Direct drive wave power system position sensorless mover motion state estimation method and system

Abstract

The application discloses a kind of direct-drive wave power generation system position sensorless mover motion state estimation method and system, first, based on the equation of motion of floater under irregular wave excitation force, permanent magnet synchronous linear generator model and power tracking control circuit, establish direct-drive wave power generation system mathematical model;Again based on the coupling relationship between the hydrodynamic, mechanical and electrical response in the system, select the key variables that can reflect the coupling characteristics as state variables, construct wave-electric whole process state equation;On this basis, the mover motion estimation method based on spherical simplex unscented Kalman filtering algorithm is proposed, the power optimization control without position sensor is realized;Finally, the simulation model of direct-drive wave power generation system is constructed in Simulink, and the method is verified offline and online, and the results show that the method can realize power optimization control without position sensor, and improve the calculation efficiency of the estimation method and the control accuracy of the system.

Description

Technical Field

[0001] This invention belongs to the technical field of wave power generation system modeling and state estimation, and relates to a method for estimating the motion state of a mover without a position sensor. Specifically, it relates to a method and system for estimating the motion state of a mover in a direct-drive wave power generation system without a position sensor. Background Technology

[0002] Wave energy, as an important marine renewable energy source, boasts significant advantages such as abundant resources, high power density, and minimal environmental impact. Among various wave energy utilization technologies, direct-drive wave energy generation systems have become a key research focus due to their high energy conversion efficiency, simple mechanical structure, and strong operational reliability. They are also considered promising for providing continuous and reliable power to offshore autonomous equipment such as marine data buoys and navigation facilities. In recent years, prototype direct-drive wave energy generation systems have been successfully deployed at sea, verifying their engineering feasibility in actual sea conditions. However, direct-drive wave energy generation systems still face challenges in practical applications, including maintenance difficulties and large fluctuations in output power. To improve the operational reliability of wave energy generation systems and achieve real-time optimization of output power, various control strategies have been proposed and validated, including model predictive control, non-causal linear optimal control, and fault-tolerant control. Although these control methods differ in principle and implementation, they all rely on a common key factor: accurate acquisition of the motion state of the mover.

[0003] In existing direct-drive wave energy generation systems, the acquisition of motion information of the mover mainly relies on mechanical position sensors. However, due to the long-term operation of such power generation devices in complex and variable marine environments, mechanical sensors face severe challenges. On the one hand, strong wave impacts can easily cause physical damage or signal interference to the sensors; on the other hand, the large diurnal temperature range and significant seasonal variations in surface seawater mean that without additional temperature compensation and protection measures, environmental factors will directly affect the accuracy and reliability of sensor measurement data. To mitigate these effects, temperature compensation devices, vibration damping structures, and shock-resistant designs are typically introduced to enhance sensor adaptability. However, this inevitably increases equipment complexity, conflicting with the design intent of direct-drive wave energy generation systems to achieve simple structure and high reliability. Furthermore, since such devices are usually deployed using seabed anchoring, the feasibility of regular sensor maintenance, calibration, and replacement is low.

[0004] Therefore, traditional mechanical position sensors are ill-suited to the long-term, reliable operation requirements of wave energy power generation devices in harsh marine environments. Accurately and reliably acquiring the motion state information of the mover has become one of the key challenges restricting the development of direct-drive wave energy power generation technology. Against this backdrop, researching mover motion state estimation methods that do not rely on position sensors has significant engineering practical value and theoretical implications for improving the control reliability and accuracy of direct-drive wave energy power generation systems, thereby achieving stable system operation and real-time optimization of output power. Summary of the Invention

[0005] This invention addresses the problems existing in the prior art by providing a sensorless motion state estimation method and system for a direct-drive wave power generation system. First, a mathematical model of the direct-drive wave power generation system is established based on the motion equation of the float under irregular wave excitation force, a permanent magnet synchronous linear generator model, and a power tracking control circuit. Then, based on the coupling relationship between the hydrodynamic, mechanical, and electrical responses in the system, key variables reflecting this coupling characteristic are selected as state variables to construct the wave-electricity full-process state equation. On this basis, a motion estimation method based on the spherical simplex unscented Kalman filter algorithm is proposed to achieve sensorless power optimization control. Finally, a simulation model of the direct-drive wave power generation system is built in Simulink to verify the proposed method offline and online. The results show that the method of this invention can achieve sensorless power optimization control, improving the computational efficiency of the estimation method and the control accuracy of the system.

[0006] To achieve the above objectives, the technical solution adopted by this invention is: a sensorless motion state estimation method for a direct-drive wave power generation system, comprising the following steps:

[0007] S1: Establish a mathematical model of a direct-drive wave power generation system. The model includes the motion equation of the float under the excitation force of irregular waves, the permanent magnet synchronous linear generator model, and the power tracking control circuit model.

[0008] S2: Based on the coupling relationship between the hydrodynamic, mechanical and electrical responses of the system in the mathematical model of the direct-drive wave power generation system established in step S1, key variables are determined as state variables, and the wave-electric whole process state equation is constructed; the key variables include the displacement and velocity of the float, the phase current of the linear motor and the internal state quantities describing the fluid memory effect in the hydrodynamic response;

[0009] S3: Based on the wave-electric whole process state equation constructed in step S2, the motion state of the mover is estimated by using the spherical simplex unscented Kalman filter algorithm. The spherical simplex unscented Kalman filter algorithm introduces the spherical simplex unscented transformation optimization strategy into the standard unscented Kalman filter framework. The spherical simplex unscented transformation is used to recursively update the state and error covariance of the nonlinear state space model, approximating the statistical characteristics of the state distribution.

[0010] A simulation model of a direct-drive wave power generation system was built in Simulink, and system operation data was collected. Based on this data, the methods in steps S1-S3 were verified offline in MATLAB to evaluate the accuracy and computational efficiency of motion state estimation. Then, based on the constructed system simulation model, the methods in steps S1-S3 were verified online. The estimation method was encapsulated and modularized using MATLAB Functions and embedded into the Simulink system simulation model. The output results were used as feedback signals for power point tracking control in closed-loop control, and the effectiveness of its application in the control loop was evaluated by assessing the stability and accuracy of power point tracking control.

[0011] As an improvement of the present invention, in step S1, the equation of motion of the float under the excitation force of irregular waves is specifically as follows:

[0012]

[0013] in, For the mass of the float, For when Additional mass at that time For wave frequency, , and They are respectively The displacement, velocity, and acceleration of the buoy at any given moment. The restoring stiffness coefficient of the hydrostatic restoring force. and They represent The wave excitation force and electromagnetic force that the buoy experiences at any given moment. This represents the radiation impulse response function.

[0014] As another improvement of the present invention, in step S1, the permanent magnet synchronous linear generator is in The model in the coordinate system is as follows:

[0015]

[0016] in, and for Phase current in coordinate system It is a three-phase inductor. It is a three-phase resistor. It is a permanent magnet flux chain. and These are the displacement and velocity of the float, respectively. For polar distance, and for Phase voltage in the coordinate system.

[0017] As another improvement of the present invention, in step S1, the power tracking control of the direct-drive wave power generation system in the power tracking control circuit model is implemented using a field-oriented control strategy. Shaft reference current for:

[0018]

[0019] in, and These represent the optimal electromagnetic force and the generator damping, respectively.

[0020] As another improvement of the present invention, in step S2, the convolution term used to describe the fluid memory effect in the hydrodynamic response is equivalently transformed into a state-space form using the Realization method:

[0021]

[0022] in, For approximate convolution terms Order state variables, , and For the parameter matrix, Represents the radiation impulse response function. The speed of the float, For the current time, The historical time variable in the convolution integral;

[0023] By combining the motion equations of the mover after replacing the convolution terms with the generator model, the state equations for the entire wave-electric process in continuous time are as follows:

[0024]

[0025] in, and for Phase current in coordinate system and These are the displacement and velocity of the float, respectively. Here are the state variables used to approximate the fluid memory effect in the hydrodynamic response, , and For the parameter matrix, The restoring stiffness coefficient of the hydrostatic restoring force. The wave excitation force experienced by the float, For the mass of the float, For when Additional mass at that time It is a three-phase inductor. It is a three-phase resistor. It is a permanent magnet flux chain. For polar distance, and for Phase voltage in coordinate system This is process noise.

[0026] As a further improvement of the present invention, the spherical simplex unscented Kalman filter algorithm in step S3 specifically includes the following steps:

[0027] S31, Sigma point selection: Sigma point set Composed of n+2 sigma points, the sigma points are selected using an optimization strategy based on the spherical simplex unscented transformation. indivual The sigma point set based on the spherical simplex unscented Kalman filter, composed of dimensional vectors, is represented as:

[0028]

[0029] in, Indicates the first 1 Sigma point, It is the mean vector. It is the covariance matrix;

[0030] S32, State Prediction: In At time t, each sigma point is transitioned via the state transition function. Mapping, generation time The predicted sigma points are used to calculate the mean of the predicted state. Covariance A new set of sigma points is generated; the mean of the predicted state. Covariance The calculation method is as follows:

[0031]

[0032]

[0033]

[0034] in, Indicates the first sigma points at Predicted value at time, express Input at any time Represents the state transition function. This represents the dimension of the state vector. express The mean of the predicted state at time t. Indicates the first The weight of each sigma point express The covariance of the predicted state at time t. Represents the process noise covariance matrix;

[0035] S33. Measurement Update: Substitute the new sigma point generated in step S32 into the measurement equation to generate the predicted measurement vector. Calculate the probability distribution of the predicted measurement and the cross-covariance matrix between the state variables and the measurement. Used to solve for Kalman gain And update the state estimate and covariance:

[0036]

[0037]

[0038]

[0039]

[0040]

[0041] in, express The predicted measurement vector at time t. Represents the measurement matrix. Indicates the first sigma points at Predicted value at time, express The mean of the predicted state at time t. This represents the dimension of the state vector. Indicates the first The weight of each sigma point Represents the measurement noise covariance matrix. and They are respectively The mean and residual covariance matrix of the predicted measurements at time points. express The cross-covariance matrix between the state variables and measurements at time points. express Kalman gain at time step;

[0042] The state and covariance update process is represented as follows:

[0043]

[0044]

[0045] in, express State estimate at time 10:00 express The mean of the predicted state at time t. express Kalman gain at time step express The actual measurement value at that moment. for The mean of the predicted measurements at time. express The covariance matrix of the posterior error at time t. express The covariance of the predicted state at time t. for The residual covariance matrix at time step;

[0046] To achieve the above objectives, the present invention also adopts the following technical solution: a sensorless motion state estimation system for a direct-drive wave power generation system, comprising a computer program, wherein the computer program, when executed by a processor, implements the steps of any of the methods described above.

[0047] Compared with the prior art, the present invention has the following advantages: The present invention proposes a sensorless motion state estimation method for a direct-drive wave power generation system, which can obtain accurate motion information of the mover under sensorless conditions. By taking into account the computational efficiency and accuracy of the estimation method, the system can ensure the reliability of operation in the marine environment while realizing real-time power optimization, which has practical application value.

[0048] (1) This invention establishes a mathematical model of a direct-drive wave power generation system that takes into account the fluid memory effect, and describes the hydrodynamic, mechanical and electrical responses and their coupling relationship in the system under real irregular wave excitation force, so as to provide a dynamic model basis for the motion state estimation of the mover under real wave excitation.

[0049] (2) By selecting the state variables that reflect the coupling characteristics of multiple physics fields in the mathematical model of the direct-drive wave power generation system and constructing the wave-electric whole process state equation covering the complete energy conversion process from wave excitation to power output, this invention can accurately characterize the coupling mechanism of hydrodynamic, mechanical and electrical response of the system and depict the kinematic characteristics of the mover under the coupling conditions of multiple physics fields, providing a model basis for the application of recursive state estimation algorithm.

[0050] (3) The present invention realizes power optimization control without position sensors by proposing a motion estimation method based on the spherical simplex unscented Kalman filter algorithm, which reduces the computational burden of sampling point generation, propagation and weighting process, and improves computational efficiency while ensuring estimation accuracy. It is suitable for real-time deployment on embedded platforms with limited computing resources.

[0051] (4) The motion state estimation method proposed in this invention is decoupled from the control strategy and has the same interface as the existing power control strategy. The proposed method has the engineering implementation advantage of being integrated into the existing direct-drive wave power generation system control architecture with minimal modifications at both the software and hardware levels.

[0052] (5) The offline and online verification mechanism proposed in this invention can verify the effectiveness of the method from the dimensions of estimation error and computational efficiency, closed-loop power tracking performance, etc., providing a basis for subsequent hardware prototype verification and engineering application and reducing development risks. Attached Figure Description

[0053] Figure 1 This is a block diagram of the control structure of the direct-drive wave power generation system without position sensors of the present invention.

[0054] Figure 2 This is a flowchart of the sensorless motion state estimation method for the direct-drive wave power generation system of the present invention. Detailed Implementation

[0055] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0056] Example 1

[0057] A sensorless motion state estimation method for a direct-drive wave power generation system is proposed. Based on the wave-electricity whole-process state equation, a motion estimation method based on the spherical simplex unscented transformation-unscented Kalman filter (SSUT-UKF) algorithm is constructed, enabling sensorless power optimization control. The system control structure block diagram is shown below. Figure 1As shown in the figure, the power point tracking control of the direct-drive wave power generation system is implemented using classic field-oriented control. The figure illustrates the integration of the mover motion estimation module based on SSUT-UKF with the control circuit. A sensorless mover motion state estimation method for a direct-drive wave power generation system is also presented. Figure 2 As shown, the specific steps include:

[0058] Step S1: Establish a mathematical model of the direct-drive wave power generation system, including the motion equation of the float under the excitation force of irregular waves, the permanent magnet synchronous linear generator model, and the power tracking control circuit model.

[0059] The forces acting on the buoy in the heave direction are analyzed based on Airy wave theory. According to fluid dynamics analysis and Newton's second law, the equation of motion of the buoy in irregular waves can be expressed as:

[0060]

[0061] in, For the mass of the float, for The acceleration of the buoy at any given moment, , , and They represent The buoy is subjected to wave excitation force, electromagnetic force, still water restoring force and radiation force at all times.

[0062] The wave excitation force experienced by the float in irregular waves is generated by the ITTC (International Towing Tank Conference) spectrum, and its spectral function is:

[0063]

[0064] in, For the sake of righteousness, the waves rise high. The peak period of the spectrum, For wave frequency, For spectral peak factor, These are the spectral peak shape parameters.

[0065] The wave excitation force, constructed from the discretized wave frequency, can be expressed as:

[0066]

[0067] in, Indicates the first Wave component The lowest frequency, For frequency intervals, the first Amplitude of each wave component , Indicates the first The random initial phase of each wave component, This is the frequency response function.

[0068] The radiative force is modeled using the Cummins method:

[0069]

[0070] in, Indicates when The added mass at that time. The convolution term describes the fluid's "memory effect," in which... The radiative impulse response function is expressed as follows:

[0071]

[0072] in, For radiation damping.

[0073] Therefore, the heaving motion of the float in irregular waves can be expressed as:

[0074]

[0075] Permanent magnet synchronous linear generator in The mathematical model in the coordinate system is expressed as:

[0076]

[0077] in, and for Phase current in coordinate system It is a three-phase inductor. It is a three-phase resistor. It is a permanent magnet flux chain. For polar distance, and for Phase voltage in the coordinate system.

[0078] Permanent magnet synchronous linear generator in Electromagnetic force in a coordinate system can be expressed as:

[0079]

[0080] generator in Mathematical models in a coordinate system can be derived from the Park transformation. Exporting the coordinate system. Electromagnetic force in the coordinate system can be further expressed as:

[0081]

[0082] The power tracking control of the direct-drive wave power generation system is achieved using a classic field-oriented control strategy. Shaft reference current for:

[0083]

[0084] in, and These represent the optimal electromagnetic force and the generator damping, respectively.

[0085] Step S2: Analyze the coupling relationship between the hydrodynamic, mechanical and electrical responses of the system in the model, select key variables that can reflect the coupling characteristics of the three as state variables, and construct the wave-electric whole process state equation.

[0086] In a direct-drive wave power generation system, the motion of the mover is influenced by the coupling effects of hydrodynamic, mechanical, and electrical multi-physics fields. To describe the dynamic response of the mover under these coupling effects, the displacement and velocity of the float, rigidly connected to the linear motor mover, are used as mechanical state variables, and the phase current of the linear motor is used as an electrical state variable. Internal state variables are also introduced to describe the fluid memory effect in the hydrodynamic response. Based on the coupling relationship between the float motion equation in the mathematical model of the direct-drive wave power generation system and the state variables established in the generator model, a wave-electricity full-process state equation is constructed. This state equation encompasses the entire process from wave excitation to electrical energy output, describing the motion characteristics of the mover under multi-physics coupling, and providing a theoretical basis for the application of recursive state estimation algorithms.

[0087] To account for the memory effect in the hydrodynamic response, internal state variables are introduced into the wave-electric whole-process state equation. Using the Realization method, the convolution terms describing the fluid memory effect in the hydrodynamic response are equivalently transformed into a state-space form:

[0088]

[0089] in, For approximate convolution terms Order state variables, , and It is a parameter matrix.

[0090] By combining the motion equations of the mover after replacing the convolution terms with the generator model, the state equation for the entire wave-electric process in continuous time can be expressed as:

[0091]

[0092] in, This is process noise.

[0093] The above continuous-time wave-electric process state equations are discretized using the Euler method at a given sampling period to obtain the discrete-time state equations:

[0094]

[0095] Among them, subscript Represents discrete time points. This is the state transition function. The state vector is defined as follows: The input vector is .

[0096] In the state vector, the measurable component is the current. , The phase current of the permanent magnet synchronous linear generator, acquired by a current sensor, is obtained through Clarke transformation. Therefore, the discrete-time measurement equation can be expressed as:

[0097]

[0098] in, To measure noise, the measurement matrix is ​​defined as follows: , and Let these represent the identity matrix and the zero matrix, respectively. The measurement vector is defined as... The superscript "meas" indicates that the variable was obtained by sensor measurement.

[0099] Step S3: Based on the above wave-electric whole-process state equations, a mover motion state estimation method based on SSUT-UKF is proposed. This method improves the computational efficiency of real-time estimation by introducing a spherical simplex unscented transformation optimization strategy into the standard unscented Kalman filter (UKF) framework. SSUT-UKF recursively updates the state and error covariance of the nonlinear state-space model through spherical simplex unscented transformation. This transformation approximates the statistical properties of the state distribution through a set of weighted sampling points (called sigma points). The execution process of SSUT-UKF mainly includes three steps:

[0100] (1) Selection of sigma points

[0101] sigma point set It consists of n+2 sigma points, and the sigma points are selected using an optimization strategy based on the spherical simplex unscented transformation. This represents the dimension of the state vector. Indicates the first sigma points The weight.

[0102] First, the initial weights The value of satisfies The weights of the remaining sigma points are:

[0103]

[0104] Auxiliary vector sequence (dimension) used for sigma point selection This can be represented as:

[0105]

[0106] in, , For the first Victor There are 1 vector components, and the initial vector sequence is:

[0107]

[0108] From the above indivual The set of sigma points of SSUT-UKF composed of dimensional vectors can be represented as:

[0109]

[0110] in, Indicates the first 1 Sigma point, It is the mean vector. Let be the covariance matrix.

[0111] (2) State prediction

[0112] exist At time t, each sigma point is transitioned via the state transition function. Mapping, generation time The predicted sigma points are then used to calculate the mean of the predicted state. Covariance Next, following the method in step (1), a new set of sigma points is generated.

[0113]

[0114]

[0115]

[0116] in, Indicates the first sigma points at Predicted value at time, This represents the process noise covariance matrix.

[0117] (3) Measurement update

[0118] The new sigma points generated in step (2) are substituted into the measurement equation to generate the predicted measurement vector. Then, the probability distribution of the predicted measurement and the cross-covariance matrix between the state variables and the measurement are calculated. Used to solve for Kalman gain And update the state estimate and covariance:

[0119]

[0120]

[0121]

[0122] in, and These are the mean vector and residual covariance matrix of the predicted measurements, respectively. This represents the measurement noise covariance matrix.

[0123] cross-covariance matrix It can be represented as:

[0124]

[0125] Kalman gain It can be represented as:

[0126]

[0127] The state and covariance update process can be represented as:

[0128]

[0129]

[0130] A simulation model of a direct-drive wave power generation system was built in Simulink, and simulation data was collected. The proposed mover motion state estimation method was then validated offline in MATLAB based on this data, and its estimation accuracy and computational efficiency were evaluated.

[0131] The simulation model of the direct-drive wave power generation system includes a wave excitation input, a float hydrodynamic response module, a permanent magnet linear synchronous generator module, a three-phase rectifier using SVPWM modulation and current feedback control, and a terminal DC load. System parameters are shown in Table 1. The spectral peak period and significant wave height of the irregular wave are 4.5 s and 1.5 m, respectively, and the sampling period of the current control loop is set to 1 µs. All simulations were performed on a computer equipped with an Intel Core i9-13900HX processor and 16 GB of memory using the MATLAB R2023b / Simulink platform.

[0132] Table 1 Parameters of Direct-Drive Wave Power Generation System

[0133] The baseline results required for comparison and verification are obtained through simulation. These baseline results are then imported into MATLAB and used as the true values ​​for subsequent current measurement generation and estimation error comparison. Measured values The motion state is obtained by superimposing measurement noise onto the real current component. Extended Kalman Filter (EKF), UKF, and SSUT-UKF methods were used for mover motion state estimation. The estimation results were compared with the real state, and the computation time of different estimation algorithms was compared to evaluate the estimation accuracy and computational efficiency of the mover motion state estimation method based on SSUT-UKF. Offline validation results based on the EKF, UKF, and SSUT-UKF methods are shown in Table 2.

[0134] Table 2 Offline verification results

[0135]

[0136] As can be seen from the mean absolute errors in Table 2, both UKF and SSUT-UKF can accurately track the motion trajectory of the mover. The comparison of computational efficiency in Table 2 shows that the EKF method has the lowest computational burden. This is because the EKF method avoids generating and propagating multiple sigma points in the unscented transformation by linearizing the state transition function. However, the estimation results further indicate that EKF produces a larger error in velocity estimation compared to UKF and SSUT-UKF. The SSUT-UKF method is comparable to the standard UKF method in terms of estimation accuracy, while further reducing the computational burden, achieving an approximately 33% improvement in computational efficiency compared to UKF.

[0137] Based on the constructed system simulation model, the proposed mover motion state estimation method was validated online to evaluate the stability and accuracy of power point tracking (PPT) control in a direct-drive wave generator system using the proposed estimation algorithm. The estimation method was encapsulated as a module using a MATLAB function and embedded into the Simulink system simulation model. Its output served as the feedback signal for PPT control in the closed-loop control. The stability and accuracy of PPT control were used as evaluation indicators to assess the effectiveness of the proposed estimation method in the control loop. The online validation results are shown in Table 3. Analysis of the mean absolute error during online estimation and the average deviation of the mover motion relative to the baseline operation when the estimation result is used as control feedback shows that the online validation results are largely consistent with the offline validation results: both UKF and SSUT-UKF can effectively track the reference motion trajectory, while EKF exhibits a relatively large motion deviation.

[0138] The results of both offline and online verification show that the proposed motion state estimation method based on SSUT-UKF improves computational efficiency while maintaining estimation accuracy comparable to that of UKF. It also demonstrates good stability and accuracy in closed-loop power point tracking control and is suitable for real-time control applications of direct-drive wave power generation systems.

[0139] Table 3 Online Validation Results

[0140] In summary, the method of this invention considers the dynamic response of the float under irregular wave excitation force and establishes a mathematical model of a direct-drive wave power generation system. Key variables reflecting the coupling characteristics between hydrodynamic, mechanical, and electrical responses are selected as state variables to construct the wave-electrical whole-process state equation. Based on this state equation, a mover motion estimation method based on the SSUT-UKF algorithm is proposed to achieve power optimization control without position sensors. Finally, a simulation model of the direct-drive wave power generation system is built in Simulink to verify the proposed method offline and online. The verification results show that the method of this invention significantly improves the control reliability and accuracy of the system and has practical application value.

[0141] It should be noted that the above content merely illustrates the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. For those skilled in the art, various improvements and modifications can be made without departing from the principle of the present invention, and all such improvements and modifications fall within the scope of protection of the claims of the present invention.

Claims

1. A sensorless motion state estimation method for a direct-drive wave power generation system, characterized in that... It includes the following steps: S1: Establish a mathematical model of a direct-drive wave power generation system. The model includes the motion equation of the float under the excitation force of irregular waves, the permanent magnet synchronous linear generator model, and the power tracking control circuit model. S2: Based on the coupling relationship between the hydrodynamic, mechanical and electrical responses of the system in the mathematical model of the direct-drive wave power generation system established in step S1, key variables are determined as state variables, and the wave-electric whole process state equation is constructed; the key variables include the displacement and velocity of the float, the phase current of the linear motor and the internal state quantities of the fluid memory effect; S3: Based on the wave-electric whole process state equation constructed in step S2, the motion state of the mover is estimated by using the spherical simplex unscented Kalman filter algorithm. The spherical simplex unscented Kalman filter algorithm introduces the spherical simplex unscented transformation optimization strategy into the standard unscented Kalman filter framework. The spherical simplex unscented transformation is used to recursively update the state and error covariance of the nonlinear state space model, approximating the statistical characteristics of the state distribution.

2. The sensorless motion state estimation method for a direct-drive wave power generation system as described in claim 1, characterized in that: In step S1, the equation of motion of the float under the excitation force of irregular waves is specifically as follows: ； in, For the mass of the float, For when Additional mass at that time For wave frequency, , and They are respectively The displacement, velocity, and acceleration of the buoy at any given moment. The restoring stiffness coefficient of the hydrostatic restoring force. and They represent The wave excitation force and electromagnetic force that the buoy experiences at any given moment. This represents the radiation impulse response function.

3. The sensorless motion state estimation method for a direct-drive wave power generation system as described in claim 2, characterized in that: In step S1, the wave excitation force experienced by the float Specifically: ； in, Indicates the first Wave component The lowest frequency, For frequency intervals, the first Amplitude of each wave component , Indicates the first The random initial phase of each wave component, This is the frequency response function.

4. The sensorless motion state estimation method for a direct-drive wave power generation system as described in claim 1, characterized in that: In step S1, the permanent magnet synchronous linear generator is The model in the coordinate system is as follows: ； in, and for Phase current in coordinate system It is a three-phase inductor. It is a three-phase resistor. It is a permanent magnet flux chain. and These are the displacement and velocity of the float, respectively. For polar distance, and for Phase voltage in the coordinate system.

5. The sensorless motion state estimation method for a direct-drive wave power generation system as described in claim 4, characterized in that: In step S1, the power tracking control of the direct-drive wave power generation system in the power tracking control circuit model is achieved using a field-oriented control strategy. Shaft reference current for: ； in, and These represent the optimal electromagnetic force and the generator damping, respectively.

6. The sensorless motion state estimation method for a direct-drive wave power generation system as described in claim 1, characterized in that: In step S2, the convolutional terms used to describe the fluid memory effect in the hydrodynamic response are transformed into a state-space form using the Realization method: ； in, For approximate convolution terms Order state variables, , and For the parameter matrix, Represents the radiation impulse response function. The speed of the float, For the current time, The historical time variable in the convolution integral; By combining the motion equations of the mover after replacing the convolution terms with the generator model, the state equations for the entire wave-electric process in continuous time are as follows: ； in, and for Phase current in coordinate system and These are the displacement and velocity of the float, respectively. Here are the state variables used to approximate the fluid memory effect in the hydrodynamic response, , and For the parameter matrix, The restoring stiffness coefficient of the hydrostatic restoring force. The wave excitation force experienced by the float, For the mass of the float, For when Additional mass at that time It is a three-phase inductor. It is a three-phase resistor. It is a permanent magnet flux chain. For polar distance, and for Phase voltage in coordinate system This is process noise.

7. The sensorless motion state estimation method for a direct-drive wave power generation system as described in claim 1, characterized in that: The spherical simplex unscented Kalman filter algorithm in step S3 specifically includes the following steps: S31, Sigma point selection: Sigma point set Composed of n+2 sigma points, the sigma points are selected using an optimization strategy based on the spherical simplex unscented transformation. indivual The sigma point set based on the spherical simplex unscented Kalman filter, composed of dimensional vectors, is represented as: ； in, Indicates the first 1 Sigma point, It is the mean vector. It is the covariance matrix; S32, State Prediction: In At time t, each sigma point is transitioned via the state transition function. Mapping, generation time The predicted sigma points are used to calculate the mean of the predicted state. Covariance A new set of sigma points is generated; the mean of the predicted state. Covariance The calculation method is as follows: ； ； ； in, Indicates the first sigma points at Predicted value at time, express Input at any time Represents the state transition function. This represents the dimension of the state vector. express The mean of the predicted state at time t. Indicates the first The weight of each sigma point express The covariance of the predicted state at time t. Represents the process noise covariance matrix; S33. Measurement Update: Substitute the new sigma point generated in step S32 into the measurement equation to generate the predicted measurement vector. Calculate the probability distribution of the predicted measurement and the cross-covariance matrix between the state variables and the measurement. Used to solve for Kalman gain And update the state estimate and covariance: ； ； ； ； ； in, express The predicted measurement vector at time t. Represents the measurement matrix. Indicates the first sigma points at Predicted value at time, express The mean of the predicted state at time t. This represents the dimension of the state vector. Indicates the first The weight of each sigma point Represents the measurement noise covariance matrix. and They are respectively The mean and residual covariance matrix of the predicted measurements at time points. express The cross-covariance matrix between the state variables and measurements at time points. express Kalman gain at time step; The state and covariance update process table is as follows: ； ； in, express State estimate at time 10:00 express The mean of the predicted state at time t. express Kalman gain at time step express The actual measurement value at that moment. for The mean of the predicted measurements at time. express The covariance matrix of the posterior error at time t. express The covariance of the predicted state at time t. for The residual covariance matrix at time step 1.

8. A sensorless motion state estimation system for a direct-drive wave power generation system, comprising a computer program, characterized in that: When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-7 above.

Images