Method for suppressing high-voltage SVG active resonance under periodic impact load of submersible motor

Abstract

The application discloses a high-voltage SVG active resonance suppression method under periodic impact load of a submersible motor, and relates to the technical field of power electronic control. In view of the problems that the resonance spectrum drifts and diffuses due to the time-varying and impact load of cable distribution parameters of a deep-well submersible motor power supply system, and the existing fixed frequency point suppression method is invalid, the application only depends on the voltage and current measurement of a point of common coupling, adopts a moving horizon estimation to online identify the equivalent parameters and transmission delay of a cable, constructs a wide-frequency positive definite damping matrix which is updated adaptively based on a port-controlled Hamilton model, and decouples high-frequency electrical resonance and low-frequency electromechanical coupling disturbance through a nonlinear disturbance observer, so as to respectively apply damping suppression and voltage transient support, realize stable suppression of wide-frequency time-varying resonance, and can be used for resonance treatment of a marine and deep-well oil and gas exploitation power supply system.

Description

Technical Field

[0001] This invention relates to the field of power electronic control technology, specifically a method for suppressing active resonance of high-voltage SVG under periodic impact loads of submersible motors. Background Technology

[0002] In offshore platforms or onshore cluster well deep oil and gas extraction scenarios, submersible motors are the core lifting equipment. Their power supply system typically consists of a ground frequency converter, a static var generator (SVA), and a flexible cable of several kilometers connecting the ground and the downhole motor. The SVA collects voltage and current signals at the common junction point, calculates and injects compensation current in real time, and achieves dynamic reactive power compensation and resonance suppression functions. Due to the significant distributed resistance, inductance, and capacitance characteristics of ultra-long downhole cables, when gas-liquid two-phase flow or slug flow impact occurs downhole, the torque and current on the motor exhibit periodic abrupt changes. At the same time, the temperature and stress state of the cable body change drastically, causing the cable distributed parameters to exhibit strong nonlinear time-varying characteristics, which in turn causes the system resonant frequency to dynamically drift.

[0003] Currently, active resonance suppression of high-voltage SVG in the industry generally adopts the active damping method based on a local fixed impedance model. That is, during the commissioning stage, the equivalent impedance parameters of the cable are measured offline and a fixed model is established. The resonant components near the fixed resonant frequency are extracted by fast Fourier transform or notch filter algorithm. A proportional resonant controller or multi-resonant controller is designed to generate a reverse suppression current command at the fixed frequency point and inject it into the common coupling point to attenuate the resonant energy.

[0004] The above-mentioned solution can achieve a certain suppression effect under the condition that the cable parameters are approximately constant and the load impact spectrum is relatively concentrated. It is the closest technical solution to the present invention. However, it relies on an offline fixed model and fails to track the time-varying characteristics of cable parameters and signal transmission delay online. It also fails to effectively cover the resonant spectrum of continuous drift and multi-frequency dispersion, and fails to distinguish between high-frequency electrical resonance and low-frequency electromechanical coupling disturbance. There are safety hazards such as cable overvoltage caused by compensation phase lag, multi-frequency resonance runaway and shaft breakage caused by subsynchronous resonance of shaft system. There is an urgent need for a new method that can adaptively track time-varying parameters online and achieve broadband resonance stability suppression. Summary of the Invention

[0005] The technical problem this invention aims to solve is that existing high-voltage SVG active resonance suppression methods rely on offline fixed cable impedance models, making it difficult to track the time-varying characteristics of cable distribution parameters and signal transmission delay without relying on downhole sensing equipment. They also struggle to stably suppress broadband resonances with continuous drift and multi-frequency dispersion. Furthermore, they fail to differentiate between high-frequency electrical resonances and low-frequency electromechanical coupling disturbances, leading to safety hazards such as cable overvoltage breakdown caused by phase lag compensation, multi-frequency resonance runaway, and subsynchronous resonance shaft breakage induced by submersible motor shaft system.

[0006] To address the aforementioned technical problems, this invention provides a method for suppressing active resonance of a high-voltage SVG under periodic impact loads in a submersible motor. This method includes: Multi-scale frequency decomposition was performed on the three-phase current signal collected at the point of common coupling to obtain the high-frequency electrical resonance disturbance component and the low-frequency electromechanical coupling disturbance component. Based on the three-phase voltage and current signals collected from the common connection point, the equivalent distributed parameters and signal transmission delay of the downhole cable are identified online by recursively using the moving horizon estimation method. Based on the equivalent distributed parameters, the signal transmission delay, and the high-frequency electrical resonance disturbance components, the desired Hamiltonian function and the desired damping matrix of the port-controlled Hamiltonian system are constructed, the interconnection and damping distribution control law is solved, and the high-frequency damping compensation current command is generated. The low-frequency electromechanical coupling disturbance component is fed into the low-frequency voltage transient support compensation controller to generate a low-frequency compensation current command. The high-frequency damping compensation current command is superimposed with the low-frequency compensation current command to obtain a composite modulation command, which drives the high-voltage SVG converter to inject compensation current into the point of common coupling.

[0007] In the above technical solution, the moving field of view estimation module can identify the equivalent parameters of the cable and the transmission delay online based solely on the electrical measurement data of the ground common connection point, without the need to install any sensing equipment underground; The port-controlled Hamiltonian model and the passive control method of interconnection damping distribution virtually generate a wideband positive definite damping matrix that is adaptively updated with cable parameters in the control loop, so that the SVG port always maintains passive dissipation characteristics throughout the entire frequency band that may drift. The nonlinear disturbance observer separates the high-frequency electrical resonance from the low-frequency electromechanical coupling disturbance by setting two observer branches with different cutoff frequencies (high-frequency and low-frequency), and applies differentiated compensation to each, thus avoiding overlap between the control bandwidth and the mechanical torsional vibration frequency band. These three elements work together to form a complete closed-loop resonance suppression scheme.

[0008] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows: Firstly, by estimating the time-varying parameters and transmission delay of the cable online and correcting the expected Hamiltonian function accordingly, and actively compensating for phase lag, the risk of the compensation current being transformed into excitation current due to phase misalignment can be avoided, which is beneficial to preventing cable overvoltage breakdown. Secondly, by constructing a wideband positive definite damping matrix that adaptively updates with cable parameters, passive suppression of the entire possible drift frequency band is achieved without relying on locking a single fixed resonant frequency point, thus having a strong adaptability to the resonant spectrum dispersion phenomenon caused by impact load. Third, multi-scale decoupling of the resonant disturbance is achieved through two observer branches, high frequency and low frequency, and broadband damping suppression and voltage transient support compensation are applied respectively. This avoids the overlap between the control bandwidth and the inherent mechanical torsional vibration frequency band of the submersible motor shaft system, which helps to reduce the risk of inducing subsynchronous resonance and shaft breakage in the shaft system and improves the overall operational safety of the system. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the overall structure of the deep well power supply system for submersible motors and the high-voltage SVG active resonance suppression control architecture of the present invention; Figure 2 A flowchart illustrating the algorithm for estimating time-varying parameters and delayed moving horizon of a long cable; Figure 3 This is a block diagram illustrating the principle of a broadband passive damping reshaping algorithm based on a port-controlled Hamiltonian model. Figure 4 This is a schematic diagram of the electromechanical multi-scale frequency decoupling algorithm based on a nonlinear perturbation observer. Detailed Implementation

[0010] The following is in conjunction with the appendix Figure 1-4 The specific embodiments of the present invention will be further described below. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0011] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0012] The high-voltage SVG active resonance suppression method provided in this embodiment is applied to a deep well power supply system for a submersible motor. This system consists of a surface power grid, a surface frequency converter, a high-voltage SVG device, an ultra-long flexible cable running downhole, and a submersible motor load. The high-voltage SVG device is equipped with a voltage and current sampling unit that connects to a common junction point to acquire three-phase voltage and current signals. It also features a DSP and FPGA main control platform to implement the three core algorithm modules and signal collaborative processing logic of this invention. The three core algorithm modules are: a long cable time-varying parameter and delay-moving horizon estimation module; a wideband passive damping reshaping module based on a port-controlled Hamiltonian model; and a... The system includes an electromechanical multi-scale frequency decoupling module based on a nonlinear disturbance observer. The three modules coordinate with each other through signal flow. As a whole, the sampling signal at the common connection point is first sent to the electromechanical multi-scale frequency decoupling module for high- and low-frequency component separation. The separated high-frequency electrical resonance component and the identification result of the moving horizon estimation module are sent to the wideband damping reshaping module to generate high-frequency compensation commands. The separated low-frequency electromechanical coupling component is sent to the low-frequency voltage transient support compensation controller to generate low-frequency compensation commands. The two commands are superimposed in the command synthesis module to drive the SVG power converter. The compensation current is injected into the common connection point through the LCL filter to form a closed-loop control loop.

[0013] I. Estimation of Time-Varying Parameters and Delayed Moving Horizon of Long Cables This module uses only the three-phase voltage and current measurement data at the common connection point to recursively estimate the equivalent distributed parameters of the downhole cable and the signal transmission delay online, without requiring any downhole sensing equipment.

[0014] The downhole cable is modeled as a multi-level π-type distributed parameter model. Its equivalent resistance, equivalent inductance, equivalent capacitance, and signal transmission delay constitute the state vector to be estimated, denoted as: ; In the formula, For the first The state vector to be estimated at each sampling time. Indicates the first Equivalent resistance of the cable at each sampling time Indicates the equivalent inductance. Indicates the equivalent capacitance. This indicates the signal transmission delay.

[0015] Let the observation equation describe the dynamic relationship between the voltage and current at the point of common coupling and the reflected voltage on the load side: ; In the formula, Indicates the first Measured voltage at the point of common coupling at any given time. , These are the estimated values ​​of the cable's equivalent resistance and inductance at the current moment, respectively. The measured current at the point of common coupling, Indicates delay The estimated load-side voltage component is then reflected in the point of common coupling.

[0016] The moving horizon estimation algorithm obtains the current state estimate within the sliding time window by solving an optimization problem of the following cost function. ; In the formula, For the first The cost function value at time t. The length of the sliding time window. For the first Real-time measured output vector, The predicted output given by the observation equation is given in the above equation. This is the inverse of the measurement noise covariance matrix, used to weight the measurement residuals. To measure the noise covariance matrix, and in conjunction with the cable equivalent resistance mentioned above... The meanings are different and should not be confused. This is the state vector at the start of the window. This is the prior estimate of the state at the start of the window. This is the inverse of the prior estimation error covariance matrix, used to weight the prior bias terms.

[0017] In this embodiment, the specific implementation steps of the algorithm are as follows: Step 1: Data acquisition and coordinate transformation. The three-phase voltage and three-phase current at the common connection point are synchronously sampled at the set sampling frequency. After Clarke transformation and Park transformation, they are converted into DC component voltage and DC component current in a synchronous rotating coordinate system, which constitute the input and output observation data sequence of this module.

[0018] Step 2: Construct a state-space prediction model for the cable. Based on the π-type equivalent circuit of the cable distributed parameters and the transmission line telegraph equations, establish a state-vector prediction model. For the parameters to be estimated, use the observation equation To output the predicted discrete-time state-space model, the state transition process is set as a slowly varying random walk model to reflect the gradual change of cable parameters relative to the sampling period.

[0019] Step 3: Cache the sliding time window data and set the window length. The most recently cached memory is stored in the main control board memory. The input and output data within each sampling period constitute the window data set used for this estimation solution.

[0020] Step four: Construct and solve the cost function optimization problem. Substitute the window data set into the cost function expression, and use the iterative solver implemented by the FPGA hardware multiplier-accumulator to obtain the estimated values ​​of the cable equivalent parameters and delay at the current moment online.

[0021] Step 5: Window sliding update. Remove the earliest set of sampled data in the window, add the latest sampled data points, update the sliding window data set, and use the current estimation result as the prior estimate for the next estimation. Repeat steps 3 and 4 to achieve continuous online recursive estimation of cable parameters and delay.

[0022] Step 6: Output the estimation results. The estimated cable equivalent parameters and delay are output to the broadband damping reshaping module in real time. This module is used to update the energy parameters in the desired Hamiltonian function online and to actively compensate for the phase lag caused by the delay.

[0023] II. Wideband Passive Damping Reshaping Based on Port-Controlled Hamiltonian Model This module models the high-voltage SVG and its grid-connected LCL filter as a port-controlled Hamiltonian system. Through interconnection and damping allocation methods, a virtual wideband positive damping matrix is ​​generated in the control loop, so that the SVG port always exhibits strictly passive characteristics in the wideband domain. To avoid symbol confusion, the interconnection matrix, damping matrix, etc. involved in this section are all in matrix form, which have different meanings from the scalar notation of the cable equivalent parameters mentioned above.

[0024] Let the state equation of the open-loop port controlled Hamiltonian system consisting of a high-voltage SVG and an LCL filter be: ; In the formula, This is the system state vector, containing components of the filter inductor current and the filter capacitor voltage. Let be the total energy storage function of the system, i.e., the Hamiltonian function, which is expressed as a quadratic energy expression of the state vector. Let be the gradient of the Hamiltonian function with respect to the state vector. The interconnection matrix satisfies the antisymmetric property and describes the lossless energy exchange relationships within the system. Let be the system's inherent damping matrix, satisfying the positive semi-definite property, describing the system's inherent energy dissipation characteristics. For the input matrix, To control the input vector.

[0025] The desired closed-loop system retains the port-controlled Hamiltonian structure, but possesses a reshaped desired interconnect matrix, desired damping matrix, and desired Hamiltonian function, namely: ; In the formula, For the desired interconnect matrix, To determine the desired damping matrix, this embodiment designs it as a block diagonal matrix that is positive definite over a wide frequency band. The desired Hamiltonian function has its energy stiffness parameters adaptively updated based on the cable equivalent parameters estimated online by the moving horizon estimation module.

[0026] To solve the control law that satisfies the above desired dynamics, the interconnection and damping distribution matching equations are established: ; The analytical expression of the control law is obtained by solving: ; In the formula, Represents the input matrix The generalized inverse.

[0027] The passive property determination criterion for a closed-loop system is: ; That is, the difference between the desired damping matrix and the inherent damping matrix is ​​a positive semi-definite matrix. At this time, the closed-loop system satisfies strict passivity. According to the passivity theorem and Lyapunov stability theory, the entire parallel system has asymptotic stability. No matter in which frequency band the external resonance spectrum drifts and disperses, the SVG port exhibits energy dissipation characteristics in that frequency band, thereby suppressing resonance.

[0028] In this embodiment, the specific implementation steps of the algorithm are as follows: Step 1: Establish port-controlled Hamiltonian mathematical models of the SVG and LCL filters, and determine the system state vector. Specific components, Hamiltonian function Quadratic expression, interconnection matrix Inherent damping matrix and input matrix The specific structural form is determined to complete the energy modeling of the open-loop system.

[0029] Step 2: Construct an adaptive desired Hamiltonian function, mapping the cable equivalent parameters estimated online by the moving horizon estimation module to the desired Hamiltonian function. In the energy stiffness parameters, the desired energy function is made to achieve online adaptive updating as the cable parameters drift.

[0030] Step 3: Design the wideband desired damping matrix. Based on the resonant frequency range that may be caused by the impact load of the submersible motor, design the desired damping matrix. It is a positive definite block diagonal matrix within this frequency band, ensuring that the dissipation characteristics are met throughout the entire wideband range, without relying on locking a single fixed resonant frequency.

[0031] Step four: Solve the matching equations for interconnection and damping distribution. Substitute the matrices and functions determined in steps one through three into the matching equations. Using the analytical expressions derived offline in advance and the parameters updated online, calculate the control law in real time.

[0032] Step 5: Control law modulation output. The calculated control law is converted into voltage or current modulation commands for the SVG converter. These commands are then used by the space vector pulse width modulation module to drive the power switching devices, forming an actual virtual broadband damping compensation current and injecting it into the point of common coupling.

[0033] Step six: Passive online verification and damping adaptive adjustment. Verify in real time whether the passive condition is met. If the condition is violated due to a sudden change in cable parameters, trigger the damping matrix online adaptive adjustment step. Based on the currently estimated inherent damping matrix, redistribute the diagonal gain of the expected damping matrix to ensure that the passive condition is maintained during parameter drift.

[0034] III. Electromechanical Multiscale Frequency Decoupling Based on Nonlinear Perturbation Observer This module performs multi-scale frequency separation on the current signal at the common connection point, separating the low-frequency synchronous or interharmonic pulsation components caused by the mechanical impact of the slender shaft system of the submersible motor from the high-frequency electrical resonance components caused by the capacitive parameters and resonance of the cable, and sending them to different compensation control paths to avoid overlap between the control bandwidth and the mechanical torsional vibration frequency band.

[0035] Suppose that the dynamics of the current signal at the point of common coupling can be described as a general nonlinear system as follows: ; In the formula, It is the current state vector. Given the known nonlinear dynamics of the system, For input channel functions, To control the input, The unmodeled disturbance component to be estimated, in this embodiment, includes a high-frequency electrical resonance component and a low-frequency electromechanical coupling component.

[0036] The high-frequency nonlinear disturbance observer is constructed as follows: ; ; In the formula, These are auxiliary state variables within the high-frequency observer. This is the gain matrix for the high-frequency observer, and its corresponding cutoff frequency is set within the frequency range where electrical resonance may occur in the system containing the submersible generator. The output signal is the actual measured current at the point of common coupling. To estimate the obtained high-frequency electrical resonance disturbance components.

[0037] The low-frequency nonlinear disturbance observer adopts the same structure as the one above, except that the gain matrix is ​​replaced with the gain matrix corresponding to the low-frequency cutoff frequency. ,in The corresponding cutoff frequency is significantly lower than The corresponding cutoff frequency is used to match the inherent subsynchronous and interharmonic torsional vibration frequency band of the submersible motor shaft system. The estimated output is denoted as the low-frequency electromechanical coupling disturbance component. .

[0038] In this embodiment, the specific implementation steps of the algorithm are as follows: Step 1: Signal Acquisition and Preprocessing. Acquire the three-phase current signal at the point of common coupling, and obtain the current component in the synchronous rotating coordinate system through coordinate transformation, which serves as the input signal for the observer. .

[0039] Step 2: Construct a high-frequency nonlinear perturbation observer and set the observer gain matrix. This ensures that the corresponding observation bandwidth covers the expected high-frequency electrical resonant frequency band, and the high-frequency disturbance components are estimated in real time using the aforementioned high-frequency observer equations. .

[0040] Step 3: Construct a low-frequency nonlinear disturbance observer and set the observer gain matrix. This ensures that the corresponding observation bandwidth covers the inherent low-frequency electromechanical torsional vibration frequency band of the submersible motor shaft system. The low-frequency disturbance component is then recursively estimated in real time using the same observer structure as in step two. .

[0041] Step four: The high-frequency components are fed into the damping reshaping module, and the estimated high-frequency disturbance components are... As the input excitation of the aforementioned broadband damping reshaping module, the corresponding high-frequency damping compensation current command is generated through broadband damping reshaping calculation. .

[0042] Step 5: The low-frequency component is fed into the voltage transient support compensation module to estimate the low-frequency disturbance component. The data is fed into a low-frequency voltage transient support compensation controller based on a proportional-integral structure to generate low-frequency compensation commands to offset voltage drops or fluctuations caused by shaft impacts of the submersible motor. This pathway does not apply damping suppression to low-frequency electromechanical coupling components in order to avoid interfering with the inherent dynamics of the mechanical system.

[0043] Step 6: Compensation command synthesis and output, which involves synthesizing and outputting the high-frequency damping compensation current command. Low-frequency voltage support compensation command The signals are superimposed to obtain the final composite modulation command of the SVG converter, which drives the power switching devices to achieve coordinated and frequency-band suppression of wideband electrical resonance and low-frequency electromechanical disturbances.

[0044] IV. Description of Overall Implementation Examples In a specific application scenario, the downhole cable of a deep well submersible motor power supply system is on the order of several kilometers in length. Under normal operating conditions, the equivalent parameters of the cable are approximately constant. However, under the impact condition of gas-liquid two-phase flow, the cable temperature and stress state change significantly, and the system resonant frequency continuously drifts within a certain bandwidth. After adopting the method described in this invention, the long cable time-varying parameter and delay-shifting horizon estimation module continuously updates the estimated values ​​of the cable's equivalent resistance, inductance, capacitance, and signal transmission delay online at a set sampling frequency. The broadband damping reshaping module continuously updates the desired Hamiltonian function and the desired damping matrix based on the above estimates, so that the SVG port always maintains positive definite damping characteristics within the drift bandwidth. The electromechanical multi-scale frequency decoupling module separates the high-frequency electrical resonant component from the low-frequency electromechanical coupling component in the common junction current. The high-frequency component is reshaped by damping to generate a compensation current command, and the low-frequency component is reshaped by the voltage transient support controller to generate a compensation current command. The two commands are superimposed to drive the SVG converter. Compared with the existing scheme that relies on an offline fixed impedance model and locks a single resonant frequency, the method described in this embodiment can continuously track the phase lag and actively compensate for it during the continuous changes in cable parameters and load impacts. It maintains a stable suppression capability for the wideband drift resonant spectrum, while avoiding inappropriate interference with the inherent low-frequency electromechanical torsional vibration of the submersible motor shaft system.

[0045] V. Other Alternative Implementation Methods In another embodiment of the present invention, the online estimation of time-varying parameters and delay of long cables can be replaced by extended Kalman filtering or unscented Kalman filtering algorithms instead of moving horizon estimation algorithms. The recursive filtering structure is used to perform online state estimation of cable equivalent parameters and signal delay. This method has relatively lower computational complexity and is suitable for field application scenarios where the computing power of the main control chip is limited.

[0046] In another embodiment of the present invention, the broadband passive damping reshaping stage can be replaced by a broadband robust damping injection method based on sliding mode variable structure control instead of the passive control method for interconnection and damping distribution based on the port-controlled Hamiltonian model. By designing the sliding surface and switching control law, similar broadband dissipation characteristics are achieved on the sliding surface, so that the system state is constrained to the vicinity of the sliding surface and resonance divergence is suppressed. This method has strong robustness to model uncertainty.

[0047] Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of the present invention. Therefore, any modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for suppressing active resonance of high-voltage SVG under periodic impact load of submersible motor, characterized in that, include: Multi-scale frequency decomposition was performed on the three-phase current signal collected at the point of common coupling to obtain the high-frequency electrical resonance disturbance component and the low-frequency electromechanical coupling disturbance component. Based on the three-phase voltage and three-phase current signals collected from the common connection point, the equivalent distributed parameters and signal transmission delay of the downhole cable are identified online by recursively using the moving horizon estimation method. Based on the equivalent distributed parameters, the signal transmission delay, and the high-frequency electrical resonance disturbance components, the desired Hamiltonian function and the desired damping matrix of the port-controlled Hamiltonian system are constructed, the interconnection and damping distribution control law is solved, and the high-frequency damping compensation current command is generated. The low-frequency electromechanical coupling disturbance component is fed into the low-frequency voltage transient support compensation controller to generate a low-frequency compensation current command. The high-frequency damping compensation current command is superimposed with the low-frequency compensation current command to obtain a composite modulation command, which drives the high-voltage SVG converter to inject compensation current into the point of common coupling.

2. The method according to claim 1, characterized in that, The steps for online recursive identification of the equivalent distributed parameters and signal transmission delay of the downhole cable using the moving horizon estimation method include: The downhole cable is equivalent to a multi-level π-type distributed parameter model, and a state vector to be estimated is set. The state vector to be estimated includes the cable's equivalent resistance, equivalent inductance, equivalent capacitance, and signal transmission delay. Establish observation equations that describe the relationship between the voltage and current at the common connection point and the reflected voltage on the load side. Within the sliding time window, a cost function is constructed based on the deviation between the measured output within the window and the predicted output given by the observation equation, as well as the deviation between the state at the beginning of the window and its prior estimate. By solving the optimization problem of the cost function, the state estimate at the current time is obtained. The earliest set of sampled data in the window is removed, the latest sampled data is added, the sliding time window is updated, and the state estimate at the current moment is used as the prior estimate of the optimization problem at the next moment, thus realizing recursive estimation.

3. The method according to claim 2, characterized in that, The state transition process of the state vector is set as a slowly varying random walk model to characterize the gradual variation of the cable's equivalent distribution parameters with respect to the sampling period.

4. The method according to claim 1, characterized in that, The steps for constructing the desired Hamiltonian function and desired damping matrix of a port-controlled Hamiltonian system, and solving the interconnection and damping distribution control law include: State equations are established for an open-loop port-controlled Hamiltonian system consisting of a high-voltage SVG and its grid-connected LCL filter. The state equations consist of an interconnect matrix, an inherent damping matrix, a Hamiltonian function gradient, and the product of the input matrix and the control input. The energy stiffness parameter of the desired Hamiltonian function is updated online based on the equivalent distribution parameters, and the desired damping matrix is ​​set as a positive definite block diagonal matrix within the frequency band according to the resonant frequency band range corresponding to the submersible motor impact load. Establish the matching equations for interconnection and damping distribution, and solve them to obtain the analytical expression of the control law; The high-frequency damping compensation current command is generated according to the control law.

5. The method according to claim 4, characterized in that, Also includes: The system verifies online whether the difference between the desired damping matrix and the inherent damping matrix is ​​a positive semi-definite matrix. If the verification result shows that the positive semi-definite condition is not met, the diagonal gain of the desired damping matrix is ​​reallocated according to the currently identified inherent damping matrix to maintain the passive dissipation characteristics of the system.

6. The method according to claim 1, characterized in that, The steps for multi-scale frequency decomposition of the three-phase current signal acquired at the point of common coupling include: A high-frequency nonlinear disturbance observer is constructed, and the observation bandwidth corresponding to its observer gain matrix is ​​set to cover the expected high-frequency electrical resonance frequency band, and the high-frequency electrical resonance disturbance component is output. A low-frequency nonlinear disturbance observer is constructed, and the observation bandwidth corresponding to its observer gain matrix is ​​set to cover the inherent low-frequency electromechanical torsional vibration frequency band of the submersible motor shaft system. The cutoff frequency corresponding to the low-frequency nonlinear disturbance observer is lower than the cutoff frequency corresponding to the high-frequency nonlinear disturbance observer, and the low-frequency electromechanical coupling disturbance component is output.

7. The method according to claim 1, characterized in that, The low-frequency voltage transient support compensation controller adopts a proportional-integral structure to offset the voltage drop and fluctuation caused by the impact of the submersible motor shaft system, and does not apply damping suppression to the low-frequency electromechanical coupling disturbance component.

8. The method according to claim 1, characterized in that, Before performing multi-scale frequency decomposition on the three-phase current signal acquired at the point of common coupling, the method further includes performing Clarke transform and Park transform on the three-phase voltage signal and three-phase current signal acquired at the point of common coupling to obtain the DC component voltage and DC component current in the synchronous rotating coordinate system.

9. A high-voltage SVG active resonance suppression system under periodic impact load of a submersible motor, characterized in that, include: The voltage and current sampling unit is used to acquire the three-phase voltage and three-phase current signals at the point of common coupling. A multi-scale frequency decoupling module, connected to the voltage and current sampling unit, is used to perform multi-scale frequency decomposition on the three-phase current signal to obtain high-frequency electrical resonance disturbance components and low-frequency electromechanical coupling disturbance components. The long cable time-varying parameter and delay identification module is connected to the voltage and current sampling unit and is used to identify the equivalent distributed parameters and signal transmission delay of the downhole cable online by using the moving horizon estimation method based on the three-phase voltage signal and the three-phase current signal. A wideband damping reshaping module is connected to the long cable time-varying parameter and delay identification module and the multi-scale frequency decoupling module. It is used to construct the desired Hamiltonian function and desired damping matrix of the port-controlled Hamiltonian system based on the equivalent distributed parameters, the signal transmission delay and the high-frequency electrical resonance disturbance component, and generate a high-frequency damping compensation current command. A low-frequency voltage transient support and compensation controller is connected to the multi-scale frequency decoupling module and is used to generate a low-frequency compensation current command based on the low-frequency electromechanical coupling disturbance component. The instruction synthesis module, connected to the wideband damping reshaping module and the low-frequency voltage transient support compensation controller, is used to superimpose the high-frequency damping compensation current instruction and the low-frequency compensation current instruction to obtain a composite modulation instruction; A high-voltage SVG power converter, connected to the instruction synthesis module, is used to inject compensation current into the point of common coupling according to the composite modulation instruction.

Images