Model-free predictive power control method and system for three-phase grid-connected inverter

Abstract

The invention discloses a model-free predictive power control method and system for a three-phase grid-connected inverter, which belongs to the field of power electronic control. Firstly, a super-local model is constructed to replace the original inverter model, so as to reduce the dependence on the system model. Secondly, a quadratic linear extended state observer is designed, and a quadratic expansion estimation is performed on the nonlinear disturbance residual through an integral negative feedback link, so as to estimate the lumped disturbance of the super-local model, thereby improving the anti-disturbance capability of the observer. Meanwhile, a Kalman filter is introduced into the integral negative feedback link, and the filter factor is automatically adjusted according to the dynamic characteristics of the system, so as to avoid the problems of current distortion and response delay caused by excessive filtering, so as to make it more effective in processing complex nonlinear disturbances. Finally, a deadbeat cost function is designed to ensure the high dynamics of the system, so as to solve the problem that it is difficult to establish an accurate physical model due to the nonlinear disturbance caused by parameter mismatch of the three-phase grid-connected inverter in actual working condition operation.

Description

Technical Field

[0001] The present invention belongs to the field of power electronic control, and in particular relates to a model-free predictive power control method and system for a three-phase grid-connected inverter based on a Kalman filter quadratic linear extended state observer. Background Art

[0002] Three-phase inverters have been widely used in the fields of renewable energy generation, microgrid energy storage and fuel cells due to their advantages of high power density, low harmonic distortion and high reliability. However, in practical applications, three-phase inverters are affected by parameter mismatch, unknown disturbances and other factors, making it difficult to establish an accurate physical model, resulting in a decrease in system control performance.

[0003] In addition, in practice, unknown disturbances usually exhibit nonlinear characteristics, and when dealing with this type of disturbance, the linear extended state observer has problems such as low estimation accuracy, phase lag and amplitude attenuation. Summary of the invention

[0004] The purpose of the present invention is to propose a model-free predictive power control method and system for a three-phase grid-connected inverter, which does not rely on an accurate model. A compensation factor is introduced into the feedback loop of a linear extended state observer disturbance estimation to compensate for the estimation residual of the linear extended state observer for nonlinear disturbances, thereby improving the utilization rate of the disturbance estimation for errors. At the same time, a Kalman filter is introduced into the integral negative feedback link, thereby improving the dynamics and robustness of the system to effectively handle complex nonlinear disturbances.

[0005] The present invention provides a model-free predictive power control method for a three-phase grid-connected inverter, comprising the following steps:

[0006] Step S1, constructing a super local model to replace the original inverter model, so as to reduce the dependence on the system model;

[0007] Step S2, designing a quadratic linear extended state observer QLESO, performing quadratic expansion estimation on the nonlinear disturbance residual through an integral negative feedback link, so as to estimate the lumped disturbance of the super-local model; at the same time, introducing a Kalman filter in the integral negative feedback link, taking into account the process noise in the system dynamic model through the Kalman filter, and automatically adjusting its parameters according to the dynamic characteristics of the system and the changes in the observed data, so as to improve the estimation accuracy of the quadratic linear extended state observer QLESO;

[0008] Step S3, substitute the discrete power super-local prediction model into the cost function of the power loop designed in combination with the deadbeat control principle, calculate the optimal voltage vector of the active power and reactive power loops, and output the optimal voltage vector to the controller to realize model-free predictive power control.

[0009] The step S1 is specifically:

[0010] Step S11: Establish a three-phase inverter Power model in coordinate system:

[0011]

[0012] in, Represents the active power of the grid, Represents the reactive power of the grid, and Respectively represent the inverter output voltage at The components in the coordinate system, and Respectively represent the grid voltage at Components in the coordinate system; L represents the filter inductance, R represents the equivalent series resistance in the power grid, Indicates the angular frequency of the power grid;

[0013] Step S12: construct a power prediction model for a three-phase inverter:

[0014] According to the forward Euler formula, the formula (1) is discretized to obtain the power prediction model of the three-phase inverter:

[0015]

[0016] in is the system sampling period; and They represent the inverter output voltage in the kth sampling period. Components in the coordinate system; and They represent the grid voltage in the kth sampling period. Components in the coordinate system; Represents the reactive power of the power grid at the moment of the kth sampling period; represents the active power of the power grid in the kth sampling period; Indicates the angular frequency of the power grid; R Represents the equivalent series resistance in the power grid;

[0017] The cost function of the power loop designed in combination with the deadbeat control principle is:

[0018] (3)

[0019] in, is the reference value of active power, The active power is The value of the moment, is the reference value of reactive power, Reactive power is The value of the moment;

[0020] when and hour, The minimum, optimal voltage vector size at time k is:

[0021]

[0022] in, and are the voltage vectors of active power and reactive power respectively, and are the optimal voltage vectors for active power and reactive power loops respectively;

[0023] Step S13: Convert the power model into a super-local model:

[0024]

[0025] in, , is the input gain of each control loop, , are the lumped disturbances in the active power and reactive power models, respectively, which can be expressed as:

[0026]

[0027] The input gain is:

[0028] (7)

[0029] According to the forward Euler formula, equation (4) is discretized as follows:

[0030]

[0031] in, is the system sampling period; , Respectively represent k Sampling period ; , Respectively represent k +1 sampling period ; , Respectively represent k Sampling period ; , Respectively represent k Sampling period .

[0032] The step S2 is specifically:

[0033] Step S21, design a quadratic linear extended state observer QLESO, and perform quadratic expansion estimation on the nonlinear disturbance residual through an integral negative feedback link to estimate the lumped disturbance of the hyperlocal model:

[0034] The quadratic linear extended state observer QLESO is:

[0035]

[0036] in, is the state vector, is the state vector The observed value of is the state vector observation The derivative of is the gain vector of the system, is the power observation loop gain, is the perturbation observation loop gain, and are the output and input of each loop respectively, for Observed value of Lumped disturbance The compensation function, Lumped disturbance The estimated value of Compensates for the unknown model function in the original system ;

[0037] make , then the quadratic linear extended state observer QLESO with integral form is:

[0038]

[0039] in, represents the filtering factor, is the control coefficient;

[0040] According to the Lyapunov stability criterion, the characteristic equation of the quadratic linear extended state observer QLESO is expressed as:

[0041] (11)

[0042] in, s represents the variable in the Laplace transform domain, is the bandwidth of QLESO;

[0043] In order to make the observer stable, the roots of the characteristic equation must all fall on the negative semi-axis, so , , ; Discretize equation (10) to get:

[0044]

[0045] Among them, only the control coefficient Different from the state variable x, in the active power loop, let , , used to substitute into the active power super-local model; in the reactive power loop, let , , used to substitute into the reactive power hyperlocal model;

[0046] Step S22, design a quadratic linear state observer KF-QLESO containing Kalman filtering, which takes into account the process noise in the system dynamic model through the Kalman filter, and automatically adjusts its parameters according to the dynamic characteristics of the system and the changes in the observed data, so as to improve the estimation accuracy of the quadratic linear extended state observer QLESO:

[0047] (1) Kalman filter prediction steps:

[0048] The Kalman filter model is as follows:

[0049]

[0050] in, X is the system state matrix, H is the state transfer matrix, J is the control input matrix, U To control the amount, W is the process noise, Z is the observed quantity of the system state matrix, I is the state observation matrix, M To measure noise;

[0051] Set the current state estimate of the unknown nonlinear disturbance to:

[0052] (14)

[0053] (2) Kalman filter update steps:

[0054] The covariance of the updated estimate error is:

[0055] (15)

[0056] in, for H The transposed matrix ofP is the estimated error covariance of the state at the previous moment, Q is the covariance of process noise;

[0057] Update the Kalman gain:

[0058] (16)

[0059] in, K g is the Kalman gain, for I The transposed matrix of R is the covariance of the observation noise;

[0060] k The optimal estimate of the unknown nonlinear perturbation at time is:

[0061] (17)

[0062] Substituting formula (14) into formula (17), we can obtain:

[0063] (18)

[0064] but k The error covariance at time is:

[0065] (19)

[0066] The quadratic linear extended state observer with Kalman filter is estimated Substituting into equation (8) in step S13, the discrete power super-local prediction model is obtained:

[0067] .

[0068] The optimal voltage vector of the active power and reactive power loop calculated in step S3 is:

[0069] .

[0070] A three-phase grid-connected inverter model-free predictive power control system comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the model-free predictive power control method of the three-phase grid-connected inverter as described above is implemented.

[0071] A computer storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned model-free predictive power control method for a three-phase grid-connected inverter.

[0072] The technical solution of the present invention is adopted. Firstly, a super-local model is constructed to replace the original inverter model, so as to reduce the dependence on the system model and improve the robustness of the system. Secondly, a quadratic linear extended state observer is designed, and the nonlinear disturbance residual is quadratically expanded through an integral negative feedback link to estimate the lumped disturbance of the super-local model, thereby improving the anti-disturbance ability of the observer. Meanwhile, a Kalman filter is introduced in the integral negative feedback link to automatically adjust the filter factor according to the dynamic characteristics of the system to avoid problems such as current distortion and response delay caused by excessive filtering, so as to make it more effective in processing complex nonlinear disturbances. Finally, a deadbeat cost function is designed to ensure the high dynamics of the system, so as to solve the problem that it is difficult to establish an accurate physical model due to the nonlinear disturbance caused by parameter mismatch of the three-phase grid-connected inverter in actual working condition operation.

[0073] The present invention can not only improve the dynamic steady-state performance of the three-phase grid-connected inverter under the influence of external factors such as temperature changes during actual operation, but also use the constructed quadratic linear extended state observer to perform quadratic expansion on the unknown nonlinear disturbance residual using the negative feedback link of the first-order filter containing the filter factor to improve the estimation accuracy of the observer. However, the quadratic linear extended state observer QLESO uses a first-order linear low-pass filter to correct the unknown disturbance in the system, and has limitations when processing rapidly changing nonlinear disturbances. To solve this problem, a Kalman filter is used to replace the first-order linear low-pass filter in the QLESO, and a Kalman filter quadratic linear extended state observer is designed. Since the Kalman filter takes into account the process noise in the system dynamic model and can automatically adjust its parameters according to the dynamic characteristics of the system and the changes in the observed data, this adaptability makes the Kalman filter more effective when processing system fluctuations, further improving the estimation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 is a control logic diagram of the present invention;

[0075] Figure 2 This is a structural diagram of the Kalman filter quadratic linear extended state observer in the present invention. DETAILED DESCRIPTION

[0076] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0077] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0078] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or scheme described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0079] like Figure 1 As shown, a model-free predictive power control method for a three-phase grid-connected inverter is shown. In the figure, L represents the filter inductor, which is used to reduce the high-frequency components in the inverter output voltage and current to improve the power quality; R represents the line impedance; Indicates bus voltage; Indicates the reference value of bus voltage; Represents the grid current, Indicates the phase voltage of the grid voltage, Indicates the grid current The weight on the axis, Indicates the phase voltage of the power grid voltage The weight on the axis, Indicates the voltage of each phase of the inverter output Components on the axis; represents the value of the lumped disturbance in the active power model at time k, Representing the value of the lumped disturbance in the reactive power model at time k; comprising the following steps:

[0080] Step S1: construct a super-local model to replace the original inverter model to reduce the dependence on the system model:

[0081] Step S11: Establish a three-phase inverter Power model in coordinate system:

[0082]

[0083] in, Represents the active power of the grid, Represents the reactive power of the grid, and Respectively represent the inverter output voltage at The components in the coordinate system, and Respectively represent the grid voltage at Components in the coordinate system; L represents the filter inductance, R represents the equivalent series resistance in the power grid, Indicates the angular frequency of the power grid. The frequency of industrial frequency AC is 50Hz;

[0084] Step S12: construct a power prediction model for a three-phase inverter:

[0085] According to the forward Euler formula, the formula (1) is discretized to obtain the power prediction model of the three-phase inverter:

[0086]

[0087] in is the system sampling period, and They represent the inverter output voltage in the kth sampling period. Components in the coordinate system; and They represent the grid voltage in the kth sampling period. Components in the coordinate system; Represents the reactive power of the power grid at the moment of the kth sampling period; represents the active power of the power grid in the kth sampling period; Indicates the angular frequency of the power grid; R Represents the equivalent series resistance in the power grid;

[0088] The cost function of the power loop designed in combination with the deadbeat control principle is:

[0089] (3)

[0090] in, is the reference value of active power, The active power is The value of the moment, is the reference value of reactive power, Reactive power is The value of the moment;

[0091] when and hour, The minimum, optimal voltage vector size at time k is:

[0092]

[0093] in, and are the voltage vectors of active power and reactive power respectively, and are the optimal voltage vectors for active power and reactive power loops respectively;

[0094] Step S13: Convert the power model into a super-local model:

[0095]

[0096] in, , is the input gain of each control loop, , are the lumped disturbances in the active power and reactive power models, respectively, which can be expressed as:

[0097]

[0098] The input gain is:

[0099] (7)

[0100] According to the forward Euler formula, equation (4) is discretized as follows:

[0101]

[0102] in, is the system sampling period; , Respectively represent k Sampling period ; , Respectively represent k +1 sampling period ; , Respectively represent k Sampling period ; , Respectively represent k Sampling period ;

[0103] Step S2, design a quadratic linear extended state observer QLESO, perform quadratic expansion estimation on the nonlinear disturbance residual through an integral negative feedback link, so as to estimate the lumped disturbance of the super-local model, thereby improving the anti-disturbance ability of the observer; at the same time, introduce a Kalman filter in the integral negative feedback link, take into account the process noise in the system dynamic model through the Kalman filter, and automatically adjust its parameters according to the dynamic characteristics of the system and the changes in the observed data, so as to improve the estimation accuracy of the quadratic linear extended state observer QLESO:

[0104] Step S21, design a quadratic linear extended state observer QLESO, such as Figure 2 As shown in the figure, the nonlinear disturbance residual is estimated by quadratic expansion through the integral negative feedback link to estimate the lumped disturbance of the hyperlocal model, thereby improving the anti-disturbance ability of the observer:

[0105] The quadratic linear extended state observer QLESO is:

[0106]

[0107] in, is the state vector, is the state vector The observed value of is the state vector observation The derivative of is the gain vector of the system, is the power observation loop gain, is the perturbation observation loop gain, and are the output and input of each loop respectively, for Observed value of Lumped disturbance The compensation function, Lumped disturbance The estimated value of Compensates for the unknown model function in the original system ;

[0108] make , then the quadratic linear extended state observer QLESO with integral form is:

[0109]

[0110] in, represents the filtering factor, is the control coefficient;

[0111] According to the Lyapunov stability criterion, the characteristic equation of the quadratic linear extended state observer QLESO is expressed as:

[0112] (11)

[0113] in, s represents the variable in the Laplace transform domain, is the bandwidth of QLESO;

[0114] In order to make the observer stable, the roots of the characteristic equation must all fall on the negative semi-axis, so , , ; Discretize equation (10) to get:

[0115]

[0116] Among them, only the control coefficient Different from the state variable x, in the active power loop, let , , used to substitute into the active power super-local model; in the reactive power loop, let , , used to substitute into the reactive power hyperlocal model;

[0117] Step S22: Design a quadratic linear state observer KF-QLESO with Kalman filtering, with a filter factor The higher the value, the higher the precision of QLESO observations, but If the value is too high, the output signal will take longer to process the input signal, resulting in a slow response to changes in the input signal, which in turn affects the response speed of the system. The Kalman filter takes into account the process noise in the system dynamic model and can automatically adjust its parameters according to the dynamic characteristics of the system and changes in the observed data. This adaptability makes the Kalman filter more effective in dealing with system fluctuations, thereby further improving the estimation accuracy of the quadratic linear extended state observer QLESO.

[0118] Kalman filtering mainly consists of two steps:

[0119] (1) Prediction step

[0120] The Kalman filter model is as follows:

[0121]

[0122] in, X is the system state matrix, H is the state transfer matrix, J is the control input matrix, U To control the amount, W is the process noise, Z is the observed quantity of the system state matrix,I is the state observation matrix, M To measure noise;

[0123] Set the current state estimate of the unknown nonlinear disturbance to:

[0124] (14)

[0125] (2) Update steps

[0126] The covariance of the updated estimate error is:

[0127] (15)

[0128] in, for H The transposed matrix of P is the estimated error covariance of the state at the previous moment, Q is the covariance of process noise;

[0129] Update the Kalman gain:

[0130] (16)

[0131] in, K g is the Kalman gain, for I The transposed matrix of R is the covariance of the observation noise;

[0132] k The optimal estimate of the unknown nonlinear perturbation at time is:

[0133] (17)

[0134] Substituting formula (14) into formula (17), we can obtain:

[0135] (18)

[0136] but k The error covariance at time is:

[0137] (19)

[0138] The quadratic linear extended state observer with Kalman filter is estimated Substituting into equation (8) in step S13, the discrete power super-local prediction model is obtained:

[0139] ;

[0140] Step S3, substitute the discrete power super-local prediction model formula (20) into the cost function (3) of the power loop designed in combination with the deadbeat control principle, calculate the optimal voltage vector of the active power and reactive power loops, and output the optimal voltage vector to the controller to realize model-free predictive power control:

[0141] The optimal voltage vector of the active power and reactive power loop obtained by the calculation is:

[0142] .

[0143] Embodiment 2

[0144] The professionals should further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented by electronic hardware, computer software or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0145] Specifically, each step of the method embodiment in the embodiment of the present application can be completed by the hardware integrated logic circuit and / or software instructions in the processor, and the steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor, or a combination of hardware and software modules in the decoding processor. Optionally, the software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps in the above method embodiment in combination with its hardware.

[0146] Embodiment 2 of the present invention provides a three-phase grid-connected inverter model-free predictive power control system, which may include: a memory and a processor, the memory is used to store a computer program and transmit the program code to the processor. In other words, the processor can call and run the computer program from the memory to implement the method of embodiment 1 of the present application. For example, the processor can be used to execute the above method embodiment according to the instructions in the computer program.

[0147] In some embodiments of the present application, the processor may include but is not limited to:

[0148] General-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc.

[0149] In some embodiments of the present application, the memory includes, but is not limited to, volatile memory and / or non-volatile memory. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus random access memory (Direct Rambus RAM, DR RAM).

[0150] In some embodiments of the present application, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the method provided by the present application. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the system.

[0151] Embodiment 3

[0152] Embodiment 3 of the present invention further provides a computer storage medium on which a computer program is stored. When the computer program is executed by a computer, the computer is enabled to execute the method of embodiment 1 above.

[0153] When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the first embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integration. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (digital video disc, DVD)), or a semiconductor medium (e.g., a solid state disk (solid state disk, SSD)), etc.

[0154] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A model-free predictive power control method for a three-phase grid-connected inverter, characterized in that The following steps are involved: Step S1: construct a super-local model to replace the original inverter model to reduce the dependence on the system model; specifically: Step S11: Establish a three-phase inverter Power model in coordinate system: in, Represents the active power of the grid, Represents the reactive power of the grid, and Respectively represent the inverter output voltage at The components in the coordinate system, and Respectively represent the grid voltage at Components in the coordinate system; L represents the filter inductance, R represents the equivalent series resistance in the power grid, Indicates the angular frequency of the power grid; Step S12: construct a power prediction model for a three-phase inverter: According to the forward Euler formula, the formula (1) is discretized to obtain the power prediction model of the three-phase inverter: in is the system sampling period, and They represent the inverter output voltage in the kth sampling period. Components in the coordinate system; and They represent the grid voltage in the kth sampling period. Components in the coordinate system; Represents the reactive power of the power grid at the moment of the kth sampling period; represents the active power of the power grid in the kth sampling period; Indicates the angular frequency of the power grid; R Represents the equivalent series resistance in the power grid; The cost function of the power loop designed in combination with the deadbeat control principle is: （3） in, is the reference value of active power, The active power is The value of the moment, is the reference value of reactive power, Reactive power is The value of the moment; when and hour, The minimum, optimal voltage vector size at time k is: in, and are the voltage vectors of active power and reactive power respectively, and are the optimal voltage vectors for active power and reactive power loops respectively; Step S13: Convert the power model into a super-local model: in, , is the input gain of each control loop, , are the lumped disturbances in the active power and reactive power models, respectively, which can be expressed as: The input gain is: （7） According to the forward Euler formula, equation (4) is discretized as follows: in, is the system sampling period; , Respectively represent k Sampling period ; , Respectively represent k +1 sampling period ; , Respectively represent k Sampling period ; , Respectively represent k Sampling period ; Step S2, designing a quadratic linear extended state observer QLESO, performing quadratic expansion estimation on the nonlinear disturbance residual through an integral negative feedback link, so as to estimate the lumped disturbance of the hyperlocal model; at the same time, introducing a Kalman filter in the integral negative feedback link, taking into account the process noise in the system dynamic model through the Kalman filter, and automatically adjusting its parameters according to the dynamic characteristics of the system and the changes in the observed data, so as to improve the estimation accuracy of the quadratic linear extended state observer QLESO; specifically: Step S21, design a quadratic linear extended state observer QLESO, and perform quadratic expansion estimation on the nonlinear disturbance residual through an integral negative feedback link to estimate the lumped disturbance of the hyperlocal model: The quadratic linear extended state observer QLESO is: in, is the state vector, is the state vector The observed value of is the state vector observation The derivative of is the gain vector of the system, is the power observation loop gain, is the perturbation observation loop gain, and are the output and input of each loop respectively, for Observed value of Lumped disturbance The compensation function, Lumped disturbance The estimated value of Compensates for the unknown model function in the original system ; make , then the quadratic linear extended state observer QLESO with integral form is: in, represents the filtering factor, is the control coefficient; According to the Lyapunov stability criterion, the characteristic equation of the quadratic linear extended state observer QLESO is expressed as: （11） in, s represents the variable in the Laplace transform domain, is the bandwidth of QLESO; In order to make the observer stable, the roots of the characteristic equation must all fall on the negative semi-axis, so , , ; Discretize equation (10) to get: Among them, only the control coefficient Different from the state variable x, in the active power loop, let , , used to substitute into the active power super-local model; in the reactive power loop, let , , used to substitute into the reactive power hyperlocal model; Step S22, design a quadratic linear state observer KF-QLESO containing Kalman filtering, which takes into account the process noise in the system dynamic model through the Kalman filter, and automatically adjusts its parameters according to the dynamic characteristics of the system and the changes in the observed data, so as to improve the estimation accuracy of the quadratic linear extended state observer QLESO: (1) Kalman filter prediction steps: The Kalman filter model is as follows: （13） in, X is the system state matrix, G is the state transfer matrix, J is the control input matrix, U To control the amount, W is the process noise, Z is the observed quantity of the system state matrix, I is the state observation matrix, M To measure noise; Set the current state estimate of the unknown nonlinear disturbance to: （14） (2) Kalman filter update steps: The covariance of the updated estimate error is: （15） in, for G The transposed matrix of P is the estimated error covariance of the state at the previous moment, Q is the covariance of process noise; Update the Kalman gain: （16） in, K g is the Kalman gain, for I The transposed matrix of R is the covariance of the observation noise; k The optimal estimate of the unknown nonlinear perturbation at time is: （17） Substituting formula (14) into formula (17), we can obtain: （18） but k The error covariance at time is: （19） The quadratic linear extended state observer with Kalman filter is estimated Substituting into equation (8) in step S13, the discrete power super-local prediction model is obtained: ； Step S3, substitute the discrete power super-local prediction model into the cost function of the power loop designed in combination with the deadbeat control principle, calculate the optimal voltage vector of the active power and reactive power loops, and output the optimal voltage vector to the controller to realize model-free predictive power control.

2. The model-free predictive power control method for a three-phase grid-connected inverter according to claim 1, characterized in that The optimal voltage vector of the active power and reactive power loop calculated in step S3 is: 。 3. A three-phase grid-connected inverter model-free predictive power control system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the model-free predictive power control method for the three-phase grid-connected inverter as claimed in claim 1 or 2 is implemented.

4. A computer storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the model-free predictive power control method for a three-phase grid-connected inverter as claimed in claim 1 or 2 is implemented.

Images