Cylindrical permanent magnet linear motor thrust characteristic optimization method based on matrix algorithm

Abstract

The invention discloses a cylindrical permanent magnet linear motor thrust characteristic optimization method based on a matrix algorithm, and the method comprises the steps: carrying out the grouping of a stator tooth part height, a stator tooth part width, a groove part height, a groove part width, an air gap width, a magnetic pole width, and a magnetic pole height parameter through an SEM measurement model, and constructing two dynamic measurement models; and performing data quantization on the SEM measurement model to respectively obtain a TPMLM average thrust value and a TPMLM thrust fluctuation value. The method comprises the following steps: establishing sub-Hankel matrixes based on a G-L fractional derivative algorithm, rotationally converting two sub-Hankel matrixes into a Hankel matrix bundle, and solving a determinant value of the Hankel matrix bundle as an optimization coefficient of motor thrust and thrust fluctuation values. Compared with the prior art, the SEM measurement model, the matrix algorithm and the G-L fractional derivative algorithm are adopted to carry out optimization design on the motor, the thrust of the cylindrical permanent magnet linear motor is improved, the thrust fluctuation of the motor is reduced, the performance of the motor is improved, and the cost of a permanent magnet material is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of motor structure parameter design, and specifically relates to a method for optimizing thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm. Background Art

[0002] In the design of permanent magnet linear motor (TPMLM), the stator core is split at both ends and the stator slots are the basic structural features to achieve its function. However, these design choices inevitably lead to two significant effects: end effect and cogging effect. These effects have a profound impact on the performance of the motor, especially in terms of thrust fluctuation.

[0003] Cogging is caused by the interaction between the stator teeth and the permanent magnets, which results in periodic changes in the cogging force, causing thrust instability. Similarly, end effects occur at the two ends of the stator core, and due to the discontinuous distribution of the magnetic field in these areas, additional end forces are generated, further increasing the thrust fluctuations. These thrust fluctuations not only reduce the operating efficiency of the motor, but also limit its applicability in precision control and high-performance applications.

[0004] In order to comprehensively improve the performance of permanent magnet linear motors and expand their application prospects, the key is to reduce thrust fluctuations and increase average thrust through optimized design. How to reduce thrust fluctuations and increase average thrust is a technical problem that urgently needs to be solved in this field. Summary of the invention

[0005] Purpose of the invention: In response to the problems pointed out in the background technology, the present invention discloses a method for optimizing the thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm. The SEM measurement model, matrix algorithm and GL fractional-order derivative algorithm are used to optimize the design of the motor, thereby improving the thrust of the cylindrical permanent magnet linear motor and reducing the thrust fluctuation of the motor.

[0006] Technical solution: The present invention discloses a method for optimizing thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm, comprising the following steps:

[0007] Step 1: Model the TPMLM parameter phasor as a mathematical model consisting of stator tooth height components h0 and h1, stator tooth width component w2, slot height component h2, slot width component w3, air gap width component w4, pole width component w1, and pole height component h3; divide the eight structural parameters into two groups: the first group is the structural parameters h0, h1, and w2 about the stator teeth; the second group is the other five structural parameters h2, h3, w1, w3, and w4; and collect multiple values ​​of the first group and the second group;

[0008] Step 2: Construct the SEM measurement model, use the collected data to quantify the SEM measurement model, and mathematically express the relationship between the latent variables and the manifest variables and latent variables of the SEM measurement model;

[0009] Step 3: Obtain the TPMLM average thrust latent variable F and the TPMLM thrust fluctuation fitting characteristic latent variable K through the SEM measurement model;

[0010] Step 4: Based on the GL fractional derivative algorithm, a sub-Hankel matrix is ​​established for the TPMLM average thrust latent variable and the TPMLM thrust fluctuation fitting characteristic latent variable data set, and the two sub-Hankel matrices are rotated and transformed into a Hankel matrix bundle;

[0011] Step 5: Solve the determinant value of the Hankel matrix bundle as the optimization coefficient of the TPMLM average thrust and TPMLM thrust fluctuation, and determine the optimal parameters by multiplying the optimization coefficient with the TPMLM average thrust latent variable F and the fitting characteristic latent variable K of TPMLM thrust fluctuation obtained in step 3.

[0012] Furthermore, the SEM measurement model is as follows:

[0013] X1=e1σ+δ

[0014] X2=e2γ+ξ

[0015] In the formula: X1 and X2 are the output indicators of the external signal source, e1 is the factor loading of the external signal source output indicator on the exogenous latent variable, σ is the exogenous latent variable, δ is the measurement error of the external signal source output indicator f1; e2 is the factor loading of the external signal source output indicator on the exogenous latent variable, γ is the exogenous latent variable, ξ is the measurement error of the external signal source output indicator f2; the measurement error range of δ is [-2.0, +2.0], and the measurement error range of ξ is [-0.9, +0.9].

[0016] Furthermore, the collected data is used to quantify the data of the SEM measurement model, and the relationship between the latent variables and the manifest variables and latent variables of the measurement model is mathematically expressed as follows:

[0017]

[0018] Among them, F is the TPMLM average thrust latent variable, K is the fitting characteristic latent variable of TPMLM thrust fluctuation, e i , i=1,2,3……n is the factor loading of the output index of the external signal source to the exogenous latent variable;

[0019] SEM measurement model i, i = 1, 2, 3...n connects the external signal source output indicators to quantify the factor loading of the exogenous latent variables:

[0020]

[0021] Furthermore, sub-Hankel matrices are established for the TPMLM average thrust latent variable and the TPMLM thrust fluctuation fitting characteristic latent variable data sets, and the two sub-Hankel matrices are combined into a Hankel matrix bundle, as follows:

[0022] S1: Set f(t) to be continuous in the interval [a,b], generate N equidistant nodes in the interval [0,t], with a step size of L, and the first-order derivative of f(t) is expressed by backward difference as follows:

[0023]

[0024] Among them, f(-1) is not continuous in the interval and is 0; t i =t0+iL,t i is the time value of the i-th equidistant node in the interval [0, t]. Since the nodes are equidistant, the time interval between each node is fixed, that is, the step length L; t0 is the first node in the interval [0, t]; i h From the first node t0 to the i-th node t i The time interval is multiplied by i. Since the step length is L, the total distance from the first node to the i-th node is i times the step length, that is, iL;

[0025] S2: The backward difference derivative expression of the Nth node of the mean thrust of one of the components of the TPMLM parameter phasor is:

[0026]

[0027] Where ▽ is the backward difference of the function f(t) at the node t0, which is used to approximate the derivative of the point; the backward difference is the difference between the function value from the current point t1 to the previous point t0, divided by the step size L; f0 is the value of the function f(t) at the first node t0, f1 is the value of the function f(t) at the second node t1, and f N is the function f(t) at the last node t N The value of

[0028] S3: The backward difference of the Nth node is expressed in vector matrix form as:

[0029]

[0030] Where M is the vector matrix of f′(t), T N is the function value vector matrix obtained at equidistant nodes, A1 N is the first-order backward difference vector matrix;

[0031] S4: The n-th order derivative backward difference vector matrix Q(L) of f(t) obtained by the above transformation is expressed as follows:

[0032]

[0033] in,

[0034] S5: Given the backward difference vector matrix Q(L) of the Nth node by S4, find the backward difference approximation matrix at the Nth node:

[0035]

[0036] Among them, α is the order value of the backward difference vector matrix Q, L -α The αth power of the reciprocal of the step length L represented by ;

[0037] S6: Extract the phasor matrix vector Q from the backward difference approximation matrix of S5 h0 :

[0038]

[0039] S7: Q in S6 h0 The matrix is ​​rotated 90° counterclockwise to obtain the fitting characteristic Hankel matrix J of the average thrust of TPMLM F , converted into matrix product for calculation;

[0040] S8: Repeat the above operation to obtain the fitting characteristic Hankel matrix J of TPMLM thrust fluctuation K .

[0041] Furthermore, the specific process of solving the determinant value of the Hankel matrix bundle is:

[0042] Step 1) Fitting characteristic Hankel matrix J of TPMLM average thrust F , solve for J F The determinant value of the TPMLM average thrust fitting characteristic Hankel matrix J F for:

[0043]

[0044] Step 2) Obtain the fitting characteristic Hankel matrix J of the PMLM average thrust F The determinant of is to multiply the diagonal elements of the lower triangular matrix, where the diagonal elements are:

[0045]

[0046] Get the determinant value:

[0047] Step 3) Since each term contains L -a , extract it and compute the product of the remainder:

[0048]

[0049] Step 4) Repeat the above steps to obtain the fitting characteristic Hankel matrix J of PMLM thrust fluctuation K The determinant value of ;

[0050] Step 5) The fitting characteristic Hankel matrix J of the PMLM average thrust is obtained F Multiply the determinant value of by the theoretical formula calculated value of the average thrust F of TPMLM, and we get |J F |·F=F new ;

[0051] Step 6) Repeat steps 1) to 5) to obtain the fitting characteristic Hankel matrix J of PMLM thrust fluctuation K Multiply the determinant value of by the theoretical formula value of TPMLM thrust fluctuation K, and we get

[0052] |J K |·K=K new .

[0053] Beneficial effects:

[0054] The present invention adopts SEM measurement model, matrix algorithm and GL fractional derivative algorithm to optimize the design of the motor, thereby improving the thrust of the cylindrical permanent magnet linear motor and reducing the thrust fluctuation of the motor.

[0055] 1. Improve motor performance: The present invention uses the SEM measurement model to group the stator tooth height, stator tooth width, slot height, slot width, air gap width, pole width, and pole height parameters to construct two dynamic measurement models, and quantify the data of the SEM measurement model to obtain the TPMLM average thrust value and the TPMLM thrust fluctuation value. Through the precise SEM measurement model and matrix algorithm, the key parameters of the motor can be carefully grouped and quantitatively analyzed, which helps to understand the distribution of the electromagnetic field inside the motor, thereby improving the operating efficiency and thrust of the motor through optimized design.

[0056] 2. Reduce thrust fluctuation: The present invention establishes a sub-Hankel matrix based on the GL fractional-order derivative algorithm, rotates and transforms the two sub-Hankel matrices into a Hankel matrix bundle, and solves the determinant value of the Hankel matrix bundle as the optimization coefficient of the motor thrust and thrust fluctuation value, thereby effectively reducing the thrust fluctuation of the motor during operation and improving the accuracy and stability of its motion control.

[0057] 3. Reduce material costs: The present invention optimizes the design of motor parameter structure data, improves the thrust characteristics of the motor, helps to reduce the amount of permanent magnet material used, and thus reduces manufacturing costs.

[0058] 4. Expand the scope of application: The present invention improves motor performance and reduces manufacturing costs, making this cylindrical permanent magnet linear motor more competitive in various precision control and high-performance application fields. The optimized motor can be widely used in industrial automation, precision machinery, electric transportation and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is the structural diagram of the TPMLM model of the present invention;

[0060] Figure 2 It is the structural parameter diagram of TPMLM of the present invention;

[0061] Figure 3 This is a schematic diagram of the SEM structural equation of the present invention;

[0062] Figure 4 Schematic diagram of the matrix algorithm structure of the present invention. DETAILED DESCRIPTION

[0063] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0064] The present invention discloses a method for optimizing thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm, which specifically comprises the following steps:

[0065] Step 1: Model the TPMLM parameter phasor as a mathematical model consisting of stator tooth height components h0 and h1, stator tooth width component w2, slot height component h2, slot width component w3, air gap width component w4, pole width component w1, and pole height component h3.

[0066] Step 2: Divide the eight structural parameters into two groups. The first group is the structural parameters of the stator teeth, the stator tooth height components h0 and h1, and the stator tooth width component w2. The second group is the other five structural parameters, the slot height component h2, the pole height component h3, the pole width component w1, the slot width component w3, and the air gap width component w4.

[0067] Step 3: Construct SEM measurement model:

[0068] X1=e1σ+δ

[0069] X2=e2γ+ξ

[0070] In the formula: X1 and X2 are the output indicators of the external signal source, e1 is the factor loading of the external signal source output indicator on the exogenous latent variable, σ is the exogenous latent variable, δ is the measurement error of the external signal source output indicator f1; e2 is the factor loading of the external signal source output indicator on the exogenous latent variable, γ is the exogenous latent variable, ξ is the measurement error of the external signal source output indicator f2. The measurement error range of δ is [-2.0, +2.0], and the measurement error range of ξ is [-0.9, +0.9].

[0071] Step 4: Collect and input the first set of values, stator tooth structural parameters stator tooth height components h0 and h1, stator tooth width component w2. Collect and input the second set of values, slot height component h2, pole height component h3, pole width component w1, slot width component w3, air gap width component w4.

[0072] The input data collected in this embodiment are as follows: Tables 1 and 2:

[0073] Table 1 Structural parameters of stator teeth

[0074]

[0075] Table 2 The remaining five structural parameters

[0076]

[0077]

[0078] Step 5: Quantify the data for the SEM measurement model, and mathematically express the relationship between the latent variables, manifest variables, and latent variables of the measurement model as follows:

[0079]

[0080] Among them, F is the TPMLM average thrust latent variable, K is the fitting characteristic latent variable of TPMLM thrust fluctuation, e i(i=1,2,3……n) is the factor loading of the output index of the external signal source to the exogenous latent variable.

[0081] Step 6: Model for SEM Measurements i (i=1,2,3……n) Connect the external signal source output indicators to quantify the factor loading of the exogenous latent variables:

[0082]

[0083] Step 7: The TPMLM average thrust value and TPMLM thrust fluctuation value are obtained through the SEM measurement model, as shown in Tables 3 and 4 below:

[0084] Table 3 Thrust values ​​for group 1

[0085]

[0086] Table 4 Thrust values ​​for Group 2

[0087]

[0088] Step 8: Create a sub-Hankel matrix for the TPMLM average thrust latent variable and the TPMLM thrust fluctuation fitting characteristic latent variable data set, and combine the two sub-Hankel matrices into a Hankel matrix bundle. The specific process is as follows:

[0089] S1: Set f(t) to be continuous in the interval [a,b], generate N equally spaced nodes in the interval [0,t], with a step size of L, where t i =t0+iL(i=0,1,2,...,N), the first-order derivative of f(t) is expressed as follows using backward difference:

[0090]

[0091] Among them, f(-1) is not continuous in the interval and is 0.

[0092] t i : The time value of the i-th equidistant node in the interval [0, t]. Since the nodes are equidistant, the time interval between each node is fixed, that is, the step size L.

[0093] t0: The first node in the interval [0, t], usually the starting point, which is equal to 0 here.

[0094] i h :From the first node (t0) to the i-th node (t i ) multiplied by i. Since the step size is L, the total distance from the first node to the i-th node is i times the step size, i.e. iL.

[0095] S2: The backward difference derivative expression of the Nth node of the mean thrust of one of the components of the TPMLM parameter phasor is:

[0096]

[0097] ▽: Backward difference of function f(t) at node t0, used to approximate the derivative at that point. Backward difference is the difference between the function value at the current point t1 and the previous point t0, divided by the step size L.

[0098] f0: The value of function f(t) at the first node t0.

[0099] f1: The value of function f(t) at the second node t1.

[0100] f N :Function f(t) at the last node t N The value of .

[0101] S3: The backward difference of the Nth node is expressed in vector matrix form as:

[0102]

[0103] Where M is the vector matrix of f′(t), T N is the function value vector matrix obtained at equidistant nodes, A 1 N is the first-order backward difference vector matrix.

[0104] S4: The n-th order derivative backward difference vector matrix Q(L) of f(t) obtained by the above transformation is expressed as follows:

[0105]

[0106] in,

[0107] S5: The backward difference vector matrix of the Nth node is given by step S4. S4 gives the backward difference vector matrix Q(L) of the nth order derivative of f(t). Find the backward difference approximation matrix at the Nth node:

[0108]

[0109] Among them, α is the order value of the backward difference vector matrix Q, L -α Represents the αth power of the reciprocal of the step size L.

[0110] S6: Extract the phasor matrix vector Q from the backward difference approximation matrix of S5 h0 :

[0111]

[0112] S7: Q in S6 h0 The matrix is ​​rotated 90° counterclockwise to obtain the fitting characteristic Hankel matrix J of the average thrust of TPMLM F , converted into matrix product for calculation, making the calculation easier.

[0113] S8: Repeat the above operation to convert the fitting characteristic Hankel matrix J of TPMLM thrust fluctuation into K .

[0114] The specific process of solving the determinant value of the Hankel matrix bundle is:

[0115] 1) Fitting characteristics of PMLM average thrust Hankel matrix J F , solve for J F The determinant value of the TPMLM average thrust fitting characteristic Hankel matrix J F :

[0116]

[0117] 2) Obtain the fitting characteristic Hankel matrix J of the PMLM average thrust F The determinant of is to multiply the diagonal elements of the lower triangular matrix, where the diagonal elements are,

[0118]

[0119] Get the determinant value:

[0120] 3) Since each term contains L -a , extract it and compute the product of the remainder:

[0121]

[0122] 4) Repeat the above steps to obtain the fitting characteristic Hankel matrix J of PMLM thrust fluctuation K The determinant value of .

[0123] 5) The fitting characteristic Hankel matrix J of the PMLM average thrust is obtained F Multiply the determinant value of by the theoretical formula calculated value of the average thrust F of TPMLM, and we get |J F |·F=F new .

[0124] 6) Repeat 1 to 5) to obtain the fitting characteristic Hankel matrix J of PMLM thrust fluctuation K Multiply the determinant value of by the theoretical formula calculated value of TPMLM thrust fluctuation K, and we get |JK |·K=K new .

[0125] F new , K new The two determinant values ​​are multiplied by F and K respectively. F and K are the specific values ​​obtained by calculation in step 7. The two determinant values ​​are optimization coefficients. The specific values ​​of F and K are further optimized.

[0126] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent transformation or modification made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm, characterized in that: The steps include: Step 1: Model the TPMLM parameter phasor as a mathematical model consisting of stator tooth height components h0 and h1, stator tooth width component w2, slot height component h2, slot width component w3, air gap width component w4, pole width component w1, and pole height component h3; divide the eight structural parameters into two groups: the first group is the structural parameters h0, h1, and w2 about the stator teeth; the second group is the other five structural parameters h2, h3, w1, w3, and w4; and collect multiple values ​​of the first group and the second group; Step 2: Construct two dynamic SEM measurement models, use the collected data to quantify the SEM measurement models, and mathematically express the relationship between the latent variables and the manifest variables and latent variables of the SEM measurement models; Step 3: Obtain the TPMLM average thrust latent variable F and the TPMLM thrust fluctuation fitting characteristic latent variable K through the SEM measurement model; Step 4: Based on the GL fractional derivative algorithm, a sub-Hankel matrix is ​​established for the TPMLM average thrust latent variable and the TPMLM thrust fluctuation fitting characteristic latent variable data set, and the two sub-Hankel matrices are rotated and transformed into a Hankel matrix bundle; Step 5: Solve the determinant value of the Hankel matrix bundle as the optimization coefficient of the TPMLM average thrust and TPMLM thrust fluctuation, and determine the optimal parameters by multiplying the optimization coefficient with the TPMLM average thrust latent variable F and the fitting characteristic latent variable K of TPMLM thrust fluctuation obtained in step 3.

2. The method for optimizing thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm according to claim 1, characterized in that: The SEM measurement model is as follows: X1=e1σ+δ X2=e2γ+ξ In the formula: X1 and X2 are the output indicators of the external signal source, e1 is the factor loading of the external signal source output indicator on the exogenous latent variable, σ is the exogenous latent variable, δ is the measurement error of the external signal source output indicator f1; e2 is the factor loading of the external signal source output indicator on the exogenous latent variable, γ is the exogenous latent variable, ξ is the measurement error of the external signal source output indicator f2; the measurement error range of δ is [-2.0, +2.0], and the measurement error range of ξ is [-0.9, +0.9].

3. The method for optimizing thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm according to claim 1, characterized in that: The collected data is used to quantify the data of the SEM measurement model, and the relationship between the latent variables, explicit variables and latent variables of the measurement model is mathematically expressed as follows: Among them, F is the TPMLM average thrust latent variable, K is the fitting characteristic latent variable of TPMLM thrust fluctuation, e i , i=1,2,3……n is the factor loading of the output index of the external signal source to the exogenous latent variable; SEM measurement model i , i = 1, 2, 3...n connects the external signal source output indicators to quantify the factor loading of the exogenous latent variables:

4. The method for optimizing thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm according to claim 3 is characterized in that: A sub-Hankel matrix is ​​established for the TPMLM average thrust latent variable and the TPMLM thrust fluctuation fitting characteristic latent variable data set, and the two sub-Hankel matrices are combined into a Hankel matrix bundle, as follows: S1: Set f(t) to be continuous in the interval [a,b], generate N equidistant nodes in the interval [0,t], with a step size of L, and the first-order derivative of f(t) is expressed by backward difference as follows: Among them, f(-1) is not continuous in the interval and is 0; t i =t0+iL,t i is the time value of the i-th equidistant node in the interval [0, t]. Since the nodes are equidistant, the time interval between each node is fixed, that is, the step length L; t0 is the first node in the interval [0, t]; i h From the first node t0 to the i-th node t i The time interval is multiplied by i. Since the step length is L, the total distance from the first node to the i-th node is i times the step length, that is, iL; S2: The backward difference derivative expression of the Nth node of the mean thrust of one of the components of the TPMLM parameter phasor is: in, is the backward difference of the function f(t) at the node t0, which is used to approximate the derivative of the point; the backward difference is the difference between the function value from the current point t1 to the previous point t0, divided by the step size L; f0 is the value of the function f(t) at the first node t0, f1 is the value of the function f(t) at the second node t1, f N is the function f(t) at the last node t N The value of S3: The backward difference of the Nth node is expressed in vector matrix form as: Where M is the vector matrix of f′(t), T N is the function value vector matrix obtained at equidistant nodes, A 1 N is the first-order backward difference vector matrix; S4: The n-th order derivative backward difference vector matrix Q(L) of f(t) obtained by the above transformation is expressed as follows: in, S5: Given the backward difference vector matrix Q(L) of the Nth node by S4, find the backward difference approximation matrix at the Nth node: Among them, α is the order value of the backward difference vector matrix Q, L -α The αth power of the reciprocal of the step length L represented by ; S6: Extract the phasor matrix vector Q from the backward difference approximation matrix of S5 h0 : S7: Q in S6 h0 The matrix is ​​rotated 90° counterclockwise to obtain the fitting characteristic Hankel matrix J of the average thrust of TPMLM F , converted into matrix product for calculation; S8: Repeat the above operation to obtain the fitting characteristic Hankel matrix J of TPMLM thrust fluctuation K .

5. The method for optimizing thrust characteristics of a cylindrical permanent magnet linear motor based on a matrix algorithm according to claim 4 is characterized in that: The specific process of solving the determinant value of the Hankel matrix bundle is: Step 1) Fitting characteristic Hankel matrix J of TPMLM average thrust F , solve for J F The determinant value of the TPMLM average thrust fitting characteristic Hankel matrix J F for: Step 2) Obtain the fitting characteristic Hankel matrix J of the PMLM average thrust F The determinant of is to multiply the diagonal elements of the lower triangular matrix, where the diagonal elements are: First line: Second line: Line N: Get the determinant value: Step 3) Since each term contains L -a , extract it and compute the product of the remainder: Step 4) Repeat the above steps to obtain the fitting characteristic Hankel matrix J of PMLM thrust fluctuation K The determinant value of ; Step 5) The fitting characteristic Hankel matrix J of the PMLM average thrust is obtained F Multiply the determinant value of by the theoretical formula calculated value of the average thrust F of TPMLM, and we get |J F |·F=F new ; Step 6) Repeat steps 1) to 5) to obtain the fitting characteristic Hankel matrix J of PMLM thrust fluctuation K Multiply the determinant value of by the theoretical formula value of TPMLM thrust fluctuation K, and we get |J K |·K=K new 。

Images