Double inductance parameter on-line identification method and system of double-fed motor MRAS position observer

Abstract

This invention discloses an online identification method and system for dual inductor parameters of a doubly-fed induction generator (DFIG) MRAS position observer, belonging to the field of sensorless control for doubly-fed variable-speed pumped storage. The method includes: a stator leakage inductance coefficient identification step and a mutual inductance coefficient identification step. The stator leakage inductance coefficient identification step includes: performing coordinate transformation on the position observed by the observer, the α' component and β' component of the rotor current to obtain the identified q-axis component of the rotor current; constructing a stator leakage inductance coefficient identification formula for identifying the stator leakage inductance coefficient to obtain the stator leakage inductance coefficient identified at the current moment; the mutual inductance coefficient identification step includes: using the stator leakage inductance coefficient identified at the current moment to perform online mutual inductance coefficient identification to obtain the mutual inductance coefficient identified at the current moment. This invention pertains to online identification of mutual inductance and stator leakage inductance coefficients under sensorless control conditions, which can improve the accuracy of the MRAS position observer and ensure the power control performance of the doubly-fed variable-speed pumped storage unit.

Description

Technical Field

[0001] This invention belongs to the field of sensorless control technology for doubly-fed variable speed pumped storage, and more specifically, relates to an online identification method and system for dual inductor parameters of a doubly-fed motor MRAS position observer. Background Technology

[0002] Doubly fed variable speed pumped storage (DFFP) units, as large-capacity energy storage devices, can significantly enhance grid frequency stability and are therefore widely used. Vector control is a common method for decoupling active and reactive power control in DFFP units. Vector control requires accurate position information of the doubly fed motor, which can be obtained through a position encoder or through sensorless control of the DFFP unit. Because position encoders are complex to install, have high maintenance costs, and are susceptible to interference, sensorless control of DFFP units has been extensively studied.

[0003] To date, sensorless control methods for doubly-fed induction motors (DFIGs) in wind turbine applications include the sliding mode observer method, Kalman filtering method, high-frequency signal injection method, and Model Reference Adaptive System (MRAS) method. Among these methods, MRAS technology has better dynamic performance, and its steady-state error is negligible in applications where the speed variation range of DFIGs is narrow. Furthermore, since the rotor speed of a DFIG variable-speed pumped storage unit is typically within ±10% of the synchronous speed, the MRAS position observer is suitable for DFIG variable-speed pumped storage units. In actual operation, DFIG variable-speed pumped storage units can operate under various conditions, such as light load, islanded grid, grid-connected, and flight start-up. The stator flux-based MRAS observer can easily adapt to various operating conditions in DFIGs, making it widely used in DFIG variable-speed pumped storage units.

[0004] However, the accuracy of MRAS position observers based on stator flux linkage is affected by the accuracy of the motor parameters used. Among these, the mutual inductance and stator leakage inductance coefficients have a significant impact on the observer's accuracy. Mismatched mutual inductance parameters or stator leakage inductance coefficients can cause errors in the position observer, leading to poor power control performance and, in severe cases, instability in doubly-fed variable-speed pumped-storage units. Existing online parameter identification technologies for position observers mainly include online mutual inductance identification methods based on rotor current amplitude using PI controllers and online mutual inductance identification methods based on motor equations using PI controllers. These methods all perform online identification of the mutual inductance coefficient, treating the stator leakage inductance coefficient as a constant. This results in a deviation between the stator leakage inductance coefficient and the actual stator leakage inductance coefficient during actual use. The mismatched stator leakage inductance coefficient not only directly affects the observer's accuracy but also impacts the accuracy of the mutual inductance coefficient identification. Furthermore, existing mutual inductance identification methods have poor convergence speed and dynamic performance. Summary of the Invention

[0005] To address the shortcomings and improvement needs of existing technologies, this invention provides an online identification method and system for dual inductor parameters of a doubly-fed motor MRAS position observer, aiming to improve the accuracy of the MRAS position observer and thus better control the operation of the doubly-fed variable speed pumped storage unit.

[0006] To achieve the above objectives, according to a first aspect of the present invention, an online identification method for dual inductor parameters of a doubly fed motor MRAS position observer is provided, comprising: a stator leakage inductance coefficient identification step and a mutual inductance coefficient identification step;

[0007] The stator leakage inductance coefficient identification step includes:

[0008] The position observed by the MRAS position observer α' component of rotor current i rα' and β' component i rβ' By performing a coordinate transformation, the q-axis component of the identified rotor current is obtained.

[0009] A formula for identifying the stator leakage inductance coefficient is constructed to identify the stator leakage inductance coefficient and obtain the stator leakage inductance coefficient at the current time. The formula for identifying the stator leakage inductance coefficient is: Among them, i sq This represents the q-axis component of the stator current.

[0010] The mutual inductance coefficient identification step includes: using the stator leakage inductance coefficient identified at the current moment. Perform online mutual inductance coefficient identification to obtain the mutual inductance coefficient at the current time.

[0011] Furthermore, the stator leakage inductance coefficient identification step also includes:

[0012] Update the stator leakage inductance coefficient identified at the current moment using the constructed enabling update conditions. The enable update condition is: |ε|≤ε0; where |ε| represents the absolute value of the generalized error ε of the MRAS position observer, and ε0 is a set first threshold.

[0013] Furthermore, the mutual inductance coefficient identification step specifically includes:

[0014] S1. Calculate the square of the actual output rotor current amplitude |i r | 2* The mutual inductance coefficient to be identified and the stator leakage inductance coefficient identified at the current moment, and the mutual inductance coefficient to be identified. Based on this, the square of the estimated rotor current amplitude is calculated.

[0015] S2, with the squared value |i r | 2* and the square value The difference e is used as the input to the adaptive law to obtain the identification generalized error ε. m The adaptive law is designed using gradient descent and the MIT rule.

[0016] S3. The generalized error ε of the identification m The mutual inductance coefficient identified at the current moment is obtained by performing integration and adding the result of the integration to the compensation term. The compensation term is used to compensate for dynamic quantities during the mutual inductance identification process;

[0017] S4. Determine whether the difference e is less than a set second threshold. If not, use the mutual inductance coefficient identified at the current time. Update the mutual inductance coefficient to be identified in S1, and recalculate the square value. Repeat steps S2-S3; if so, output the mutual inductance coefficient identified at the current time.

[0018] Furthermore, the adaptive law is:

[0019]

[0020] Where, ψ sd ψ sq Let i represent the d-axis and q-axis components of the stator flux linkage during stator flux linkage orientation, respectively. sd i sq These represent the d-axis and q-axis components of the stator current when the stator flux linkage is oriented. γ is the gradient coefficient, which is a constant.

[0021] Furthermore, the compensation term C is:

[0022]

[0023] Furthermore, in S1, the square value is calculated by constructing an adjustable model for the square of the rotor current amplitude. By constructing a reference model for the squared value of the rotor current amplitude, the squared value |i is calculated. r | 2* ;

[0024] The adjustable model is:

[0025]

[0026] The reference model is:

[0027]

[0028] Where, ψ sd ψ sq Let i represent the d-axis and q-axis components of the stator flux linkage during stator flux linkage orientation, respectively. sd i sq These represent the d-axis and q-axis components of the stator current when the stator flux linkage is oriented.

[0029] Furthermore, it also includes: quantitatively analyzing the relationship between the identified mutual inductance coefficient, the identified stator leakage inductance coefficient, and the observation error Δθ2 of the MRAS position observer:

[0030]

[0031]

[0032] Among them, L m ΔL represents the actual mutual inductance coefficient. m σ represents the difference between the identified mutual inductance coefficient and the true mutual inductance coefficient, σ represents the true stator leakage inductance coefficient, and Δσ represents the difference between the identified stator leakage inductance coefficient and the true stator leakage inductance coefficient. sq i represents the q-axis component of the stator current when the stator flux linkage is oriented. rd i rq These represent the d-axis and q-axis components of the rotor current, respectively.

[0033] According to a second aspect of the present invention, an online identification system for dual inductor parameters of a doubly fed motor MRAS position observer is provided, for performing the method described in any one of the first aspects, the system comprising: a stator leakage inductance coefficient identification module and a mutual inductance coefficient identification module;

[0034] The stator leakage inductance coefficient identification module is used to identify the position observed by the MRAS position observer. α' component of rotor current i rα' and β' component i rβ' By performing a coordinate transformation, the q-axis component of the identified rotor current is obtained.

[0035] The stator leakage inductance coefficient identification module is also used to construct a stator leakage inductance coefficient identification formula, which is used to identify the stator leakage inductance coefficient to obtain the stator leakage inductance coefficient identified at the current time. The formula for identifying the stator leakage inductance coefficient is: Among them, i sq This represents the q-axis component of the stator current.

[0036] The mutual inductance coefficient identification module is used to identify the stator leakage inductance coefficient at the current time. Perform online mutual inductance coefficient identification to obtain the mutual inductance coefficient at the current time.

[0037] Furthermore, the mutual inductance coefficient identification module includes:

[0038] Reference model used to calculate the squared value of the actual output rotor current amplitude |i r | 2* ;

[0039] An adjustable model is used to identify the mutual inductance coefficient to be identified and the stator leakage inductance coefficient identified at the current time. Based on this, the square of the estimated rotor current amplitude is calculated.

[0040] Adaptive law, used to express the squared value |i r | 2* and the square value The difference e is used as input to obtain the generalized identification error ε. m The adaptive law is designed using gradient descent and the MIT rule.

[0041] An integral controller is used to integrate the identification generalized error ε m Perform integration;

[0042] The compensation term is used to compensate for dynamic quantities during the mutual inductance identification process;

[0043] The mutual inductance coefficient identification unit is used to add the result of the integral operation to the compensation term to obtain the mutual inductance coefficient identified at the current time.

[0044] The judgment unit is used to determine whether the difference e is less than a set second threshold. If not, it uses the mutual inductance coefficient identified at the current time. Update the mutual inductance coefficient to be identified in the adjustable model and recalculate the square value. The adaptive law, integral controller, compensation term, and mutual inductance coefficient identification unit are repeatedly executed; if so, the mutual inductance coefficient identified at the current time is output.

[0045] According to a third aspect of the invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method as described in any of the first aspects.

[0046] In summary, the above-described technical solutions conceived in this invention can achieve the following beneficial effects:

[0047] (1) The online identification method of dual inductor parameters of the doubly fed motor MRAS position observer of the present invention includes identification of stator leakage inductance coefficient and mutual inductance coefficient, and identification based on the position observed by the observer. α' component of rotor current i rα' and β' component i rβ' A stator leakage inductance coefficient identification formula is designed to identify the stator leakage inductance coefficient. The identified stator leakage inductance coefficient is used to identify the mutual inductance coefficient, instead of directly setting the stator leakage inductance coefficient as a constant, which improves the accuracy of online mutual inductance identification. At the same time, when identifying the stator leakage inductance coefficient, the online mutual inductance identification result does not need to be used. The stator leakage inductance coefficient identification and the online mutual inductance identification are decoupled, which further improves the accuracy of online mutual inductance identification. This can effectively reduce the observation error of the position observer, thereby improving the accuracy of the MRAS position observer and ensuring the power control performance of the doubly-fed variable speed pumped storage unit.

[0048] (2) Further, the generalized error ε in the position observer is used to construct the enable update condition. Since the online identification of the stator leakage inductance coefficient is coupled with the position observer, when the generalized error in the position observer is less than a small value, the stator leakage inductance coefficient identified at the current time is updated. At this time, the accuracy of the stator leakage inductance coefficient identified is higher.

[0049] (3) Furthermore, when identifying mutual inductance coefficient based on the stator leakage inductance coefficient obtained by identification, an adaptive law and a compensation term are designed specifically based on the stator leakage inductance coefficient obtained by identification, the difference between the actual value and the estimated value of the square value of the rotor current amplitude, the dq axis component of the stator flux linkage when the stator flux linkage is oriented, and the dq axis component of the stator current. The adaptive law and the compensation term work together to improve the convergence speed and dynamic performance of mutual inductance identification.

[0050] (4) Furthermore, the adjustable model and reference model designed in this invention have the same physical quantity (square value of rotor current amplitude) output. One outputs the true value of the square value of rotor current amplitude, and the other outputs the estimated value of the square value of rotor current amplitude. The two models work simultaneously. When the difference between the outputs of the two models is small, the purpose of tracking the actual value of mutual inductance identification value can be achieved.

[0051] (5) Furthermore, compared with the existing technology, which analyzes the influence of mismatched motor parameters (mutual inductance and stator leakage inductance coefficients) on motor speed observation and then transforms it into the influence on observer position observation, the parameter sensitivity analysis of the present invention directly analyzes the relationship between mismatched motor parameters and observer position observation error, directly and quantitatively reflects the influence of mismatched motor parameters on observer position observation, which facilitates the more targeted design of the observer-based parameter identification system.

[0052] In summary, this invention enables online identification of dual inductor parameters, which can improve the accuracy of the MRAS position observer and thus better control the operation of the doubly fed variable speed pumped storage unit. Attached Figure Description

[0053] Figure 1 This is a control block diagram for the online identification method of dual inductor parameters of the doubly fed motor MRAS position observer in this embodiment of the invention, which performs position-free control.

[0054] Figure 2 This is a control block diagram for online identification of stator leakage inductance coefficient in an embodiment of the present invention.

[0055] Figure 3 This is a control block diagram for online identification of mutual inductance coefficients in an embodiment of the present invention.

[0056] Figure 4 This is a control block diagram for the MRAS position observer of the stator flux linkage. Detailed Implementation

[0057] 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. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0058] In this invention, the terms "first," "second," etc., used in the invention and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0059] like Figure 1 As shown, and in combination Figure 2and Figure 3 The online identification method for dual inductor parameters of the doubly fed motor MRAS position observer of the present invention mainly includes: stator leakage inductance coefficient identification step and mutual inductance coefficient identification step;

[0060] The stator leakage inductance coefficient identification steps include:

[0061] Obtain the position observed by the MRAS position observer The α' component i of the measured rotor current rα' and the β' component i of the measured rotor current rβ' ;

[0062] The observed position α' component of rotor current i rα' and β' component i rβ' By performing a coordinate transformation, the q-axis component of the identified rotor current is obtained.

[0063] A formula for identifying the stator leakage inductance coefficient is constructed to identify the stator leakage inductance coefficient and obtain the stator leakage inductance coefficient at the current time. The formula for identifying the stator leakage inductance coefficient is as follows: i sq The q-axis component of the stator current is given; the stator leakage inductance coefficient identification formula is designed based on the flux linkage equation when the stator flux linkage is oriented.

[0064] Mutual inductance coefficient identification steps: Use the stator leakage inductance coefficient identified at the current time. Perform mutual inductance coefficient identification to obtain the mutual inductance coefficient at the current time.

[0065] As a further design of the present invention, since the input for online identification of the stator leakage inductance coefficient includes the position observed by the position observer, the input includes the position observed by the position observer. In other words, the online identification of the stator leakage inductance coefficient is coupled with the position observer. When the position error observed by the position observer is small, the stator leakage inductance coefficient identified at the current moment is updated. At this point, the stator leakage inductance coefficient is identified. It has higher accuracy.

[0066] Therefore, the stator leakage inductance coefficient identification step also includes updating the stator leakage inductance coefficient identified at the current moment using the constructed enabling update condition. The enable update condition is: |ε|≤ε0, where |ε| represents the absolute value of the generalized error ε in the position observer, and ε0 is the set first threshold, which is a constant and is selected based on experience in practical applications.

[0067] When the absolute value of the generalized error ε in the position observer does not exceed the set threshold, the error Δθ2 of the position observer will also be relatively small.

[0068] Preferably, the stator leakage inductance coefficient identification step also includes using a low-pass filter to analyze the stator leakage inductance coefficient identified at the current moment. Perform filtering.

[0069] Specifically, the mutual inductance coefficient identification steps include:

[0070] S1. Calculate the square of the actual output rotor current amplitude |i r | 2* And the mutual inductance coefficient to be identified and the stator leakage inductance coefficient identified at the current moment. Based on this, the square of the estimated rotor current amplitude is calculated.

[0071] S2, using the square of the actual output rotor current amplitude |i r | 2* and the square of the estimated rotor current amplitude The difference e is used as the input to the adaptive law to obtain the identification generalized error ε. m Among them, adaptive laws are designed using gradient descent and the MIT rule;

[0072] S3, Identify the generalized error ε m The mutual inductance coefficient identified at the current moment is obtained by performing an integral operation and adding the result of the integral operation to the compensation term C. The compensation term is used to compensate for dynamic quantities during the mutual inductance identification process;

[0073] S4. Determine if the difference e is less than the set second threshold. If not, use the mutual inductance coefficient identified at the current moment. Update the mutual inductance coefficient to be identified in S1, and recalculate the square of the estimated rotor current amplitude. Repeat steps S2-S3; if so, output the mutual inductance coefficient identified at the current time. The goal is to achieve the purpose of tracking the actual value of the mutual inductance identification value.

[0074] Preferably, the second threshold is set to 0, at which point the mutual inductance identification value can fully track the actual value; in this embodiment of the invention, the second threshold is set to 0.

[0075] Specifically, by constructing a reference model for the square of the rotor current amplitude, the square of the actual output rotor current amplitude |i r | 2* The reference model is as follows:

[0076]

[0077] In equation (1), i r Indicates rotor current;

[0078] Specifically, by constructing an adjustable model for the square of the rotor current amplitude, the estimated square of the rotor current amplitude is calculated. The adjustable model is:

[0079]

[0080] In equation (2), For the mutual inductance coefficients to be identified, during the iteration process, the coefficients are updated in real time based on the mutual inductance coefficients identified at the current moment. Until the difference e is less than the set second threshold; ψ sd ψ represents the d-axis component of the stator flux linkage when the stator flux linkage is oriented. sq i represents the q-axis component of the stator flux linkage when the stator flux linkage is oriented. sd i represents the d-axis component of the stator current when the stator flux linkage is oriented. sq This represents the q-axis component of the stator current when the stator flux linkage is oriented. This represents the stator leakage inductance coefficient identified at the current moment.

[0081] The input to the reference model does not include the mutual inductance coefficients to be identified. The input to the adjustable model includes the mutual inductance coefficients to be identified.

[0082] Specifically, adaptive laws are designed using gradient descent and the MIT rule, including:

[0083] Define the performance index function J of the controlled object. In this embodiment of the invention, the performance index function J is defined as follows:

[0084]

[0085] in, As the controlled object, For ease of analysis, the identification of the reciprocal of mutual inductance is used instead of the identification of mutual inductance.

[0086] Since the goal of online mutual inductance identification is to adjust one of the input parameters of the adjustable model: the mutual inductance coefficient to be identified. This makes e = 0, at which point the corresponding performance index function J is 0, taking its minimum value across the entire range; where,

[0087] Based on the principle of gradient descent, a negative gradient is used to make the performance index function change towards its minimum value, so that J converges to the minimum value quickly. Specifically, J is taken as a pair of... Partial derivatives:

[0088]

[0089] Based on formula (4), an additional quantity is added to control the gradient descent speed, namely the gradient coefficient γ. If the descent is too fast, it is easy to get trapped in a local optimum. γ can be set according to experience.

[0090] Formula (4) above can be transformed into:

[0091]

[0092] According to the MIT rule, the identified parameters The adjustment direction is the negative gradient direction of J:

[0093]

[0094] After sorting, we can conclude that:

[0095]

[0096] Therefore, the adaptive law designed in this invention is:

[0097]

[0098] Meanwhile, based on the approximate value of the m identification value during stator flux linkage orientation, the corresponding design compensation term C is:

[0099]

[0100] The online identification method for dual-inductance parameters of the doubly-fed motor MRAS position observer of the present invention includes: a stator leakage inductance coefficient identification step and a mutual inductance coefficient identification step, based on the position observed by the observer. α' component of rotor current i rα' and β' component i rβ' A stator leakage inductance coefficient identification formula is designed to identify the stator leakage inductance coefficient. The identified stator leakage inductance coefficient is used to identify the mutual inductance coefficient, instead of directly setting the stator leakage inductance coefficient as a constant, which improves the accuracy of online mutual inductance identification. At the same time, when identifying the stator leakage inductance coefficient, it is not necessary to use the online mutual inductance identification result, that is, the stator leakage inductance coefficient identification and the online mutual inductance identification are decoupled, which further improves the accuracy of online mutual inductance identification. This can effectively reduce the observation error of the position observer, thereby improving the accuracy of the MRAS position observer and ensuring the power control performance of the doubly-fed variable speed pumped storage unit.

[0101] Furthermore, the generalized error ε in the position observer is used to construct the enabling update condition. Since the online identification of the stator leakage inductance coefficient is coupled with the position observer, when the generalized error in the position observer is smaller than a certain value, the stator leakage inductance coefficient identified at the current time is updated. At this time, the accuracy of the identified stator leakage inductance coefficient is higher.

[0102] Furthermore, when identifying mutual inductance coefficients based on the identified stator leakage inductance coefficient, an adaptive law and a compensation term were specifically designed based on the difference between the actual and estimated values ​​of the actual and estimated values ​​of the square of the rotor current amplitude, the dq-axis components of the stator flux linkage during stator flux linkage orientation, and the dq-axis components of the stator current. The adaptive law and the compensation term work together to improve the convergence speed and dynamic performance of mutual inductance identification.

[0103] Meanwhile, this online mutual inductance identification step does not require rotor position information, which can improve the reliability of online mutual inductance identification under sensorless control conditions.

[0104] As a further design of the present invention, the online identification method for dual inductor parameters of the doubly fed motor MRAS position observer of the present invention also includes parameter sensitivity analysis of the MRAS position observer based on stator flux linkage, which is used to quantitatively analyze the influence of mutual inductance coefficient and stator leakage inductance coefficient on the accuracy of the position observer. Figure 4 This is a control block diagram of a stator flux linkage-based MRAS position observer. The stator flux linkage-based MRAS position observer includes a reference model of the stator flux linkage, an adjustable model of the stator flux linkage, an adaptive law, and a PI controller. Figure 1 and Figure 2 The physical quantity parameter R involved in s L represents the stator resistance of a doubly-fed induction generator. s L represents the stator inductance of a doubly-fed induction generator. s =L m +L ls R r L represents the rotor resistance of a doubly-fed induction generator. r Let α represent the rotor inductance of the doubly-fed induction generator, αβ represent the stator stationary two-phase coordinate system, dq represent the synchronous rotating coordinate system, α'β' represent the rotor stationary two-phase coordinate system, and θ represent the rotor inductance. r The rotor position, specifically the mechanical angle ω between stator phase A and rotor phase A. r The values ​​represent the rotor's mechanical angular velocity, where θ1 represents the angle between the synchronous rotating coordinate system and the stator stationary coordinate system, ω1 represents the angular velocity of the synchronous rotating coordinate system relative to the stator stationary coordinate system, and ω2 represents the angular velocity of the synchronous rotating coordinate system relative to the rotor stationary coordinate system. The superscript (^) indicates that the physical quantity is an identified or estimated value.

[0105] The parameter sensitivity analysis in this embodiment of the invention includes:

[0106] Assume that the mutual inductance coefficients of the observers are mismatched, that is, the mutual inductance coefficient L' actually used by the observers is mismatched. m With the actual mutual inductance coefficient L m There is an error ΔL between them m L'm =L m +ΔL m ;

[0107] When the observer is stable, the stator flux linkage and the mismatched mutual inductance coefficient L' in the adjustable model of the observer are constructed. m The relationship between the position observer error Δθ2 and the position observer error:

[0108]

[0109] in, Δθ2 represents the observed (identified) value of the q-axis component of the stator flux linkage during stator flux linkage orientation; Δθ2 represents the position observation error of the position observer; i rd i represents the d-axis component of the rotor current. rq Represents the q-axis component of the rotor current; where, θ2 represents the position observed by the MRAS position observer, and θ2 represents the angle between the synchronous rotating coordinate system of the position observer and the rotor stationary coordinate system.

[0110] Represent the stator flux of the reference model using the equations of the adjustable stator flux model:

[0111] ψ sq =(1+σ)L m i sq +L m i rq (11)

[0112] Where σ represents the actual stator leakage inductance coefficient of the doubly-fed motor, specifically, σ = L ls / L m L ls This indicates the stator leakage inductance of a doubly-fed motor.

[0113] In this embodiment of the invention, the control system employs stator flux linkage orientation:

[0114] ψ sq =0 (12)

[0115] Once the observer is stable, the adaptive law is:

[0116]

[0117] Where ε represents the generalized error of the observer; This represents the observed value of the d-axis component of the stator flux linkage during stator flux linkage orientation.

[0118] According to the combined formula (12):

[0119]

[0120] Subtracting equation (11) from equation (10) constructs the equation for mismatched mutual inductance and position observer error:

[0121] 0=(1+σ)L' m i sq -(1+σ)L m i sq +L' m (-i rd sinΔθ2+i rq cosΔθ2)-L m i rq (15)

[0122] Taylor expansions of sin(Δθ2) and cos(Δθ2) with respect to Δθ2 = 0:

[0123]

[0124] Substituting formula (16) into equation (15), and ignoring the higher-order terms of Δθ2, we obtain the error ΔL using the mutual inductance coefficient. m The formula for representing the position observer error Δθ2 is:

[0125]

[0126] It can be seen that formula (17) quantitatively analyzes the mutual inductance coefficient L' actually used by the observer. m With the actual mutual inductance coefficient L m There is an error ΔL between them m In this case, the error will affect the accuracy of the position observer.

[0127] Similarly, when the stator leakage inductance coefficient of the observer is mismatched, it is assumed that the stator leakage inductance coefficient used by the observer is σ'=σ+Δσ, where Δσ is the difference between the used value and the true value of the stator leakage inductance coefficient.

[0128] When the observer is stable, construct the relationship between the stator flux linkage, the mismatched stator leakage inductance coefficient σ', and the position observer error Δθ2 in the adjustable model of the observer:

[0129]

[0130] Based on formulas (11)-(14), the equations for the mismatched stator leakage inductance coefficient and the position observer error are constructed as follows:

[0131] 0=(1+σ')L m i sq -(1+σ)L m i sq +L m (-i rdsinΔθ2+i rq cosΔθ2)-L m i rq (19)

[0132] Similarly, performing a Taylor expansion of sin(Δθ2) and cos(Δθ2) with Δθ2 = 0, substituting the expansion into equation (19), and neglecting the higher-order terms of Δθ2, we obtain the formula for the position observer error Δθ2 expressed in terms of the stator leakage inductance coefficient usage error Δσ:

[0133]

[0134] It can be seen that formula (20) quantitatively analyzes that if there is an error Δσ between the actual stator leakage inductance coefficient σ' used by the observer and the true stator leakage inductance coefficient σ, this error will affect the accuracy of the position observer.

[0135] It can be seen that mismatched mutual inductance and stator leakage inductance coefficients both affect the accuracy of the position observer. Furthermore, the MRAS position observer based on stator flux linkage is relatively more sensitive to mutual inductance coefficient errors. In addition, since stator leakage inductance coefficients are required when identifying mutual inductance coefficients, mismatched stator leakage inductance coefficients will also affect the accuracy of mutual inductance coefficient identification.

[0136] Compared to existing technologies that analyze the impact of mismatched motor parameters (mutual inductance and stator leakage inductance coefficients) on motor speed observation and then convert this into an impact on observer position observation, the parameter sensitivity analysis of this invention directly analyzes the relationship between mismatched motor parameters and observer position observation error. It directly and quantitatively reflects the impact of mismatched motor parameters on observer position observation, which is more instructive for designing observer-based parameter identification systems.

[0137] According to a second aspect of the present invention, an online identification system for dual inductor parameters of a doubly fed motor MRAS position observer is also provided, for performing the steps corresponding to the online identification method for dual inductor parameters of a doubly fed motor MRAS position observer in the above embodiments; the system includes: a stator leakage inductance coefficient identification module and a mutual inductance coefficient identification module;

[0138] The stator leakage inductance coefficient identification module is used to identify the position observed by the MRAS position observer. α' component of rotor current i rα' and β' component i rβ' By performing a coordinate transformation, the q-axis component of the identified rotor current is obtained.

[0139] The stator leakage inductance coefficient identification module is also used to construct a stator leakage inductance coefficient identification formula. This formula is used to identify the stator leakage inductance coefficient to obtain the stator leakage inductance coefficient identified at the current time. The formula for identifying the stator leakage inductance coefficient is: Among them, i sq This represents the q-axis component of the stator current.

[0140] The mutual inductance coefficient identification module is used to identify the stator leakage inductance coefficient at the current time. Perform online mutual inductance coefficient identification to obtain the mutual inductance coefficient at the current time.

[0141] Specifically, the mutual inductance coefficient identification module includes: a reference model of the square of the rotor current amplitude, an adjustable model of the square of the rotor current amplitude, a compensation term, an adaptive law, an integral controller, a mutual inductance coefficient identification unit, and a judgment unit.

[0142] The reference model is used to calculate the squared value of the actual output rotor current amplitude |i r | 2* ;

[0143] The adjustable model is used to identify the mutual inductance coefficient to be identified and the stator leakage inductance coefficient identified at the current time. Based on this, the square of the estimated rotor current amplitude is calculated.

[0144] Adaptive laws are used to express squared values ​​|i r | 2* Sum of squares The difference e is used as input to obtain the generalized identification error ε. m Among them, adaptive laws are designed using gradient descent and the MIT rule;

[0145] The integral controller is used to identify the generalized error ε m Perform integration;

[0146] The compensation term is used to compensate for dynamic quantities during the mutual inductance identification process;

[0147] The mutual inductance coefficient identification unit is used to add the result of the integral operation to the compensation term to obtain the mutual inductance coefficient identified at the current time.

[0148] The judgment unit is used to determine whether the difference e is less than the set second threshold. If not, it uses the mutual inductance coefficient identified at the current time. Update the mutual inductance coefficients to be identified in the adjustable model and recalculate the squared values. The adaptive law, integral controller, compensation term, and mutual inductance coefficient identification unit are repeatedly executed; if so, the mutual inductance coefficient identified at the current time is output.

[0149] According to a third aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the online identification method for dual inductor parameters of a doubly fed motor MRAS position observer as described in the above embodiments.

[0150] 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 online identification of dual inductor parameters of a doubly-fed motor MRAS position observer, characterized in that, include: Steps for identifying stator leakage inductance coefficient and mutual inductance coefficient; The stator leakage inductance coefficient identification step includes: The position observed by the MRAS position observer α' component of rotor current i rα' and β' component i rβ' By performing a coordinate transformation, the q-axis component of the identified rotor current is obtained. A formula for identifying the stator leakage inductance coefficient is constructed to identify the stator leakage inductance coefficient and obtain the stator leakage inductance coefficient at the current time. The formula for identifying the stator leakage inductance coefficient is: Among them, i sq This represents the q-axis component of the stator current. The mutual inductance coefficient identification step includes: using the stator leakage inductance coefficient identified at the current moment. Perform online mutual inductance coefficient identification to obtain the mutual inductance coefficient at the current time.

2. The method according to claim 1, characterized in that, The stator leakage inductance coefficient identification step also includes: Update the stator leakage inductance coefficient identified at the current moment using the constructed enabling update conditions. The enable update condition is: ε≤ε0; where ε represents the absolute value of the generalized error ε of the MRAS position observer, and ε0 is a set first threshold.

3. The method according to claim 1 or 2, characterized in that, The mutual inductance coefficient identification step specifically includes: S1. Calculate the square of the actual output rotor current amplitude i. r 2* The mutual inductance coefficient to be identified and the stator leakage inductance coefficient identified at the current moment, and the mutual inductance coefficient to be identified. Based on this, the square of the estimated rotor current amplitude is calculated. S2, with the squared value i r 2* and the square value The difference e is used as the input to the adaptive law to obtain the identification generalized error ε. m The adaptive law is designed using gradient descent and the MIT rule. S3. The generalized error ε of the identification m The mutual inductance coefficient identified at the current moment is obtained by performing integration and adding the result of the integration to the compensation term. The compensation term is used to compensate for dynamic quantities during the mutual inductance identification process; S4. Determine whether the difference e is less than a set second threshold. If not, use the mutual inductance coefficient identified at the current time. Update the mutual inductance coefficient to be identified in S1, and recalculate the square value. Repeat steps S2-S3; if so, output the mutual inductance coefficient identified at the current time.

4. The method according to claim 3, characterized in that, The adaptive law is: Where, ψ sd ψ sq Let i represent the d-axis and q-axis components of the stator flux linkage during stator flux linkage orientation, respectively. sd i sq These represent the d-axis and q-axis components of the stator current when the stator flux linkage is oriented. γ is the gradient coefficient, which is a constant.

5. The method according to claim 4, characterized in that, The compensation item C is:

6. The method according to claim 3, characterized in that, In S1, the square value is calculated by constructing an adjustable model for the square of the rotor current amplitude. By constructing a reference model for the squared value of the rotor current amplitude, the squared value i is calculated. r 2* ; The adjustable model is: The reference model is: Where, ψ sd ψ sq Let i represent the d-axis and q-axis components of the stator flux linkage during stator flux linkage orientation, respectively. sd i sq These represent the d-axis and q-axis components of the stator current when the stator flux linkage is oriented.

7. The method according to claim 1, characterized in that, It also includes: quantitative analysis of the relationship between the identified mutual inductance coefficient, the identified stator leakage inductance coefficient, and the observation error Δθ2 of the MRAS position observer: Among them, L m ΔL represents the actual mutual inductance coefficient. m σ represents the difference between the identified mutual inductance coefficient and the true mutual inductance coefficient, σ represents the true stator leakage inductance coefficient, and Δσ represents the difference between the identified stator leakage inductance coefficient and the true stator leakage inductance coefficient. sq i represents the q-axis component of the stator current when the stator flux linkage is oriented. rd i rq These represent the d-axis and q-axis components of the rotor current, respectively.

8. An online identification system for dual-inductor parameters of a doubly-fed motor MRAS position observer, characterized in that, The system for performing the method according to any one of claims 1-7, the system comprising: a stator leakage inductance coefficient identification module and a mutual inductance coefficient identification module; The stator leakage inductance coefficient identification module is used to identify the position observed by the MRAS position observer. α' component of rotor current i rα' and β' component i rβ' By performing a coordinate transformation, the q-axis component of the identified rotor current is obtained. The stator leakage inductance coefficient identification module is also used to construct a stator leakage inductance coefficient identification formula, which is used to identify the stator leakage inductance coefficient to obtain the stator leakage inductance coefficient identified at the current time. The formula for identifying the stator leakage inductance coefficient is: Among them, i sq This represents the q-axis component of the stator current. The mutual inductance coefficient identification module is used to identify the stator leakage inductance coefficient at the current time. Perform online mutual inductance coefficient identification to obtain the mutual inductance coefficient at the current time.

9. The online identification system for dual inductor parameters according to claim 8, characterized in that, The mutual inductance coefficient identification module includes: Reference model used to calculate the squared value of the actual output rotor current amplitude i r 2* ; An adjustable model is used to identify the mutual inductance coefficient to be identified and the stator leakage inductance coefficient identified at the current time. Based on this, the square of the estimated rotor current amplitude is calculated. Adaptive law, used to express the squared value i r 2* and the square value The difference e is used as input to obtain the generalized identification error ε. m The adaptive law is designed using gradient descent and the MIT rule. An integral controller is used to integrate the identification generalized error ε m Perform integration; The compensation term is used to compensate for dynamic quantities during the mutual inductance identification process; The mutual inductance coefficient identification unit is used to add the result of the integral operation to the compensation term to obtain the mutual inductance coefficient identified at the current time. The judgment unit is used to determine whether the difference e is less than a set second threshold. If not, it uses the mutual inductance coefficient identified at the current time. Update the mutual inductance coefficient to be identified in the adjustable model and recalculate the square value. The adaptive law, integral controller, compensation term, and mutual inductance coefficient identification unit are repeatedly executed; if so, the mutual inductance coefficient identified at the current time is output.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-7.

Images