A method for VSR current estimation based on unscented Kalman filter

Abstract

This invention discloses a method for VSR current estimation based on an unscented Kalman filter, comprising: establishing a state model of a three-phase voltage-source PWM rectifier; and estimating the input current of the three-phase voltage-source PWM rectifier using an unscented Kalman filter. This invention constructs a state model of the three-phase voltage-source PWM rectifier and uses an unscented Kalman filter to predict the input current of the three-phase voltage-source PWM rectifier. Based on the strong coupling and nonlinearity characteristics of the three-phase voltage-source PWM rectifier, this invention couples unscented Kalman filters together, and then uses the unscented Kalman filter to predict the input current of the three-phase voltage-source PWM rectifier. By leveraging the nonlinear characteristics of the three-phase voltage-source PWM rectifier, this invention achieves accurate measurement of nonlinear systems.

Description

Technical Field

[0001] This invention relates to the field of three-phase voltage-source rectifier controllers, and in particular to a method for VSR current estimation based on an unscented Kalman filter. Background Technology

[0002] Unscented Kalman filters, as a mathematical tool, approximate the probability density distribution of nonlinear functions by using a series of deterministic samples to approximate the posterior probability density of the state. Traditional Kalman filters perform Taylor expansions on the nonlinear system equations or observation equations and retain first-order approximation terms, which easily introduces linearization errors, resulting in extremely poor filtering performance for nonlinear systems.

[0003] Therefore, overcoming the shortcomings of the existing technology is an urgent problem to be solved in this technical field. Summary of the Invention

[0004] The technical problem this invention aims to solve is how to address the issue that traditional unscented Kalman filters, when performing Taylor expansions on nonlinear system equations or observation equations and retaining first-order approximation terms, result in extremely poor filtering performance for nonlinear systems, leading to significant errors in the measurement of nonlinear systems.

[0005] This invention provides a method for VSR current estimation based on an unscented Kalman filter, comprising:

[0006] Establish a state model for a three-phase voltage-source PWM rectifier;

[0007] The input current of a three-phase voltage-source PWM rectifier is estimated using an unscented Kalman filter.

[0008] Preferably, establishing the state model of the three-phase voltage-source PWM rectifier specifically includes:

[0009] Based on the decoupling mathematical model of the three-phase PWM rectifier and the discretization of the continuous state equation, the mathematical ideal model of the rectifier in the discrete time domain is obtained.

[0010] The mathematical ideal model of the discrete-time rectifier is analyzed, and the state model of the three-phase voltage-source PWM rectifier is established.

[0011] Preferably, the state model formula for the three-phase voltage-source PWM rectifier is:

[0012] x(k+1)=Gx(k)+Hu(k)+We g (k)+ω(k)

[0013] in, x=(i a ib i c ) T e g =(e a e b e c ) T , u = (u 1a u 1b u 1c ) T e a e b e c The three-phase input grid voltage; i a i b i c V is the three-phase input grid current; o R is the DC output voltage of the rectifier; L is the grid-side filter inductance; R is the grid-side equivalent resistance. u a Let u be the duty cycle of rectifier a. b U is the duty cycle of rectifier b. c The duty cycle of rectifier c is , ranging from 0 to 1; T s Let x(K) be the control cycle of the control system, x(K) be the time, and x(K+1) be the state at the next time.

[0014] Preferably, it also includes acquiring observed values ​​of the three-phase input current, specifically including:

[0015] Based on the correspondence between voltage and current in the same phase of the power grid, an ideal output equation between power grid voltage and current is constructed;

[0016] Increase the compensation amount of the ideal output equation to obtain the observed values ​​of the three-phase input current based on the voltage values ​​within the unscented Kalman filter.

[0017] Preferably, the formula for the observed value of the three-phase input current is:

[0018] y(k+1)=C(k+1)x(k+1)+v(k+1)

[0019] Where, x(k)=(i a (k) i b (k) i c (k)) T y(k+1)=(e a (k+1) e b (k+1) e c (k+1)) T Q = ω(k) is the process noise of the covariance matrix, and R = v(k+1) is the observation noise of the covariance matrix.

[0020] Preferably, the estimation of the input current of the three-phase voltage-source PWM rectifier using an unscented Kalman filter specifically includes:

[0021] Based on the mean of the original state distribution and the dimension of the state variables in the three-phase PWM rectifier, obtain the current value, the weight corresponding to the sampling point, the mean of the sampling point, and the covariance of the sampling point for each sampling point in the sampling point set;

[0022] Calculate the one-step prediction value of the sample point set, and obtain the one-step prediction and covariance matrix of the system state variables by weighted summation;

[0023] Substitute the predicted value into the formula for the observed value to obtain the predicted observed value, and then obtain the mean and covariance of the system prediction by weighting based on the predicted observed value.

[0024] Calculate the Kalman gain and update the system state to output the input current of the three-phase voltage-source PWM rectifier based on the updated system state.

[0025] Preferably, after updating the system state, the method further includes updating the covariance formula and obtaining the sampling current of each sampling point in the next round based on the covariance formula updated in the previous round, specifically including:

[0026] Update the covariance formula and calculate the matrix equation value for the next round of sampling based on the covariance formula of the previous round.

[0027] When calculating the sampling current of the sampling points in the next round, the matrix equation value of the sampling in the next round is used for calculation to obtain the sampling current of each sampling point in the next round.

[0028] Preferably, the step of obtaining the current value, the weight corresponding to the sampling point, the mean of the sampling point, and the covariance of the sampling point within the sampling point set based on the mean of the original state distribution and the dimension of the state variables in the three-phase PWM rectifier specifically includes:

[0029] Based on the circuit schematic of the three-phase PWM rectifier, the load and voltage under the original state distribution are obtained, and the sampling current of the first sampling point in the sampling point set is calculated according to the ideal output equation between the grid voltage and current.

[0030] Obtain the dimension n of the three-phase PWM rectifier, calculate the total number of sampling points 2n+1 based on the dimension of the three-phase PWM rectifier, and calculate the sampling current value corresponding to each sampling point according to the current sampling formula of the sampling point.

[0031] The mean and covariance of each sampling point are calculated using the formulas for mean and covariance.

[0032] Preferably, the current sampling formula is:

[0033]

[0034] The formulas for the mean and covariance are as follows:

[0035]

[0036] Where n = 3, This is the sample mean, which is the sampled current at the first sampling point. Let be the i-th column of the matrix equation, and λ be the scaling factor to reduce the overall expected error. The mean, Let λ be the covariance, α be a constant, β be a constant, and P(0) = e.

[0037] Preferably, the system status is updated as follows:

[0038]

[0039] The covariance is updated as follows:

[0040]

[0041] This invention constructs a state model of a three-phase voltage-source PWM rectifier and uses an unscented Kalman filter to predict the input current of the three-phase voltage-source PWM rectifier. Based on the strong coupling and nonlinear characteristics of the three-phase voltage-source PWM rectifier, this invention couples unscented Kalman filters together, and then uses the unscented Kalman filters to predict the input current of the three-phase voltage-source PWM rectifier. By leveraging the nonlinear characteristics of the three-phase voltage-source PWM rectifier, accurate measurement of nonlinear systems can be achieved. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0043] Figure 1 This is a flowchart of a method for VSR current estimation based on an unscented Kalman filter provided in an embodiment of the present invention;

[0044] Figure 2 This is a flowchart of a method for establishing a state model of a three-phase voltage-source PWM rectifier based on a VSR current estimation method using an unscented Kalman filter, provided in an embodiment of the present invention.

[0045] Figure 3 This is a circuit schematic diagram of a three-phase PWM rectifier provided in an embodiment of the present invention;

[0046] Figure 4 This is a flowchart of a method for obtaining observed values ​​of three-phase input current within a VSR current estimation method based on an unscented Kalman filter, provided in an embodiment of the present invention.

[0047] Figure 5 This is a detailed flowchart of step 202 in a method for VSR current estimation based on an unscented Kalman filter provided in an embodiment of the present invention.

[0048] Figure 6 This is a flowchart of a method for obtaining the sampling current of the next round of sampling points provided in an embodiment of the present invention;

[0049] Figure 7 This is a detailed flowchart of step 501 within a method for VSR current estimation based on an unscented Kalman filter provided in an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0051] In the description of this invention, the terms "inner", "outer", "longitudinal", "lateral", "upper", "lower", "top", "bottom", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and do not require that this invention must be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0052] In this invention, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0053] In this application, unless otherwise expressly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral part; it can be a direct connection or an indirect connection through an intermediate medium. Furthermore, the term "coupled" can refer to an electrical connection that enables signal transmission.

[0054] 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.

[0055] Example 1:

[0056] Embodiment 1 of this invention provides a method for VSR current estimation based on an unscented Kalman filter, such as... Figure 1 As shown, it includes:

[0057] Step 201: Establish the state model of the three-phase voltage-source PWM rectifier.

[0058] The three-phase PWM rectifiers in this invention are divided into two types: voltage-source (VSR) and current-source (CRS). The current-source rectifier contains a large inductor, and its impedance component leads to significant energy loss. The inductance and capacitance on the AC side also cause current distortion and oscillation. The voltage-source rectifier is widely used due to its high energy storage efficiency, small size, and simple structure. The control strategies for the three-phase PWM rectifier are divided into direct control and indirect control. Direct current control uses the input current as both feedback and controlled variables, offering advantages such as fast dynamic response and high control accuracy. However, it requires three current sensors, increasing application cost and product size. Indirect current control is based on the controller's low-frequency steady-state mathematical model and indirectly controls the input voltage through amplitude and phase, offering advantages such as small size and low cost. Because three-phase voltage-source PWM rectifiers have strong coupling and nonlinear characteristics...

[0059] Step 202: Estimate the input current of the three-phase voltage-source PWM rectifier using an unscented Kalman filter.

[0060] In this embodiment of the invention, an unscented Kalman filter is used to predict the input current of a three-phase voltage-source PWM rectifier. Based on the strong coupling and nonlinear characteristics of the three-phase voltage-source PWM rectifier, this embodiment couples unscented Kalman filters together, and then uses these filters to predict the input current of the rectifier. By leveraging the nonlinear characteristics of the three-phase voltage-source PWM rectifier, accurate measurement of nonlinear systems can be achieved, avoiding the large measurement errors that can occur when simply using an unscented Kalman filter to measure the input current.

[0061] The following is a detailed explanation of the embodiments of the present invention. Specifically, the establishment of the state model of the three-phase voltage-source PWM rectifier is as follows: Figure 2 As shown, it specifically includes:

[0062] Step 301: Based on the decoupling mathematical model of the three-phase PWM rectifier, and by discretizing the continuous state equation, the mathematical ideal model of the rectifier in the discrete time domain is obtained.

[0063] like Figure 3 The diagram shown illustrates the circuit schematic of a three-phase PWM rectifier. It is worth noting that Embodiment 3 of this invention uses a conventional circuit diagram of a conventional three-phase PWM rectifier. Based on Figure 3 Based on the schematic diagram, a decoupling mathematical ideal model of a three-phase PWM rectifier is constructed. The formula for the decoupling mathematical ideal model of the three-phase voltage-source PWM rectifier is as follows:

[0064] x(k+1)=Gx(k)+Hu(k)+We g (k)

[0065] in, x=(i a i b i c ) T e g =(e a e b e c ) T , u = (u 1a u 1b u 1c ) T e a e b e c The three-phase input grid voltage; i a i b i c V is the three-phase input grid current; o R is the DC output voltage of the rectifier; L is the grid-side filter inductance; R is the grid-side equivalent resistance. u a Let u be the duty cycle of rectifier a. b U is the duty cycle of rectifier b. c The duty cycle of rectifier c is , ranging from 0 to 1; T s Let x(K) be the control cycle of the control system, x(K) be the time, and x(K+1) be the state at the next time.

[0066] Step 302: Analyze the mathematical ideal model of the discrete-time rectifier and establish the state model of the three-phase voltage-type PWM rectifier.

[0067] In this embodiment of the invention, the three-phase voltage-source PWM rectifier employs an indirect control strategy, enabling the embodiment to operate without current sensors, thus eliminating the need for at least three current sensors (see [reference]). Figure 3As shown in the diagram, theoretically, each current would correspond to a current sensor, making the measurement model of this embodiment smaller. Based on the above analysis of the mathematical ideal model of the discrete-time rectifier in this embodiment, a continuous-state model of the three-phase voltage-source PWM rectifier is constructed. The corresponding formula for the continuous-state model of the three-phase voltage-source PWM rectifier is:

[0068] x(k+1)=Gx(k)+Hu(k)+We g (k)+ω(k)

[0069] in, x=(i a i b i c ) T e g =(e a e b e c ) T , u = (u 1a u 1b u 1c ) T e a e b e c The three-phase input grid voltage; i a i b i c V is the three-phase input grid current; o R is the DC output voltage of the rectifier; L is the grid-side filter inductance; R is the grid-side equivalent resistance. u a Let u be the duty cycle of rectifier a. b U is the duty cycle of rectifier b. c The duty cycle of rectifier c is , ranging from 0 to 1; T s Let x(K) represent the control cycle of the control system, x(K) be the state at time k, and x(K+1) be the state at the next time. In the calculation, K represents the current state time, in units of 1 (a positive integer). In actual calculations, the sampling time is relatively short, possibly measured in milliseconds (ms). For example, if the sampling period is 100 ms, then Ts is 100 ms, and Ts is the actual sampling period of the controller.

[0070] Furthermore, x(k)=(i a (k) i b (k) i c (k)) T y(k+1)=(e a (k+1) e b(k+1) e c (k+1)) T ω(k) represents the process noise of the continuous state model of the three-phase voltage-source PWM rectifier.

[0071] The VSR current estimation method based on an unscented Kalman filter in this invention embodiment further includes acquiring observed values ​​of the three-phase input current, such as... Figure 4 As shown, it specifically includes:

[0072] Step 401: Based on the correspondence between voltage and current in the same phase of the power grid, construct the ideal output equation between the power grid voltage and current.

[0073] In this embodiment of the invention, the observed value of the single-phase input current is obtained by using the measured value of the grid voltage through an unscented Kalman filter. Since the grid voltage and the output current are in phase, the instantaneous values ​​of the voltage and current in the same phase are proportional. An ideal output equation is constructed: y(k)=C(k)x(k), where C(k) is the parameter value of the output equation, y(k) is the voltage value, x(k) is the current value, and k is a certain moment.

[0074] Step 402: Increase the compensation amount of the ideal output equation to obtain the observed values ​​of the three-phase input current based on the voltage values ​​within the unscented Kalman filter.

[0075] The ideal output equation of this invention is only applicable to the calculation of voltage output under ideal conditions. To make this invention applicable to actual measurements, a compensation amount is added to the output equation based on the ideal output equation. The observed input current of the three-phase PWM rectifier is obtained through the voltage value within the unscented Kalman filter. The formula for the observed three-phase input current is:

[0076] y(k+1)=C(k+1)x(k+1)+v(k+1)

[0077] Where, x(k)=(i a (k) i b (k) i c (k)) T y(k+1)=(e a (k+1) e b (k+1) e c (k+1)) T Q = ω(k) is the process noise of the covariance matrix, and R = v(k+1) is the observation noise of the covariance matrix.

[0078] The following is a detailed explanation of step 202, which involves estimating the input current of the three-phase voltage-source PWM rectifier using an unscented Kalman filter, as follows: Figure 5 As shown, it specifically includes:

[0079] Step 501: Based on the mean of the original state distribution and the dimension of the state variables within the three-phase PWM rectifier, obtain the current value, the weight corresponding to the sampling point, the mean of the sampling point, and the covariance of the sampling point for each sampling point in the sampling point set.

[0080] In this embodiment of the invention, an unscented Kalman filter is used to measure a three-phase voltage-type PWM rectifier. Therefore, the dimension of the state variables in the three-phase PWM rectifier of this embodiment is 3, i.e., n = 3, and 2n + 1 sampling points are obtained. The sample current sampling value corresponding to the specific sampling point is:

[0081] x (0) x (1) x (2) x (3) x (4) x (5) and x (6) In this embodiment of the invention, the sampling point at i=0 is defined as the sample mean point, that is: The current sampling formula corresponding to the embodiment of the present invention is as follows:

[0082]

[0083] The formulas for the mean and covariance are as follows:

[0084]

[0085] Where n = 3, This is the sample mean, which is the sampled current at the first sampling point. Let be the i-th column of the matrix equation, and λ be the scaling factor to reduce the overall expected error. The mean, Let λ be the covariance, α be a constant, β be a constant, and P(0) = e.

[0086] Step 502: Calculate the one-step prediction value of the sampling point set, and obtain the one-step prediction and covariance matrix of the system state variables by weighted summation.

[0087] To better understand the meaning of a one-step prediction value, let's take a specific example. Suppose that I want to calculate the process from x1 to x3, which requires passing through the x2 step, i.e., x1->x2->x3. The process of calculating x2 is a one-step prediction value, which can be understood as an intermediate process value.

[0088] In this embodiment of the invention, the sample point set is predicted using a formula for a one-step prediction value. The specific formula for the one-step prediction value is as follows:

[0089]

[0090] Where i = 0, 1, 2, 3, 4, 5, and 6, k represents a certain time, k+1 represents the next time, and k+1|k represents a one-step prediction of the k+1 state based on the k state. (i) (k+1|k) represents the one-step prediction obtained by performing a one-step prediction on a set of 7 points (i = 0, 1, 2, 3, 4, 5, and 6) in state k. By utilizing the one-step predictions of the collected point sets and performing a weighted summation, the one-step predictions and covariance matrix of the system state variables are obtained. The corresponding system state variable equation is:

[0091]

[0092] The corresponding covariance matrix equation is:

[0093]

[0094] Where Q = ω(k).

[0095] Step 503: Substitute the predicted value into the formula for the observed value to obtain the predicted observed value, and obtain the mean and covariance of the system prediction by weighting based on the predicted observed value.

[0096] The predicted values ​​are substituted into the formula for the three-phase input current observations to obtain the predicted observations. Based on the predicted observations, the mean and covariance of the system predictions are obtained through a weighted average. The corresponding formula for obtaining the predicted observations is as follows:

[0097] y (i) (k+1|k)=C (i) (k)x (i) (k+1|k)

[0098] The formula for obtaining the weighted mean of the system predictions is:

[0099]

[0100] The formula for covariance is:

[0101]

[0102]

[0103] Where R = v(k+1) is the observation noise of the covariance matrix, and T represents the transpose of the matrix.

[0104] Step 504: Calculate the Kalman gain and update the system state to output the input current of the three-phase voltage-source PWM rectifier based on the updated system state.

[0105] The formula for calculating the Kalman gain in this embodiment of the invention is as follows:

[0106]

[0107] The updated system state output input current equation of the three-phase voltage-type PWM rectifier in this embodiment of the invention. The formula for the input current equation has already been listed above and will not be repeated here. It is worth noting that the system state variables in this embodiment of the invention have three values: the three-phase current values. The system state is a 1x3 matrix composed of the above three state variables.

[0108] In this embodiment of the invention, during the current sampling process at 2n+1 sampling points, the matrix equation in the sampling formula needs to be updated to obtain the matrix equation under different states. Then, the sampling current at the 7 points under that state is obtained through the matrix equation in the i-th column. It is worth noting that, in this embodiment of the invention, time k and time k+1 can be understood as times under two different states. Based on this, the VSR current estimation method based on an unscented Kalman filter in this embodiment of the invention, after updating the system state, also includes updating the covariance formula and obtaining the sampling current at each sampling point in the next round based on the covariance formula updated in the previous round, such as... Figure 6 As shown, it specifically includes:

[0109] Step 601: Update the covariance formula and calculate the matrix equation value for the next round of sampling based on the covariance formula of the previous round.

[0110] The mean value predicted by the system is substituted into the covariance formula to obtain the covariance equation for the next round of states. The corresponding updated covariance equation formula is as follows:

[0111] Step 602: When calculating the sampling current of the sampling points in the next round, use the matrix equation value of the sampling in the next round to calculate and obtain the sampling current of each sampling point in the next round.

[0112] This invention provides an embodiment of the invention that obtains the expression of the matrix equation under different states by updating the covariance formula equation. The matrix equation under state k is P(K), and the matrix equation under state k+1 is P(K+1). This embodiment starts from 0 and obtains the corresponding matrix equation P(0). Through continuous updates, it obtains the matrix equation under the current state, and then obtains the sampling current of the corresponding sampling point using the current formula for the sampling point. For example, when it is necessary to obtain the sampling current of the sampling point under state k+1, the matrix equation under state k+1 is updated using the matrix equation under state k, and the i-th column of the matrix equation under state k+1 is calculated. Then, the i-th column of the matrix equation is substituted into the current formula for the sampling current to calculate the sampling current of each sampling point under state k+1.

[0113] Step 501 of this embodiment of the invention is further refined. This embodiment of the invention obtains the current value, the corresponding weight, the mean of the sampling points, and the covariance of the sampling points within the sampling point set based on the mean of the original state distribution and the dimension of the state variables within the three-phase PWM rectifier. Figure 7 As shown, it specifically includes:

[0114] Step 701: Based on the circuit schematic of the three-phase PWM rectifier, obtain the load and voltage under the original state distribution, and calculate the sampling current x of the first sampling point in the sampling point set according to the ideal output equation between the grid voltage and current.

[0115] like Figure 3 The diagram shows the circuit schematic of a three-phase PWM rectifier, which obtains the load and voltage under the original state distribution. The load and voltage under the original state distribution in this embodiment represent empirical values ​​of the three-phase PWM rectifier, obtained through sampling of the three-phase PWM rectifier.

[0116] Step 702: Obtain the dimension of the three-phase PWM rectifier, calculate the total number of sampling points 2n+1 based on the dimension of the three-phase PWM rectifier, and calculate the sampling current value corresponding to each sampling point according to the current sampling formula of the sampling point.

[0117] This embodiment of the invention employs a three-phase PWM rectifier, therefore its dimension n = 3. Based on the dimension of the three-phase PWM rectifier, the total number of sampling points is calculated to be 2n + 1 = 2 × 3 + 1 = 7. The sampling current value corresponding to each sampling point is then calculated using the corresponding current sampling formula. The sampling formula used for the sampling points in this embodiment of the invention has already been presented above and will not be repeated here.

[0118] Step 703: Calculate the mean and covariance of each sampling point according to the formulas for mean and covariance.

[0119] Once the sampling point formula for the corresponding state is obtained, the mean and covariance of each sampling point in that state are obtained according to the mean and covariance formulas. The formulas for the mean and covariance of each sampling point have been explained above and will not be repeated here.

[0120] This invention employs an unscented Kalman filter to predict the input current of a three-phase voltage-source PWM rectifier. Based on the strong coupling and nonlinear characteristics of three-phase voltage-source PWM rectifiers, this invention couples unscented Kalman filters together, thereby predicting the input current. By leveraging the nonlinear characteristics of the three-phase voltage-source PWM rectifier, accurate measurement of nonlinear systems can be achieved, avoiding the large measurement errors that can occur when using only an unscented Kalman filter for input current measurement. Furthermore, this invention uses an indirect control method, calculating the corresponding current value through a voltage sensor. This reduces the need for at least three current sensors found in traditional three-phase voltage-source PWM rectifiers, resulting in a smaller and simpler structure.

[0121] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for VSR current estimation based on an unscented Kalman filter, characterized in that, include: Establish a state model for a three-phase voltage-source PWM rectifier; including: based on the decoupled mathematical model of the three-phase PWM rectifier, and discretizing the continuous state equations to obtain the mathematical ideal model of the rectifier in the discrete time domain; analyze the mathematical ideal model of the rectifier in the discrete time domain, and establish a state model for the three-phase voltage-source PWM rectifier. Estimating the input current of a three-phase voltage-source PWM rectifier using an unscented Kalman filter; The formula for the state model of the three-phase voltage-source PWM rectifier is: in, , , , , , , , , , ; , , This refers to the three-phase input grid voltage; , , This refers to the three-phase input grid current; R is the DC output voltage of the rectifier; L is the grid-side filter inductance; R is the grid-side equivalent resistance. , , , Let be the duty cycle of rectifier a. Let b be the duty cycle of rectifier b. The duty cycle of rectifier c is , ranging from 0 to 1; Let x(K) be the control period of the control system, x(K) be the state at time k, and x(K+1) be the state at the next time step. This is process noise; It also includes: constructing an ideal output equation between grid voltage and current based on the correspondence between voltage and current in the same phase of the grid; increasing the compensation amount of the ideal output equation to obtain the observed values ​​of the three-phase input current based on the voltage values ​​in the unscented Kalman filter; the formula for the observed values ​​of the three-phase input current is: in, , Q= For the process noise of the covariance matrix, R = The observation noise is the covariance matrix. The method of estimating the input current of a three-phase voltage-source PWM rectifier using an unscented Kalman filter specifically includes: obtaining the load and voltage under the original state distribution according to the circuit schematic of the three-phase PWM rectifier, and calculating the sampling current of the first sampling point in the sampling point set according to the ideal output equation between the grid voltage and current. The process involves: obtaining the dimension n of the three-phase PWM rectifier; calculating the total number of sampling points (2n+1) based on the dimension; calculating the sampling current value corresponding to each sampling point according to the current sampling formula; calculating the mean and covariance of each sampling point according to the mean and covariance formulas; calculating the one-step prediction value of the sampling point set and obtaining the one-step prediction and covariance matrix of the system state variables through weighted summation; substituting the one-step prediction value into the formula of the observation value to obtain the predicted observation value; and obtaining the mean and covariance of the system prediction through weighted summation based on the predicted observation value; calculating the Kalman gain and updating the system state to output the input current of the three-phase voltage-type PWM rectifier according to the updated system state. After updating the system state, the method further includes: updating the covariance formula, calculating the matrix equation value of the next round of sampling based on the previous round of covariance formula; when calculating the sampling current of the sampling points in the next round, using the matrix equation value of the next round of sampling to obtain the sampling current of each sampling point in the next round. The current sampling formula is as follows: The formulas for the mean and covariance are as follows: Where n=3, This is the sample mean, which is the sampled current at the first sampling point. The first of the matrix equations i List, To scale the ratio and reduce the overall expected error, The mean, For covariance, It is a constant. It is a constant. As a constant, P(0) = e.

Images