _f , a _f ) at the measurement point X _f is abnormal. The value is recognized (step S12).

If all the most probable circle error amounts Δ _{i, a} (i = 1 to m, j = a) are within the reference values, it is determined that normal measurement has been performed, and the most probable circle misalignment data is obtained. The misalignment calculation work is completed (step S10). The most probable circle misalignment data is the misalignment length between the center of the most probable circle and the center of rotation, and includes the misalignment amount (eccentric distance e) and the misalignment angle. Specifically, the misalignment amount corresponds to the amplitude of the sine curve of the most probable circle, misalignment angle corresponds to the initial angle theta _a.

When the measured value P _f is recognized as an abnormal value, the shake amount detection unit 62 re-measures the shake amount (radial displacement amount and measurement angle) for all measurement points X _i (i = 1 to m). Then, the remeasurement value is acquired from the storage unit 63, and the measurement value P _i (i = 1 to m) is replaced with the remeasurement value (step S13).

When the replacement of each measurement value at the measurement point X _i (i = 1 to m) is completed, the process returns to step S2 and recalculation is performed. The calculation procedure of steps S2 to S9 shown in FIG. 5 is repeated until the most probable circle error amount Δ _{i, a} (i = 1 to m, j = a) falls within the reference value. The deviation calculation means 641 ends.

  When selecting a simpler axis bending calculation method, the steps S11, S12, and S13 are omitted, and the misalignment amount and misalignment angle are calculated regardless of the presence or absence of an abnormal value (step S10). The next rotor disk misalignment calculating operation (step S14) may be performed. Even in such a case, the basic idea of the present invention remains unchanged and is included in the scope of the present invention.

  Subsequently, the most probable circle misalignment data is calculated for the other rotor disks by repetitive calculation (step S14).

  With reference to the most probable circular misalignment data of all the rotor disks, the misalignment amount indicating the maximum value is selected as the maximum misalignment amount (step S15).

  The misalignment determining means 642 further determines whether or not the maximum misalignment amount is within a reference value (reference value 2) (step S16). If the maximum misalignment amount is within the reference value, it is determined that the rotor shaft bending is within the appropriate range, the shaft bending distribution is calculated (step S17), and the shaft bending calculation operation is completed.

  When the maximum misalignment amount exceeds the reference value, it is necessary to correct the shaft bending. The method of correcting the shaft bending is determined by the correction disk selection means 644 and the correction amount determination means 645.

  The correction disk selection means 644 selects a rotor disk that indicates the maximum misalignment among the rotor disks as a correction target disk (step S18). The reason why the rotor disk having the maximum misalignment amount is the correction target disk is that the shaft bending of the turbine rotor after the correction is the smallest.

  The correction amount determination means 645 determines the disk correction amount to be applied to the correction target rotor disk from the calculated rotor disk misalignment amount and misalignment angle. Specifically, the maximum misalignment amount Z calculated in step S15 is selected as the misalignment correction amount, and the cross-sectional shapes corresponding to the lengths L1 and L2 that can be calculated from Equations 5 and 6 (cross-section FGR in FIG. 8). , The cutting amount of the cross section FGT) is the disc correction amount to be obtained. In FIG. 8, the outer surface of the rotor disk may be cut with a cutting amount corresponding to the correction amount from the joint surface. Thus, the reason why the cutting amount calculated from the maximum misalignment amount is selected as the disk correction amount is that the correction method is the simplest and the shaft bending of the corrected rotor is the smallest (step S19). ).

  The disk to be corrected is cut to correct the rotor disk 50 (step S20). Next, after the corrected rotor disk is integrated, the shake amount detection unit 62 remeasures the shake amount of the rotor disk 50. The remeasurement value is taken into the storage unit 63, and the shake amount measurement value is switched from the initial measurement value to the remeasurement value (step S21). Thereafter, the process returns to the start, and the calculation restarts from the calculation of the number of combinations of the three measurement points (step S2).

  This calculation work is repeated in the misalignment determining means 642 until the maximum misalignment amount falls within the reference value (reference value 2) (step S16). If the maximum misalignment amount falls within the reference value, the axis bending distribution is calculated (step S17), and the axis bending calculation work is completed.

  When the shaft bending calculation system of the present invention is applied, the shaft bending of the rotor can be calculated by a simpler method than the conventional method. In addition, when a measurement including an abnormal value is performed, the abnormal value can be easily eliminated, and the operator can immediately determine whether or not to re-measure, so that the reliability of the shaft bending calculation work is improved. Further, in the correction work of the rotor disk, it is easy to specify the correction target disk and determine the correction amount, and it is also easy to correct the shaft bending.

The relationship of the measured value, reference | standard circle, and calculation circle concerning this invention is shown. The relationship between the measured value on the XY coordinates, the reference circle, and the calculated circle is shown. The conceptual diagram of an eccentric disk cam is shown. The relationship between the rotation angle of an eccentric disk cam and the displacement of a contact is shown. 1 shows a configuration of a turbine rotor shaft bending calculation system according to the present invention. The calculation procedure of the axis | shaft bending of the turbine rotor concerning this invention is shown. An example of the input / output screen of the axis bending calculation system is shown. The shaft bending generation | occurrence | production state figure of a gas turbine rotor is shown. The structure of a gas turbine rotor is shown. An example of axial curve distribution of a gas turbine rotor is shown.

Explanation of symbols

1 Gas turbine rotor
DESCRIPTION OF SYMBOLS 10 Compressor rotor part 11 Rotor blade 20 Turbine rotor part 25 Intermediate shaft 30 Spindle bolt 50 Rotor disk 50b, 50c Contact surface 51 Adjacent rotor disk 51b, 51c Contact surface S1, S2 Bearing RC Rotor rotation center line CC1, CC2 Rotor axis L1, L2 Reference length of outer surface of rotor disk corresponding to cutting amount LL1, LL2 Distance between bearing and rotor disk contact surface DD Rotor disk diameter DM Number of rotor disks M Rotor disk rank number X Measurement point P Measurement value Q Calculated circle value θ Measurement angle α Contact angle of rotor disk α1, α2 Rotor shaft center tilt angle β Rotation angle a Measured value b Calculated circle value Δ Error amount ΔS Total error amount e Between the figure center and rotation center of the turbine rotor Eccentric distance m Number of measurement points in the circumferential direction of the rotor disk n Combination of 3 measurement points

Claims (4)

A turbine rotor shaft bending calculation system,
An input unit for setting a measurement point of a rotor disk constituting the turbine rotor;
A deflection amount detector for deriving a radial displacement amount of the turbine rotor based on a measurement value measured by a displacement meter for at least four measurement points along the outer circumferential surface of the turbine rotor;
A storage unit for storing the radial displacement amount derived by the shake amount detection unit and the measurement angle of the measurement point;
A calculation unit that calculates the most probable circle misalignment data of the turbine rotor based on the data stored in the storage unit;
The calculation unit is
Call all the measurement points stored in the storage unit, select any three points from the radial displacement amount and the measurement angle to calculate a calculation circle, and calculate the calculation circle for the measurement point from the calculation circle Value is calculated, the difference between the calculated circle value and the radial displacement amount is calculated as an error amount for each measurement point, and the error amount is summed to derive a total error amount value for all measurement points. Repeat the calculation for each of the three combinations of measurement points to calculate each error amount total value, and make the smallest calculated circle from the total error amount values for all the obtained combinations as the most probable circle. A misalignment calculating means for selecting and calculating a deviation between the center of the most probable circle and the rotation center of the turbine rotor as the most probable circle misalignment data;
Misalignment determining means for calculating a maximum misalignment amount with reference to the most probable circular misalignment data of all the rotor disks, and determining whether the maximum misalignment amount is within a reference value;
An axis bending distribution calculating means for calculating an axis bending distribution from the most probable circular misalignment data;
An axial bending calculation system for a turbine rotor, comprising: The calculation unit includes a correction disk selection means for selecting a correction target rotor disk, and
Correction amount determining means for determining the correction amount of the correction target rotor disk;
The turbine rotor shaft bending calculation system according to claim 1, comprising:   3. The turbine rotor shaft bending calculation system according to claim 2, wherein the correction disk selection means selects a rotor disk having the maximum misalignment amount as a correction target rotor disk.   4. The turbine rotor shaft bending calculation system according to claim 2, wherein the correction amount determining means selects a disk correction amount calculated from the maximum misalignment amount as a correction amount.

Images