Power control method for wave power generation based on magnetic screw composite generator

Abstract

The invention relates to a power control method for wave power generation based on a magnetic screw composite generator. First, an improved M-type unscented Kalman filter is used to estimate the mechanical angular velocity of the magnetic screw composite generator rotor, and its optimal speed is obtained by the instantaneous active power of the generator terminal; then the rotor speed is adjusted to an optimal value through a PI controller, and finally the maximum power is obtained; this control method also integrates low voltage ride-through control, and meets the requirements of the grid specification by injecting reactive current during grid disturbance; the method solves the problems of friction loss and self-locking failure during the operation of the mechanical screw, the problem of unstable linearization process when the traditional filter is used, and the problem of huge calculation amount of the general maximum power control algorithm. The method has the advantages of being able to improve the energy conversion efficiency, adjust the rotor speed of the magnetic screw composite motor and generate optimal power, and reduce the output power fluctuation, thereby reducing the size of the wave energy converter.

Description

Technical Field

[0001] The invention relates to the technical field of wave power generation, and in particular to a power control method for wave power generation based on a magnetic screw composite generator. Background Art

[0002] Since the Industrial Revolution, the earth's ecological environment has been severely damaged by the greenhouse gases produced by the consumption of fossil energy such as oil, coal, and natural gas. At the same time, the non-renewable nature of fossil energy has caused countries around the world to face energy shortages, which seriously threatens the sustainable development of human society. Improving the energy consumption structure of society is imminent. Countries around the world are paying more and more attention to clean, renewable new energy. In the past 50 years, new energy such as solar energy and wind energy have been strongly supported by governments of various countries and have achieved large-scale commercialization. However, the development of ocean energy has only emerged in recent years, and ocean energy has broad prospects for development and application.

[0003] The ocean area accounts for about three quarters of the total area of ​​the earth, and it stores a wealth of new renewable energy. The energy contained in the ocean appears in many forms, such as tidal energy, wave energy, salinity difference energy, temperature difference energy, etc., which are collectively referred to as ocean energy. Among them, the reserves of salinity difference energy and temperature difference energy are the most abundant; the largest reserve of exploitable ocean energy is wave energy, about 300GW, which is mainly distributed in the 40-60° latitude area on the east coast of the Pacific and Atlantic Oceans.

[0004] While my country's economy is developing rapidly, it is also facing the challenge of adjusting its energy structure. As an important component of clean and renewable ocean energy, the rational development of ocean wave energy has emerged with the needs of society. However, in the existing wave power generation, there is no precise optimization control in the process of power conversion control, which makes the conversion of wave energy not ideal. Therefore, it is necessary to optimize the power converted from wave energy to maximize the power obtained. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a power control method for wave power generation based on a magnetic screw composite generator, so as to optimize and control the electric energy converted from wave energy to maximize the obtained electric energy.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A power control method for wave power generation based on a magnetic screw composite generator comprises the following steps:

[0008] S1: Design an improved wave energy conversion power generation system, including using a magnetic screw composite generator to couple with the existing Archimedean wave swing model: AWS, so that when AWS moves with the waves, the magnetic screw composite generator converts the low-speed linear motion of AWS into high-speed rotational motion, and provides the electrical energy obtained in the high-speed motion to the power grid through a back-to-back converter;

[0009] S2: In the process of providing electric energy to the grid through the back-to-back converter, a power control algorithm is used to control and adjust the operation of the magnetic screw composite motor, thereby optimizing the grid to obtain electric energy at the maximum power point.

[0010] To optimize the above technical solutions, the specific measures taken also include:

[0011] Furthermore, the power control algorithm described in step S2 includes a generator-side converter MPPT control algorithm, and the specific contents of controlling and adjusting the operation of the magnetic screw composite motor by the algorithm are as follows:

[0012] S2.1: Get the terminal three-phase voltage v of the magnetic screw compound generator gen and three-phase current i gen ;

[0013] S2.2: Use the instantaneous active power calculator to calculate the instantaneous active power P generated by the magnetic screw compound generator gen , the specific calculation formula is:

[0014] P gen =v a i a +v b i b +v c i c

[0015] In the formula, v a 、v b 、v c is the terminal three-phase voltage v of the magnetic screw compound generator gen The parameter content in i a 、i b 、i c Terminal three-phase voltage i of magnetic screw compound generator gen Parameter content in;

[0016] S2.3: Parameter P calculated according to step S2.2 gen , determine the optimal linear stator speed v of the magnetic screw hybrid generator in maximum power point tracking opt , v opt The specific calculation formula is:

[0017]

[0018] Where P g is the parameter P gen Abbreviation, β h is the hydrodynamic damping resistance effect coefficient of the AWS model;

[0019] S2.4: Adjust the linear stator speed of the magnetic screw composite generator to the optimal linear stator speed v through closed-loop PI control opt , while processing and estimating the error between the predicted speed and the optimal rotor speed through a PI controller to generate the converter VSC current on the magnetic screw compound generator side The orthogonal components of

[0020] S2.5: Set the d-axis command value of the stator current to zero, that is, To reduce the power loss in the stator winding of the magnetic screw composite generator; at the same time, the rotor position y is estimated est , the estimation method is to obtain it by integrating the rotor mechanical speed;

[0021] S2.6: Based on the parameters in step S2.4 and the parameters in step S2.5 Parameter y est Determine the control current By controlling the current Realize the optimization of power grid to obtain the maximum power point.

[0022] Furthermore, the hydrodynamic damping resistance effect coefficient β of the AWS model described in step S2.3 is h The specific calculation formula is:

[0023]

[0024] In the formula, ω mest represents the mechanical angular velocity of the rotor of the magnetic screw compound generator, which is estimated by an algorithm included in the power control algorithm: an improved M-type unscented Kalman filter rotor mechanical angular velocity prediction algorithm, λ represents the pole width of the magnetic screw compound generator, L d and L q They represent the dq inductance of the stator winding in the dq coordinate system, i d and i q They represent the dq components of the stator current in the dq coordinate system, ψ pm Represents the flux linkage of a permanent magnet.

[0025] Furthermore, the parameter ω mestThe specific process of estimating the rotor mechanical angular velocity prediction algorithm by improving the M-type unscented Kalman filter is as follows:

[0026] (1) Construct a Sigma point set, that is, construct a point set of data sample points, by selecting 2n+1 sample points from the rotor angular velocity data of historical information and using the mean and covariance to represent these 2n+1 sample points, specifically:

[0027] Assume that the state variable x has a mean of x, a covariance of P, and a dimension of n. The set formula for the 2n+1 points representing the true mean and covariance of each state variable x is as follows:

[0028]

[0029] In the formula, χ i represents any point among 2n+1 sample points, λ is the scale factor, P x,i is the covariance matrix element at the xth original value and the ith column corresponding to the state variable x; 2n+1 sample points of different states: Sigma points are combined in the matrix χ=[χ0χ1…χ 2n ], construct the Sigma point set;

[0030] (2) Time domain update, that is, update the point states of all Sigma points in the Sigma point set and preliminarily calculate the predicted state vector of each Sigma point and the error covariance matrix prediction Specifically:

[0031] Update the point status of all Sigma using the following formula:

[0032]

[0033]

[0034] In the formula, χ k is the updated Sigma point, f(χ k-1 ,u k-1 ) is the nonlinear state dynamic function f(χ k ,u k ) deformation, f(χ k ,u k ) is determined by the following formula: in is the vector derivative of the state variable, and x(t) is x=[i d i q ω mest ] T , v k is a zero-mean Gaussian noise vector;

[0035] Where Y k To solve the output point, H(χ k ) is to bring the updated Sigma point into the output function, the output function H(χ k ) is determined by the following formula: Where y(t) is the output vector, which is y=[i d i q ] T , μ k It is also a zero-mean Gaussian noise vector;

[0036] Preliminary calculation of the predicted state vector for each sample point and the corresponding observation vector

[0037]

[0038] In the formula, and They are and The i-th row of It means that for the average value, each sample point χ i The corresponding weight is

[0039] Preliminary calculation of the error covariance matrix prediction for each sample point :

[0040]

[0041] In the formula, diag is used to construct a diagonal matrix, that is, a square matrix whose elements are not on the diagonal are all zero. The operator diag(w (c) ) generates a zero square matrix where diag(w (c) ) is located on the main diagonal, which is Indicates that for covariance, each sample point χ i The corresponding weight is ; Represented as Sigma point matrix χ k Subtract the predicted state vector from each column of Q is the zero-mean Gaussian noise vector v k The covariance matrix of

[0042] (3) Measurement update, that is, the predicted state vector of each Sigma point calculated initially in (2) and the error covariance matrix prediction Perform calibration, specifically:

[0043] Calculate the innovation covariance matrix P yy and the cross-covariance matrix P xy :

[0044] In the formula, is the Sigma point matrix Subtract the predicted state vector from each column of Output point matrix Subtract the observation vector from each column of R is the zero-mean Gaussian noise vector μ k The covariance matrix of

[0045] Finally, the M-type unscented Kalman filter gain is updated and the predicted state vector is corrected according to the following formula: get and the corrected error covariance Get P k :

[0046]

[0047] In the formula, K k is the Kalman gain, is the corrected predicted state vector, P k is the corrected error covariance, Subtract the observation vector from the observation vector The i-th element of ;

[0048] Since the prediction vector and error covariance can represent the characteristics of a sample point, the calculated parameters and parameter P k Select a specific sample point with the same characteristics from the Sigma point set, and estimate the mechanical angular velocity ω of the rotor of the magnetic screw composite generator based on the angular velocity of this sample point in the historical information. mest .

[0049] Furthermore, for the averaging, each sample point χ i The corresponding weight is The specific calculation formula is:

[0050]

[0051] For the covariance, each sample point χ i The corresponding weight is w i (c) ; The specific calculation formula is:

[0052]

[0053] In the formula, λ is the scaling factor, and factors α and β are the distribution and mean of the surrounding Sigma points, respectively. Prior parameters of the distribution.

[0054] The beneficial effects of the present invention are:

[0055] 1. Magnetic screws can convert low-speed linear motion into high-speed rotary motion, providing high gear effect and low torque requirements. Compared with mechanical screws, magnetic screws avoid friction loss and self-locking failure; compared with ordinary direct drives using permanent magnet linear generators, the force density of the system can be increased several times, and torque density can be transmitted at a greater speed.

[0056] 2. The present application provides a power control method based on a magnetic screw-coupled generator, which solves the problems of friction loss and self-locking failure during the operation of the mechanical screw, the unstable linearization process when using traditional filters, and the huge amount of calculation of the general maximum power control algorithm; by controlling the power of the VSC on the generator side, its advantage lies in the ability to improve the energy conversion efficiency, adjust the rotor speed of the magnetic screw composite motor and generate optimal power, while reducing the output power fluctuation, thereby reducing the size of the wave energy converter. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of the framework of the wave energy conversion system of the present invention.

[0058] Figure 2 It is a schematic diagram of the MPPT control system framework of the generator-side VSC of the present invention.

[0059] Figure 3 It is a schematic diagram of the control framework of the LVRT controller of the present invention. DETAILED DESCRIPTION

[0060] The present application scheme is now described in detail with reference to the accompanying drawings.

[0061] like Figure 1 The present invention proposes a direct-drive wave energy conversion system based on a magnetic screw composite generator, including a hydrodynamic model of a wave energy conversion device, a magnetic screw composite generator model, a maximum power tracking strategy, and a magnetic screw composite generator rotor speed prediction algorithm.

[0062] The hydrodynamic model of the wave energy conversion device is the Archimedean wave swing (AWS) model;

[0063] The AWS is the first wave energy conversion device to feature a direct drive power output. The AWS consists of two parts: a hollow cylinder and its cap, called a float. The hollow cylinder is fixed to the seabed and filled with air, while the float is vibrated vertically by the waves. When the waves are higher than the AWS, the float moves downward due to the weight of the water above the AWS, which pressurizes the air inside the device. In contrast, when the trough of the wave is higher than the AWS, the weight of the water is reduced. Therefore, the float moves upward in proportion to the amount of air expanded in the cylinder. A generator is directly connected to the AWS, converting the linear motion of the float into electrical energy. As the wave excitation force changes, the generator output also changes in amplitude and frequency, so a back-to-back converter is used to transfer the oscillating power from the generator to the grid.

[0064] The hydrodynamic equation of motion for AWS is:

[0065] F w =F e +m tot dv / dt+β h v+k s y (1)

[0066] Where v is the linear velocity of the rotor, v = dy / dt, y is the distance moved by the float and the rotor, m tot is the total mass of the moving parts of the AWS and the additional mass of the water above the AWS that must be accelerated, β h is the hydrodynamic damping coefficient of AWS, k s is the spring constant of AWS, F w is the exciting force of the wave on the float, F e It is the electromagnetic force of the magnetic screw compound generator.

[0067] However, if AWS is directly coupled to a direct-drive wave generator, its volume, weight and cost will be greatly increased because a large force is required to directly drive the generator. To avoid the above problems, a magnetic screw composite generator model is introduced.

[0068] The magnetic screw composite generator model is to directly couple the magnetic screw (MLS) composite motor with AWS; similar to the mechanical screw, MLS can well convert low-speed linear motion into high-speed rotational motion, provide high gear effect and low torque demand, and is suitable for medium-speed generators. Compared with the mechanical screw, MLS avoids friction loss and self-locking failure because there is no physical contact between the parts transmitting force. Compared with ordinary direct-drive wave energy converters (WECs) using permanent magnet linear generators, the force density of WEC systems using MLS can be increased several times, and torque density can be transmitted at a greater speed.

[0069] The magnetic screw compound generator is directly coupled to the AWS. When the buoy moves with the waves, the magnetic screw compound generator converts the low-speed linear motion of the AWS into high-speed rotation. The converted electrical energy is provided to the grid through a back-to-back converter.

[0070] The magnetic field period of the magnetic screw composite generator can be expressed by the lead (expressed as τ), which is equal to twice the magnetic pole spacing of the magnetic screw composite generator measured along the axial direction, that is, the magnetic field period of the magnetic screw. Due to magnetic coupling, the movement of the mover and rotor of the magnetic screw composite motor is synchronized. When the mover moves one magnetic field period in the axial direction, the rotor will rotate 2π to achieve the same magnetic field change. According to the working principle of the magnetic screw, ω m (rad / s) is the mechanical angular velocity of the rotor, and the linear velocity of the stator, v (m / s), can be expressed by equation (2). According to the law of conservation of energy, the electromagnetic force F e and rotor torque T mls It can be obtained in formula (3).

[0071]

[0072]

[0073] The rotor of the magnetic screw compound generator reciprocates, so the magnitude and direction of the rotor speed change all the time. Therefore, the induced voltage and current of the magnetic screw compound generator change in amplitude and frequency. The vector model of the generator fixed on the rotor in the dq coordinate system is as follows:

[0074] v d =-Ri d +L d d d / dt-ω e L q i q (4)

[0075] v q =-Ri q +L q d q / dt+ω e L d i d +ω e ψ pm (5)

[0076]

[0077]

[0078]

[0079] In the formula, vd and v q are the dq components of the stator voltage respectively; i d and i q are the dq components of the stator current respectively; R is the stator resistance, L d and L q are the dq inductance of the stator winding; pm is the flux linkage of the permanent magnet, F e is the electromagnetic force of the magnetic screw compound generator, λ is the pole width of the magnetic screw compound generator, ω e is the stator electromagnetic angular velocity, ω e =2πv / λ, T e is the torque of the magnetic screw compound generator, T mls is the rotor torque, J is the moment of inertia, ω m is the rotor mechanical angular velocity.

[0080] The Park transformation given by (9) is used to transform from the abc reference coordinate system to the dq rotor reference coordinate system, that is, to project the stator's a, b, c three-phase currents onto the direct axis (d axis), the quadrature axis (q axis) and the zero axis perpendicular to the dq plane as the rotor rotates, thereby diagonalizing the stator inductance matrix.

[0081]

[0082] Where θ is the phase angle, θ = ∫ω e dt-π / 2, because the speed is relatively ahead of the position of AWS by π / 2, θ is obtained by the phase-locked loop (PLL) circuit.

[0083] Furthermore, as the wave excitation force changes, the output of the generator will also change in amplitude and frequency.

[0084] In order to obtain the maximum power from the waveform, a back-to-back converter is used to transfer the oscillating power from the generator to the grid. The magnetic screw compound generator is integrated into the grid through a back-to-back converter, which consists of two current-controlled voltage source converters (VSCs), such as Figure 1As shown. The back-to-back converter is connected to the grid at the PCC through a step-up transformer. The main function of the back-to-back converter is to transfer the power extracted by the direct drive wave energy converter based on the AWS-based screw composite generator to the grid. The generator-side VSC is controlled to extract maximum power from the AWS-based screw composite generator. The MPPT control algorithm is controlled by the speed sensorless control of the screw composite generator to track the waveform. The optimal rotor speed is obtained by the measured voltage and current at the generator terminals. The main goal of the grid-side VSC is to deliver power from the generator-side VSC to the grid through the DC link. This function can be achieved by simply regulating the DC link voltage to a constant value. The PI controller is used to process the measured DC link voltage, V dc and the required voltage level The error between Figure 2 The output of the PI controller is the d-axis current command value required for DC link voltage regulation. To prevent exceeding the rated power of the grid-side VSC, The limiter is designed as follows:

[0085]

[0086] Where P g | rated is the rated power of the magnetic screw compound motor, V PCC is the grid voltage.

[0087] During normal operation, the quadrature current command is set to zero to achieve zero reactive power injection. This is not the case during a grid fault, as the LVRT requires the generator set to provide reactive power to support grid voltage recovery according to the grid code. Finally, the inverse Park transformation is used to construct the reference value current of the VSC three-phase voltage. The phase angle θ required for the Park transformation is obtained by the phase-locked loop (PLL) circuit.

[0088] During a grid fault, the grid-side VSC cannot transfer the active power of the generator to the grid due to insufficient PCC voltage. As a result, the DC link voltage increases and may exceed the rated voltage of the back-to-back VSC. Figure 1 As shown, the brake chopper is connected to the external resistor R b The back-to-back VSCs are protected from overvoltage by discharging the DC link capacitors. Once the DC link voltage reaches a predetermined limit, the chopper will be activated to consume active power.

[0089] The generator side maximum power control algorithm includes a generator side current conversion MPPT control algorithm and an improved M-type unscented Kalman filter speed prediction algorithm.

[0090] Furthermore, the maximum power is captured from the wave when the AWS is in a state of mechanical resonance. To operate in a resonant state, the following two conditions must be met:

[0091] m tot dv / dt+K s y=0 (11)

[0092] F e =β h v (12)

[0093] Condition (11) represents the resonant angular frequency of the float, Equal to the angular frequency ω of the wave wave ,

[0094]

[0095] To maintain resonance at different wave periods (frequencies), the spring constant K s This can be changed by adjusting the air pressure and volume within AWS to meet the requirements of (13).

[0096] Furthermore, (12) implies that the generator should be controlled in such a way that its electromagnetic force is equal to the hydrodynamic damping force.

[0097] Figure 2 The flow chart of the generator side inverter MPPT control algorithm is given. The instantaneous active power P generated gen The calculation is as follows:

[0098] P gen =v a i a +v b i b +v c i c (14)

[0099] In the formula, v gen (v a ,v b ,v c ) and i gen (i a ,i b ,i c ) are the terminal phase voltage and current of the magnetic screw compound generator respectively. Due to the switching action of the VSC, a first-order low-pass filter (LPF) is used to filter out P gen High frequency ripples in the

[0100] Ignore the internal loss of the magnetic screw compound generator and replace P in (12) g =F e Replace v with i gen (ia ,i b ,i c )

[0101]

[0102] where v opt is the optimal linear stator speed for maximum power point tracking, β h is the AWS hydrodynamic damping coefficient, P g is the active power of the magnetic screw compound generator. In order to make the linear stator speed of the magnetic screw compound generator reach the optimal value v opt , the actual value is required to participate in the operation, and the actual value is obtained by the improved M-type unscented Kalman filter rotor mechanical angular velocity prediction algorithm described below. The rotor position abc / dq required by the Park transform is estimated by the mechanical angular velocity ω of the rotor of the magnetic screw compound generator m Discrete integration implementation of .

[0103]

[0104] Substituting equation (12) into equation (11), the optimal linear stator speed of the magnetic screw composite generator can be obtained from the actual value obtained by the prediction algorithm.

[0105] The linear stator speed of the magnetic screw composite generator is adjusted to the optimal linear stator speed v by closed-loop PI control. opt , the PI controller processes and estimates the error between the predicted speed and the optimal rotor speed to generate the generator side VSC current orthogonal components of .

[0106] The d-axis command value of the stator current is set to zero, To reduce the power loss in the stator winding of the magnetic screw hybrid generator and maximize the power to current operation. Finally, the reference value of the three-phase current of the VSI on the generator side is constructed using the inverse Park transform. Using the estimated rotor position y est The use of a phase-locked loop (PLL) circuit is avoided. A limiter is designed for active power.

[0107] Further, in order to obtain the optimal rotor speed for maximum power point tracking, the improved M-type unscented Kalman filter magnetic screw composite generator rotor speed prediction algorithm is used to estimate the rotor speed based on the measured magnetic screw composite generator voltage and current, rather than measuring the rotor speed. The improved M-type unscented Kalman filter gives the least squares optimal estimate in the case of a nonlinear system model (for the state we want to estimate, the mean value is x at time k). k , the variance is Pk , such a Gaussian distribution, this is our posterior at time k. Now we use this posterior as the starting point and combine it with certain prior knowledge to estimate the mean and variance of our state at time k+1. This process is the principle of Kalman filter prediction. The unscented Kalman filter prediction is to find a Gaussian distribution that is close to the true distribution, that is, to find a Gaussian distribution with the same mean and covariance as the true distribution).

[0108] The improved M-type unscented Kalman filter prediction algorithm is based on the implementation of the unscented transformation (UT) of the nonlinear state distribution. A set of deterministic sample points are selected around the state and propagated through the nonlinear model of the magnetic screw compound generator.

[0109] To obtain the estimated speed of the rotor, (4) and (5) are written in the following third-order vector form as follows:

[0110]

[0111] y(t)=Hx(t)=H(x k )+μ k (18)

[0112] Where x=[i d i q ω mest ] T The vector of state variables, u = [v d v q ] T is the input vector, y=[i d i q ] T is the output vector, the superscript T is the matrix transpose operation, v k and μ k is a zero-mean Gaussian noise vector, G is a vector describing the noise of each state, f(x k ,u k ) is the nonlinear state equation at t = k, H is the measurement matrix, H(x k ) is the output state equation when t = k. At the instant t = k, its covariance matrix is ​​Q and R respectively. The system state matrix is ​​expressed as follows:

[0113]

[0114] Where, R is the stator resistance, L d and L q are the dq inductance of the stator winding, ω mest is the estimated mechanical speed of the magnetic screw compound generator rotor, ψ pmis the flux linkage of the permanent magnet, and λ is the pole width of the magnetic screw compound generator. In formula (14), if it is assumed that the rotor speed of the magnetic screw compound generator remains unchanged during the sampling time, its state variable ω mest The derivative of is zero.

[0115] The algorithm implementation steps are:

[0116] (1) Sigma point selection:

[0117] Assume that the mean and covariance P of dimension n state variables x, representing the 2n+1 points χ of the true mean and covariance of each state x i The set formula is as follows:

[0118]

[0119] Where λ is the scaling factor, P x,i is the covariance matrix element at the xth original value and the i-th column corresponding to the state variable x. The Sigma point combinations of different states are in the matrix χ=[χ0 χ1 … χ 2n ] In the case of averaging, each sample point χ i The corresponding weight is For covariance, each sample point χ i The corresponding weight w i (c) As shown below:

[0120]

[0121]

[0122] Factors α and β represent the distribution of surrounding Sigma points and Prior knowledge of the distribution.

[0123] (2) Time domain update:

[0124] Using the traceless transformation and the magnetic screw compound generator model of equations (13) and (14), the states of all Sigma points are updated as follows:

[0125]

[0126]

[0127] in, is the updated Sigma point, f(χ k-1 ,u k-1 ) is to convert χ k-1 and u k-1 Substitute the nonlinear state equation into To solve the output point, To bring the updated Sigma point into the output equation.

[0128] The weighted average of any state variable and observation is given by,

[0129]

[0130]

[0131] Get. and They are and In addition, and Corresponding to the predicted state vector and the observation vector The error covariance matrix is ​​predicted as follows:

[0132] The error covariance matrix is ​​predicted as follows:

[0133]

[0134] in, is the error covariance matrix, diag is used to construct a diagonal matrix (a square matrix whose elements not on the diagonal are all zero), and the operator diag(w (c) ) generates a zero square matrix where diag(w (c) ) are located on the main diagonal, is the Sigma point matrix Subtract the predicted state vector from each column of

[0135] (3) Measurement update:

[0136] The formula for calculating the covariance matrix is ​​as follows:

[0137]

[0138]

[0139] P yy is the innovation covariance matrix, P xy is the cross-covariance matrix, is the Sigma point matrix Subtract the predicted state vector from each column of Output point matrix Subtract the observation vector from each column of

[0140] Finally, the M-type unscented Kalman filter gains are updated and the best estimate and error covariance are corrected as follows:

[0141]

[0142]

[0143]

[0144] Among them, K k is the Kalman gain, is the predicted state vector, Subtract the observation vector from the observation vector The ith element of .

[0145] The above prediction algorithm can be used to obtain the mechanical angular velocity ω of the rotor of the magnetic screw composite generator. mest .

[0146] Furthermore, since grid codes for renewable energy sources require them to remain connected during grid disturbances to ride through faults.

[0147] The LVRT controller described herein, LVRT control as follows Figure 3 Once a fault is detected, the two VSCs The selector switch will switch from position A for normal operation to position B for LVRT control. During a fault, the generator side VSC is set to zero, reducing the active power of the magneto-screw compound generator to zero. This action prevents DC link overvoltage during the fault. However, at the moment of the fault, DC link overvoltage may occur until the generator-side VSC is demagnetized. Therefore, a braking chopper is still required to protect the back-to-back VSCs, but the rated energy is reduced. In addition, contrary to normal operation, the grid-side VSC is controlled to inject reactive current i into the grid according to the grid connection criteria. reactive For PCC voltage VPCC below 0.5pu, many grid codes require the grid-side VSC to support a grid with a grid reactive current equal to the rated current. While meeting the LVRT capability, attention must be paid to other current components of the grid-side VSC, namely the i d To prevent overcurrent tripping. Therefore, the dynamic limit The reactive current taken from the grid code sets a limit to prevent the grid-side VSC current from exceeding its rated value. This limit is customized as follows:

[0148]

[0149] Given that the grid-side VSC current closely tracks its reference signal, assuming

[0150] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A power control method for wave power generation based on a magnetic screw composite generator, characterized in that: The following steps are involved: S1: Design an improved wave energy conversion power generation system, including using a magnetic screw composite generator to couple with the existing Archimedean wave swing model: AWS, so that when AWS moves with the waves, the magnetic screw composite generator converts the low-speed linear motion of AWS into high-speed rotational motion, and provides the electrical energy obtained in the high-speed motion to the power grid through a back-to-back converter; S2: In the process of providing electric energy to the grid through the back-to-back converter, a power control algorithm is used to control and adjust the operation of the magnetic screw composite motor, thereby optimizing the grid to obtain electric energy at the maximum power point; The power control algorithm described in step S2 includes a generator-side converter MPPT control algorithm, and the specific contents of controlling and adjusting the operation of the magnetic screw composite motor through the algorithm are as follows: S2.1: Get the terminal three-phase voltage v of the magnetic screw compound generator gen and three-phase current i gen ; S2.2: Use the instantaneous active power calculator to calculate the instantaneous active power P generated by the magnetic screw compound generator gen , the specific calculation formula is: P gen =v a i a +v b i b +v c i c In the formula, v a 、v b 、v c is the terminal three-phase voltage v of the magnetic screw compound generator gen The parameter content in i a 、i b 、i c Terminal three-phase voltage i of magnetic screw compound generator gen Parameter content in; S2.3: Parameter P calculated according to step S2.2 gen , determine the optimal linear stator speed v of the magnetic screw hybrid generator in maximum power point tracking opt , v opt The specific calculation formula is: Where P g is the parameter P gen Abbreviation, β h is the hydrodynamic damping resistance effect coefficient of the AWS model; S2.4: Adjust the linear stator speed of the magnetic screw composite generator to the optimal linear stator speed v through closed-loop PI control opt , while processing and estimating the error between the predicted speed and the optimal rotor speed through a PI controller to generate the converter VSC current on the magnetic screw compound generator side The orthogonal components of S2.5: Set the d-axis command value of the stator current to zero, that is, To reduce the power loss in the stator winding of the magnetic screw composite generator; at the same time, the rotor position y is estimated est , the estimation method is to obtain it by integrating the rotor mechanical speed; S2.6: Based on the parameters in step S2.4 and the parameters in step S2.5 Parameter y est Determine the control current By controlling the current Realize the optimization of power grid to obtain the maximum power point.

2. The power control method for wave power generation based on a magnetic screw composite generator according to claim 1 is characterized in that: The hydrodynamic damping resistance effect coefficient β of the AWS model described in step S2.3 h The specific calculation formula is: In the formula, ω mest represents the mechanical angular velocity of the rotor of the magnetic screw compound generator, which is estimated by an algorithm included in the power control algorithm: an improved M-type unscented Kalman filter rotor mechanical angular velocity prediction algorithm, λ represents the pole width of the magnetic screw compound generator, L d and L q They represent the dq inductance of the stator winding in the dq coordinate system, i d and i q They represent the dq components of the stator current in the dq coordinate system, ψ pm Represents the flux linkage of a permanent magnet.

3. The power control method for wave power generation based on a magnetic screw composite generator according to claim 2 is characterized in that: The parameter ω mest The specific process of estimating the rotor mechanical angular velocity prediction algorithm by improving the M-type unscented Kalman filter is as follows: (1) Construct a Sigma point set, that is, construct a point set of data sample points, by selecting 2n+1 sample points from the rotor angular velocity data of historical information and using the mean and covariance to represent these 2n+1 sample points, specifically: Set the mean to be The set formula for the 2n+1 points representing the true mean and covariance of each state variable x with covariance P and dimension n is as follows: In the formula, χ i represents any point among 2n+1 sample points, λ is the scale factor, P x,i is the covariance matrix element at the xth original value and the ith column corresponding to the state variable x; 2n+1 sample points of different states: Sigma points are combined in the matrix χ=[χ0χ1…χ 2n ], construct the Sigma point set; (2) Time domain update, that is, update the point states of all Sigma points in the Sigma point set and preliminarily calculate the predicted state vector of each Sigma point and the error covariance matrix prediction Specifically: Update the point status of all Sigma using the following formula: In the formula, is the updated Sigma point, f(χ k-1 ,u k-1 ) is the nonlinear state dynamic function f(χ k ,u k ) deformation, f(χ k ,u k ) is determined by the following formula: in is the vector derivative of the state variable x, v k is a zero-mean Gaussian noise vector; In the formula, To solve the output point, To bring the updated Sigma point into the output function, the output function Determined by the following formula: Where y(t) is the output vector, which is y=[i d i q ] T , μ k It is also a zero-mean Gaussian noise vector; Preliminary calculation of the predicted state vector for each sample point and the corresponding observation vector In the formula, and They are and The i-th row of It means that for the average value, each sample point χ i The corresponding weight is Preliminary calculation of the error covariance matrix prediction for each sample point In the formula, diag is used to construct a diagonal matrix, that is, a square matrix whose elements are not on the diagonal are all zero. The operator diag(w (c) ) generates a zero square matrix where diag(w (c) ) is located on the main diagonal, which is Indicates that for covariance, each sample point χ i The corresponding weight is Represented as Sigma point matrix Subtract the predicted state vector from each column of Q is the zero-mean Gaussian noise vector v k The covariance matrix of (3) Measurement update, that is, the predicted state vector of each Sigma point calculated initially in (2) and the error covariance matrix prediction Perform calibration, specifically: Calculate the innovation covariance matrix P yy and the cross-covariance matrix P xy : In the formula, is the Sigma point matrix Subtract the predicted state vector from each column of Output point matrix Subtract the observation vector from each column of R is the zero-mean Gaussian noise vector μ k The covariance matrix of Finally, the M-type unscented Kalman filter gain is updated and the predicted state vector is corrected according to the following formula: get and the corrected error covariance Get P k : In the formula, K k is the Kalman gain, is the corrected predicted state vector, P k is the corrected error covariance, Subtract the observation vector from the observation vector The i-th element of ; Since the prediction vector and error covariance can represent the characteristics of a sample point, the calculated parameters and parameter P k Select a specific sample point with the same characteristics from the Sigma point set, and estimate the mechanical angular velocity ω of the rotor of the magnetic screw composite generator based on the angular velocity of this sample point in the historical information. mest .

4. The power control method for wave power generation based on a magnetic screw composite generator according to claim 3 is characterized in that: For the averaging, each sample point χ i The corresponding weight is The specific calculation formula is: For the covariance, each sample point χ i The corresponding weight is The specific calculation formula is: In the formula, λ is the scaling factor, and factors α and β are the distribution and mean of the surrounding Sigma points, respectively. Prior parameters of the distribution.

Images